PTH Immunoassay Interference Due to Human Anti-Mouse Antibodies in a Subject With Obesity With Normal Parathyroid Function

Abstract

Context

Immunoassay interference has been most often found with prolactin measurement. However, only few data exist on immunoassay interference for other hormones.

Case Description

A 36-year-old woman with obesity (body mass index, 31 kg/m²) had regularly attended our endocrine unit for type 2 diabetes therapy. When she was included as a control subject in a study for obesity management, detailed laboratory testing was performed, including PTH. In the absence of clinical symptoms, she presented with normal calcium, phosphate, and vitamin D levels. However, the PTH levels were >5000 ng/L. These results were obtained using the Roche Elecsys electrochemiluminescence assay. Repeated measurements with this assay (mouse antibody) led to the same findings. However, using an Euroimmun assay (goat antibody), the exact PTH values were measured at 18.0 ng/L. After pretreatment with a heterophilic antibody blocking reagent, the results of the Roche assay had decreased to a normal level. This phenomenon was explained by the detection of human anti-mouse antibodies in the proband’s serum.

Conclusions

In cases of prolactin immunoassay interference, endogenous antibodies will bind to the hormone in vivo, resulting in complexes of a high molecular weight that are less efficiently cleared by the kidneys and, thus, accumulate in the blood. In contrast, the PTH values >5000 ng/L detected in our subject most likely had resulted from the specific interference of the human anti-mouse antibodies present in the proband’s serum with the assay antibody, resulting in artificial stimulation of the Roche assay detection system ex vivo.

In clinical endocrinology, immunoassay interference has been most often reported for prolactin measurements, with antibodies binding to the hormone and resulting in complexes of a high molecular weight (1). Usually, these macromolecular complexes, referred to as “macro-prolactin,” will not be functionally active and, thus, will not result in clinical symptoms of hyperprolactinemia (2). However, they will be measurable by immunoassays and, thereby, result in the detection of apparently high prolactin serum concentrations. The prevalence of macro-prolactin has varied from 9% to 42% and, thus, is a common phenomenon of which clinical endocrinologist must be aware (3). In contrast to prolactin, such antibody-mediated immunoassay interferences for other hormones have been rare.

Case Report

We have presented the case of a 36-year-old white woman who had regularly attended our endocrine unit for obesity treatment. Her initial body weight had been 132 kg in 2010 (body mass index, 46 kg/m²). She was undergoing treatment with a lifestyle intervention program in accordance with the guidelines of the German Obesity Association. This had resulted in body weight reduction to 90 kg in 2014, which she was able to stabilize until 2017. At her last visit in April 2019, she had presented with a weight of 94 kg. She also had type 2 diabetes, which had been treated initially with metformin and the DPP4 inhibitor saxagliptin. This pharmacotherapy was switched to the GLP-1 analog liraglutide in 2017 to support her weight stabilization. However, in addition to this dramatic reduction in her body weight in the interval from 2017 to 2019, insulin therapy had to be initiated with insulin aspart (prandial) and insulin glargine (basal) resulting in an HbA1c of 6.2% (44 mmol/mL) in April 2019. Antibodies against glutamate-decarboxylase and antibodies against islet cell anti-gene were negative on repeated measurements. However, the course of her diabetes progression suggested progressive β-cell failure in addition to insulin resistance.

Because of the positive response to the obesity therapy program, she had been included as a control subject in an observational study on the potential role of the microbiome in weight maintenance. Because in a previous study we had found that genetic polymorphisms in the human vitamin D receptor influence the composition of the microbiome (4), in that study, we were also interested in characterizing the vitamin D metabolism. Thus, we were also measuring PTH. In April 2019, her PTH serum concentrations were measured at >5000 ng/L (normal range, 15.0 to 65.0; Table 1). This measurement was repeated twice, with the same result. Her laboratory measurements for creatinine (66 µmol/L; normal range, 45 to 84), calcium (2.34 mmol/L; normal range, 2.15 to 2.50), albumin-corrected calcium (2.30 mmol/L; normal range, 2.15 to 2.50), phosphate (1.12 mmol/L; normal range, 0.87 to 1.45), blood urea nitrogen (5.30 mmol/L; normal range, 2.76 to 8.07), albumin (43 g/L; normal range, 35 to 53), α1-globuline (2.52 g/L; normal range, 0.90 to 2.60), α2-globuline (10.0 g/L; normal range, 4.4 to 10.3), β-globuline (8.49 g/l; normal range, 5.00 to 11.20), γ-globuline (9.41 g/L; normal range, 4.40 to 13.90), sodium (141 mmol/L; normal range, 136 to 145), and potassium (3.84 mmol/L; normal range, 3.40 to 4.40) were all in the normal range. Serum electrophoresis revealed no pathological features. In addition, we measured the biomarkers for bone metabolism and detected normal levels for β-crosslaps and N-MID-osteocalcin serum concentrations. In a 24-hour urine sample, we found normal calcium and creatine excretion. Ultrasound examination of the neck did not show any typical lesions for parathyroid adenoma.

