Impact of Laboratory Standards on the Accuracy of Blood ......Dr. Jochen Sieber (Sanofi, Frankfurt,...
Transcript of Impact of Laboratory Standards on the Accuracy of Blood ......Dr. Jochen Sieber (Sanofi, Frankfurt,...
Créditosfotografiadecapa:DialysisTechnicianSalary
AnaCarolinaQueijoFernandes
ImpactofLaboratoryStandardsontheAccuracyofBloodGlucoseMetersandItsClinicalEffecton
InsulinDosing.
DissertaçãoapresentadaàUniversidadedeCoimbraparacumprimento dos requisitos necessários à obtenção dograudeMestreemEngenhariaBiomédica
Orientador(es):Prof.Dr.MiguelMorgado(UniversidadedeCoimbra,Coimbra,Portugal)Dr.JochenSieber(Sanofi,Frankfurt,Germany.)
Coimbra,2016
i
Estetrabalhofoidesenvolvidoemcolaboraçãocom:
Sanofi
InstitutfürDiabetes-Technologie(IDT)
iii
Estacópiadateseéfornecidanacondiçãodequequemaconsultareconhecequeosdireitos de autor são pertença do autor da tese e que nenhuma citação ouinformaçãoobtidaapartirdelapodeserpublicadasemareferênciaapropriada.Thiscopyofthethesishasbeensuppliedonconditionthatanyonewhoconsultsitisunderstood to recognize that its copyright rests with its author and that noquotation from the thesis and no information derived from it may be publishedwithoutproperacknowledgement.
v
Acknowledgments
I am sincerely grateful for the opportunity provided by Sanofi, instigated bytherelentlessDr.GaborBoka.
Iwould like to thankDr. FrankFlacke forwelcomingme since the first day,andprovidingknowledgebeyondmyacademicwork.TomysupervisorDr. JochenSieber for valuable insights on diabetes mellitus and Dr. Alexandra Beer for hermentoring, not only did she give technical knowledge but also very muchappreciatedguiding.ToDr.Freckmannforkindlyprovidingthedata for thisworkandDr.Pleus forpatientlyansweringallmyquestions.A special thankyou tomysupervisorProfessorMorgadoforreviewingthiswork.
Thankyou tomy friendsat Sanofi forgreat conversationsduring lunchbreakandmakingiteasierbeingawayfromhome.
TomybrotherforbeingthedrivingforcetofinishthisworkandtoKatiforthemoralsupport.
Aosmeuspaispeloapoioincondicional
vii
Abstract
Diabetes Mellitus is characterized by inexistent or insufficient insulinproduction, with consequent hypoglycemia. To overcome this deficiency patientsneedtoadministerexogenousinsulintocoverhighbloodglucoselevels.Asaresulttheamountofdruginjectedisdependentonbloodglucoseconcentration,measuredbypatientswithhandheldbloodglucosemeters.Thesedevices,however,canhavedifferent accuracy depending on various properties and also the laboratorystandardusedtocalibratethem.Therefore itwas firstmadeareviewofBGMs, itstechnology,history,regulatoryrequirements,qualityrequirementsandlimitations,suchas,userhandlingandtechnicalerrors.
Secondly, and since it isnot clearwhat is the impacton insulindosewhenmeasuring glucose with BGMs performing and calibrated in a different way, thequantitativeeffectofthesedifferencesoncorrectinsulindosingwasinvestigated.
With that purpose, insulin doses based on results for each BGM and tworeferencemethodswerecalculatedandcompared.Aseparateanalysiswasmadeforlow,normalandhighbloodglucosetodistinguishmeterswithbetterperformancesineachlevel.
The basis of this work was an accuracy study where 10 meters werecompared with two enzymatic methods following ISO standards. Accuracycompliance to ISO:15197:2013 criteria for the evaluated meters was alsosummarized.
ix
Resumo
Adiabetesmellituséumadoençacaracterizadapelainexistenteouinsuficienteproduçãode insulina, comconsequentehiperglicemia.Demodoaultrapassarestalimitação,osdoentesnecessitamdeadministrarinsulinaparacobrirníveiselevadosdeglicemia.Assim,aquantidadedeinsulinaadministradadependedaconcentraçãodeglicosenosangue,queémedidacomaparelhosportáteischamadosmedidoresdeglicose no sangue. Estes dispositivos, no entanto, possuem valores diferentes deexactidãoresultadodediferençastecnológicasequímicase,também,doaparelhodecalibração.Primeiro, para perceber o funcionamento dosmedidores de glicose, foi feita umarevisãodahistória,tecnologiaregulamentação,requisitosdequalidadeelimitações,taiscomo,errostécnicosedoutilizador.
Seguidamente, foi investigado as diferenças no doseamento de insulinabaseado em valores adquiridos com dispositivos com tecnologias e desempenhosdiferentes. Para isso, doses de insulina foram calculadas e comparadas utilizandovaloresdeglicosenosanguemedidoscom10dispositivosdiferentesedoismétodoslaboratoriaisdereferência.Umaoutraanálisefoifeitademodoaperceberquaisosmedidorescommelhordesempenhoemváriosníveisdeglicemia.
Este trabalho é fundamentado num estudo laboratorial que analisou aexactidão de 10 medidores de glicose utilizando dois métodos laboratoriais dereferência. O estudo referido seguiu os procedimentos referidos na normainternacionalISO15197.
xi
TableofContentsAcknowledgments....................................................................................................................vAbstract.....................................................................................................................................viiResumo ...................................................................................................................................... ixTableofContents.................................................................................................................... xi1 Introduction ....................................................................................................................... 11.1 Diabetesmellitus....................................................................................................................11.1.1 Insulintherapy ................................................................................................................................. 2
1.2 Guidelinesandrecommendationsondiabetesdiagnosisandmanagement.....51.3 Selfmonitoringofbloodglucose(SMBG).................................................................... 111.3.1 History................................................................................................................................................111.3.2 Technologyandchemistry ........................................................................................................121.3.3 Requirements..................................................................................................................................181.3.4 ImportanceofAccuracy..............................................................................................................20
2 IntroductiontotheProblem.......................................................................................293 Methods .............................................................................................................................334 Results ................................................................................................................................395 DiscussionandConclusion..........................................................................................415.1 Significanceofthisstudy................................................................................................... 415.2 Methodology.......................................................................................................................... 415.3 Summaryofresults ............................................................................................................. 425.4 Limitationsofthestudy..................................................................................................... 435.5 FutureOutlook ..................................................................................................................... 44
References ...............................................................................................................................47Appendix ..................................................................................................................................51
xii
1
1 Introduction
1.1 Diabetesmellitus
Diabetes Mellitus is a chronic disease that affects millions of peopleworldwide. According to the World Health Organization’s (WHO) “Global StatusReport on Noncommunicable Diseases” of 2014, the prevalence of diabetes wasestimatedtobe9%inpeopleolderthan18.Thesamereportsays1.5milliondeathsin2012wererelatedtodiabetes[1].By2040,theInternationalDiabetesFederation(IDF)predicts642millionpeoplewilllivewiththisdisease[2].
Definitionandclassification
It is a metabolic disorder characterized by chronic hyperglycemia due toinsufficient or inexistent insulin production by pancreatic β-cells in the islets ofLangerhansoranacquiredresistancetothehormone’seffect.
Etiologicallydiabetescanbeclassifiedintwomajorcategories,calledtype1and type 2 diabetes [3]. Further clinical cases are gestational diabetes and otherspecifictypes.
Type 1 diabetes is caused by cellular-mediated autoimmune β-celldestructioninpancreaticislets.Asaresultinsulinisnotproducedandpatientsmustrelyonexogenousinsulinadministrationtosurvive[3]. It isfrequentlydetectedinearlyteens.
On the contrary, type 2 characterized by “adult-onset” (age≥45), though inrecent years increasing number of children have been diagnosed [4], is stronglyassociatedwithobesity (about70 to80%ofdiabeticpatients areobese), lifestyleandgeneticpredisposition[5].
Insulinactioninitstargettissuescanbeimpairedduetoobesity.Therefore,today’selevatedrateofobesity(15%ofwomenand11%ofmenolderthan18areobese), has increased theprevalence of diabetes type2. Globally, 90%of diabetic
2
patientsaretype2[1].Althoughinearlystagesofdiseaseoverallinsulinproductionisnormal,thereisadelayinitsreleaseanditisrelativelylowtotissuessensitivity.This factor puts stress in pancreatic cells that have to compensate by producingmore insulin. Over time, due to insulin resistance, there is progressive β-celldysfunctionandlowerinsulinavailability.Thusinthefirststagesitcanbemanagedwith healthy diet and regular exercise but, inmost cases, insulin therapywill berequired in later stages [6]. However, patients’ unwillingness to change lifestylehabits leads to an accelerated decline in patients’ health and needing insulinadministration, even earlier. Increasingly sedentary lifestyle is indicated as oneofthereasons for thegrowingnumberofnewtype2diabetesdiagnoses[1].Coupledwith diabetes onset at a younger age, it is possible that the number of insulindependanttype2patientswillgrow.
Insulin is an anabolic hormone that facilitates glucose transport tomusclecells and adipose tissue and, in the form of glycogen, liver storage. In the pre-prandial state insulin levels are low and there is protein and lipid catabolism,formingketonebodies,hepaticgluconeogenesisandglycogenolisisinordertokeepplasma glucose concentration stable and available for the brain and other tissues(suchasredbloodcells).Whenthereisaspikeinglucoseafterameal if insulinisnotsufficientorthereisadecreasedresponsivenesstoitseffect,metabolismisthesameasinfastingwithincreasinghyperglycemia[5].Theseelevatedglucoselevelslead to microvascular complications such as retinopathy, neuropathy, andnephropathy.Eventuallymacrovascularcomplicationsappear,throughtheprocessofatherosclerosis,namelycoronaryarterydisease,peripheralarterialdiseaseandstroke[7].
1.1.1 Insulintherapy
Insulintherapyisessentialwhenthereisinsulindeficiency[8].Insuchcasesdaily insulindosesareadministered. Initialsourcesof insulin forclinicalusewereanimal pancreases (cow; pig) [9]. In 1972 pharmaceutical companies startedproduction of synthetic “human” insulin manufactured using recombinant DNA
3
technique [9]. In 1996 FDA approved the first insulin analog [10]. Insulin analogshaveamodifiedaminoacidsequencethatallowsshorter/longeractionprofiles[11].
Insulinisinjectedthroughsubcutaneousinjectionsensuringthatitarrivestothebloodstream.Otherformulations,lesspopular,allowalternativeadministrationroutesthroughinhalers[9].
One type of insulin is not sufficient tomimic insulin secretion of a healthypersonthroughouttheday.Varioustypesofinsulin,accordingtotheiractionprofile,areavailable.Bolusinsulins(rapid-;short-acting)haveapromptonsetofactionandshort duration. They are administered before meals or snacks. Basal insulins(intermediate-,long-acting)havealongeractingprofilelastingtoupto30hours[8].
Rapid-actinginsulinstartsactingwithin15minofinjectionwithpeakafter2handisclearedafter4to6hours.Itisusedforhighglucosecorrectionortocoverglycemicspikessuchasthoseafterameal.Thereforeitisadministeredbeforemealsorsnacks[12].
Short-actinginsulinreachesbloodstreamafter30min,withapeakofactionbetween2to3hours.Itlastfrom3to6hours[12].
Intermediate-acting insulinhasaprofileof actionmoreextendedwithanonsetbetweensecondandfourthhourafterinjection.Itspeakisaroundhour4to12.Itseffectivenesscanlastupto18hours[12].
