Case report: False-positive B-type natriuretic peptide results with the cyclic amplified fluorescence immunoassay

Abstract

B-type natriuretic peptide, which was initially isolated from porcine brain tissue, has been used in the clinic as an important indicator for the diagnosis, risk stratification, and prognosis of heart failure. In recent years, many commercial immunoassays to assess B-type natriuretic peptide have been reported to undergo interference due to various factors. At present, the commonly used assays to assess B-type natriuretic peptide are enzyme-linked immunosorbent assay and direct chemiluminescence, which are based on a “sandwich structure” formed by dual antibodies; both assays have been reported to undergo interference. We report the case of a patient whose B-type natriuretic peptide results, which were detected using a cyclic amplified fluorescence immunoassay, were false-positive. The patient had a low-risk pulmonary embolism without right heart failure; however, her B-type natriuretic peptide results were surprisingly high. Further study confirmed that the abnormal B-type natriuretic peptide result was caused by the interference of the human anti-mouse antibodies. Therefore, when encountering B-type natriuretic peptide results that are not consistent with the clinical situation, the possibility of interference in the assay should be considered.

Background

B-type natriuretic peptide (BNP) is the primarily recommended biomarker used for the diagnosis of heart failure (HF) ¹ ; however, it should always be used in combination with other clinical information. In addition to HF, any disease that can result in ventricular stretching, increase in wall tension, or a reduction in BNP excretion can result in an increase in BNP, such as myocardial hypertrophy, arrhythmia, pericardial effusion, acute coronary syndrome, pulmonary embolism (PE), pulmonary hypertension, renal insufficiency, and so on. ² BNP can also be used in the risk stratification of PE ³ ; the higher the level of BNP, the higher the risk stratification.

Researchers have found that many results of assays to assess endogenous substances, such as tumor markers, endocrine hormones, infectious markers, and myocardial markers (troponin, myoglobin, creatine kinase-MB, etc.), can undergo interference due to multiple factors, including heterotropic antibodies (HAs) and rheumatoid factor (RF), leading to false-positive results.^4,5 A false-positive result for a cardio-specific marker carries a high risk of jeopardizing clinical decision-making. The type and degree of interference may vary with different immunoassays. New forms of interference are also emerging with the development of new diagnostic methods and therapies.

We report the case of a patient who neither had right HF nor any of the above other diseases, aside from a low-risk PE. The results of BNP were surprisingly high when assessed using a cyclic amplified fluorescence immunoassay (CAFIA; Xingtong Pylon IRIS analyzer), yet normal when assessed using a direct chemiluminescence assay (Siemens ADVIA Centaur BNP analyzer). Therefore, it was considered that there were interferences in the CAFIA. Finally, using a sample interference screening test, we confirmed that the interference may have been caused by the high concentration of human anti-mouse antibodies (HAMAs) in the patient’s samples. The reporting of this study conforms to the CARE guidelines. ⁶

Case presentation

A female patient in her early 60s had a cough following COVID-19, which was accompanied by expectoration and fever (the highest temperature was 37.8°C). She was hospitalized at a local hospital with an elevated BNP measurement of 4002 ng/L (reference range: 0–100 ng/L, Xingtong Pylon 3 D analyzer) and subsequently received anti-HF treatments, including cardiotonic agents and diuresis. However, the patient showed no significant improvement.

The patient had experienced paroxysmal chest tightness for 1 week, which was aggravated after activity, accompanied by palpitations and dry cough; therefore, in December 2022, she went to Jinhua Hospital of Zhejiang University for further treatment. There was no special medical history or vaccinations. Her heart rate was 124 beats per minute, with no other positive physical signs. Laboratory examination showed elevated levels of D-dimer (1523 μg/L; reference range: <500 μg/L), BNP (>5000 ng/L; reference range: 0–100 ng/L, Xingtong Pylon IRIS analyzer), leukocytes (11.17 ×10⁹/L; reference range: 3.5–9.5 × 10⁹/L), and neutrophils (83.6%; reference range: 40%–75%). The other markers (hemoglobin, blood lipids, bilirubin, creatinine, thyroid hormone, etc.) were all normal. Each laboratory value presented in the manuscript represents the mean value calculated from three independent observations, and there were no statistically significant differences among these three replicates. Electrocardiogram indicates sinus rhythm and normal electrocardiogram. Transthoracic echocardiography showed that the mitral and tricuspid valves were slightly regurgitated, and cardiac function was normal. Coronary angiography found that the proximal end of the anterior descending branch was narrowed by 30%; there was no evidence of stenosis in the other vessels. Computed tomography pulmonary angiography showed a slight PE in the pulmonary branch of the lower lobe of the right lung (Figure 1(a)), with a low level of infection in both lungs (Figure 1(b)). There was no evidence of right HF or pulmonary infarction.

