Interference in immunoassays: avoiding erroneous results

types of immunoassay interference are similar to interferences in other chemistry assays and some types are unique to immunoassays. The interfering substance (interferent) in the sample may be exogenous to the patient, for example a drug, or endogenous, for example antibodies produced by the patient. The bias caused by interference may be positive or negative and, in some cases, the degree and direction of interference may vary with the concentration of the interferent or the concentration of the analyte. Types of interference in immunoassays include: 1. crossreactivity; 2. the hook effect; 3. antibody interference; 4. signal interference; and, 5. matrix effects. Some immunoassay designs are especially prone to particular types of interference. These are summarised in Table 1. Several publications review these interferences in detail [1-4]. Crossreactivity Crossreactivity is non-specificity, whereby a substance in the sample with structural similarity to the analyte competes for antibody binding. Crossreactivity is probably the most common type of interference in immunoassays. The crossreactant may be a metabolite or precursor of the analyte, e.g., conjugated cortisol metabolites in urinary cortisol assays; or a co-administered drug of similar structure, e.g., tricyclic antidepressants. Crossreactivity usually causes positive interference, but negative interference is possible with certain assay designs. For example, Figure 1 shows the crossreactivity of oleandrin, a cardiac glycoside similar to digoxin, in the AxSYM digoxin assay at different digoxin concentrations [5]. At low digoxin concentration, oleandrin causes the typical positive interference, but at high digoxin concentration, the interference is negative. Crossreactivity can affect any type of immunoassay, but it is most problematic in competitive assays. In two-site immunometric ("sandwich") assays, two reagent antibodies must bind the analyte at the same time, making these assays much more specific.