Long-acting insulin or basal insulin has duration up to 24h (newerformulationscanlastover30h[13])anddoesnothaveadefinedpeakofaction.Itisadministeredintheeveningbeforebedorinthemorningbeforebreakfast,usuallyreachesbloodstreamafter90minutes[12](newformulations’onsetdevelopsoveraperiod of 6h [13]). It is designed to mimic small amounts of insulin secretionthroughoutthedayinresponsetoglucosereleasebytheliverandtolowermorningfastingplasmaglucoselevels.
WHO Expert Committee of Biological Standardization established aninternationalStandardforhumaninsulin.Bydefinition1InternationalUnit(IU) is
4
the activity of 0.03846mgof human insulin [14]. Since insulin analogsmaydifferfromhumaninsulintheyarelabeledinunits(U)[15].InsulinUnitsarerelatedtoitsbiological action,which is the blood glucose-lowering activity. Insulin is stored invialswithdifferentconcentrations(numberofinsulinunitsperml;U-100;U-500).Vialswithsmalleramountsarealsoavailable(U-40).
Syringes or insulin pens can be used to inject insulin. Insulin pens arecommonlyusedduetoeasinessofuse.Avarietyofinsulinpensareavailabletodayfor adults and children. Pre-filled pens already come with one vial and are non-reusable after all unitswere administered. Reusable pens can be usedwithmorethan one vial. Pens are graduated to measure insulin in units. Patients have tochoose,inthepen,theintendeddoseinunits,insertthelancetsubcutaneouslyandtheinsulinpenwillinjecttheselecteddose.Typically,insulinpensmeasureinsulinin2,1or0.5Uincrements.
Continuous subcutaneous insulin infusion is another way of deliveringinsulin that closely resemblesphysiological insulin release.Rapid acting insulin iscontinuously being injected in small doses and in response to measured glucose[16].Insulinpumpscandeliverdosesinevensmallerincrementsthaninsulinpens.
Insulin regimens are usually 40%-50% basal insulin and 50%-60% bolusinsulin[8][17].Basalinsulindosesarecalculatedbasedonglucoseconcentrationinthemorning (fasting). Insulin titrationmay followa titration schemeprovidedbytheHCP(HealthCarePractitioners)wherenumberofunitswillincreaseuntilbloodglucosetargetrangeisreached.
Mealtime insulin doses are intended to cover carbohydrate intake and tocorrectforhighbloodsugar.Dosestocovercarbohydrateintakearedeterminedbytheamountofcarbohydrate inthemealandtheamountofcarbohydratedisposedby1Uofinsulin.Oneunitusuallycoversfrom6-30gofcarbohydrates.Thisvaluemaydependontimeofday,exerciseandvaryfrompersontoperson[8].Dosestocorrect for high blood sugar are calculated based on pre-prandial blood glucose,target BG and insulin sensitivity factor (ISF). ISF expresses what glucoseconcentrationdecreaseisachievedbyinjectingoneunitof insulin.Thisparameter
5
is,atleastfordiabetespatientswithpumptherapyorintensifiedtherapy,onemajorparameterindiabetestherapy.
Important factors that affect insulin dose and insulin availability are,carbohydratecount;differences in insulinabsorptionandmeasuredbloodglucose[18].Erroreffectswillbeassociatedtoinsulinactionprofile.Anerrorinbolusdosecan represent an immediate emergency (30min)whereas in basal insulin errorswillhaveamorelastingeffect.Animportantcauseofconcernishypoglycemiaandinparticularnocturnal hypoglycemia (related tobasal insulin)hence the fear andreluctanceofmanypatientsinusinginsulinparticularlyinintensifiedregimens.
1.2 Guidelinesandrecommendationsondiabetesdiagnosisandmanagement
Many internationally recognized organizations are committed to diabetesresearch and patient care. The European Association for the Study of Diabetes(EASD), American Diabetes Association (ADA), American Association of ClinicalEndocrinologists(AACE)andInternationalDiabetesFederation(IDF)aresomewithmoreexposure.
To help patients andHCPsmanage diabetes and inform them of the latestdevelopment in treatment they publish in websites and scientific journalsinformation,guidelinesandrecommendationstoimprovequalityofcare.
Recommendationsandguidelinesarewrittenbygroupsofpeopleconsideredspecialists in the field, supported by clinical evidence. They reflect a consensusopinionamongthegroup.Diagnosis,monitoringandtherapeuticactionsindiabetesaresomeofthetopicsdiscussed.
DiagnosisofDiabetesMellitus
Diabetes is diagnosed by testing the presence of hyperglycemia either byplasmaglucosecriteriaorglycatedhemoglobin (HbA1c) criteria.Cut-offvaluesare
6
establishedbasedonthresholdsofglycemiaassociatedwithmicrovasculardisease.Testsareperformedwithlaboratorymethodsusingvenoussamples.
Table 1.1 represents parameters and respective cut-off values for diabetesdiagnosis.
Table1.1:DiagnosisofDiabetes.FPG,fastingplasmaglucose;2hPG,2-hourplasmaglucose;OGTT,oralglucosetolerancetest;HbA1c,glycatedhemoglobin;PG,plasmaglucose
Parameter Cut-offValue
FPG ≥126mg/dL(7mmol/l)2-hourPGOGTT ≥200mg/dL(11.1mmol/l)
HbA1c ≥6.5%(48mmol/mol)RandomPG ≥200mg/dL(11.1mmol/l)In the presence of classic hyperglycemia symptoms (polyuria, polydipsia
unexplainedweightloss),asingletestindiabetesrangeisenoughforclassification.If not, the test previously used should be repeated on a different day, except ifrandomPGwasmeasured,inwhichcaseanalternativemethodisrecommended.
HbA1ctestrepresentsthepercentageofglycatedhemoglobininerythrocytesand is ameasure of glycemic control of the past 2 to 3months (average time forerythrocyte turnover).When usingHbA1c to diagnose diabetes it is important totakeintoconsiderationfactorsthatcanleadtomisleadingresults.Suchfactorsareage,ethnicityandmedicalconditions(hemolyticanemias,hemoglobinopathies,irondeficiency).Thistestisnotintendedfordiagnosisinchildren,adolescents,pregnantwomenorwhensuspectedT1DM.Itmustalsobeperformedusingastandardized,validatedassay.
In patients with likely T1DM, confirmation by laboratory test should notdelaytreatmentinitiationinorderforpreventrapiddeterioration.
GlycemicTargets
ADA,IDF,CDA(CanadianDiabetesAssociation)recommendanHbA1ctarget< 7% or less for adult non-pregnant patients. Patients with no significant risk ofhypoglycemia or adverse events can target to < 6,5%. Conversely a less stringent
7
goal (≤ 8%) can be set for patients with risk of hypoglycemia, limited lifeexpectancy, advanced micro and macrovascular complications and othercomorbidities. Premeal capillary blood glucose should be targeted between 80-130mg/dLandthepostprandialpeakbelow180mg/dL [8].AACEproposes fastingandpremealbloodglucosebelow110mg/dLand2hpost-prandial<140mg/dL.
Even though these recommendations are based on clinical evidence ofimproved outcomes, they represent a guide for glycemic control. Goals should betailoredtoindividualneeds.
ManagingDiabetes-Lifestyleinterventions
Diabetes care must comprise lifestyle changes, some of which are dietmodification, increase inexercise, sufficientamountof sleep, smokecessationandmoderationinalcoholconsumption.
Medicalnutrition therapy isbeneficial forglycemiccontrol inDM.Glycatedhemoglobin can be reduced by 0.3-1% in T1DM and 0.5-2% in T2DM [8][19].Patients should be educated about nutrition therapy at the time of diagnosis.Strategies for meal planning, grocery shopping, healthy food choices should beaddressed.Regardingrelativedistributionofcaloriesacrossmacronutrients, thereisnotoneperfectproportionforeachone.Energysourcesmayrangefrom45-60%carbohydrate,15-20%protein,and20-35%fat.Foodchoicescanbeindividualizedconsidering preference, religion and geographic region with metabolic goals inmind.Emphasis shouldbegiven to foodwithhigh fiber contentand lowglycemicload in detriment of high sugar content.Whole grains, vegetables, fruits, legumesanddairyproductsshouldbethepreferredsourceofcarbohydrates[8].Preferenceto food with high content in polyunsaturated and monounsaturated fatty acidsshouldbegivenwithlimitedintakeofsaturatedfattyacidsandavoidanceoftransfats.
Educationoncarbohydratecountisessentialtopatientswithflexibleinsulinregimens since administered dose will depend on carbohydrate amount. Fixed
8
insulinregimensdependonsteadycarbohydrate intake.Patientswillbenefit fromeducationonmealplanningandportioncontrol.
In overweight or obese T2DM patients, weight loss has been shown to bebeneficial to reach glycemic targets and reduce need for pharmacologicinterventions [8]. Lower healthier body weight can be achieved following anutritionallybalanceddiet,reducedenergyintakeandregularexercise[20].
Regular physical activity is associated with increased cardiorespiratoryfitness, improved glycemic control, improved insulin sensitivity, blood pressurereduction, improved lipid profile and maintenance of weight loss [21]. It isrecommendedforpeoplewithdiabetestodomoderate-intensityaerobicexerciseatleast3daysaweek(atleast150mincumulatively)withnomorethan2consecutivedaysbetweenexercisedays[8][21].
Resistance training is also related with improved glycemic control anddecreased insulin resistance. Patients should perform resistance exercise twice aweek[21].
Exercisemayprompthypoglycemiainpatientsusinginsulin.Dosereductionor carbohydrate intake adjustment can prevent dangerously low blood glucose.Blood glucose should be checked before exercise and carbohydrates should beingestedifmeasurementislow(<100mg/dL)
Anotherimportantpartofcareindiabetesiseducationonself-management.It helps peoplewith diabetes tomake informed choices regarding treatment andfacilitateseffectiveself-managementthroughouttheirlife[22].
Pharmacologicaltherapy
T1DMrequiresinsulintherapyimmediatelyatdiagnosis.Patientsshouldbetreatedwithmultipledoseinsulininjections(3ormoreinjectionsperdayofbasalandprandial insulin)orcontinuoussubcutaneous insulin infusion(pumptherapy)[8]. Diabetes Control and Complication Trial established the advantages of
9
intensified insulin therapy in reducingmicrovascular complications.Although riskof severe hypoglycemia was approximately three times higher, new developedinsulinanalogsareassociatedwithlowerrateofhypoglycemia[23].
Bolusinsulincomprises50-60%oftotaldailyinsulin(TDI)distributedbythenecessary premeal doses. Doses must account for carbohydrate intake, glycemicindexofeachfoodandmeasuredBG.
WhenlifestyleinterventionsalonecannotmaintainglycemicgoalsadditionaloralpharmacologictherapyisonechoiceforT2DMmanagement.Thereareseveralagentswithdifferentphysiologicaleffects.Theyacttoreduceglycemiclevelmainlyby decreasing hepatic glucose production, elevating insulin sensitivity in targettissuesorimprovinginsulinsecretion[5].
Biguanides (Metformin) are generally the first recommended oral anti-hyperglycemicagent[8][24][25].Theyacttosuppresshepaticglucoseproductionandraiseperipherytissuessensitivitytoinsulin.Ifafterapproximately3months,athighestpossibledose,HbA1ctargetisnotachieved,combinationtherapyshouldbeconsidered with a second oral agent. Subsequently a third agent can be added.Choiceoforal agentneeds tobe individualized consideringpatient characteristics(degree of hypoglycemia, height, comorbidities), agent effects (blood glucoselowering efficacy, effect on height, side effect) and costs to provide best possiblecarewhileminimizingsideeffects[8].
Insulin therapy inT2DM isadvantageousandrecommendedwhenpatientspresent hyperglycemic symptoms and elevated glycemia or high HbA1c or whenothermethods fail to help patients achieve glycemic goals. If insulin is necessarytherapyshouldstartwithbasalinsulinoncedaily.
