Trade-offs in mass spectrometry.

When should a new instrumental technique or analytical method be adopted for use in clinical chemical analyses? Many clinical methods are approved by the FDA based on the 510 k process, which requires demonstration that a new test is at least as good as those already in the marketplace. This is justified from both a scientific and free-market regulatory perspective when replacement of a sound method or technique is being considered. It is particularly true when considering adopting new instrumentation. Bowers and Borts in this issue [1] provide a powerful example of the need for circumspection when considering replacement of an existing instrumental technique with a new and supposedly better one. Their paper also presents a strong case for the perennial need for good sample preparation techniques. Clearly, advances in instrumentation, combined with reduced costs, have established mass spectrometry (MS) as an essential tool in bioanalytical research. The increased availability of modern instrumentation has had and will continue to have a positive influence on clinical chemistry—particularly for analytes that are not amenable to quantification by more conventional methods. At the same time, there is often a continuing temptation to attribute aspects of the panacea—the universal curative—to a new or improved type of instrument. Users of MS often seem to be particularly susceptible to this lure of an instrumental cure-all that makes all measurements simpler, faster, and more sensitive than earlier instruments. The possibility of a new instrumental approach led Bowers and Borts to compare results of measurements made in a quadrupole system (QMF), a mature technology, with those from a quadrupole ion trap (QIT), a technology that is winning wide acceptance rapidly. For a model system, the authors monitored 1 to 10 ions from tetrachlorobenzene simultaneously. They showed that a QIT, operated in a frequency-modulated selected-ionstorage mode, had lower limits of detection than a QMF, operated in selected-ion-monitoring mode. This result with a model compound led Bowers and Borts to hypothesize that application of the same scanning paradigms to the determination of anabolic steroid metabolites and the b-agonist clenbuterol in a urine extract would give lower limits of detection for the QIT. However, the results showed no improvement in detection limits for the QIT in comparison with the QMF. Why did the ion trap fail to give the expected improvement for analysis of the complex system? The authors give an instrumentally correct explanation. Bowers and Borts argue that the lack of improvement for the QIT when presented with an analyte present in a complex matrix arises because the width of the window for ion admittance to the trap allows matrix or column bleed ions (or both) to enter with the analyte ions. The width of the ion admittance window is set by the frequency modulation limits and cannot be made smaller in the mass range of interest for their study. Such problems are not unique to QIT: Coelution of a compound with the analyte of interest was shown to decrease ionization efficiency in a QMF [2]. The inability of the QIT to demonstrate the expected improvement in detection limit is thus not a failure of the instrument per se, but a failure brought about by chemical noise. Bowers and Borts recognize this particular failure. In this Editorial we name it by using an undergraduate physical chemistry student’s expression for the 2nd Law of Thermodynamics: “You can’t get something for nothing.” In the report of Bowers and Borts, the principle is manifest by an inability to attain maximum sensitivity without eliminating matrix and contaminant ions. Because MS is generally considered to be a technique well suited to reduction of matrix interferences, some may be surprised that, as applied by Bowers and Borts, it was unsuccessful in that task despite their using sound extraction and derivatization procedures. Is it possible to put limits on the expectations of success? Bowers and Borts’ work defines one limit in terms of analytical sensitivity and specificity. Improvements in detection limit for analysis of steroids in their complex matrix clearly may require further cleanup or use of additional analytical dimensions such as improved chromatography, higherresolution MS, or even MS/MS. What do we conclude from this example? First, new instrumentation and methods must be continually under development, for it is from such developments that improvements become possible. Second, no analytical technique or method can be considered as an ideal solution: Every new development must be rigorously evaluated on its merits and carefully compared with existing procedures.