Comparative Analysis of Commercial Immunoassays for the Determination of Total, Intact, and Nonintact Luteinizing Hormone in Urine

ABSTRACT

Background

In our recent publications, we reported the identification of three different molecular forms of total luteinizing hormone (LH) in urine, the intact LH, the free beta‐subunit (LHβ), and its core fragment of LHβ (LHβcf), the latter two establishing the nonintact portion of LH. Following the discontinuation of the Delfia immunofluorometric assay (IFMA) (Wallac, PerkinElmer Finland, Finland), a leading method for detecting urinary LH for 30 years, this study seeks to assess the efficacy of three alternative commercial immunoassays in identifying various forms of U‐LH.

Methods

Diluted urine samples underwent gel filtration to separate them into fractions, each containing different forms of LH. These were then assayed using Delfia IFMA, Architect LH (Abbott, USA), Elecsys LH Cobas (Roche, Switzerland), and Immulite 2000 LH (Siemens, Germany) immunoassays.

Results

Both Delfia and Immulite assays detected total U‐LH, that is, all three forms of U‐LH, including intact LH, LHβ, and LHβcf. Cobas detected only intact LH and LHβ, whereas Architect detected solely the intact LH.

Conclusions

Immulite assay can be an alternative tool to detect all forms of urinary LH, a feature likely to be instrumental in developing noninvasive, practical, and scalable solutions for evaluating total U‐LH changes during minipuberty in neonates, during the onset of central puberty in peripubertal children, puberty‐associated disorders in adolescents, and the fertility window in women, with a special focus on postpeak changes.

In our study, we evaluated alternative assays to replace a discontinued method for detecting all forms of urinary LH, including its degradation products. By analyzing LH in gel‐filtrated urine samples, we assessed the comprehensive detection capabilities of various diagnostic tests. Our novel findings revealed diverse fields of utility for different assays, offering a nuanced understanding of their specific clinical applications as choosing the correct assay with an informed decision would facilitate enhanced clinical practice in investigating minipuberty in neonates, the onset of central puberty in peripubertal children, puberty‐associated disorders in adolescents, and the fertility window in women, with a particular emphasis on postpeak changes.

1. Introduction

During the prepubertal period, there appears to be a long period of priming at the anterior pituitary before the initial increases in nocturnal luteinizing hormone (LH) secretion are finally detectable at around Age 9 or 10 in both sexes [1]. Nocturnal hypothalamic–pituitary–gonadal (HPG) activation periods are rather short‐lived and do not resume during the daytime, and the resulting LH secretory bursts are momentary and transitionary during the first weeks and months of the onset of central (hormonal) puberty. Thus, nocturnal LH can be recovered from first‐morning‐voided (FMV) urine samples reflecting the integrated nighttime LH excretion, but not from daytime spot urine or serum samples because the distribution half‐life of LH in serum is approximately 1 h only, whereas the elimination half‐life for LH in urine is approximately 12 h [2, 3]. The half‐lives of the wild‐type and variants of LH differ to some extent, but not in amplitudes that could affect the detectability of LH from morning serum or urine samples [4].

Detecting the initially miniscule bursts of hypophyseal LH secretion during nighttime sleep requires the ability to determine even low amounts of immunoreactivity emerging from any remnant of LH, that is, the degradation products of LH in FMV urine, which is the filtered form of nocturnal sleeptime blood circulation, reflecting the integrated remnants of the initial nighttime LH bursts, even if they are short‐lived and low in amplitude. Indeed, integrated gonadotropin activity during overnight sleep was shown to be reflected as an increase in total LH and follicle‐stimulating hormone (FSH) in nocturnal sleeptime and FMV urine samples [1, 5, 6]. Therefore, particularly the purpose of detecting an imminent puberty or the onset of central puberty, it is of utmost importance not to miss any degradation form of LH in FMV urine, even in the absence of intact LH. Total LH assays that detect not only the intact LH but also its degraded forms may therefore be the solution to detect any remnant of nighttime LH immunoreactivity (LH‐ir) activation in FMV urine samples for predicting or detecting the onset of central puberty long before the clinical signs of puberty [1, 7].