ADA/EASDguidelinespropose initialbasal insulindoseof10U/dayor0.1-0.2U/Kg/dayandadding10-15%or2-4UtopreviousdoseonceortwiceaweekfordosetitrationwhileBGisabovetarget.Ifglycemiccontrolisnotreachedguidelinesrecommendaddingpreferablyoneoralagent(Glucagon-likepeptide-1,GLP-1)thatstimulatesglucosereleaseorprandialinsulin.
10
Several schemes are available for prandial insulin initiation and titration.Initially only one dose before largest meal and if targets are not met addinginjectionsbefore2or3meals[8][17].Adjustmentscanbedonebyadding10-15%previousdoseor1-2U,twoorthreetimesaweekwhilepostprandialglucoseisnotattarget[8][24].
MonitoringDiabetes
Monitoring diabetes is needed to evaluate disease progression andeffectiveness of treatment. Glycemic control may be assessed with two principaltools:SMBGorlaboratorytestsforglycatedhemoglobin.
Glycatedhemoglobinisameasureofaverageglycemiccontrolandshouldbetestedevery3monthsor6ifvaluesareconsistentlyintargetrange.
SMBG is used as an aid to guide and assess interventions and detecthypoglycemia.UnderstandinghowtoperformSMBG,what resultsmeanandwhatare the appropriate actions is essential for optimal use of SMBG. Regularmeasurements give patients immediate feedback about intervention effects(exercise, food, medication) on glycemic control. If performed and recordedregularly helps establish glycemic patterns that can be correlated to therapeuticactions.
Frequent self-testing of blood glucose in insulin dependant patients onintensified insulin therapy is of the upmost importance and has been related toHbA1c reduction. Patients are advised to test BG beforemeals and snacks and attimes after meals, before exercise, at bedtime, when they think BG is low, aftertreating low BG until values normalize and before dangerous/serious tasks likedriving.T2DMpatientstreatedwithoncedailybasalinsulinplusoralagentsshouldtest at least once every day at different times [26]. If patients are on non-insulintherapyormedicalnutrition therapy,SMBG is recommended tocontrol treatmenteffectiveness[8].
11
1.3 Selfmonitoringofbloodglucose(SMBG)
The Diabetes Control and Complications Trial showed that tight glycemiccontrol, achieved with intensive insulin therapy, can slow the progression ofmicrovascular complications for type 1 patients [23]. Similarly, another study,conducted by UK Prospective Diabetes Study (UKPDS) Group, reached identicalconclusionsfortype2patients[27].Bothstudiesareatestimonytotheimportanceofintensifiedglycemiccontrolandinsulintherapyforwhichdailyself-monitoringofbloodglucose(SMBG)iskey.
1.3.1 History
Self-monitoringofbloodglucosestartedin1963withthedevelopmentofdrychemistryteststripsDextrostix®(MilesLaboratories,Elkhart,IN,USA;nowpartofBayer),byErnestAdamsandhisresearchteam,whichdisplayedabluecolourwithintensityproportional toglucose concentration. It requiredadropofbloodof50-100µl that had to be wiped after 1 minute. Glucose concentration was thenestimatedcomparingthepaperstrip’scolourtoacolour-concentrationchart[28].In1971, Anton Clemens at Miles Laboratories, patented the first device for self-monitoringofbloodglucose[29].Thedevice,calledAmesReflectanceMeter,usingreflectance photometry was able to detect and quantify reflected light fromDextrostix®teststrips.Thismeter,fortoday’sstandards,wasbulky,heavy(1.2kg)and expensive; however, it was an improvement from visual evaluation of teststrips. Further improvements were made releasing meters easier to handle, lessexpensiveandcapableofstoringdata.
Although Clarke and Lyons proposed the first glucose biosensor in 1962usinganamperometricenzymemethod[30],biosensorswereonlyavailabletoendusersaround1987whenMediSence(Waltham,MA,USA;nowAbbottDiabetesCare,Alameda, CA) launched ExacTech® [31]. It was very innovative in terms ofportability and appearance with two types offered. The customer could choose
12
betweenapenandacreditcardsizedmeter.Detectionandquantificationofglucosewasdonethroughanenzymecoupledwithanelectrontransfermolecule.
From then on the market of glucose sensing gradually moved fromphotometrictoelectrochemicaltechnology.
Operator-dependant steps that were potentially error sources were alsominimizedorremoved.Forexample,wipingofteststripsandtimingwasnolongernecessary;samplesizewasreducedwiththeintroductionofcapillarityfilling[28].
It is worth mentioning another generation of glucose meters used forcontinuous invivobloodglucosemonitoring.Theelectrochemical sensor isplacedsubcutaneouslythoughaflexiblecatheterintheformofaneedleoflessthan1mmin diameter [32]. Continuous glucose monitoring (CGM) permits a betterunderstanding of glucose levels’ progression with real-time values every 1 to 5minutes.
Since that first blood glucosemonitor (BGM), technology for blood glucosemeasurement evolved substantially. Todaymeters for self-measurement fit in thepalmofahand,requiresamplesofabout1µLorevensmaller,arefast(5s)andareeasytouserequiringverylittleinputfromtheuser.
1.3.2 Technologyandchemistry
BGMs need to be able to detect glucose from a complex blood sample andconvert its concentration into a measurable signal. Themajority of commerciallyavailable SMBG systems use electrochemistry and have two fundamentalcomponents: a biorecognition agent and a transducer. Bioreconignition is donethroughanenzymespecificforglucoseandaredoxmediatoractsastransductor.
Glucose measurement in BGMs is based on oxidation-reduction reactions.Thereiselectrontransferencefromglucosemoleculesfirsttotheenzyme,thenthemediator and lastly picked up by an electrode. Sowhen glucose is present in thesampleafluxofelectronsisgeneratedproportionaltoglucoseconcentration.Figure1.1 is a schematic representation of reactions happening in a GOD test strip. The
13
enzymesandmediatorsarecapableofparticipatinginseveralreactionsbecause,asseeninFigure1.1,aftereachreductionthereisanoxidationtoreturnthemoleculestotheirinitialstate.
Figure1.1:Oxidationandreductionreactions thatoccur ina teststripGODbased.OXandRedrepresent theoxidisedandreducedstaterespectively.Reprintedfrom[33]Copyright2008,AmericanChemicalSociety
Enzymes
They are oxidoreductases that act in the reducing end of glucose, thehydroxyl group, to form gluconolactone. In handheld BGMs those enzymes areglucose oxidase (GOD) or glucose dehydrogenase (GDH) [34]. In laboratoryanalysersHexokinase(HK)isalsousedbutnotGDH.
ThereactioncatalyzedbyGODisdescribedineq.(1.1)
(1.1)
GOD oxidises glucose transferring two electrons to the enzyme cofactor,flavin adenine dinucleotide (FAD) reducing it. Oxygen is the natural acceptor ofelectronsfromGOD,forminghydrogenperoxide(eq.(1.1)).H2O2isanactiveoxidantthatcannonspecificallyoxidizemetabolitesfromthesample,suchasuricacidandbilirubin, causing interferences in the measurement. For that reason, othermoleculesareusedasmediators(eq.(1.3)),butoxygencanstillcompetewiththemforelectrons.WhenusingaGODbasedmeter, sample type, i.e. venous, arterialorcapillary,willrenderdifferentresultsintermsofglucoseconcentrationexactlyduetodifferencesinoxygenpartialpressure.
14
SinceGDHisunabletouseO2astheelectronacceptoroxygencontentinthesample no longer interferes with measurement. While GOD only uses FAD ascoenzyme,GDHalsousesnicotineadeninedinucleotide(NAD)ornicotineadeninedinucleotidephosphate(NADP)andpyrroloquinolinequinone(PQQ).Thecofactorplays an important role in enzyme specificity. GDH(PQQ) can react withmaltose,xyloseandgalactose[35].GDH(NAD)reactswithxyloseandGDH(FAD)reactswithmaltose, mannose, galactose and lactose but in very low percentages [36]. Whenusing ameter, health carepractitioners (HCP) or patients shouldbe aware of thepossibleinterferencesandchoosedevicesaccordingly.
Mediators
Mediators are organic or inorganic molecules capable of existing in theoxidizedorreduced form. Inotherwords, theyreact rapidly toacceptanddonateelectrons[36].
Subsequent to glucose oxidation by the enzyme (eq. (1.2)), electrons aretransferred from thecofactor to themediator reducing it (eq.(1.3)).Theenzyme,now in the reduced form, canagain takepart inmore reactionswithglucose.Themediatorcanalsoberegeneratedwhenoxidizedbytheelectrode(eq.(1.4)). It isthis reoxidation that gives the signal current needed to measure glucoseconcentration.
(1.2)
(1.3) (1.4)
As seen in eq. 1 the pairO2/H2O2 can act as amediator.However, in BGMothermoleculesareusedas,forexample,hexacyanoferrateIII/hexacyanoferrateII;Hexaammineruthenium(III) chloride/Hexaammineruthenium(II) chloride [37]. Thelower the redox potential, less interference from other bioactive molecules will
€
glucose +GOD(ox ) →gluconolactone +GOD(red )
€
GOD(red ) + 2M(ox ) →GOD(ox ) + 2M(red )
€
2M(red ) →2M(ox ) + 2e−
15
occur [36]. Mediators are chosen according to their redox potential along withsolubilityandrateofdissolution,stabilityinmixtureswithproteins,redoxpotential,availabilityandcost[32].
Teststripdesign
Test strips are small thin multilayered plastic strips where the chemicalcomponents for detecting glucose are housed. Each manufacturer has its ownspecificdesign,butoverallteststripsaremaderoughlyinthesameway.
Twolayersofplasticinthebottomandtopgivesupport.Eachteststriphasatleast2or3electrodes.Usuallythereisaworkingelectrodeandareferenceandauxiliary electrode. Fill detection electrodes can also be present to detect whensufficient amountof sample is introduced.One commonmaterial for electrodes iscarbon ink that is screenprinted on test strips. Reference and auxiliary electrodecanbemadeofothermaterialse.g.screenprintedinkofAg/AgCl.
To start the reaction in the test strip themeter has to apply an electricalpotential to the electrodes.After analyzing the current time responseof the strip,the meter then converts the signal into a glucose concentration shown in thedisplay.
Figure1.2:Squematicviewofteststriplayers;(A)electrodesystem;(B)hydrophobiclayer.Reprintedfrom[33]
SMBGdevices takewhole blood samples tomeasure glucose concentrationbutprovideaglucoseresultthatisplasmaequivalent.Duetodifferencesinglucose
16
concentrationbetweensamplesaconversionfactorneedstobeapplied. Insystemaccuracytestingthemostcommonconversionfactorisgivenbyeq(1.5):
(1.5)
Gplasma refers to glucose concentration in plasma and Gwhole_blood glucoseconcentration in whole blood. The value 1.11 is the International Federation ofClinicalChemistry(IFCC)recommendedconversionfactoralthoughthereareotherconversionfactors,someofthemhematocritdependent.
Interferences
The complex enzymatic reaction in BGMs can be altered by a series ofinterferingagents.They canbebrought inby environment factors,manufacturingprocessorbloodsamplecomposition.
AltitudeaffectsmetersperformanceduetodifferencesinO2partialpressure.Asdescribedbefore,O2isthenaturalacceptorofelectronsinGODbasedteststrips.Since the electrode can only pick up electrons from the mediator, glucosemeasurement is compromised by changes in O2 partial pressure. For instance, athighaltitude,wherepO2islower,glucosevalueswillbeincreased[38].
Temperature and humidity change reaction kinetics with differentconsequences for different types of test strip [38]. Furthermore, mediator can bereducedathighertemperatures.Amediatorreducedbytemperaturewill,therefore,leadtoanelevatedsignalnotrelatedwithglucoseconcentration[39].Somemetershave internal thermometers to correct temperature differences. It is necessary,however,toavoidmeasurementswhentemperaturechangesrapidly.