Figure 1.

Computed tomography pulmonary angiography. (a) Sagittal image shows slight pulmonary embolism in the pulmonary branch of the lower lobe of the right lung and (b) cross-sectional image shows a low level of infection in both lungs.

After obtaining the patient’s consent, she was treated with oral rivaroxaban 15 mg twice a day, metoprolol succinate 23.75 mg once a day, intravenous levofloxacin 100 mL once a day, and intravenous dexamethasone 5 mg once a day. After a week, her symptoms had obviously improved; hence, she no longer took all above medications except rivaroxaban, but BNP result remained >5000 ng/L. Subsequently, the BNP level was reexamined three times within 1 month; the results were 3407 ng/L, 3207 ng/L, and >5000 ng/L, respectively, all of which were detected using CAFIA (Xingtong Pylon IRIS analyzer). However, the patient had no symptoms of cardiac dysfunction, such as chest tightness, dyspnea, and limitations in activity. Although the patient had a slight PE, it was a low-risk type; thus, it did not affect right ventricular function. Therefore, the BNP results were not consistent with the clinical symptoms. We suspected interference had occurred with our BNP assays. We assessed the BNP using the CAFIA again, and the result was 3481 ng/L. At the same time, the same sample underwent BNP assessment using the direct chemiluminescence assay (Siemens ADVIA Centaur BNP analyzer), and the result was 26 ng/L, which was within the reference range. Therefore, it was considered that the BNP results were false-positive, and further investigation was required to determine the interference.

There was no hemolysis and lipemia in the samples, and the addition of a high concentration of heterophilic antibody-blocking reagent (HBR) did not avoid the interference. The sample was subsequently screened for interference: 1 mg/mL polymer, 100 μg/mL sheep IgG, and 400 μg/mL IgM were added, respectively, and the results were consistent with the original sample. Therefore, interference of polysaccharide polymer antibodies, human anti-sheep antibodies, and RF was excluded. However, when 100 μg/mL mouse IgG was added, the detection result was 2469 ng/L, which was significantly lower than that of the original sample. In addition, 300 μg/mL and 500 μg/mL mouse IgG were subsequently added, and the BNP results were 1727 ng/L and 993 ng/L. Therefore, the interference was considered to be due to the presence of HAMAs in the sample. The patient denied a definite history of a rat bite, blood transfusion, and so on.

In this case, the capture antibodies of BNP determined by CAFIA were mouse anti-human antibodies. When there is a high titer of HAMAs in the sample, it can still bridge the two antibodies in the absence of antigens, thus interfering with the double antibody sandwich immunoassay system, resulting in false-positive results (Figure 2(a) to (c)). Therefore, the reason for the false-positive BNP results in this patient was considered to be the concentration of anti-mouse IgG in the sample, which exceeded the concentration range blocked by the reagent.

Figure 2.

(a) True positive: Antigen binds to the labeled antibody (Ab) and captures Ab without human anti-mouse antibodies (HAMAs) interference. (b) False-positive: In the absence of the antigen, HAMAs bind to the labeled Ab and capture Ab, resulting in a false-positive result and (c) False-negative: HAMAs bind to the labeled Ab and capture Ab, and interfere with the binding of the target antigen, resulting in false-negative result.

