Combined Interpretations of the 2003, 2009, and 2016 Standards that apply to Volume 1 of the 2016 TNI Standard

MODULE 4: CHEMISTRY TECHNICAL REQUIREMENTS
Section: 1.7.1.1

Question:  ICVs have historically been treated as CCVs, in that high-biased failures and samples with no target compounds did not require reanalysis. It has been noted in a previous audit that this is not acceptable even though there would be qualifications on the final client report. Why would this practice not be acceptable?

TNI Response:  If the Initial Calibration Verification (ICV) does not pass the established acceptance criteria for the calibration technique used, the calibration curve is not valid and corrective actions must be documented and samples reanalyzed. An ICV is not is not the same as a Continued Calibration Verification (CCV) and should not be evaluated against the CCV acceptance criteria.

Question:  What input variables (analyte lot, solvent lot, balance, operator, etc.) must change in order for a second lot of standard to be considered to be prepared independently?

TNI Response:  5.5.5.2.2.1 d states that a standard obtained from a second lot may be used if the lot can be demonstrated from the manufacturer as prepared independently from other lots. In our mind, this is analogous to a preparation batch being prepared with the same process, personnel and reagents during a set timeframe. Assuming the manufacturer prepares a specific quantity of standard and declares that quantity a "lot", another lot prepared at another time would be independent of the first. The laboratory would be expected to contact the manufacturer to verify that one lot is not the same as another.  The 2003 and 2009 standards have identical language.  This section was slightly reworded in 2016 and a definition of "lot" was added to V1M2: Lot: A definite amount of material produced during a single manufacturing cycle, and intended to have uniform character and quality.  The SIR still has some relevance in further explaining this issue.

Question:  Does the requirement for second source standard include calibration curves for surrogate compounds?

TNI Response:  Surrogates are intended to provide a measure of recovery for every sample matrix (D.1.1.3.3 a). A second source check is designed to assure that the analytes of concern are being correctly identified and quantified. Since surrogates are not analytes of concern, and may be held at a constant level in a calibration curve, they are not required to be verified by a second source.  The 2003 and 2009 standards have identical language.  This section was slightly reworded in 2016 and a definition of "lot" was added to V1M2: Lot: A definite amount of material produced during a single manufacturing cycle, and intended to have uniform character and quality.  The SIR still has some relevance in further explaining this issue.

Question:  We have been informed that Environmental lab auditors are now requiring labs to use a second vendor for their second source calibration standards even when use of a second manufactured lot from the same vendor is allowed by both the 2003 NELAC and 2009 TNI standards.   We have even heard that second source from a second manufacturer raw material prepared by the same vendor is being noted as non-compliant in some instances.  What is the interpretation of what this standard requires of labs where it applies to non-DoD programs with regard to the following options:
1. Same vendor/supplier - two independently prepared lots from the same raw material.
2. Same vendor/supplier - two independently prepared lots from different manufacturer raw materials when available.
3. Two different vendor/suppliers for each of the primary and secondary lot standards.

TNI Response:  The purpose is to verify that the standards used for calibration have been properly prepared. The verification standard must be prepared independently from the calibration standard(s). The best option is standards from two different vendors; alternatively, standards from the same vendor but different lot numbers. While the original source (manufacturer) of the neat standard is important, the standard stresses independent preparation.  The 2003 and 2009 standards have identical language.  This section was slightly reworded in 2016 and a definition of "lot" was added to V1M2: Lot: A definite amount of material produced during a single manufacturing cycle, and intended to have uniform character and quality.  The SIR still has some relevance in further explaining this issue.

Question:  Section 1.7.1.1(f) refers to the minimum number of calibration standards required for each calibration type. Section 1.7.1.1(l) provides an exception for test procedures/methods that specify the allowance of a single calibration standard and zero point within the linear range.

Do the metals methods (i.e., EPA 200.7 and EPA 200.8) that specify the use of a calibration blank and a minimum of one or more calibration standards fall under section 1.7.1.1(l)? ICP methods are mentioned as examples for this clause in the TNI Calibration Guidance Document.