Test stripmanufacturing process can prompt changes in enzyme coverageandteststriplotdifferencesthatbringvariability.Enzymecoverageisproportionaltotheproducedcurrent.Lossofenzymeareawillleadtounderestimationofglucose[38].
17
Blood compositionof the sample canalsobring interferences.Mostmetershave a percentage range of hematocrit 20-60%where they canmeasure glucose.Outsidethoseboundaries,valuesarenolongerreliable.Hematocritinterfereswithdiffusionofglucoseinthesampletothesiteofreaction.BGMsmeasureglucoseinwholebloodandarecalibratedtoprovidearesultthat isplasmaequivalent.Sinceerythrocytescontainintracellularglucoseatadifferentconcentrationthanplasma,variationsinhematocritcancauseerrors[38].Somemeterscancompensateforthiseffect. Anemia, certain types of cancer, chronic and end-stage renal disease,malnutritionorspecificdietdeficiencies,rheumatoidarthritis,andotherconditionslower HCT [40]. Patients with underlying conditions ormedication that can alterhematocritshouldnotusehandheldmeterssensitivetohematocrit.
PartialpressureofO2 in the sample, as in atmosphere, also interfereswithmeasurements[41].SincepO2isdifferentinarterial,venousorcapillaryblood,onlycapillarybloodshouldbeusedtomeasurebloodglucoseconcentrationwithSMBGsystems intended for capillary samples. Alternate site testing should only beperformedwhenexplicitlyapprovedinBGMs’labelling.
Certainchemicalsubstancespresentinthebloodsample,eitherendogenousorduetomedicationorpathologies,canhaveacompetitivereactioninthreestepsofthereactiononateststrip.Competingwiththeenzymesubtract(1),competingwiththemediator(2)orcompetingwiththeelectrode(3).
Table1.2:Effectofdifferentsubstancesonglucosereadings1–GOD;2–GDH.Adaptedfrom[38]
Substance Effectonreading Step
O2 Variable 2Uricacid Increase1 3Galactose Increase2 1Xylose Increase2 1
Acetaminophen Decrease1 3L-dopa Variable1 3
Tolazamine Variable1 3Ascorbicacid Variable1 3Icodextrin Increase2 3
18
Due torapidchanges inhematocrit,pO2,medicationSMBG isnotapprovedfor patients in critical care or neonatal care. Also when choosing a device HCPshouldtakeintoaccountthemedicationofthepatientandcheckmanufacturerlabelforpossibleinterferences.
Although technical limitations can lower accuracy, user handling is also animportant part in obtaining good results. Some basic steps such as handwashingandcontrol testingarevery importantwhenmeasuringbloodglucosewithBGMS.Thesewillbediscussedlaterin“Importanceofaccuracy”section.
1.3.3 Requirements
IntheEuropeanUnion(EU)BGMsforself-testingareregulatedbydirective98/79/EConinvitrodiagnostic(IVD)medicaldevices.IVDsareaspecificcategoryof medical devices “intended by the manufacturer to be used in vitro for theexaminationof specimens, includingbloodand tissuedonations,derived from thehuman body, solely or principally for the purpose of providing information:concerning a physiological or pathological state; (…) or to monitor therapeuticmeasures.”[42].BGMsareclassifiedasmoderateriskdevices(listB,AnnexII).
Demonstration of compliance with essential requirements described indirective98/79/ECwill allowapplicationof CE (ConformitéEuropéene)Mark andfreemarketing inEU.Notifiedbodiesareapartof theapprovalprocess forBGMS.Theseindependentthirdpartiesevaluatethequalityofdocumentationprovidedbythemanufacturerandcanaskforadditionalinformation.
Thedirectivedoesnotdescribespecific technical factors. Instead, itdefinesbroad essential requirements for any IVD regarding safety for all users. MoredetailedtechnicalevaluationsaredescribedinEuropeanharmonizedstandards(e.g.ISO13485-QualityManagementSystems).
The InternationalOrganization forStandardization(ISO) isan internationalbodythatspecifiesstandardstoensuresafetyandqualityofproductsandservices.
19
Performance requirements regardingsystemaccuracyofBGMs for self-testingarespecifiedinISO15197.Thisstandardwasrevisedin2013.
AnalyticalSystemaccuracy
Previousminimumsystemaccuracywasdefinedas:95%ofvaluesshouldbewithin either ±15mg/dL at glucose concentration <75mg/dL orwithin ±20% atglucoseconcentrations≥75mg/dL.Therewasnorestraintregardingoutliers(5%ofresultscouldfallanywhere).
Minimumsystemaccuracycriteriaarenowdefinedas:
“95%ofmeasuredglucosevaluesshallfallwithineither:
• ±0,83 mmol/l (±15 mg/dL) of the average measured values of thereference measurement procedure at glucose concentrations < 5,55mmol/l(<100mg/dL)or
• ±15%atglucoseconcentrations≥5,55mmol/l(≥100mg/dL).”[43]
Glucose concentration cutoff from absolute value to percentagewas raisedfrom75mg/dLto100mg/dL.This,however,doesnotmeanarelaxingincriteria.Figure1.3evidencesthatnewcriteriaaremorestringentforvalueslargerthan75mg/dL.Forvaluesbelow75mg/dL,existingaccuracyrequirementsaremaintained.
20
Figure1.3:Differenceplotwithsystemaccuracylimits.AccordingtoISO15197:2003(fullline)atleast95%ofresultsshallbewithin±15mg/dLatBGconcentrations<100mg/dLandwithin20%atBGconcentrations≥100mg/dL.The2013revision(dotedline)stipulatesthatat least95%ofresultsshallbewithin±15mg/dLatBGconcentrations<75mg/dLandwithin±20%atBGconcentrations≥75mg/dL).
Therevisedstandardadoptsarisk-basedapproachwiththeConsensusErrorGrid(CEG)forT1DM.Whereasbeforetherewerenorequirementsfor5%ofresults,now 99% of measured values are required to fall within zones A and B of CEG(explainedinnextsection).
1.3.4 Importanceofaccuracy
Measurement variability in BGM can be related to numerous factorsintroducedsincethemanufacturingprocesstothetimeofuserhandling.Ingeneral,sources of interferences can be related to the monitoring system, as describedearlier, calibration process or user errors [44]. All three can be present andcontributetothefinalsystemaccuracy.
BGMSystemaccuracycriteriaandproceduresforitsassessmentaredefinedby the International Organization for Standardization in ISO 15197. Thestandardizationoftheseproceduresisimportanttodefineminimumrequirements
21
and to have comparability between laboratory studies. These studies determinesystem accuracy and demonstrate meter’s compliance to the internationalstandards,whichismandatoryforCEmarking.Butregularandindependentstudiesare also important to ensure constant adherence of BGMS to internationalstandards. However, such studies are not compulsory and there is no EU-wideindependent institution that evaluates BGMS quality. Manufacturers themselvesgenerally do these studies. But variability between test strip batches may affectmeasurementquality [45]. Individual test strip lots fromEUmarketedBGMShavebeenreportedtonotfulfilminimumsystemaccuracycriteria[46][47].
It is important to point out that an analytically accurate meter does notassureoptimalperformanceindailyuse.User’shandlingproficiencyisabigpartofthe measurement. In a standardized laboratory study, performed by trainedpersonnel,wherethereisacontrolledenvironmentsetforoptimalperformanceandwhere interferences can be reduced to a minimum, analytical accuracy can beevaluated exclusively. However, these conditions do not reflect patients’ dailymeasurements where sampling conditions and device handling may influencemeasurement results. Figure 1.4 summarizes conditions in daily life that mayinfluencesystemaccuracy.
Usererrors
Amongthemostfrequentusererrorsarefailingtocleansitebeforedrawingthesampleandincorrectteststripandmeterhandling[48].
An unclean hand containing traces of glucose-containing products cansubstantially increase glucose concentration values [49] [50]. More so withmicrosamplemeterssincesmallamountsofcontaminatingsubstanceshavegreaterinfluencewhensamplevolumeissmaller[38].Inastudywheresubjects’fingershadbeen exposed to fruit previous tomeasurement showed that for 88%of subjects’valuesweremorethan10%highercomparedtocontrolmeasurementusingthefistdropofblood.Usingtheseconddropofblood improvedresults.However,11%ofpatientsstillobtainedvalues10%higherthancontrol[51].Onlywashinghandswith
22
soap and water and drying them provided satisfactory results. Traces of othersubstances, forexamplehydrogenperoxide, found inhandsanitizersandascorbicacid,foundinfruit,mayactasareducingagentsandaffectmeasurementsaswell.
IncorrectteststriphandlingTeststripsduetotheircomplexenzymaticreactionareaverysensitivepart
of BGM. Its lifetime is about 2 years when stored in appropriate humidity andtemperature conditions, as per manufacturer recommendations, for optimalperformance [38]. A study comparing performance between test strips of openversus closed vials in different conditions of humidity, temperature and lightexposure concluded that test strips fromopen vials deteriorated faster than teststrips in closed vials. Even test strips from closed vials stored incorrectly underdirectlightorhighhumiditydidnotremain“analyticallystable”lasting28ofthe50study days [39]. Patients should be aware of and respect manufacturer’s storagerecommendations.
Figure1.4:Sourcesoferrorinroutinebloodglucosetesting.Reprintedfrom[53].
23
Mechanicalstressappliedtotheteststripandteststripreusehavealsobeenreported as commonmistakes patientsmake [52][53]. The later is, inmost cases,owedtoteststripcost.
Incorrectmeterhandling
When performing the measurement some aspects need to be taken inconsideration. Sample evaporation, with consequent elevation of glucose, andcontactwith oxygen can alter sample composition. Correct sample application onthe test stripandswiftmeasurement reduce thoseerrors.These simplemeasurescanbeimprovedwithtrainingandmakeobtainingreliableresultseasier.
Alternatesitetestingfromforearm,palm,thighandearlobeisavailablewithsomemeters.Theseoptionsaredescribedtobelesspainfulandsomoreappealingfor frequent testing.Howeverdue todifferences in skinblood flow,bloodglucoseresults can have a lag time especially during rapid glucose excursions delayinghyper-orhypoglycemiadetection[54].
Adding up meter’s technical limitations and user errors can make adifference in the final result. It is both important to have an analytically accuratesystemandcorrectmeasurementtechnique.
To minimize user error SMBG manufacturers must provide clear labelingwith instructions for use, understandable to patients, so that goodmeasurementscanbeobtainedwithoutfurthertraining.Evidentlypatientsshouldtakethetimetoreadthemcarefullyandadheretothem.AtthetimeoffirstSMBGcontacttrainingshouldalsobeprovidedby theHCPwhoshouldalsobe familiarwith instructionsfor use. Another strategy, adopted bymanufacturers, is to reduce the number ofoperator-dependantineachmeasurementand,consequently,diminishingmistakeslikelihood.Forexamplewhen test stripcodingwasabsent,user relatederrorwaslowercomparedwithmetersthatstillemployedthattechnology[55].
24
When using blood glucose meters patients should be aware of itsperformance under different environmental conditions, know meters’ limits,technical limitations and possible interferences. Knowing these users can takemeasurestomitigateerrorsandmakeaconsciousinterpretationoftheresultsgivenby themeter.Frequent testingwithcontrolsolutioncanhelpestablish ifdevice isperformingcorrectly.
At home a patient has to rely on values given by BGM systems to make
appropriate therapeutic decisions. Blood glucose values,measuredwith handheldBGMs,areasourceof information tooptimizeglycemiccontrolandpreventacutechroniccomplicationsofdiabetes.Inmanagingdiabetesthemeasuredvalueistakenasthe“true”glucoseconcentrationandtreatmentdecisionsaremadebasedonthatsupposition.Preprandial glycemia for example is akeyvalue to calculateprandialinsulin dose. This dosewill determine postprandial glycemic excursions and overtimeglobalmetaboliccontrol[56].