LH shares epitopes with LH‐beta‐subunit (LHβ) and its core fragment (LHβcf) as well as with human chorionic gonadotropin (hCG), its beta‐subunit (hCGβ), and beta core fragment (hCGβcf) [8, 9], thus, immunoassays that can specifically detect total LH‐ir in urine are also essential for the noninvasive assessment of LH‐ir during the first 3 weeks of life, as cross‐reactivity with hCG can lead to incorrect test results in neonates of this age group [10]. The structure of LHβcf is not, in fact, precisely known, but, on the basis of its reactivity with different LHβ and hCGβ antibodies, it resembles hCGβcf. Birken et al. isolated 12 kDA‐sized LHβ core fragments from the pituitary gland and showed that they are formed by proteolytic digestion of LHβ, which causes losses in the N‐ and C‐terminal parts of LHβ [11]. Birken and Kovalevskaja developed monoclonal antibodies and based on these, specific sandwich immunoassays, which enabled them to demonstrate the presence of multiple forms of LH in urine [8, 12].

We recently demonstrated the occurrence of three distinct forms of LH‐ir, that is, intact LH and its degradation products forming the nonintact portion of the LH‐ir, namely LHβ, and a 10–12 kD fragment of LHβ, called core fragment (LHβcf) in urine samples obtained from fertile women using a commercially available diagnostic method [13, 14, 15]. However, specific antibodies against LHβ or LHβcf are not available, thus there is no commercially available or in‐house assay to detect any of these degradation products separately. Only total LH can be determined, and this is achievable solely through specific diagnostic kit products, such as the Delfia immunofluorometric assay method (IFMA), which has been used in most research studies on gonadotropin determinations in the evaluation of puberty over the past 3 decades [6, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29]. Unfortunately, production of this widely used assay has been discontinued recently [30]. Therefore, it is not currently possible to determine LH concentrations specifically for LHβ or LHβcf by any commercially available assay, nor is it possible to detect the total LH concentrations or specify the ratios of concentrations derived from these individual degradation products. Some earlier studies reported that the Immulite 2000 LH immunochemiluminometric assay (ICMA) (Siemens, Munich, Germany) detected intact LH and LHβ subunit [31, 32], but there was only an indirect, unconfirmed report about the ability of this assay to measure LHβcf concentrations [33]. In addition, no information has been available about its ability to detect the smaller subunits and fragments at lower concentrations of LH as in children's urine.

In the current study, in a quest to find an alternative assay method to determine total LH‐ir in urine samples, we measured LH concentrations in gel‐filtrated fragments of urine samples to reveal which diagnostic test products can be used for detecting all the possible sources of LH‐ir in urine as broadly as possible, including the degradation products of LH.

