Competitive immunoassay versus immunometric (sandwich) assay

There are two main immunoassay formats relevant to endocrinology laboratory testing, and understanding some of the basic differences between these formats helps both interpretation of results and troubleshooting of unexpected situations.⁵ The first class of assays is termed competitive, with the classic radioimmunoassay (RIA) as the archetypal example. A primary antibody against the substance (“analyte”) of interest is added to the patient’s sample, together with a radiolabeled version of the analyte (“tracer”) that competes with the endogenous analyte for binding to the primary antibody (Figure 4-2, upper panel). After a sufficiently long incubation time, the primary antibody is precipitated using a second anti-IgG antibody, polyethylene glycol, or (most commonly nowadays) by using primary antibody attached to a solid support such as a bead, which allows a simple centrifugation step to collect the primary antibody. Any unbound tracer or analyte is washed away, and the amount of tracer in the precipitate is then quantified.

To convert this tracer signal into a concentration, a standard curve is prepared from samples where varying concentrations of a known amount of analyte have been added. For a competitive immunoassay, the amount of signal detected (amount of tracer bound to antibody) decreases as the analyte in the patient’s sample increases. Importantly, anything that prevents the tracer from binding to the primary antibody would also decrease signal and increase apparent analyte concentration. For example, taking an immunoassay optimized for serum and using it for an extremely concentrated urine sample or in a body fluid where the protein concentration is extremely high can significantly alter the results. Because high salt or high protein concentration can inhibit tracer binding to antibody, tracer signal will be low and the apparent analyte concentration very high, even in the complete absence of any analyte.

The second class of immunoassays is termed immunometric, with the most important example being the noncompetitive, so-called sandwich assay. Typically, as shown in Figure 4-2 (lower panel), an antibody attached to a solid support (plate, wall of a tube, or a bead) is used to capture the analyte of interest, followed by the addition of a second, labeled antibody that binds to a different site on the analyte, creating an antibody-analyte-antibody sandwich. After unbound detection antibody is washed away, that which remains generates a radioactive (immunoradiometric [IRMA]), chemiluminescent (immunochemiluminometric [ICMA]), colorimetric (enzyme-linked immunosorbent assay [ELISA]), or fluorescent (immunofluorescence assay [IFMA]) signal, depending on the label chosen.

The immunometric assays offer certain inherent advantages over competitive immunoassays. First of all, the fact that two antibodies are involved, each binding to a separate epitope on the analyte, greatly increases analytic specificity, including the ability to select very specific isoforms of an analyte.⁶ For example, one antibody can be directed at the N-terminus of a peptide, whereas the other antibody can be directed at the C-terminus, ensuring that only full-length peptide is detected. A second advantage is that immunometric assays tend to be more analytically sensitive than their competitive assay counterparts.⁷ This arises from the fact that the sensitivity of a competitive immunoassay largely depends on the affinity of the antibody used in the assay—and developing very high affinity antibodies is not an easy or predictable task. In contrast, the sensitivity of an immunometric assay can be improved by use of a higher activity signal, which is more easily detected even when analyte concentrations are very low. As signal detection technology continues to improve, increasingly sensitive immunometric assays can be developed. A third advantage of immunometric assays arises from the ability to use monoclonal antibodies, which can be more easily produced in quantity and have more predictable characteristics than polyclonal antibodies.

Potential confounders in immunoassays

Sandwich immunometric assays are potentially vulnerable to what has been termed the “hook” or “prozone” effect that can lead to falsely normal or low values in the presence of large amounts of the analyte⁸ (Figure 4-3). Generally, as analyte concentration increases, more antibody-analyte-antibody sandwiches are formed, leading to more signal detected. At extremely high analyte concentrations, however, it is possible for both capture and detection antibodies to be saturated with analyte prior to formation of a sandwich complex, leading to the detection antibody being washed away and the low signal being misinterpreted as a low concentration of analyte. This phenomenon is not seen with competitive immunoassays. To minimize the chances of a hook effect, most immunometric assays have now modified their methods as follows: analyte is allowed to bind to capture antibody, a thorough wash step removes excess unbound analyte, and only then is the detection antibody added. In the now extremely rare situation where a hook effect is still suspected, serial dilutions of the sample may be performed to see if the apparent concentration of the analyte actually rises with increasing sample dilution.

There is some inconsistency in the literature around the terms heterophilic antibody and human antianimal antibody (HAAA) or human antimouse antibody (HAMA). Some authors use these terms interchangeably; others reserve the term heterophilic for low-affinity antibodies that spontaneously arise against multiple, poorly defined antigens and use HAMA or HAAA only for those patients who develop specific high-affinity antibodies due to treatment with a murine monoclonal antibody or due to recurrent exposure to a particular animal species (e.g., scratches endured by a rabbit handler). Regardless of the term used, any given patient may have antibodies present that bind to and interfere with animal (mouse monoclonal; rabbit, goat, or donkey polyclonal) antibodies used in a particular immunoassay,⁹ even without previous exposure to that particular animal. These antibodies can bridge the capture and detection antibodies in an immunometric assay, as shown in Figure 4-4 (left panel), creating a sandwich and a high signal (false elevation) even in the absence of analyte. Alternatively, they could attach at the binding site of the capture antibody and prevent binding of the analyte (Figure 4-4, right panel), eliminating formation of the sandwich, and causing a falsely low value. Competitive immunoassays are less often affected, although they may show falsely high values if the interfering antibody outcompetes the assay antibody for tracer, or falsely low values if they bind preferentially to the analyte over the tracer.

