Limitations of direct estradiol and testosterone immunoassay kits

Abstract

Estradiol (E₂) and testosterone (T) are biologically active hormones that serve as important diagnostic markers in serum of premenopausal and postmenopausal women and in men. These hormones are measured frequently by immunoassay in clinical laboratories and the test results are used in the diagnosis and treatment of patients. For measuring the hormones by immunoassay, most laboratories utilize commercially available reagents that are packaged in the form of a kit and are used either in an automated instrument or manually. However, both the diagnostic kit manufacturer and testing laboratory seldom thoroughly validate the assay methods generated with these kits. This deficiency may lead to unreliable test results that could affect clinical evaluation and treatment of patients. The purpose of the present study was to assess the reliability of immunoassays that quantify serum E₂ and T levels with commercial diagnostic kits. The data generally show wide differences in the apparent levels of each hormone in a given sample obtained with kits from different manufacturers. This was especially true when measuring postmenopausal E₂ and T levels. However, a purification step, which included organic solvent extraction, prior to radioimmunoassay (RIA) of E₂ gave values that compared well with those obtained by conventional RIA (with preceding extraction/chromatographic steps). Our results point out the importance of more thoroughly validating assays performed with commercial immunoassay kits, especially with respect to sensitivity and specificity, prior to their use for measuring hormone levels in patient samples.

Introduction

Estradiol (E₂) and testosterone (T) levels in serum serve as important diagnostic markers in women and men [1]. The absolute serum E₂ level and its temporal change are followed closely to monitor follicle development and maturation, predict ovulation and prevent ovarian hyperstimulation in anovulatory women undergoing induction of ovulation and assisted reproductive techniques, such as in vitro fertilization or embryo transfer. Also, serum E₂ levels are used to determine postmenopausal status and for clinical evaluation of men with gynecomastia. As for T, serum concentrations of this hormone are useful in evaluating patients with polycystic ovary syndrome and hyperandrogenic women with clinical manifestations of hirsutism, acne and alopecia, in postmenopausal women with androgen deficiency, and in men with hypogonadism. In addition, E₂ and T are measured in many studies, e.g. in epidemiologic studies of breast and prostate cancers.

A variety of immunoassay methods are used to measure E₂ and T levels [2]. In general, E₂ and T immunoassays utilize either a radioactive or nonradioactive marker, belong to the excess-antigen/limited-antibody type of immunoassay (competitive immunoassay), and are performed with or without preceding purification of the analyte. E₂ and T radioimmunoassays (RIAs) usually employ an iodinated radioligand, whereas the nonradioactive assays use primarily either chemiluminescent or colorimetric markers, and sometimes a fluorescent marker. The non-radioactive assays are now used widely in automated systems.

Laboratories that perform measurements of serum E₂ and T levels for diagnostic purposes generally utilize direct immunoassay methodology, i.e. without a purification step [1]. The reagents required to carry out the assays are most often obtained from manufacturers that produce ready-to-use immunoassay reagents, packaged as kits, for automated instruments. Immunoassays carried out manually are now used less frequently. Although commercial diagnostic kits are widely used, they are not thoroughly validated by the kit manufacturer with respect to assay sensitivity, accuracy, precision and specificity.

Notable deficiencies in validation of immunoassays generated with commercial kits involve assay sensitivity and specificity. Assay sensitivity is usually poorly defined, leading kit users to believe that they can measure very low levels of an analyte, e.g. postmenopausal serum E₂ levels (<20 pg/ml). Similarly, validation of assay specificity is often limited to showing cross-reactivity of the antiserum with possible interfering compounds. However, quantitatively important metabolite standards, especially conjugated ones, are usually not included primarily because they are not readily available. Also, no attempt is made to compare hormone levels obtained by direct immunoassay with those obtained by conventional immunoassay utilizing extraction and chromatographic steps, or by a classical method such as gas or liquid chromatography–mass spectrometry (GC–MS or LC–MS). Lack of adequate validation of assay sensitivity and specificity may lead to unreliable patient test results.

Several studies have evaluated the reliability of measuring endogenous E₂ and T levels directly in serum or plasma by use of commercial immunoassay kits, either manually or on an automated instrument [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14]. However, none of the studies compared E₂ or T values determined by direct immunoassay with corresponding values obtained by immunoassay, following both organic solvent extraction and chromatography. Some of the studies used an extraction step in the assay [4], [5], [7], [8], [9], [14]. Three of the studies [3], [10], [11] used GC–MS as the standard for comparison, but the studies were limited to either use of a small number of samples, pooled serum and/or evaluation of only one or two commercial kits.

The objective of the present study was to evaluate E₂ and T immunoassay methods utilizing kits obtained from a variety of different manufacturers. Specifically, our aim was to compare the E₂ and T values measured by use of direct immunoassays with corresponding values determined by an immunoassay method that utilizes extraction and chromatographic steps prior to quantification.