2. Materials and Methods

2.1. Subject, Samples, and Storage Conditions

This study was conducted at the Department of Clinical Chemistry, Helsinki University Hospital and University of Helsinki, Helsinki, Finland. FMV urine samples, presumably with high LH concentrations, were obtained from a postmenopausal woman who had not consumed water for at least 12 h prior to sampling, and were subsequently subjected to centrifugal concentration. Concentrated fresh urine samples were then processed by gel filtration to obtain separate filtrates with fractions representing decreasing molecular weights of LH degradation products corresponding to increasing fraction numbers. Urine samples were fractionated by gel filtration on a 10/300 mm Superdex G‐75 and 16/600 mm Sephacryl S‐100 columns (both from Sigma‐Aldrich, USA) eluted with 0.1 M ammonium bicarbonate buffer (15 mmol/L, pH 8) at a flow rate of 0.5 mL/min. Samples of 0.5 mL were collected, and the filtrates of various fractions were kept in 0.1% BSA/TBS solution at 4°C until assayed. LH concentrations in the filtrates were measured by the following assays: Architect LH chemiluminescent microparticle immunoassay (CMIA) (Abbott, Chicago, Illinois, USA), Elecsys LH Cobas electrochemiluminescence immunoassay (ECLIA) (Roche, Basel, Switzerland), Delfia IFMA (Wallac, PerkinElmer Finland, Turku, Finland), and Immulite 2000 LH ICMA (Siemens, Munich, Germany). Immunoreactivity emerging from the intact LH, LHβ, and LHβcf, or any combination of these, was measured separately by the respective assays in urine. There was no need to adjust the urinary concentrations for specific gravity or creatinine because the main focus was to compare the LH‐ir obtained by different assays in gel‐filtrated fractions [34, 35], as confirmed also by our earlier studies [26, 36].

2.2. The Immunofluorometric Method

Delfia LHspec IFMA is a commercial sandwich assay using monoclonal antibodies (Wallac Oy, PerkinElmer Finland Oy, Turku, Finland). One antibody is immobilized onto a microtiter strip well, whereas the other is labeled with a europium chelate. In the Delfia assay, both capture and detection antibodies are directed to β‐subunit targeting specifically different epitopes [37]. This assay was designed to detect total LH, that is, intact LH, LHβ, and LHβcf. It was possible to detect only the intact LH with some minor modification by replacing the detection (tracer) antibody with the FSH tracer [38]. The assays were performed according to the manufacturer's instructions. A sample volume of 25 μL was used for analysis. The total assay volume was 225 μL. The calibrators were standardized against the WHO 2nd International Standard for pituitary LH for immunoassay (code 80/552) [26].

The limits of detection calculated by using both the measured limits of blank and test replicates of a sample known to contain a low concentration of the analyte for the U‐LH assays was 0.012 IU/L [6, 39]. The intra‐ and interassay CVs for the U‐LH assay were 5.7% and 6.4%, respectively [6]. The intra‐assay coefficients of variation (CV) for the said assay were <2% at levels between 3 and 250 IU/L and approximately 10% at 0.3 IU/L. The interassay coefficient of variation was <3% at 4–18 IU/L for LH [5]. Hormone concentrations were not corrected for variations in urine excretion rate (such as urinary density or creatinine), because the correlation with serum levels was not improved but was impaired because of overcorrection in very dilute urine samples [26].

2.3. Other Immunoassays

For the Immulite assay, the intra‐assay variability exhibited CV of 13.1%, 3.0%, 3.7%, and 3.6% at LH concentrations of 0.15, 1.04, 1.89, and 8.7 IU/L, respectively. Correspondingly, interassay CVs were observed at 10.8%, 3.6%, 2.5%, and 3.1% for these concentrations. The detection limit for LH measurement, as per the manufacturer's documentation, is 0.05 IU/L. For Architect, the intra‐assay CV were ≤3.5%, ≤ 3.1%, and ≤2.9% for LH concentrations ≤1.00, 5–15, and >15 IU/L, respectively, and the interassay CV were ≤4.1%, ≤ 4.4%, and ≤3.2% for LH concentrations ≤1.00, 5–15, and >15 IU/L, respectively; the detection limit was 0.03 IU/L for LH determinations according to the manufacturer's kit insert. For Cobas, the intra‐ and interassay CV were 2.2%, 1.1%, and 1.0% for 3 different LH concentrations of 0.99, 10.7, and 63.4 IU/L, respectively, and the interassay CV were 2.3%, 1.6%, and 1.1%, respectively, for the same three LH concentrations, and the detection limit was 0.03 IU/L according to the manufacturer's kit insert.

