SEARCH

SEARCH BY CITATION

Keywords:

  • biochemical recurrence;
  • heterophile antibody;
  • immunoassay;
  • prostate specific antigen

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Case report
  5. Discussion
  6. References

Abstract:  Human antimouse heterophile antibodies (HAMA) are naturally occurring antibodies that can interfere with modern serum assays. We report a case of HAMA interference with a commonly used prostate-specific antigen (PSA) assay, leading to false elevation (2.17–2.46 ng/mL) after radical prostatectomy. Pre-operative PSA was 4.4 ng/mL, and final pathology was Gleason 3 + 3, pT2cNXMX. This markedly elevated postoperative PSA led to unnecessary imaging for metastasis and psychological distress to the patient. Direct measurement of HAMA in the patient’s serum yielded a value of 440 ng/mL (<74 ng/mL). An alternate PSA assay using goat detection antibody eliminated interference, with all values 0.05 ng/mL. When a patient’s PSA is inconsistent with the clinical scenario, one should consider immunological interference by HAMA in PSA assays.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Case report
  5. Discussion
  6. References

Since 1986, serum prostate-specific antigen (PSA) has been used for prostate cancer detection and follow up after therapy.1–3 Current assays apply serum onto monoclonal PSA capture antibodies (Ab) which have been immobilized on a solid phase, followed by detector Ab introduction (Fig. 1a).

image

Figure 1. (a) Normal prostrate-specific antigen (PSA) immunoassay, in the presence of analyte. (b) Heterophilic interference of PSA immunoassay, in the absence of analyte.

Download figure to PowerPoint

In this report, we present an unusual case of human antimouse heterophile antibodies (HAMA) causing spuriously elevated PSA after radical prostatectomy (RP).

Case report

  1. Top of page
  2. Abstract
  3. Introduction
  4. Case report
  5. Discussion
  6. References

During annual screening, a 67-year-old man’s PSA rose from 3.8 ng/mL (2004) to 4.4 ng/mL (2005), using the Bayer ADVIA Centaur (minimum detectable level 0.05 ng/mL) chemiluminescent method (Bayer Diagnostics, Tarrytown, NY, USA). His digital rectal exam revealed an unremarkable prostate without induration, measuring approximately 35 cc. He had well-controlled hypertension, and a brother with prostate cancer who died of metastatic disease. He was a retired salesman, without occupational or medical exposure to mice or humanized murine antibodies.

After appropriate counseling, biopsy demonstrated Gleason 3 + 3 adenocarcinoma in one of 12 cores. He underwent uncomplicated laparoscopic RP, and pathology demonstrated Gleason 6 cancer (pT2cNXMX) in 10% of the gland. There was a focal positive margin at the left apex but no extracapsular extension.

Surprisingly, his first postoperative total PSA (TPSA) was 2.25 ng/mL. A repeat TPSA was 2.17 ng/mL. A third TPSA 4 months after surgery was 2.46 ng/mL, and free PSA (FPSA) was 2.11 ng/mL. Each postoperative PSA was measured at our institution using the Hybritech PSA assay from Beckman Coulter (minimum detectable level 0.05 ng/mL, Beckman Coulter, Brea, CA, USA; Fig. 2).

image

Figure 2. Total PSA level was dependent on Beckman-Coulter reagent lot used in the assay.

Download figure to PowerPoint

He subsequently underwent bone scan and abdominopelvic computed tomography (CT), confirming only Paget’s disease in the left ilium. Prostascint images were focally positive in the left supraclavicular area but follow-up neck and chest CT demonstrated no lymphadenopathy or masses.

Immunoassay interference was suspected when TPSA was lower than FPSA. Re-testing the January 2006 sample led to FPSA of 2.11 ng/mL and TPSA of 0.57 ng/mL. Due to this illogical result (FPSA > TPSA), serial dilutions of the specimen were performed, and another bizarre finding was that higher FPSA levels were detected after dilution. A 1:2 dilution demonstrated FPSA of 2.20 ng/mL and TPSA of 0.35 ng/mL. A 1:4 dilution demonstrated FPSA of 2.12 ng/mL and TPSA of 0.24 ng/mL, respectively (Fig. 2).