Resultspositivelybiasedcanpreventpatients fromdetectinghypoglycemiaorleadtotheadministrationofanexcessiveamountofanti-hyperglycemicdrugs.Apatientwhoseglucoseconcentrationisalreadylowcanhaveaseverehypoglycemicepisode if low glycemic levels are undetected. On the other hand, if a meter isnegatively biased hyperglycemia can remain concealed allowing chronicallyelevated BG levels associated with risk of developing diabetes-relatedcomplications.Evenwithinaccuracylimitsresultsmayvarysubstantially.Accordingto revised ISO of 2013 standards the cut-off is ±15 mg/dL for glucose levels of<100mg/dLand±15%forglucoselevelsof≥100mg/dL.Thismeansthataresultof60mg/dL, for example, couldbebetween45mg/dLand75mg/dL. ADA sets analertvalueforhypoglycemiaatplasmaglucoseconcentrationof≤70mg/dL,this,ofcourse, varies between patients and may shift depending on patients’ glucoseconcentrationhistory.FollowingADAhypoglycemiccutoff,theformerresultwouldbeidentifiedashypoglycemicwhereasthelatterwouldn’t.Fearofhypoglycemiaisoneofthemostthreateningfactorsinpatients’perspectiveandabarrierforinsulintherapy adherence [56]. A hypoglycemic event may prompt confusion, loss of
25
consciousness and, at the extreme, death. Therefore, it is a concern to accuratelymeasureBGinlowglycemicranges.
On the other side of the scale a value of 400 mg/dL could be reportedbetween340mg/dLand460mg/dLpromptingdifferenttherapeuticdecisions[48].
OnewayofmeasuringtheclinicaleffectofbasingtreatmentdecisionsonBGmeasurements isbyerrorgridanalysis. InthisqualitativeapproachBGMmeasurevalues are comparedwith adesignated comparisonmethodandplottedona griddividedby5zonesaccordingtoclinicaloutcomes(Figure1.5).ZonesrangefromA(no effect on clinical action) to E (Altered clinical action, could have dangerousconsequences). Table 1.3 describes potential clinical actions of errors falling inzonesA,B,C,DandE.Thisassessment reflectsexpertopinionbasedonevidenceavailable previous to 1994 and may be obsolete with new analytical accuracycriteria, insulin types and clinical practice. It is however a good tool to classifyoutlierdatapointsfromBGMaccordingtotheseriousnessofalteredclinicalaction.
Figure1.5:ParkserrorgriddevelopedforT1DM.
26
Table1.3:Definitionsoftheerrorgridzones
Risklevel(CEGzone)
Risktodiabeticpatient
A NoeffectonclinicalactionB Alteredclinicalaction–littleornoeffectonclinicaloutcomeC Alteredclinicalaction–likelytoaffectclinicaloutcomeD Alteredclinicalaction–couldhavesignificantmedicalriskE Alteredclinicalaction–couldhavedangerousconsequences
WhenlookingataccuracyoneshouldhaveinmindthepurposeSMBG.SMBGisrecommendedtodetermineinsulindoses,toachieveglycemicgoalsanddetectionofhypoglycemia[57].Howrelevanterrorsareandwhichlevelofaccuracyisneededwilldependonwhatwayvaluesareusedforintherapeuticdecisions.
A diverse population of patients with varied therapeutic regimens anddifferent diabetes types and acuteness uses BGMs. Evidently they have differentneeds in terms of SMBG system accuracy. For patients in insulin therapy earlydetection of hypoglycemia is essential. These patients need especially accuracy inthehypoglycemicrange.Patientsinintensifiedinsulinandinsulinpumptherapyareadvisedtocheckbloodglucosebeforemealsandexercise,atbedtime,sporadicallyposprandially,whenhypoglycemia is suspectedamongothers [8].Accuracyacrossallglucoserangesisexpected.
NotonlytheBGvalueisimportantininsulindosing,alsocarbohydrateintakeestimation,accuracyofinsulinpens,differencesininsulin’smetaboliceffect,amountofinsulinstillactivefrompreviousdosesandsubsequentphysicalexercise.Allthesefactors affect postprandial glycemic excursions. Nevertheless, an error in the firststep, i.e. BG value, will be amplified by possible subsequent errors or variations[56][58].
Calibratingmethodalsoplaysanimportantroleinmetersaccuracy.Thereisnotastandardizedmethodtocalibratemeters.Manufacturersmaychoosethemostconvenientone.Itistypicallybasedoneitherhexokinaseorglucoseoxidase.StudieshaveshowedmethodologicaldifferencesbetweenlaboratorymethodsusingHKand
27
GOD of about 8% [59]. This discrepancy yields different results with meterscalibrated with either method. Nevertheless, many BGMS exhibit largermeasurement error than the bias between HK and GOD based methods. ThuslaboratorymethodplaysavitalroleonlyinhighqualityBGMS.
Hematocritwasincludedintheinterferingsubstancestobetested.
New analytical accuracy criteria were motivated by an understanding ofSMBG importance in supporting diabetes management. Lifestyle and therapeuticdecisions are based on BGM values. Many diabetes patients have hypoglycemiaunawareness and SMBG is the only practical means for detecting asymptomatichypoglycemia. Additionally, BGM technology improvements, since publication offirstedition,ensuredthatmanufacturerscouldcomplywithstrictercriteria.
BGMSarecompared to laboratorydevicesbut there isnotoneharmonizedmethod for accuracy evaluation. Typically, laboratory devices are based either onthehexokinaseorglucoseoxidasemethod.
Laboratorydevices forcomparativeaccuracystudies(following ISO15197)needtocomplywithtraceabilityrequirementsestablishedinISO17511.Figure1.6showsoneexampleofatraceabilitychainforglucosemeasurementinbodyfluids.Atraceabilitychainusuallystartswiththedefinitionofameasurementunit,followedbyaprimarymeasurementmethodandaprimarycalibratormaterial.Forglucose,theacceptedprimarymethod is ID-GC/MS(IsotopeDilutionGasChromatography-Mass Spectrometry), but this does not exclude other methods. The primarycalibratormaterialcanbe,forexample,StandardReferenceMaterial917c(glucosepowder)fromtheUS-basedNationalInstituteforStandardsandTechnology(NISTSRM917c)orNISTSRM965b(glucose inhumanserum).Thetraceabilitychain isthencomposedofsecondary(andsuccessive)measurementmethodsandsecondary(andsuccessive)calibratormaterials.
28
Figure1.6:Traceabilitychainforthemeasurementofglucoseinbodyfluids.Reprintedfrom[60]
ID-GC/MS is considered the most accurate reference method for glucose.HoweveronlyafewlaboratoriesperformthismethodandmanufacturerscalibrateBGMS to alternative methods that are easier to handle. These methods should,nonetheless,haveID-GC/MSintheirtraceabilitychain.
An internationally recognized reference method would improvecomparabilityofresultsandhelpinterpretperformanceofBGMS.
29
2 IntroductiontotheProblem
Severalstudieshavedemonstratedconsiderablevariations inmeasurement
quality between BGM systems. Even amongmarketed systems some studies have
reportedBGMsnotcompliantwithISOsystemaccuracycriteria[61][46].Sincethe
revisionofISO15197in2013,morestringentrequirementswereproposed.Aftera
transitioningperiodof3years,in2016,inordertobemarketedinEU,allsystems
willhavetoprovecompliancetothenewinternationalstandard.
Systemaccuracyof10marketavailableBGMSwasevaluatedinalaboratory
study conducted byDr. Guido Freckmann at the Institut fürDiabetes-Technologie
Forschungs- und Entwicklungsgesellschaft mbH an der Universität Ulm (IDT),
Germany. Following ISO 15197:2013, which allows any method conforming to
establishedtraceabilityrequirementstobeusedforreferencemeasurement,system
accuracy was assessed against two laboratory methods based on Hexokinase or
GlucoseOxidase.
For one system considerable number of values were measured with test
stripsfromtwovials.Elevatedresultsfromcontrolsolutionmeasurementswithtest
strips from these two vials were detected during the study. Manufacturer’s
investigationattributedthis topossiblevialexposuretomoisture.Forthisreason,
even though ten systems were evaluated, only results with respect to 9 systems
weredescribed.
In the study, 7 systems showed compliance with system accuracy
specifications.Morethan95%ofresultswerewithinaccuracylimits,withall3lots
tested, independentof thecomparisonmethodused.Twosystemsdidnot comply
with minimum accuracy requirements irrespective of comparison method used.
Lookingatindividuallotsofthesetwosystems,90%to94%ofresultswerewithin
the established system accuracy limits when evaluated against the manufacturer
comparison method (GOD) and 84% to 99% if compared with the alternative
laboratorymethod(HK).
30
Regarding clinical accuracy all systems showed 100% of results within
consensuserrorgridzonesAandB.
Relativebiasofindividuallotstoeachlaboratorymethodwasalsocalculated.
A difference in relative bias between lots of each system bigger than 5% (5.4%;
7.6%)wasfoundin2systems.
Previousstudieshavereporteddifferencesbetweenlaboratorymethods[59].
Yet, few studies have weighted analytical accuracy results against two different
laboratory comparison methods. Now that any traceable laboratory reference
methodcanbeusedincomparisonaccuracystudies, independentofmanufacturer
calibrationmethod, the questionwhether calibrationmethod can have impact on
BGM accuracy is relevant. Variations were observed, in the referred laboratory
study, between results obtained with both laboratory methods. Differences were
reflectedinasystematicmeandifferenceofapproximately3%,indicatingthatthese
differences may have “considerable impact” on results from system accuracy
studies.
ThevariabilityfoundinmarketedBGMsinEuropebringstheissueofclinical
implicationsofusingpoorqualityBGM.
BoydandBruns’simulationmodeloftheuseofBGMtoadjustinsulindoses
weightedinsulindoseerrorsagainstdeviceimprecisionandbias.Variabilityof10%
ledto16%to45%ofincorrectinsulindoses.Toachieve95%correctinsulindoses,
precisionandbiaswouldhavetobelessthan2%.Anothersimulationevaluatedthe
percentageofinsulinerrorswithvariousBGMs,fordifferentdegreesofbias(-1.35%
–4%)andimprecision(CV:4.84%–7%).Insulindoseswhereontargetin64%to
82%ofcases.
Correct glycemic measurement and insulin dosing help achieve glycemic
targetswithsmalleramountofglycemicexcursions.Thiscanbetranslatedinbetter
metaboliccontrolandlessdiabetesrelatedcomplications.SMBGistheprimarytool
toachievetheseaims.Yet,itisagreedthatdifferentpatientgroupsrequiredifferent
levels of BGM accuracy within the clinically relevant BG ranges, hypoglycemia,
euglycemia, and hyperglycemia [62]. Among those patient groups are type 2
31
diabetes patients, type 1 diabetes on intensified insulin therapy and patients in
intensivecare.
PatientsfollowinganintensiveinsulinregimenarerecommendedtotestBG
levelsatleast6timesaday[8].BGMaccuracyforthesepatientsisimportantacross
allglucoseconcentrationrangesbutspecificallyinthehypoglycemicrange.Type2
diabetes mellitus patients do not necessarily need accuracy in low ranges. A
distinctionbetweenBGMperformancesisessentialtohelppatientsandHCPchoose
systemsthatwillofferbetterperformance,accordingtoclinicalneed.
There is concern whether BGM performance can be a source of insulin
errors even with meters compliant with accuracy criteria. In addition, a
categorizationofBGMaccordingtoaccuracyinseveralglucoseranges(low,normal
andhigh)isdesirable.Withthatintentinsulin,dosingerrorswerecalculatedbased
ondatafromBGMsmeasurements.