2.4. Statistics

The detection limit was defined as the concentration corresponding to the value of the mean plus 2 SD of 12 duplicates of the zero standard. For statistical evaluation, concentrations below the detection limit were given a value of 0.01 IU/L. The relationship between the LH concentrations obtained from different assays from the same pool of fractions of gel‐filtrated urine samples was analyzed using correlation and regression analyses. The nonparametric Spearman's test was used to check the statistical significance of the correlation. The Wilcoxon signed‐rank test was used to analyze the significance of the difference between the LH concentrations obtained by different assays from the same pool of fractions of gel‐filtrated urine samples.

3. Results

All the three forms of LH in urine corresponding to intact LH, LHβ, and LHβcf, represented by fractions 24, 27, and 31, respectively (Figure 1), were detected by Delfia and Immulite assays (Table 1). The modified Delfia assay, using an FSH tracer, confirmed the presence of intact LH at fraction 24, further establishing the detectability of LHβ and LHβcf in urine by both Delfia and Immulite assays (Figure 2). The other two assays were not able to detect all these different LH forms in FMV urine. The Cobas assay detected the LH‐ir derived from the fractions representing intact LH and LHβ, but not the LHβcf, whereas the Architect assay measured only the intact LH in urine (Table 1, Figure 1).

FIGURE 1.

Recovered concentrations of luteinizing hormone (LH), the beta‐subunit (LHβ) and its core fragment (LHβcf) that were detected in different fractions on gel filtration by the following assays: Delfia immunofluorometric assay (IFMA) (Wallac Oy, PerkinElmer Finland, Finland); Architect LH chemiluminescent microparticle immunoassay (CMIA) (Abbott, USA); Elecsys LH Cobas electrochemiluminescence immunoassay (ECLIA) (Roche, Switzerland) and Immulite 2000 LH immunochemiluminometric assay (ICMA) (Siemens, Germany). The readings at fractions 24, 27, and 31 correspond to the respective concentrations of intact LH, the LHβ, and LHβcf in urine.

TABLE 1.

Detection methods for luteinizing hormone immunoreactivity.

Assay/detected LH‐ir	Intact LH	LHβ	LHβcf
IFMA, Delfia LHspec	+	+	+
ICMA, Immulite LH	+	+	+
ECLIA, Elecsys LH Cobas	+	+	−
CMIA, Architect LH	+	−	−

Note: All or some of the different forms of luteinizing hormone (LH) immunoreactivity (LH‐ir), namely the intact LH, the beta‐subunit (LHβ) and its core fragment (LHβcf) could be detected by the following assays: Delfia immunofluorometric assay (IFMA) (Wallac, PerkinElmer Finland, Finland); Architect LH chemiluminescent microparticle immunoassay (CMIA) (Abbott, USA); Elecsys LH Cobas electrochemiluminescence immunoassay (ECLIA) (Roche, Switzerland), and Immulite 2000 LH immunochemiluminometric assay (ICMA) (Siemens, Germany).

FIGURE 2.

Recovered concentrations of luteinizing hormone (LH), the beta‐subunit (LHβ) and its core fragment (LHβcf) that were detected in different fractions on gel filtration by Delfia immunofluorometric assay (IFMA) (Wallac Oy, PerkinElmer Finland, Finland), a modified Delfia IFMA using FSH tracer and Immulite 2000 LH immunochemiluminometric assay (ICMA) (Siemens, Germany). The readings at fractions 24, 27, and 31 correspond to the respective concentrations of intact LH, the LHβ, and LHβcf in urine.

The Immulite assay was further compared with Delfia assays, employing both the original and modified protocols of the latter. This confirmed that the Immulite assay as well as the original and modified Delfia assays detected the intact LH in the same filtrate representing the fraction 24 (Figure 1). Furthermore, the Immulite as well as the original Delfia assays were able to detect also the degradation products of LH in urine, the LHβ at fraction 27, and U‐LHβcf at fraction 31 (Figure 1).

