The content of this page may not be copied or used in any manner without permission from Triads Scientific Solutions, LLC.
January, 2024 – Current Topic
Note: This document expresses Dr. Jenke’s personal views and opinions and is not a position established and supported by Triad Scientific Solutions, LLC. This document is not professional advice and does not constitute a professional service provided by either Dr. Jenke or Triad Scientific Solutions.
Does "Quantitative" and "LC/MS" Belong in the Same Sentence when Applied to Non-targeted Analysis (NTA) for E&L?
An analytical chemist performing non-targeted analysis (NTA) of samples for organic extractables or leachables (E&L) has two objectives:
Correctly identify all E&Ls above the analytical evaluation threshold (AET).
Conservatively (protectively) estimate the identified compound’s true concentration in the sample.
Liquid chromatography with mass spectrometric detection (LC/MS) is an important tool used in E&L NTA to address generally non-volatile compounds for which gas chromatographic (GC) methods are sub-optimal. Although LC/MS performs admirably as an identification tool, its ability to conservatively estimate an analyte’s true concentration in a sample is lacking. This inability to be quantitative stems from two essential circumstances, the nature of NTA and the nature of the LC/MS response. Considering the nature of NTA, perhaps the most significant aspect of NTA that affects quantitation is the fact that NTA is exploratory in nature, meaning that the compounds present in the sample cannot be specified upfront. Thus, the LC/MS analysis cannot be made quantitative upfront by establishing a compound’s response function because the compounds present in the sample are not revealed until the analysis is completed. LC/MS analysis would only be quantitative in NTA under three circumstances, none of which are realized in LC/MS applied to E&L:
The response functions for all possible E&Ls have been experimentally established,
All E&Ls have the same response function, or
A specific E&L can be linked to a surrogate standard which shares its response function.
Considering the first point, it has been estimated that there are tens of thousands of possible E&Ls. Thus, it is a herculean task for a laboratory to establish response functions for all possible E&Ls not only because of the shear number of possible E&Ls but also the virtual impossibility of securing a reference standard for all possible E&Ls. Although laboratories have complied LC/MS response factor (or relative response factor) databases, even in the best cases these databases include only a small fraction of all possible E&Ls. Furthermore, the use of response factors for quantitation presumes that the response function is linear over its entire dynamic range, a circumstance that is not always encountered in LC/MS.
Considering the second point, it is well established in the E&L literature that response factors (and therefore presumably response functions) in LC/MS are quite variable, compounds to compound, with % relative standard deviations (% RSD) for LC/MS response factors reaching, and often exceeding, 100%. Thus, quantitation of a given E&L using the response function of a surrogate standard (or a universal response function such as the average response function) is prone to inaccuracy, leading to estimated concentrations that widely underestimate or overestimate the analyte’s true concentration.
Considering the third point, while the concept of using a well-matched surrogate standard for effective quantitation is fine in theory, in practice its effective use requires that there be a means of linking an analyte with its surrogate (for example, same responses based on structural similarities, same responses based on similar retention times, etc.). At the time I write this commentary, no means of linking an E&L analyte to a response-comparable surrogate have been successfully defended.
By definition, a conservative (protective) concentration estimate is one which does not underestimate an analyte’s true concentration (that is, an analyte’s conservative and prootective concentration is always equal to or greater than its true concentration). Conservative concentration estimates are appropriate in quantitative toxicological risk assessment (qTRA), as their use in qTRA means that patient exposure to an E&L will not be underestimated (and thus that the patient will be fully protected from a potential adverse health risk due to exposure to the E&L).
If the role of the toxicological risk assessor was merely to protect patients from harm, then the use of conservative (likely over-estimated) concentration estimates is acceptable practice. However, if the toxicological risk assessor also accepts the responsibility for ensuring that patients have access to safe life-saving or life-enhancing pharmaceutical products or medical devices, then the use of overly conservative concentration estimates in qTRA is unacceptable, as so doing may lead one to conclude that a product or device is unsafe (based on over-estimated patient exposure to leachables), when in fact the product is perfectly safe when used under its prescribed conditions. In this way a patient’s health is harmed because they do not have access to a product or device that eliminates or alleviates a harmful medical condition.
One commonly employed means of making a concentration estimate conservative is to use an uncertainty factor (UF), linked to the magnitude of the compound-to-compound variation in response characteristics, to “adjust” the estimated concentration up; that is the estimated concentration is multiplied by the UF to produce a conservative concentration. As noted previously, the consequence of adjusting the estimated concentration in this manner is that the conservative concentration likely over-estimates the analyte’s true concentration. The larger the UF, the greater the number of compounds whose true concentrations are over-estimated and the greater is the magnitude of the over-estimation.
Unfortunately, the UF for LC/MS is large, with a value of 10 commonly being advocated. Thus, conservative concentrations reported for LC/MS are typically over-estimated by factors of 5 or greater, increasing the likelihood that the qTRA is over-conservative and increasing the likelihood that safe items could be assessed as being unsafe.
Thus, my answer to the title question of “is concentration estimation in LC/MS quantitative?” is a definitive “No”. However, this does not mean that “quantitation” by LC/MS is an inappropriate point of departure for qTRA. For example, a qTRA performed using conservative concentrations secured by LC/MS is appropriate and proper if the outcome of the qTRA is that the substance in question is safe. That is, if a compound is assessed as safe based on an over-estimated concentration, then surely it will also be assessed as being safe using its true concentration.
When a compound is assessed as potentially unsafe based on an over-estimated conservative concentration, then we acknowledge the possibility that in fact the compound is safe and it is the over-estimation that is unsafe. In this case, the only way to reach the proper conclusion on the safety of the compound of interest is to redo the quantitation with a more accurate approach. Thus, a compound that is assessed as being potentially unsafe based on a conservative concentration becomes a target compound which is quantified by establishing its specific response function (which is no longer NTA). The qTRA is repeated using the estimated (and more accurate) concentration and the outcome of the repeated assessment becomes the definitive outcome. In some cases, such an approach will simply re-affirm that the compound of interest is potentially unsafe. But especially in the case of a strongly responding compound (that is, a compound whose conservative concentration significantly over-estimates its true concentration), securing a more accurate concentration can reverse the conclusion of the initial assessment and support a final conclusion that the compound is safe.