InordertoevaluateBGMSperformanceindifferentglycemiclevels,errors
separatedbyglucoserangeswereconsidered.
33
3 Methods
A laboratory study was performed at IDT under principal investigator Dr.
GuidoFreckmannwithtrialregistrynumber(ClinicalTrials.gov):NCT01909687.
Study duration was between 03.07.2013 to 22.10.2013. 164 different
subjects entered the study. Subjects were type 1 or 2 diabetes mellitus or non-
diabetic.Previoustothestudy,eachpatientwascheckedforinterferingsubstances
as well as inclusion and exclusion criteria p.e. pregnancy, lactation period, acute
disease,chronicdisease.
Investigated systems were: Accu-Chek® Aviva, Contour
® XT, FreeStyle
InsuLinx,Contour®nextUSB,BGStar
®,OneTouch
®Verio
®IQ,Accu-Chek
®Performa,
mylifeTMPura
TM,Glucocard
TMG+andMyStar
TMExtra. Itwasused for each system
three test strip lots. All test strips lots and 9 of the investigated systems were
purchasedatpharmacies.MyStarExtra®wasnotfreelyavailableinGermanyandso
itwasobtainedfromthedistributor(Sanofi-AventisDeutschlandGmbH,Germany).
Controlmeasurementsweredoneeachdaytoensuresystemsproperfunctioning.
Twolaboratorymethodswereusedtoperformcomparisonmeasurement:a
glucose oxidase based method (YSI 2300 STAT Plus™ glucose analyzer, YSI
Incorporated, Yellow Springs, OH, USA) and a Hexokinase based method (Cobas
Integra®400plus,RocheInstrumentCenter,Rotkreuz,Switzerland)referredtoas
YSIandCobasrespectively.TraceabilityrequirementsaccordingtoISO17511were
confirmedbythemanufacturers.
Studywasperformedinthreeparts: inthefirstandsecondpart3different
systemsineachpartwereevaluatedchangingtheorderoftheinvestigatedsystems
every 1/3 of subjects. In the third, 4 systemswere assessed. Orderwas changed
every¼ofsubjects.
Temperatureatwhichmeasurementsweremadewas23°C±5°C.Humidity
wasmaintainedbetween37,5%–57,8%. Samplehematocrit had tobebetween
20%-60%andglucoseconcentrationmeasuredwithlaboratorymethodhadtobe
34
within systems’ measurement rage 10 mg/dL – 600mg/dL (or according to
manufacturerspecifications).
Procedures specified in international standard ISO 15197:2013 were
followedforsystemaccuracyevaluation.
Procedure
Each patientwas asked towash handswith soap andwater and dry them
beforesamplecollection.Sampleswerecollected fromfingertipcapillariesbyskin
puncture.Individualsamplesweretestedinduplicateforeachsystemlotusingtest
strips from the same vial and two devices. To ensure that test strips from 10
differentvialswereused,vialswerechangedevery10subjects,approximately.
Figure3.1:Testingsequence
Aliquotswere collected fromeach sample immediatelybefore the first and
immediately after the last measurement with up to 4 systems for duplicate
measurementwiththecomparisonmethods.
Samplestabilitywasconfirmedbycheckingthatthedifferencebetweenthe
firstaliquotandsecondwas≤4mg/dLatBGconcentrations≤100mg/dLand≤4%at
BGconcentrations>100mg/dL.
Eachsamplewasallocatedtoabinaccordingtoglucoseconcentrationmean
valuemeasuredwiththerespectivereferencemethodasspecifiedinTable3section
6.3.5ofISO15197anddescribedinTable3.1.
Skin
puncture
1streference
value(both
lab
methods)
System1
bothdevices
System2
bothdevices
System3
bothdevices
2nd
reference
value(both
labmethods
35
Table3.1:BloodglucoseconcentrationofsamplesaccordingtoISO15197:2013[1]
Bin#
Percentageofsamples[%]
Glucoseconcentration[mg/dL]
15 ≤50
215 >50–80
320 >80–120
430 >120–200
520 >200–300
615 >300–400
75 >400
Toattainthedefineddistributionofglucoseconcentrationsonlyintherange
<50 mg/dL and >400 mg/dL samples can be adjusted. Sample adjustment was
performed by incubation to allow glycolysis to take place or by glucose
supplementationwithastocksolutionof40%glucosein0,9%NaCl.
At least 100 fresh capillary blood samples were collected and prepared
according to device manufacturer’ instructions. This ensured 200 data points for
eachsystemlotwitchis600datapointsperBGMS.
Insulindoseerror
Insulin doses were calculated based on BGMSmeasurements. A simplified
modelofpre-mealinsulindosewasused.
Beforemealsinsulindosesneedtocovercarbohydrateintakeandcorrectfor
high blood sugar. Insulin dose to cover carbohydrates is calculated taking meal
carbohydratecontentdividedbyinsulin-to-carbohydrateratio.Tocorrectglycemic
levelsdifferencebetweenmeasuredBGandglycemicgloaldividedbytheirInsulin
Sensitivity Factor (ISF) is used. Carbohydrate intake was excluded from the
proposed model in order to relate BGM’s values to insulin doses. Carbohydrate
intakeandcarbohydrateestimationerrorwasconsequentlydismissed.
36
True valueof glucose concentration is not accessibledue to inherent error
when using a laboratory method. Results from the manufacturer designated
comparisonmethodwereassumedtrueandinsulindosescalculatedbasedonthose
resultswereconsideredcorrect.
Intendedinsulindoseswerethencalculatedusingthefollowingequation:
(3.1)
BGreference is the blood glucose concentration measured with a
laboratory/reference method; BGgoal is the target value for blood glucose after
insulinadministration;ISFistheinsulinsensitivityfactor.IfwesubstituteBGreference
byBGBGM,bloodglucosemeasuredwithaBGM(eq.2.2),wethenhaveinsulindose
relatedtoBGMvalues,IDBGM.
(3.2)
Thisisconsideredthedoseapatientwouldinject.
ErrorininsulindoseisthenthedifferencebetweenIDreferenceandIDBGM.
(3.3)
Wherethesignaloftheequation,plus(+)orminus(–),translatestooveror
underdosing,respectively.
IDreferenceandIDBGMwereroundedtothenearest0,5IUgiventhattypicalfor
insulinpensmeasurein0,5IUincrements.
€
IDreference =BGreference − BGgoal
ISF
€
IDBGM =BGBGM − BGgoal
ISF
€
Error = IDreference − IDBGM
37
For each system and each test trip lot insulin dose errorswere calculated,
takingbothlaboratorymethodsasreference,withtheobjectiveofdeterminingthe
percentageofinsulindoseswithin0,5IUoftheintendeddose.Percentageofinsulin
above1IUand2IUwasalsodetermined.Percentageofunderdosingwascalculated
fordoses2IUormorebelowintendeddose.
It was checked, as well, percentages of target insulin doses, overdose and
underdosebasedonlyonunadulteratedsamples.
Furthermore, dose errors were calculated per glycemic level following
glucose concentration distribution of Table 3.1 to assesswitchmeters performed
better for each glycemic range. In this assessment only the manufacturer
comparisonmethodwasusedtocalculateintendedinsulindoseforeachsystem.
39
4 Results
PleaseseeAppendixAthroughE.
41
5 DiscussionandConclusion
Theaimofthisthesisworkwastoassesserrorsininsulindosingthatmarket
availableBGMsystemsmightcauseandalsotoinvestigatewhichsystemsmightbe
moreappropriatetotype1ortype2diabetes.
5.1 Significanceofthisstudy
Patients suffering from diabetes and on insulin therapy depend on blood
glucosemonitoringsystemsdaily.Glucoseconcentration is themainparameter to
calculate insulin doses. Patient’s metabolic control is linked with correct insulin
levelinthebloodstreamandthereforewiththecorrectdosageinjected.Themajor
risk of incorrect under dosage is hypoglycemia. This can be a consequence of
incorrectly high BG value. Symptoms of hypoglycemia are confusion, anxiety,
palpitations and others. Severe hypoglycemiamay cause coma and death. Insulin
overdosing,ontheotherhand,preventstheachievementofglycemicgoalsandisan
impedimenttothedelayofmicroandmacrovascularcomplications.Thislimitsthe
years of healthy life. Also the cost in healthcare increases substantially due to
diabetesrelatedcomplications[63]. It is, therefore, inthebest interestofHCPand
patients to achieve balancedmetabolic control. For these reasons, it is important
that BGMsmeasureBG concentration accurately for correct insulin dose and that
patientsusethemappropriately.
5.2 Methodology
ThisworkestablishesarelationshipbetweenBGMperformanceandinsulin
dosingerrorswithtwolaboratorymethods(HKandGOD)asreference.
42
Inthisstudy,itwasobtaineddataregardingglucoseconcentrationfrom164
patients.Allpatientswere testedwith10systems twice (twodevicesper system)
and using 3 different test strip lots. At least 100 values per device per lot were
obtained,inconformitywithISOspecifications.Intheend,600differentdatapoints
per BGM system were generated. In addition, samples were measured with two
laboratorymethodstobeusedasreference.
In order to determine the accuracy of devices under study, BGM
measurements were compared with reference values. In the study led by Dr.
Freckman, two systems did not fulfil with ISO 15197: 2013 system accuracy
(OneTouch Verio IQ and GlucoCard G+). Those systems showed less than 95% of
results within stipulated criteria. In fact, GlucoCard G+ did not fulfill system
accuracy criteria with none of the 3 tested lots, irrespective of the laboratory
referencemethodused [64]. ContourNext showed100%of resultswithin system
accuracycriteriairrespectiveofcomparisonmethod.
Insulindoseswerecalculated fromtheoriginal600BGdatapoints. Itwere
alsocalculatedinsulindosesfromtwolaboratoryreferencemethods.Results from
BGMinsulindoseswerethencomparedwithresultsfrombothlaboratorymethods
todetermineinsulindosingerrors
5.3 Summaryofresults
In the present study, it was system 3 that showed on target insulin doses
closesttotherespectivemanufacturerreferencemethod.Bothsystemsthatdidnot
fulfil accuracy criteria were among the ones with lowest percentage of on target
insulinalongwithsystem2.
Overdoseequalorabove2Uwasrelativelylow,notsurpassing5%.Overall
thestudyshowedthatthe10BGMsunderstudygiveBGconcentrationvalueswith
limitedimpactininsulindoses.
ThisstudyrevealeddifferencesininsulindoseerrorwithinBGconcentration
rangesallowingtodistinguishbetweenBGMssuitablefordiabetestype1ortype2.
43
It can be argued that type 1 patients need small error across all ranges and
especially in the hypoglycemic range. On the contrary, type 2 patients do not
necessarilyneedthesameinthelowglycemicrange.
System5,system3andsystem7showedinsulindoseerrorsnotsuperiorto
1.5 U across all ranges and 100% of on target insulin doses in the hypoglycemic
range.
Inparticularcases,suchaspatientstreatedwithintensiveinsulintherapyor
insulinpump,moreaccuratemeasuringdevicesareneeded.Inthesegroups,insulin
doseerrorsof1Ucanhaveamorepronouncedimpactthaninotherpatients.Inthis
study,overdoseby1Uormorewasashighas12.5%(YSIreference).
Alsothethresholdof1Uwasstudiedbecauseofthesignificantphysicaland
psychological consequences of hypoglycemia. A benchmark for diabetes therapy
assessment is HbA1c, established by the publication of the Diabetes Control and
Complications (DCCT). The study showed a direct relationship between glycemic
controlandthedevelopmentofmicroandmacrovascularcomplications.Itshowed
aswell that intensively treated patients reachedHbA1c levels faster and vascular
complicationsweredelayed.This,however,camewithhigherrateofhypoglycemic
events.As itwas referredpreviously,new insulin formulationshave lower riskof
hypoglycemia.
