Specificity
Aspart of theSLV, a singleelement standardof eachanalytewas checked for apparent signal fromany
of theother 11 concomitant analytes. Table4 shows the results. Noneof the50mg/L standards
producedany significant (abovePQL) concentrationsof analytes, proving the specificityof themethod.
The internal standardswerenot checkedbecause theyarealwaysusedat a low concentrationof 50
µg/L.
Linearity
To check linearity, several standard solutionswereprepared independently from the calibration
standards, and runas samples against the current calibration. Thenormal calibration curveuses ablank
and four standards at 1x, 5x, 20x, and40x the concentrationof the lowworking standard (thehigh
standard is 40-foldhigher than the low standard). The linearity-check standardswere0.5x, 1x, 1.5x, 2.5x,
3x, 10x, 20x, 30x, and60x the concentrationof the lowworking standard. Thus, points50%below the
lowest calibration standardand50%above thehighest calibration standardwere checked for linearity.
Eachof these check standardswereanalyzed twiceonDay1, and thedataaveraged toproduceone
result. Theprocesswas repeatedwith fresh standardsonDay2andonDay3. The three-daymeans
were thenaveragedand reported inTable6. At a level of 50%of the low calibration standard, all
analytesdemonstrated sufficient agreement (95-105%,with rounding)with thenominal value. Thus,
50%of the low standard concentration is set as thePractical Limit ofQuantitation (PLOQ) of this
method,which is sufficient forour intendedworkand sample types.
LOQ
Table5 summarizes the standard concentrationsused in themethod, the solutionPLOQs (mg/L)
determined from linearityexperiments (above, Table6), thePLOQs converted toaweight basis
(mg/100g for a typical dilutionof 1.0gRTF to50mL), anda comparison to theSMPR requirements. As
themethod is set upnow, thePLOQsmeet SMPR requirements for all elementsexcept Fe, Cu, andMn.
TheSMPR requiresQLsof about 10ppb (byweight) forMnandCu inanRTF, andabout 100ppbFe
(about one tenthof the level of inherent iron inour experience). These levels area factorof 5-10 lower
thanany formulaAbbottNutrition currentlymakesand so the calibration rangeand linearity checks
wouldneed tobe redone tomeet these requirements. Lookingat the linearitydata inTable6, it is
reasonable to think that the curves couldbeextended lower and still haveadequate recoveryof these
threeelements. Also, sample sizesof 2-3gof RTF (limitedby safemicrowave conditions) couldbeused
with the currentmethod to improve the LOQ2-3x. The lowest concentrationsofMn, Cu, andFe found in
theSPIFANmatriceswas150ng/g, 580ng/gand14000ng/g, respectively, all in theSPIFAN controlMilk.
Conclusion
Themethod, as is,meets all SMPR requirements except for the LOQof Fe,Mn, andCu. Additional
linearitywork, spikingplacebos at low levels, increasing sample sizes, and/or additional low-level
standardswouldbeneeded toproveaccuracyat the requested10ppb levels.
MTE-02
FORWORKINGGROUP/ERPUSEONLY
DONOTDISTRIBUTE