The LH concentrations measured by the Immulite and Delfia assays from the same pool of fractions of gel‐filtrated urine samples correlated strongly (r = 0.99, p < 0.001) (Table 2). There was no statistically significant difference between the LH concentrations measured by Immulite and Delfia assays from the same pool of fractions of gel‐filtrated urine samples (Table 3).

TABLE 2.

Correlation between the LH concentrations obtained by different assays from the same pool of fractions of gel‐filtrated urine samples.

	Correlation (r)	p
Immulite ICMA vs. Delfia IFMA	0.99	<0.001
Cobas ECLIA vs. Delfia IFMA	0.53	0.139
Architect CMIA vs. Delfia IFMA	0.49	0.177
Architect CMIA vs. Cobas ECLIA	0.99	<0.001
Cobas ECLIA vs. Immulite ICMA	0.38	0.117
Architect CMIA vs. Immulite ICMA	0.34	0.174

Note: Correlation between the LH concentrations obtained from all 20 individual fractions (19 to 38) of gel‐filtrated urine samples by different assays, namely the Delfia immunofluorometric assay (IFMA) (Wallac Oy, PerkinElmer Finland, Finland); Architect LH chemiluminescent microparticle immunoassay (CMIA) (Abbott, USA); Elecsys LH Cobas electrochemiluminescence immunoassay (ECLIA) (Roche, Switzerland), and Immulite 2000 LH immunochemiluminometric assay (ICMA) (Siemens, Germany). Bold values emphasise statistical significance.

TABLE 3.

Difference between the LH concentrations obtained by different assays from the same pool of fractions of gel‐filtrated urine samples.

	Z‐statistics	p
Immulite ICMA vs. Delfia IFMA	−0.561	0.575
Cobas ECLIA vs. Delfia IFMA	−2.310	0.021
Architect CMIA vs. Delfia IFMA	−2.666	0.008
Architect CMIA vs. Cobas ECLIA	−3.621	<0.001
Cobas ECLIA vs. Immulite ICMA	−1.328	0.184
Architect CMIA vs. Immulite ICMA	−3.724	<0.001

Note: Difference between the LH concentrations obtained from all 20 individual fractions (19 to 38) of gel‐filtrated urine samples by different assays, namely the Delfia immunofluorometric assay (IFMA) (Wallac Oy, PerkinElmer Finland, Finland); Architect LH chemiluminescent microparticle immunoassay (CMIA) (Abbott, USA); Elecsys LH Cobas electrochemiluminescence immunoassay (ECLIA) (Roche, Switzerland), and Immulite 2000 LH immunochemiluminometric assay (ICMA) (Siemens, Germany).

There was a statistically significant difference between the LH concentrations measured by the Architect and Cobas assays from the same pool of fractions of gel‐filtrated urine samples (Z = −3.621, p < 0.001); thus, the correlation between the LH concentrations determined by the Architect and Cobas assays indicates merely an association between these results (Table 2). In addition, we found significant differences among all other paired assay result comparisons, and as expected, no significant correlation was identified among them (Tables 2 and 3).

4. Discussion

We have shown that the Immulite assay can be used to detect the total LH concentrations in urine samples of prepubertal children presenting with low urinary concentrations of LH and its molecular components. Specific antibodies against LHβ or U‐LHβcf are currently not available; thus, it is not possible to directly detect the nonintact LH concentrations in urine using any commercially available assay. Thus far, we have determined the nonintact LH concentrations using an indirect method that was described previously [14]. Accordingly, the nonintact LH concentration was calculated as the arithmetic difference in the concentration between total and intact LH. In this study, it was possible to determine the intact LH concentrations in urine by a modified Delfia assay using the FSH tracer directed toward the common α‐subunit of LH. Interestingly, we observed that the Architect assay could well be used for determining intact LH concentrations in urine with no need to modify the assay.