When lookingatresults fromunadulteratedsamples,errorsare lower than
comparedwithall samples.This is theconsequenceofexcludingsampleswithBG
concentrationabove400mg/dLwhereingeneralBGMsarelessaccurate.
5.4 Limitationsofthestudy
Thereweresomelimitationsinthedevelopmentofthisstudy,whichshould
be pointed out. These limitations refer to the study methods and to the
simplificationofthedoseerrormodel.Alaboratorystudydoesnotmimiceveryday
conditions inwhichpatientscarryoutthemeasurements.Asdiscussed,usererror
plays a significant part in the quality of results, as do ambient conditions. It is
44
nonetheless agreed that errors derived by poor analytical performance will be
propagatedwithsubsequentinterferingfactors.
Another limitation is knowing the exact value of glucose concentration.
Inherent to the measurement of a biological substance is the impossibility of
knowingitstruevalue.Inthisstudy,BGMresultswerecomparedtotwolaboratory
methodsandtherespectiveresultswereassumedasthetruevalue.
Finally,thelastlimitationinthisstudywasthesimplificationoferrormodel.
Insulin dose calculation was based on the equation used by patients to calculate
mealtimeinsulindoses.Thisequationincludesacarbohydratecomponenttocover
carbohydrateintake.ThiscomponentwasdisregardedtolinkBGMperformanceto
insulindose.
In conclusion, accuracy of BGM systems must be guaranteed before it is
availabletocustomersaswellasinpost-marketingstages.HCPandpatientsshould
beinformedabouttechnicalaspectsofBGMsanditsperformanceinordertomake
informedchoices regarding itsuse. Special consideration shouldbegiven in cases
whereBGMisusedtocalculateinsulindosesandifitwillbeusedbytype1ortype2
diabetespatients.
5.5 Futureoutlook
Diabetes is a chronic disease, which requires constant management and is
associated with several complications such as retinopathy, nephropathy,
neuropathyandcardiovasculardisease.Reducedpatientcomplianceorinadequate
treatmentregimesmaycontributetoafasterdeteriorationofpatient’shealth.Upon
diagnosis, patientsmust frequently visit thedoctor in order to reach an adequate
treatmentplan.Periodicappointmentsfollowthisinitialphaseapproximatelyevery
6months.Duringthoseperiods,patients followasetofwrittenrulesgivenbythe
doctortoreachormaintainglycemictargetandmanageglycemiclevels.
Patients with diabetes mellitus are faced with a daily cumbersome routine.
Theymustpayattentiontowhatandhowmuchtoeat,whatdoseofinsulintotake
45
and howmuch exercise to practice. Patients need help and information to make
thosedecisions.Butnotonlypatientsstruggle.Physiciansoftenhaveashortamount
of time with each patient and need to analyze all the information that a patient
bringstoconsultsandreevaluatetreatmentifnecessary.Thereisanapparentneed
forclearandaccessibleinformation.
Tools that enable easier contact between patient and HCP can improve
treatmentadjustmentandassistinmakingpromptalterationswhennecessary.
Manyappsareavailablewhichmostlyoffersimplefunctionalities.Themain
benefittousersofsuchappsforregulatingbloodglucoselevelsistheavailabilityof
quickinformationwhichgivespatientsgreaterlevelofautonomy.
One solution comes in the form of software for computers, smartphones or
tablets, capable of receiving and storing patients’ necessary data (carbohydrate
intake, insulindose,FPG,BG,HbA1c)anddocalculationsinordertogivefeedback
on the development of a given parameter or treatment suggestions. Apps for cell
phones have increasingly been promoted for patient use in day-to-day care.
Ultimately, patients are able to make informed lifestyle choices and at the same
time, if treatment is effective, see improvements in metabolic control. The visual
feedback provided by these apps acts as an incentive to behavioral changes. If a
patientisabletoseehisdailyachievementsandthateachstepbringshimcloserto
goal,hewillbemoremotivatedtocontinuetreatment.
Storingdata is not the only important component of these apps. Sharing it
withtherightpartiesalsoplaysasignificantrole inmanagingthepatients’health.
Forexample,HCPshouldhaveaccess toselected information,ensuringbetterand
safer care. Stored and shared data of blood glucose levels, insulin doses,
hypoglycemic events and pertinent graphs can give HCP a fast way to evaluate
treatmentevolutionandeventuallyactremotelyifsomevalueiscauseofconcern.
Likewise,themarketofSMBGhasbeenevolvingtobringnewfunctionalitiesto
the user, in the form of smarter BGMs that can store BG values, calculate insulin
46
dosesandstoretimeandamountofpreviousdoses.
It is important to note that these new advances inmanaging diabetes give
freedom to the patient with lessmedical interaction. These tools should be used
withcautioninparticulariftheyaredesignedtogivetreatmentsuggestions,asfor
exampleinsulindoses.HCPsshouldprescribewithcautionandonlytotrainedand
informedpatients.
Regardless of these newer Smartphone applications and the added
functionalities theyprovide, anaccuratemeasurementofbloodglucose is still the
most fundamental piece of information for a proper management of diabetes.
Therefore, BGMs’ system accuracy must be guaranteed to ensure safety for the
patient.
Studies have suggested that SMBG frequency as an important role in
metaboliccontrol(HbA1c)[65][66].Yet it is importanttoemphasizethatSMBGin
itself isnotan intervention. It isa tool toassist treatmentdecisionandultimately
achieve glycemic targets. Quality of measurement is nonetheless linked with
metabolicoutcomes.Qualityofmeasurementdoesnotreferexclusivelytoanalytical
accuracy, but its assurance is important if only to prevent propagation of error if
BGMismishandled.
AlthoughSMBGisparamounttoensureproperinsulindosage,literaturehas
shownthatthereareotherfactorswhichmaycontributetoerrors.Forpatientson
fixed insulin regimens, dose calculation is obviously not necessary. For these
patients,mealplanningwithcomparablecarbohydrateportionsismoreimportant.
Amoreadaptableregimen,however,maybepreferabletosomepatientswhowant
to have choice on type of meal and exercise. In such cases, education on insulin
calculationandcarbohydratecountingisextremelyimportant.
47
References[1] WorldHealthOrganization,"Globalstatusreportonnoncommunicablediseases,"Geneva,2014.
[2] InternationalDiabetesFederation.(2015)AnnualReport.[Online].
http://www.idf.org/sites/default/files/IDF_AnnualReport_2015_WEB.pdf
[3] P.BennetandW.Knowler,"Diabetes:Definition,Genetics,andPathogenesis,"inJoslin'sDiabetesMellitus,14thed.Philadelphia,Pa:LippincottWilliams&Wilkins,2005,pp.332-339.
[4] F.R.Kaufman,"Type2DiabetesinChildrenandYoungAdults:A“NewEpidemic”,"ClinicalDiabetes,vol.20,no.4,pp.217-218,Oct2002.
[5] ClinicadeDiabeteseNutrição,TópicosSobreDiabetes,13thed.,J.Caldeira,Ed.Lisboa,2009.
[6] I.B.Hirsch,R.M.Bergenstal,C.G.Parkin,E.WrightJr,andJ.B.Buse,"AReal-WorldApproachtoInsulin
TherapyinPrimaryCarePractice,"ClinDiabetes,vol.23,no.2,pp.78-86,Apr2005.
[7] M.J.Fowler,"MicrovascularandMacrovascularComplicationsofDiabetes,"ClinicalDiabetes,vol.26,no.2,pp.77-82,Apr2008.
[8] AmericanDiabetesAssociation,"StandarsofmedicalCareinDiabetes—2016,"DiabetesCare,vol.39,no.Suppl.1,pp.S1-S112,Jan2016.
[9] C.C.QuianzonandI.Cheikh,"HistoryofInsulin,"JCommunityHospInternMedPerspect,vol.2,no.2,Jul2012,doi:10.3402/jchimp.v2i2.18701.[Online].
http://www.jchimp.net/index.php/jchimp/article/view/18701
[10] U.S.FoodandDrugAdministration.[Online].
http://www.accessdata.fda.gov/scripts/cder/drugsatfda/index.cfm?fuseaction=Search.DrugDetails
[11] I.B.Hirsch,"Insulinanalogues.,"NEnglJMed,vol.352,no.2,pp.174-183,Jan2005.
[12] A.McGibbon,C.Richardson,C.Hernandez,andJ.Dornan,"CanadianDiabetesAssociation2013Clinical
PracticeGuidelinesforthePreventionandManagementofDiabetesinCanada:Pharmacotherapyin
Type1Diabetes,"CanJDiabetes,vol.27,no.Suppl.1,pp.S56-S60,Apr2013.
[13] SanofiBridgewater,NJ.(2015,February)Sanofi-AventisUS.[Online].
http://products.sanofi.us/toujeo/toujeo.pdf
[14] WHOInternationalLaboratoryforBiologicalStandards.(2010)NationalInstituteforBiological
StandardsandControl.[Online].https://www.nibsc.org/documents/ifu/83-500.pdf
[15] EuropeanMedicinesAgency.(2005,April)[Online].
http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/09/WC5000036
54.pdf
[16] M.J.LenhardandG.DReeves,"ContinuousSubcutaneousInsulinInfusionAComprehensiveReviewof
InsulinPumpTherapy,"ArchInternMed.,vol.161,no.19,pp.2293-2300,Oct2001.
[17] A.J.Garberetal.,"ConsensusStatementbytheAmericanAssociationofClinicalEndocrinologistsand
AmericanCollegeofEndocrinologyonTheComprehensiveType2DiabetesManagementAlgorithm–
2016ExecutiveSummary,"EndocrPract,vol.22,no.1,pp.84-113,Jan2016.
[18] AmericanDiabetesAssociation,"InsulinAdministration,"DiabetesCare,vol.26,no.Suppl.1,pp.S121-S124,2003.
[19] G.Scavoneetal.,"Effectofcarbohydratecountingandmedicalnutritionaltherapyonglycaemiccontrol
intype1diabeticsubjects:apilotstudy.,"DiabetMed.,vol.27,no.4,pp.477-479,Apr2010.
[20] P.D.Dworatzeketal.,"CanadianDiabetesAssociation2013ClinicalPracticeGuidelinesforthe
PreventionandManagementofDiabetesinCanada:NutritionTherapy,"CanJDiabetes,vol.37,no.Suppl.1,pp.S45-S55,Apr2013.
[21] R.J.Sigaletal.,"CanadianDiabetesAssociation2013ClinicalPracticeGuidelinesforthePreventionand
ManagementofDiabetesinCanada:PhysicalActivityandDiabetes,"CanJDiabtes,vol.37,no.Suppl.1,pp.S40-S44,Apr2013.
[22] AmericanDiabetesAssociation,"StandardsofMedicalCareinDiabetes-2015,"DiabetesCare,vol.38,no.Suppl.1,pp.S1-S93,Jan2015.
[23] DCCTResearchGroup,"Theeffectofintensivetreatmentofdiabetesonthedevelopmentand
progressionoflong-termcomplicationsininsulin-dependantdiabetesmellitus,"NEnglJMed,vol.329,
48
no.14,pp.977-86,1993.
[24] S.E.Inzucchi,R.M.Bergenstal,J.B.Buse,M.Diamant,andE.Ferrannini,"Managementofhyperglycaemia
intype2diabetes,2015:apatient-centredapproach.UpdatetoaPositionStatementoftheAmerican
DiabetesAssociationandtheEuropeanAssociationfortheStudyofDiabetes,"Diabetologia.,vol.58,no.49,pp.429-442,Mar2015.
