1. Matrix Effect
Matrix effects can be a problem for immunoassays, especially for a method without any prior sample clean-up. It can be caused by either nonspecific or specific interferences from the sample matrix and reagents. Possible matrix effects can Copyright 2002 by Marcel Dekker. All Rights Reserved.
be assessed by comparing a standard curve with calibrators prepared in assay buffer versus those in the intended sample matrix. The amount of nonspecific binding (NSB), maximum percent of binding in the absence of analyte (Bmax), and the slope will provide an early indication of possible matrix interferences. High NSB, low percent binding to the antibody, and a shallow slope often indicate problems that could be caused by matrix effect.
Selectivity against the endogenous compounds in the sample matrix is tested on separate matrix samples from at least six undosed individuals (or various lots of control matrix from commercial sources). If there are structural similarities between the drug and endogenous compounds, a larger number of control lots (e.g., > 20) should be tested. Control matrix is tested for NSB, Bmax, and recovery of a known amount of analyte added at or near the limit of quantitation. If the NSB and Bmax from various control matrix lots are similar, the standard deviation of B0 (Bmax - NSB) is an indication of the variability of the noise level, which can be used for the estimation of the limit of detection. If a sample from an undosed individual exhibits a response deviating more than one standard deviation from mean response, interfering material is most likely present. To determine whether interference is specific or nonspecific, a parallelism test should be conducted.
A useful tool to test matrix effect is parallelism. Test samples from a clinical trial and/or samples from various control batches, with known amounts of analytes added, are diluted with control samples containing no analyte, and these are used as standard calibrator preparations. Various dilutions (e.g., 2-, 4-, 6-, 8-, 10-, and 20-fold) are prepared and analyzed against the standard calibrators. The dose-response curves of the diluted samples are compared to those of the standard calibrators. A parallel line of the test (or spiked) sample shows that the compound present in the sample has the same antigen-antibody binding response as the analyte and, therefore, is very probably the analyte itself. If the line is not parallel to the standard curve line, and the concentrations at higher dilutions agree with one another, the matrix effect is nonspecific and could be overcome by dilution.
Sometimes, matrix effects can be corrected by normalization of the NSB/ B0. This must be tested by showing that the interference is consistent in multiple samples taken at different times from the untreated individual. In other instances, predose samples from each patient (or test animal) are used to construct the calibrator standards for quantitation. In such cases, it may be better to do sample clean-up to eliminate matrix interferences, because it is not always possible to obtain adequate predose sample volumes from each subject.
The simplest processing step to overcome matrix effect is to dilute the sample threefold or more in buffers containing chaotropic or chelating agents such as
Copyright 2002 by Marcel Dekker. All Rights Reserved.
Tween-20, Triton-X 100, and EDTA. In other cases, matrix effect and potentially interfering substances of drug metabolites and analogs should be removed by sample clean-up preceding the IA. Sample extraction as applied to other analytical methods can be applied to IA. Therefore, these methods will not be elaborated here. Briefly, liquid/liquid and solid-phase extractions are most commonly used. HPLC-IA has the combined advantages of the selectivity of HPLC and the sensitivity of IA (IA becomes a sensitive detector for the HPLC method). However, sample clean-up by HPLC is labor-intensive and has not been popular for large numbers of samples. New techniques in perfusion chromatography using porous immunoaffinity columns coupled with enzyme chemiluminescence detectors may be a novel way to address this problem (174).
For every sample clean-up method, it is important to investigate the recovery of the parent drug as well as the potentially interfering compounds. Any organic solvent must be evaporated completely and the residue reconstituted in the assay buffer to avoid altering the antibody-binding activity.
The sample extraction step can also serve as a concentration step to increase the assay sensitivity. Direct IAs generally use a sample volume of 0.1 mL or less, while it is common to extract 1.0 mL of human biological fluid.
Because IA methods are generally more economical and/or faster, a strategy in drug development is to use IA method as a main method to analyze large numbers of samples and to use another method to analyze a subset of these samples. Correlation plots of one method against another over various concentrations are constructed for method comparison. For example, cyclosporine IA methods using polyclonal and monoclonal antibodies had been compared with HPLC methods on samples from various types of patients (175-179). Several authors have concluded that the correlations were excellent for samples from normal volunteers and some patient types. An investigator in a drug-development and monitoring program should consider using comparative methods; a method that is fast and easy to handle can be chosen for high-volume analysis, while a second, more elaborate method can be used to validate the first.
During the course of drug development, several analogs may be considered as candidates for investigation. Immunogens designed not to discriminate among epitope differences of the analogs can be used to produce nonspecific antibodies. The antibodies will recognize the common structure, and one method can be used for several drug analyses. This will save a lot of time and cost. If certain analogs are possible precursors/metabolites of one another, separation methods prior to IA could be used to provide the required selectivity.
Broughton et al. (180) have demonstrated that the use of nonspecific antisera to gentamicin could also be applied to the analysis of sisomicin if gentamicin were not present in the samples, because cross-reactivity with other substances was minimal. We have successfully used anti-prednisone antiserum to measure cortisone in serum in the absence of prednisone. This was possible because cross-reactivity with other compounds was not significant (181).
Drug-monitoring programs in clinical chemistry also use nonspecific antibodies to detect drugs of abuse, such as cannabinoids, opiates, and bezodiazep-ines. References are listed in Table 4 in Sec. VIII. Investigators deliberately pooled and mixed antisera against several analytes to develop a common IA method for multiple analytes. For example, antiserum against testosterone (T) and antiserum against 5a-dihydrotestosterone (DHT) were used to analyze samples for both T and DHT simultaneously (182). A multivariable (three-dimensional) standard curve was created which allowed the independent estimation of T and DHT concentrations when both T and DHT were present in samples. The method is valid as long as both assays are precise, and the procedure avoided the need for tedious, time-consuming chromatographic separation.
Was this article helpful?
Diabetes is a disease that affects the way your body uses food. Normally, your body converts sugars, starches and other foods into a form of sugar called glucose. Your body uses glucose for fuel. The cells receive the glucose through the bloodstream. They then use insulin a hormone made by the pancreas to absorb the glucose, convert it into energy, and either use it or store it for later use. Learn more...