As specified within EPA 200.7 and EPA 200.8, more than one calibration standard is allowed, both methods just require a minimum of one calibration blank and a single calibration point. Linear regression is also acceptable in both methods.

For EPA 200.7 and EPA 200.8, is the lab allowed to analyze more than one calibration standard and still fall under the requirements in 1.7.1.1 (l) or, when more than one calibration standard is analyzed, does it default to the linear regression requirement in 1.7.1.1(f) which requires a minimum of five calibration standards? In other words, is it acceptable to use 2-4 calibration standards for EPA Methods 200.7 and 200.8 and maintain compliance with the 2016 TNI Standards?

TNI Response:  When test procedures specify and the laboratory chooses to utilize the single calibration point and a zero point, then they fall under the requirements of 1.7.1.1 (l). However, when test procedures specify, or the laboratory chooses to utilize more standards that would then fall under the restrictions of 1.7.1.1(f).

Question:  The 2016 TNI Standard states that for regression or average response/calibration factor calibrations, the minimum number of non-zero calibration standards shall be specified as in the table included.

According to the manufacturer, the acceptance criterion for pH, Conductivity, and Fluoride and Ammonia calibrations utilizing ISE electrodes is based on the slope of the instrument. As known from pH measurement and calibration, the slope is an indicator of ISE performance. For ISE curves, there is both a linear section and a non-linear section.

Therefore, my question is, for ISE calibrations such as Ammonia and Fluoride, and other meter based tests such as Conductivity and pH, is the minimum number of calibration standards listed in the TNI table applicable?

File Submitted with SIR: ISE Calibrations.pdf

TNI Response:  Yes, they are applicable if the reference method describes a calibration procedure that includes average response/calibration factor or regression. Footnote b of V1M4 1.7.1.1.f states, "Fewer calibration standards may be used only if equipment firmware or software cannot accommodate the specified number of standards. Documentation detailing that limitation shall be maintained by the laboratory."

Question:  k) The laboratory shall use and document a measure of relative error in the calibration. i, ii-a and ii-b discuss running calculations on the calibration standard results used in the curve. For metals analysis, we use ICP/ICPMS and a zero point and single calibration standard, which makes it impossible to calculate %RE or %RSE of a mid point and low level calibration standard. We are interpreting the standard to read that when a zero point, single calibration standard is used, the requirements of 1.7.1.1.k do not apply and are replaced by the requirements in the next paragraph - section 1.7.1.1.L (which specifically discuss requirements for single point calibrations.) Please confirm that we are interpreting these requirements correctly.

TNI Response:  Your interpretation is confirmed as correct.

Question:  The Standard states that all initial calibrations must be verified by a second source. If a method requires a surrogate or internal standard, do the surrogates and/or internal standards also need to be verified by a second source?

TNI Response:  No, V1M4 1.7.1.1. does not apply to internal standards or surrogates. Section 1.7.1.1 addresses the calibration of analytes. An analyte is defined in V1M2 as "A substance, organism, physical parameter, property, or chemical constituent(s) for which an environmental sample is being analyzed." Internal standards and surrogates do not meet the definition of an analyte.

Question:  For 'k.ii.a.' - Wondering if you intended to use the absolute value of this calculation? If measured value is < true value the error will be negative. Can you have a negative % error?

TNI Response:  Yes, the percent relative error calculation as written in the standard may result in a negative value.

Question:  Some instrumentation, such as turbidimeters, spectrophotometers, etc. are purchased and received with an internal calibration performed by the manufacturer. Can these internal calibration be used to calculate test results?

TNI Response:  It is recognized that the calibration capabilities and sample analysis applications for this type of equipment varies. Yes, the manufacturer established initial calibration may be used for an accredited analysis if the requirements of V1M4 1.7, the method, and any applicable regulations can be met and documented by the laboratory.

If the manufacturer established initial calibration is used to calculate test results the manufacturer shall meet all of the requirements for initial calibration as listed in V1M4 1.7.1.1 and shall provide the accredited laboratory applicable data and records for the calibration in order to meet the requirements of V1M4 1.7.1.1. The laboratory must maintain these records per V1M4 1.7.1.1 b).