When a heterophilic/HAAA/HAMA effect is suspected, the sample should be assayed on a different immunoassay platform, as the interference is often specific for one manufacturer’s assay but not another’s. Such a maneuver is often the quickest to accomplish, but the sending lab must ensure that the second lab is truly using a different platform. Another option is to use heterophilic antibody blocking agents, which are a mixture of animal immunoglobulin fragments or other proprietary agents that soak up the interfering antibodies. A disadvantage of this approach for pediatrics is that a significantly larger volume of sample is required to assay both with and without use of the blocking agent. A third, more labor-intensive option is to assay the sample on serial dilutions, as those showing antibody interference effects tend not to show linear responses on dilution.

Specific autoantibody interferences appear to be rare compared to the heterophilic/HAAA/HAMA interferences described in the previous section, with the notable exception of TgAb in patients with thyroid cancer. TgAb is present in about 20% to 25% of thyroid cancer patients and leads to false depression of thyroglobulin levels by immunometric assay (e.g., Tg-ICMA). In addition, different thyroglobulin antibody assays may report discordant (negative versus positive) results in some cases. Although a competitive assay such as Tg-RIA is less often affected by TgAb interference than is the Tg-ICMA, if the autoantibody is of sufficiently high affinity to outcompete the assay antibody, a Tg-RIA may show false elevation in the presence of such TgAb.¹⁰ Overall, any Tg immunoassay result, whether performed by Tg-ICMA or Tg-RIA, must be viewed as potentially affected in a patient with TgAb pressent (and even possibly in patients whose TgAb is reportedly negative). Helpful proxies for Tg include following TgAb titer, which should fall over time in patients without cancer recurrence.¹¹ Imaging studies may also help assess the probability of recurrence. Although rare, autoantibodies may be of sufficiently high affinity to affect both competitive and immunometric assays alike. Consequently, they would be less likely to show differences when assayed on different platforms and are not affected by heterophilic antibody blocking agents. In these scenarios, diagnosis may require measurement of specific autoantibody (e.g., a parathyroid hormone [PTH] antibody assay in someone with a discordantly high or low PTH level).

Limitations of immunoassays

The ability of an antibody to bind a specific target with high affinity is remarkable, yet there are limitations. The small size and relatively poor immunogenicity of steroid molecules make it difficult to obtain an antibody that can clearly distinguish a specific steroid from other similar steroids. For example, an antibody raised against testosterone may show significant cross-reactivity with a conjugated form such as testosterone glucuronide, or with the structurally similar dihydrotestosterone molecule. A solvent extraction step to remove water-soluble conjugates such as glucuronides, coupled with a chromatographic step to separate out structurally similar molecules, will greatly improve the analytic specificity of the assay¹² (Figure 4-5A). The extraction step also separates testosterone from binding proteins that can interfere with accurate measurement.

Testosterone immunoassays that do not involve such extraction and chromatography steps, including the majority of assays found on most automated laboratory platforms, may perform adequately at the higher levels found in adult males (> 300 ng/dL) but will frequently measure inappropriately high values in the range most pertinent to women and prepubertal children (Figure 4-5B). Because in practice the lower values are more concordant with the clinical picture, endocrinologists caring for women and children have traditionally relied on immunoassays that involve extraction and chromatography steps, or assays involving chromatography and tandem mass spectrometry.

Free and bioavailable hormone tests

The free hormone hypothesis states that free, or unbound, steroids or thyronines, vitamins, or peptides are available to enter cells or interact with cellular proteins by virtue of their small size and sometimes their hydrophobic nature.¹³ Free hormone in this section means not liganded to a binding protein. As free and bound hormone pass through a tissue bed, the free hormone may be taken up by the tissues, additional free hormone may become available from the bound fraction, or the complexed hormone may be taken up by cells.¹⁴ The systems may be further complicated by multiple binding proteins. Clearly, the activities of several steroids, thyronines, and even protein hormones are reduced by binding to specific high–affinity binding proteins and albumin. Conversely, the absence of binding proteins can reduce the total hormone level, but the free hormone level may be unaffected. For example, the majority of thyroid hormone in circulation is carried by thyroxine-binding globulin (TBG). TBG deficiency is characterized by low total thyroid hormone levels, normal free thyroid hormone levels, normal TSH (in equilibrium with the free T4), and clinical euthyroidism.

Testosterone binds to sex hormone binding globulin (SHBG) and serum albumin. Testosterone binds SHBG with much higher relative affinity, but SHBG is present in much lower concentrations than albumin. Free testosterone may be measured by equilibrium dialysis, which is accepted as the best indication of the physiologic availability of testosterone (Figure 4-6A). The free fraction may also be estimated mathematically from the testosterone, SHBG, and albumin concentration; however, not all assumptions may be valid in calculated free testosterone.^{15, 16} Analog methods measure free testosterone directly using a competitive test in automated assay instruments (Figure 4-6B), but the analog methods may at best weakly reflect free hormone activity. Similarly, free T4 measured by analog methods can perform poorly in some settings.

Bioanalytic testosterone refers to the fractions of testosterone bound to serum albumin plus the free hormone; for this reason this fraction is also called free and weakly bound testosterone. The dissociation rate of albumin bound is very fast relative to that of the SHBG bound, and for that reason it is thought to be available to the tissues for actions. This idea, supported by empirical evidence, is in contrast to the free hormone hypothesis, as free and protein-associated hormone may be available to the tissues.¹⁴ In practicality, measurement of free or bioavailable hormones provides the same information to the clinical encounter.