Discussion

Assays to detect the level of BNP are frequently used to establish the presence and severity of HF. A substantial evidence base supports the use of natriuretic peptide biomarkers to exclude HF as a cause of symptoms in ambulatory and emergency department settings. ⁷ In our case, by carefully evaluating her symptoms and the results of examinations such as echocardiogram, we were able to rule out the possibility that the elevated BNP level was attributable to HF. Simultaneously, through a thorough review of the patient’s routine examinations during hospitalization, we excluded a series of non-HF conditions that could potentially lead to an elevated BNP level. These conditions included renal insufficiency, atrial fibrillation, the influence of advanced age, hyperthyroidism, the presence of inflammation and tumors, as well as the impact of medications. ⁸ Given the abovementioned exclusions, we reasonably considered the existence of interference factors in the BNP test results. Interferences may induce false-positive, false-negative, or both results, which can drive unnecessary explorations, inappropriate treatments, or a missed diagnosis, increasing the economical and psychological burden of patients, and even causing medical disputes. Thus, clinicians should be aware of these interferences.

The factors that interfere with immunoassays can be divided into exogenous and endogenous factors. Exogenous factors mainly include hemolysis, contamination, lysozymes, etc., ⁹ and endogenous factors include RF, autoantibodies, biotin, human anti-animal antibodies (HAAAs), HAs, and so on. ¹⁰ In daily laboratory practice, HAs are used as a broad term, which basically includes two specific circumstances. The so-called “true” HAs are defined as antibodies produced against ambiguous antigens, are generally weak and with multispecific activities. Conversely, HAAAs usually develop after exposure to animal immunoglobulins, are generated against well-defined antigens, and are characterized by strong avidities. Some researchers have used the term “HAs” to refer to any endogenous interfering antibodies, regardless of the mechanism. The first report of antibody interference was published in the early 1970s. ¹¹ Almost 50 years later, new interference reports are still being submitted from all over the world.^12,13

The incidence of HAAAs is 30%–40% in patients and 3%–15% in normal subjects.^14,15 HAMAs are of particular importance as murine antibodies are most commonly used in commercial immunoassays, and mice are ubiquitous in the environment. Other antibodies include human anti-rabbit antibodies, human anti-sheep antibodies, human anti-bovine antibodies, human anti-pig antibodies, and mixed specific antibodies. Furthermore, HAAAs can be of the IgG, IgM, or IgA isotype. Although they bind mostly to the Fc region of the assay antibodies, there are reports of these interfering antibodies binding to other parts of the assay antibodies (such as the idiotope or the hinge region). ¹⁶ The causes of the presence of HAAAs include both iatrogenic and non-iatrogenic factors. Iatrogenic factors include blood transfusion, immunization, animal-derived proteins, thymosin, target antibody contrast agents, target antibody drugs, etc. Non-iatrogenic factors include mother-to-child transmission, history of accidental or occupational contact with animal proteins (such as veterinarians, pastoral areas, food processors, and pet breeding), multiple births, intake of milk and animal protein foods by patients with selective IgA deficiency, etc. ¹⁷ The stimulation of an individual by foreign protein antigens (such as mice, cattle, sheep, rabbits, pigs, etc.) triggers the immune response, which can lead to the production of anti-animal protein antibodies. ¹⁸ HAMAs can persist in the blood for months to years. The degree of interference caused by it varies from person to person and will also change over time.

At present, commonly used immunoassays to assess BNP include the enzyme-linked immunosorbent assay and direct chemiluminescence assay. They are all based on a double antibody “sandwich structure” consisting of two monoclonal antibodies or a complex of monoclonal and polyclonal antibodies. In addition, as the BNP reagents from different manufacturers are likely to be directed against different epitopes and potentially use antibodies originating from different animal species, the differences observed between BNP assays can range from 15% to 50%. ¹⁹ When HAMAs exist in the human body, in addition to forming antigen–antibody complexes, they will also form HAMAs-mediated capture antibody/labeled antibody complexes, giving false-positive results. If HAMAs only bind to capture antibodies or labeled antibodies, this reduces the number of antigen-binding sites to be tested in the sample, resulting in false-negative results.