Essentially, our study protocol was designed to account for the low concentrations of some LH forms due to gel filtration. This scenario is comparable to detection limitations confronted in pre‐ and peripubertal children, with assay sensitivity reaching as low as 0.01 IU/L for LH in immunofluorometric assays—a significant concern for potential clinical utility in prepubertal children [29]. To ensure that gel filtration processing yielded fractions with sufficient concentrations of degraded LH products, we utilized urine samples from a postmenopausal woman. This approach was crucial for demonstrating the feasibility of detecting various forms of degraded LH despite the concentration reductions caused by the gel filtration process, and it provides a strong foundation for further studies involving prepubertal and peripubertal children. Upon confirming the detectability of these LH forms, any sufficiently sensitive assay could be employed for detection, as additional processes that compromise the concentrations of various LH forms are not a concern in routine urinary LH determinations.

The observed patterns of measured LH concentrations in different fractions of the gel‐filtrated urine samples constitute the main pillar of the results of this study, showing which assays can be used for determining different components of urine LH‐ir. The findings of this study have contributed to seamlessly bridging a vital gap, preventing a return to the pre‐Delfia era when the detection of LH degradation products in urine was confined to specialized in‐house assays [12, 40] as commercially available tests primarily identified LH‐ir as monomeric and of a single variant [28].

Statistical analysis was used to confirm and further analyze our findings in paired comparisons of the assays in question. The confirmed strong correlation and the rejected difference between the U‐LH concentrations measured by Immulite and Delfia assays from the same pool of fractions of gel‐filtrated urine samples was an expected finding. On the other hand, the statistically significant strong correlation and simultaneously confirmed difference between the U‐LH concentrations obtained by the Architect and Cobas assays may seem contradictory at first glance. Nevertheless, this discrepancy can be explained by the fact that these assays functioned quite similarly in determining intact LH concentrations, with the single but significant difference being that the Cobas assay also detects the immunoreactivity emerging from the beta‐subunit. Consequently, the Immulite assay detected all the three forms of LH, and the Cobas assay was able to detect also the LHβ in addition to the intact LH, whereas Architect was able to detect only the intact LH.

The current study demonstrates that the Immulite assay can be a proper alternative to replace the Delfia assay, as both of them are able to detect the immunoreactivity originating from the intact LH as well as LHβ and LHβcf in urine. Once harmonized with Immulite in terms of equivalent reactivity and comparable concentrations for intact LH, the Architect assay can be used as a complementary method to determine the total concentration of LHβ and LHβcf, that is, the nonintact LH‐ir in urine, by calculating the arithmetic difference between total and intact LH concentrations. The Cobas assay, which yields intact LH‐ir results closer to those of Delfia, and LHβ immunoreactivity similar to that of Immulite, may be used in conjunction with the Immulite assay to determine the U‐LH concentrations derived from the LHβcf alone, subsequent to a similar harmonization and arithmetic difference calculations. In principle, clinical laboratories do not routinely produce parts of the same test using two different methods; therefore, the current method is likely to remain in use for research investigations only, until a unified method is developed that employs different specific antibodies to detect various degradation products of LH.

In conclusion, the Immulite assay stands out as one of the few commercially available tests capable of detecting all forms of urinary LH, a feature likely to be instrumental in developing noninvasive, practical, and scalable solutions for evaluating hormonal changes during minipuberty in neonates, during the onset of central puberty in prepubertal children, puberty‐associated disorders in adolescents, and the periovulatory fertility window in women, with a special focus on postpeak changes.

Author Contributions

All the authors have accepted responsibility for the entire content of this submitted manuscript and have approved submission.

Ethics Statement

The study complied with the World Medical Association Declaration of Helsinki regarding ethical conduct of research involving human subjects. The laboratory personnel working for this research study provided her own urine samples and gave informed consent for the use of the samples. Ethical committee approval (HUS/2426/2018) and research permit (HUS/54/2019) and have been granted by the Hospital District of Helsinki and Uusimaa.

Consent

Informed consent was obtained from all subjects (no minors involved).

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.