Subsequently, the patient’s samples were sent directly to Beckman Coulter for further testing. When the sera were pretreated with proprietary HAMA blocking reagents A, B and C, TPSA measurements were 0.25 ng/mL, 0.96 ng/mL and 1.03 ng/mL, respectively (original TPSA 2.25 ng/mL). These specific blocking reagents are not commercially available, and they are proprietary mouse antigens that serve to adsorb HAMA in the serum.

Human antimouse heterophile antibody interference was finally confirmed by direct HAMA measurement using an enzyme immunoassay (Human Anti-Mouse IgG Abs EIA, Specialty Laboratories, Santa Monica, CA, USA). This yielded a value of 440 ng/mL, with the upper limit of normal being 74 ng/mL. Subsequently, all three postoperative sera were sent to another hospital for Bayer TPSA assay (which uses a goat detection antibody), and all TPSA values were 0.05 ng/mL, confirming specificity of HAMA interference by murine Ab alone.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Case report
  5. Discussion
  6. References

The postoperative PSA greater than 2 ng/mL was inconsistent with the low-risk features of this patient’s disease. This caused significant psychological distress to the patient, led to extensive radiographic studies to exclude metastasis, and was puzzling for the treating physician.

Fortunately, it was determined that HAMA was the reason for falsely elevated PSA after RP. This phenomenon occurs due to intrinsic Ab binding to murine Ab used in immunometric assays (Fig. 1b). This may lead to false positive test results even in the absence of analyte,4 and occurs for markers as diverse as human chorionic gonadotropin, cancer antigen 125 (CA-125), PSA and troponin I.5,6 While HAMA more commonly leads to false elevation in test results, falsely lower values can also occur if the interfering antibody blocks the detector Ab binding site.7

In one study evaluating the incidence of HAMA, eight commonly used immunoassays were tested in 500 serum samples. They found that the incidence of HAMA ranged 0.2–3.7%, with a 1.0–1.2% incidence for PSA.8 These Ab can occur naturally or, in some cases, can arise after exposure to murine or humanized murine Ab used for medical purposes.

In the clinical chemistry published work, this interference has been long recognized.5 However, an easy solution to detect and eliminate this interference has been elusive. Various proprietary or commercially available ‘blocking agents’ have been tested, but none are uniformly successful, as in our case.

One explanation is the polyclonal nature of HAMA, such that while some Ab can be blocked, others are not. Another difficulty is that HAMA are moving targets. That is, immunometric Ab and reagent lots constantly evolve over time, even for a given manufacturer. This is demonstrated clearly in Figure 2. Lots C and D demonstrated similar TPSA results, but TPSA was very much lower using lot F. Clearly, there was reagent lot-to-lot variability in the TPSA results due to the presence of HAMA that does not occur when HAMA is not present. Because differences in antibody affinity and avidity toward HAMA can occur between different lots of monoclonal and, especially, polyclonal antibody preparations, some reagents are more affected than others.

Finally, a serial dilutional method has been reported to detect and control for interference, but again, consistency has not been demonstrable. In this specific patient, serial dilution helped raise suspicion for HAMA interference (Fig. 3) because FPSA > TPSA is impossible. In Figure 3, differences in the assay design of the FPSA and TPSA tests, and differences in affinities across reagent lots, accounts for the relative stability of the FPSA value but a large decrease in TPSA. Second, there was non-linearity of FPSA values with serial dilution, whereas the expected result is linearly decreasing values with serial dilution. However, approximately 50% of the time, samples with HAMA will still dilute linearly.9 Confirmatory testing by measuring HAMA levels was required to make the final diagnosis.

image

Figure 3. Total PSA values with serial dilution of serum specimen from January 2006, using reagent lot C.

Download figure to PowerPoint

One of the widely accepted benefits of RP over radiation is undetectable postoperative PSA. This case illustrates that while such an outcome can be expected in the vast majority of cases, artifactual PSA elevations can occur. Luckily, our patient received no unnecessary adjuvant therapy, but a published work review reveals a much more serious consequence of HAMA. In that case, a 49-year-old patient with PSA 3.3, Gleason 6 disease underwent surgery, followed by radiation and androgen ablation for 22 months before HAMA was diagnosed.10

In summary, when a patient’s PSA does not match the clinical scenario, one should consider immunological interference in the PSA assay. HAMA levels can be directly measured, and tests based on goat detector antibodies can be used to monitor for biochemical recurrence.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Case report
  5. Discussion
  6. References