In this case, the Pylon 3D and Pylon IRIS BNP detection analyzers are employed. These analyzers are found on an independently developed cyclic-enhanced fluorescence immunoassay platform and utilize a novel single-epitope sandwich assay technology. This innovative technique functions by specifically binding a single-epitope capture antibody to the stable cyclic structure of the BNP molecule. This binding event leads to the formation of a highly stable immune complex. Subsequently, a labeled antibody is introduced to bind with this complex. This sequential process effectively mitigates the influence of BNP hydrolysis on the test outcomes. Notably, the capture antibody utilized in this system is a mouse-anti-human antibody. In samples containing high-titer HAMAs, a concerning phenomenon can occur. Even in the absence of the antigen, these HAMAs can act as a bridge between the two antibodies in the assay, thereby resulting in false-positive results. In the analysis of this particular sample, attempts to circumvent the interference through sample dilution and the addition of a higher-concentration HBR-blocking agent proved unsuccessful. However, when mouse IgG was added, a remarkable decrease in the BNP level was observed. This indicates that the interfering substance within the sample is likely HAMAs. Upon further investigation, the patient firmly denied any definite history of blood transfusion, rat bites, or other relevant exposures. Consequently, it is highly probable that the presence of HAMAs in the patient’s body is of innate origin. Turning to the ADVIA Centaur BNP analyzer, although it also employs a mouse-anti-human antibody as the capture antibody, it diverges in terms of the BNP epitopes it targets. One of its antibodies is designed to recognize the epitope within the cyclic region, which represents the active form of BNP. The other antibody is specific for the C-terminal of BNP. In scenarios where the epitope targeted by the capture antibody is either masked on the HAMA present in the sample or exhibits weak immunoreactivity, the analyzer may be unable to effectively detect the HAMA. As a result, false-positive results are avoided.

There have also been individual reports of false-positive BNP results in the past. In 2007, Collin-Chavagnac et al. ²⁰ reported, for the first time, a 78-year-old male patient with a false-positive BNP result due to an increase in the monoclonal antibody IgM-kappa. Pan et al. ²¹ reported a case of a false-positive BNP result using the Abbott Architect chemiluminescence assay (Abbott AxSYM BNP Analyzer), which was confirmed to be caused by HAMAs. Janssen et al. ²² reported a false-positive BNP result possibly caused by a macro-BNP, which was only immunoreactive in the Abbott Architect BNP immunoassay. In a recent research endeavor, scholars reported on an elderly female patient with strikingly elevated plasma levels of both BNP and NT-proBNP. Their investigations clearly revealed that the pronounced elevation was due to macro-proBNP, an immune complex formed by the combination of proBNP and autoantibodies against it. ²³ Regrettably, in our study, we failed to conduct additional tests to exclude the potential interference from macro-proBNP. In the subsequent work, we will make the research complete and more refined.

When interference is suspected in BNP assay results, it can be confirmed by the following methods: 1. There are currently many kinds of HAMAs testing kits on the market, which can directly prove the existence of HAMAs. 2. Perform dilution and recovery tests on the samples; most interference due to HAMAs is nonlinear, which is mainly due to its high specificity and affinity. ²⁴ 3. Use another assay with different epitopes of BNP or serum antibodies from another animal; if there is a significant difference between the two assays, this indicates that there are interferences. 4. Application of antibody-blocking reagent; pretreat the samples with HBR or use a heterotopic antibody-blocking tube (HBT), and if there is a significant difference in the results before and after pretreatment, there is interference.

To mitigate HAMAs interference in assays, methods such as sample pretreatment (heating, adding HBR/HBT or polyethylene glycol precipitation, etc.) and HAMAs structural modification have been explored. However, no single test can identify or rule out HAMAs interference. Immunotherapies for cancers and autoimmune diseases are advancing, increasing antibody-related interferences as therapeutic antibodies can trigger HAMAs production. When treatment monoclonal antibodies are not fully humanized, HAAAs detection probabilities vary: 90% for mouse monoclonal, 40% for chimeric, and 10% for humanized ones. ²⁵ Clinicians should suspect interference when BNP is high but manifestations or test results do not suggest cardiac insufficiency. Laboratory specialists should acknowledge immunoassays’ limitations and communicate with clinicians. Reagent manufacturers should reduce HAMAs interference in sandwich immunoassay kits and note it in the manual.

Conclusion

In this patient with low-risk PE, the BNP results obtained using the CAFIA were false-positive, which was caused by the high concentration of HAMAs and had nothing to do with pathological conditions in the patient, such as cardiac insufficiency. When clinicians and laboratory specialists encounter obvious discrepancies between BNP results and patients’ clinical conditions, they need to be further evaluated in combination with other relevant examinations, and the possibility of interference in the immunoassays should be considered.