[25] W.Harperetal.,"CanadianDiabetesAssociation2013ClinicalPracticeGuidelinesforthePrevention
andManagementofDiabetesinCanada:PharmacologicManagementofType2Diabetes.,"CanJDiabetes,vol.37,no.Suppl.1,pp.S61-S68,2013.
[26] L.D.Berard,I.IanBlumer,R.Houlden,D.Miller,andV.Woo,"CanadianDiabetesAssociation2013
ClinicalPracticeGuidelinesforthePreventionandManagementofDiabetesinCanada:Monitoring
GlycemicControl,"CanJDiabetes,vol.37,no.Suppl.1,pp.S35-S39,Apr2013.
[27] UKProspectiveDiabetesStudyGroup(UKPDS),"Intensivebloodglucosecontrolwithsulphonylureasor
insulincomparedwithconventionaltreatmentandriskofcomplicationsinpatientswithtype2diabetes
(UKPDS33),"TheLancet,vol.352,no.9131,pp.837-53,1998.
[28] S.F.ClarkeandJ.R.Foster,"AHistoryOfBloodGlucoseMetersAndTheirRoleInSelf-MonitoringOf
DiabetesMellitus,"BrJBiomedSci,vol.69,no.2,pp.83-93,Mar2012.
[29] A.H.Clemens,"ReflectanceMeter,"U.S.Patent3,604,815,September14,1971.
[30] L.C.ClarkandC.Lyons,"ElectrodeSystemsForContinuousMonitoringInCardiovascularSurgery,"AnnNYAcadSci,vol.102,no.1,pp.29-49,1962.
[31] K.O.Kyvik,J.Traulsen,B.Reinholdt,andA.Frøland,"TheExacTechBloodGlucoseTestingSystem.,"
DiabetesResClinPract.,vol.10,no.1,pp.85-90,Aug-Sep1990.
[32] BHellerandA.Feldman,"ElectrochemicalGlucoseSensorsandTheirApplicationsinDiabetes
Management,"Chem.Rev.,vol.108,no.7,pp.2482-505,Jul2008.
[33] J.Wang,"Electrochemicalglucosebiosensors.,"ChemRev.,vol.108,no.2,pp.814-825,Feb2008.
[34] S.Ferri,K.Kojima,andK.Sode,"ReviewofGlucoseOxidasesandGlucoseDehydrogenases:ABird’sEye
ViewofGlucoseSensingEnzymes,"JDiabetesSciTechnol,vol.5,no.5,pp.1068-1076,Sep2011.
[35] FoodandDrugAdministration.(2009,Aug)FDAPublicHealthNotification:PotentiallyFatalErrorswith
GDH-PQQ*GlucoseMonitoringTechnology.[Online].http://www.fda.gov/MedicalDevices/Safety
/AlertsandNotices/PublicHealthNotifications/ucm176992.htm
[36] S.K.Vahist,D.Zheng,K.Al-Rubeaan,J.H.Luong,andF.S.Sheu,"Technologybehindcommercialdevices
forbloodglucosemonitoringindiabetesmanagement:areview.,"AnalChimActa,vol.703,no.2,pp.124-136,Oct2011.
[37] StremChemicalsInc.Hexaammineruthenium(III)ChlorideasanElectronMediatorforRapidWhole
BloodGlucoseDetection.[Online].http://www.strem.com/uploads/technical_notes/44-0620tech.pdf
[38] B.H.Ginsberg,"FactorsAffectingBloodGlucoseMonitoring:SourcesofErrorsinMeasurement,"JDiabetesSciTechnol,vol.3,no.4,pp.903-913,jul2009.
[39] R.Bamberg,K.Schulman,M.MacKenzie,J.Moore,andS.Olchesky,"Effectofadversestorageconditions
onperformanceofglucometerteststrips.,"ClinLabSci.,vol.18,no.4,pp.203-209,Fall2005.
[40] A.Pfützneretal.,"BloodGlucoseMetersEmployingDynamicElectrochemistryAreStableagainst
HematocritInterferenceinaLaboratorySetting,"JDiabetesSciTechnol.,vol.7,no.6,pp.1530-1537,Nov2013.
[41] A.Baumstark,C.Schmid,S.Pleus,C.Haug,andG.Freckmann,"InuenceofPartialPressureofOxygenin
BloodSamplesonMeasurementPerformanceinGlucose-Oxidase-BasedSystemsforSelf-Monitoringof
BloodGlucose,"JDiabetesSciTechnol.,vol.7,no.6,pp.1513-1521,Nov2013.
[42] "Directive98/79/ECoftheEuropeanParliamentandoftheCouncilof27October1998oninvitro
diagnosticmedicaldevices,"OfficialJournaloftheEuropeanCommunitiesL331,vol.41,December1998.
[43] InternationalOrganizationforStandardization,"Invitrodiagnostictestsystems—Requirementsfor
blood-glucosemonitoringsystemsforself-testinginmanagingdiabetesmellitus,"Geneva,International
StandardISO15197:2013(E),2013.
[44] H.W.VesperandG.L.Meyers,"ApproachesforImprovingGlucoseMonitorMeasurementsforSelf-
MonitoringofBloodGlucose:FromMeasurementHarmonizationtoExternalQualityAssessment
Programs,"JDiabetesSciTechnol,vol.1,no.2,pp.153-157,Mar2007.
49
[45] A.Baumstarketal.,"Lot-to-lotvariabilityofteststripsandaccuracyassessmentofsystemsforself-
monitoringofbloodglucoseaccordingtoISO15197.,"JDiabetesSciTechnol.,vol.6,no.5,pp.1076-1086,Sep2012.
[46] G.Freckmannetal.,"Systemaccuracyevaluationof27bloodglucosemonitoringsystemsaccordingto
DINENISO15197.,"DiabetesTechnolTher.,vol.12,no.3,pp.221-231,Mar2010.
[47] C.Tacketal.,"Accuracyevaluationoffivebloodglucosemonitoringsystemsobtainedfromthe
pharmacy:aEuropeanmulticenterstudywith453subjects.,"DiabetesTechnolTher.,vol.14,no.4,pp.1-8,Apr2011.
[48] K.TonyushkinaandJ.H.Nichols,"GlucoseMeters:AReviewofTechnicalChallengestoObtaining
AccurateResults,"JDiabetesSciTechnol,vol.3,no.4,pp.971-980,Jul2009.
[49] M.ArakawaandC.Ebato,"Influenceoffruitjuiceonfingertipsandpatientbehavioronself-monitoring
ofbloodglucose.,"DiabetesResClinPract.,vol.96,no.2,pp.e50-e52,May2012.
[50] T.Hirose,T.Mita,Y.Fujitani,R.Kawamori,andH.Watada,"Glucosemonitoringafterfruitpeeling:
pseudohyperglycemiawhenneglectinghandwashingbeforefingertipbloodsampling:washyourhands
withtapwaterbeforeyoucheckbloodglucoselevel.,"DiabetesCare,vol.34,no.3,pp.596-597,2011.
[51] J.Hortensiusetal.,"Self-monitoringofbloodglucose:theuseofthefirstortheseconddropofblood.,"
DiabetesCare,vol.34,no.3,pp.556-560,Mar2011.
[52] G.B.Kristensen,K.Nerhus,G.Thue,andS.Sandberg,"Standardizedevaluationofinstrumentsforself-
monitoringofbloodglucosebypatientsandatechnologist.,"ClinChem.,vol.50,no.6,pp.1068-1071,Jun2004.
[53] S.Skeie,G.Thue,K.Nerhus,andS.Sandberg,"Instrumentsforself-monitoringofbloodglucose:
comparisonsoftestingqualityachievedbypatientsandatechnician.,"ClinChem,vol.48,no.7,pp.994-1003,Jul2002.
[54] C.Schmid,C.Haug,L.Heinemann,andG.Freckmann,"SystemAccuracyofBloodGlucoseMonitoring
Systems:ImpactofUsebyPatientsandAmbientConditions,"DiabetesTechnolTher,vol.15,no.10,pp.889-896,Oct2013.
[55] A.Rao,M.Wiley,S.Iyengar,D.Nadeau,andJ.Carnevale,"IndividualsAchieveMoreAccurateResults
withMetersThatAreCodelessandEmployDynamicElectrochemistry,"JDiabetesSciTechnol.,vol.4,no.1,pp.145-150,Jan2010.
[56] L.Heinemann,V.Lodwig,andG.Freckmann,"AccuracyinBloodGlucoseMeasurement:WhatWilla
TighteningofRequirementsYield?,"JDiabetesSciTechnol,vol.6,no.2,pp.435-443,Mar2012.
[57] AmericanDiabetesAssociation,"StandardsofCareinDiabetes-2010,"DiabetesCare,vol.33,no.Suppl.1,pp.S11-S61,Jan2010.
[58] T.Koschinsky,S.Heckermann,andL.Heinemann,"Parametersaffectingpostprandialbloodglucose:
effectsofbloodglucosemeasurementerrors.,"JDiabetesSciTechnol.,vol.2,no.1,pp.58-66,Jan2008.
[59] P.J.Towmey,"PlasmaglucosemeasurementwiththeYellowSpringsGlucose2300STATandthe
OlympusAU640,"JClinPathol,vol.57,no.7,pp.752-754,Jul2004.
[60] R.I.Wielgosz.(2007)BureuInternationaldesPoindsetMesures.[Online].
http://www.bipm.org/ws/JCTLM/LAB-MED/Allowed/Documents/JCTLMSymp07-04.pdf
[61] G.Freckmannetal.,"SystemAccuracyEvaluationof43BloodGlucoseMonitoringSystemsforSelf-
MonitoringofBloodGlucoseaccordingtoDINENISO15197.,"JDiabetesSciTechnol.,vol.6,no.5,pp.1060-1075,Sep2012.
[62] D.C.Klonoff,"Regulatorycontroversiessurroundbloodglucosemonitoringdevices.,"JDiabetesSciTechnol.,vol.4,no.2,pp.231-235,Mar2010.
[63] R.Li,P.Zhang,L.E.Barker,F.M.Chowdhury,andX.Zhang,"Cost-effectivenessofinterventionsto
preventandcontroldiabetesmellitus:asystematicreview,"DiabetesCare.,vol.33,no.8,pp.1872-1894,Aug2010.
[64] G.Freckmannetal.,"SystemAccuracyEvaluationofDifferentBloodGlucoseMonitoringSystems
FollowingISO15197:2013byUsingTwoDifferentComparisonMethods.,"DiabetesTechnolTher.,vol.17,no.9,pp.635-648,Sep2015.
[65] W.H.Polonskyetal.,"Structuredself-monitoringofbloodglucosesignicantlyreducesA1Clevelsin
poorlycontrolled,noninsulin-treatedtype2diabetes:resultsfromtheStructuredTestingProgram
study,"DiabetesCare.,vol.34,no.2,pp.262-267,Feb2011.
50
[66] QuEDStudyGroup,"TheImpactofBloodGlucoseSelf-MonitoringonMetabolicControlandQualityof
LifeinType2DiabeticPatients,"DiabetesCare,vol.24,no.11,pp.1870-1877,Nov2001.
[67] J.Hönes,P.Müller,andN.Surridge,"TheTechnologyBehindGlucoseMeters:TestStrips,"DiabetesTechnolTher,vol.10,no.Suppl.1,pp.S10-S26,2008.
[68] I.B.Hirsch,"Glycaemicvariability:It'snotaboutA1Canymore!,"DiabetesTechnolTher,vol.7,no.5,pp.780-783,Oct2005.
[69] E.R.Seaquistetal.,"HypoglycemiaandDiabetes:AReportofaWorkgroupoftheAmericanDiabetes
AssociationandTheEndocrineSociety,"DiabetesCare,vol.36,no.5,pp.1384-1395,May2013.
51
Appendix
Appendixhasbeenremovedduetoconfidentiality