From an immunological viewpoint, she had no history of immunological disease, and antinuclear antibodies were absent. However, we detected human anti-mouse IgG antibodies with 1159 ng/mL (normal range, <40). Having obtained these results, we speculated that an interference with the immunoassay might have been the cause of our patient’s dramatically elevated PTH levels. To determine whether other hormones could also have been affected, we measured TSH and prolactin but found normal levels for both hormones.

The initial measurement for PTH had been performed using the Roche Elecsysy electrochemiluminescence assay (no. 07251068190). PTH values of >5000 ng/L were found in the measurement of both plasma and serum. In addition, we performed a PTH measurement using a different immunoassay (Euroimmun PTH ELISA; no. EQ 6421-9601). The analysis revealed a normal PTH value of 18 ng/L (range, 11 to 55), indicating a specific interference for the Roche immunoassay. A polyethylene glycol precipitation before the use of the Roche assay resulted in a PTH value of 70.11 ng/L, again suggesting immunoassay interference. To further clarify this finding, pretreatment with a heterophilic antibody blocking reagent (no. 3IX762; Scantibodies Laboratory) was performed, resulting in detection of PTH values within the normal range using the Roche assay (27.14 pg/mL; normal range, 15 to 65). Subsequently, we were able to detect human anti-mouse IgG antibodies in the proband’s serum, revealing a mechanistic explanation for the interference, because the antibody in the Roche assay is of murine origin and the antibody in the Euroimmun assay originates from goat.

Discussion

The findings from the present case have clearly indicated that interference of immunoassays resulting in apparently high hormone concentrations, not only occurs for prolactin, but also in the measurement of PTH. An asymptomatic elevated PTH level from immunoassay interference has also been reported by Prodan et al. (5). However, in their study, they only reported a mild increase in the PTH levels of approximately two to three times the elevated serum concentrations (neat sample, 28.5 pmol/L; after polyethylene glycol treatment, 10.2 pmol/l) (5). In contrast, in our patient, we found elevated PTH concentrations far greater than the normal range, with a more than 100-fold apparent increase. This dimension is more comparable to the results usually found in macroprolactinemia (1). In 2015, a case was reported in which immunoassay interference was found for several hormones simultaneously, including prolactin, TSH, ACTH, PTH, FSH, and β-human chorionic gonadotropin (6). In that report, the PTH concentrations were measured at 566 pg/mL with an Abbot immunoassay and 44 pg/mL with a Roche immunoassay. Because of these findings, we also performed measurements for TSH and prolactin and found that the interference in our patient was specific for PTH.

From a mechanistic viewpoint, our findings suggest a different type of interference for the apparently elevated PTH levels found in our proband. In the case of macroprolactin, endogenous antibodies will bind to the hormone in vivo, resulting in complexes of a high molecular weight, which are less efficiently cleared by the kidneys and, thus, accumulate in the blood (2). In contrast, the PTH values greater than 5000 ng/L detected in our subject were most likely not due to endogenous macromolecular complexes but, rather, a specific interference of the human anti-mouse antibodies (HAMAs) present in the proband’s serum with the antibody of the Roche detection system, such that the ex vivo binding of the HAMAs to the mouse antibody of the assay induced an artificial activation.

In conclusion, our data have indicated that immunoassay interference can occur in PTH measurements owing to the presence of HAMAs and that in such cases, an analysis should be performed using an assay with a different detection system.

Acknowledgments

We gratefully thank Dr. V. Herbst from Euroimmun and Dr. S. Pitz from Roche Diagnostics for their analytical support.

Financial Support: The present study was supported by a grant from the Bundesministerium für Bildung und Forschung [grant 01ZX1306A (to M.L.)].

Data Availability: All data generated or analyzed during this study are included in this published article or in the data repositories listed in References.

Abbreviation:

Abbreviation:
 
• HAMA

  human anti-mouse antibody

References and Notes