Return to the hCG Reference Service HOME PAGE

In: Gestational Trophoblastic Diseases, 2nd edition (eds: Hancock, Newland, Berkowitz, Cole), web publication, in press, 2002.
Trophoblast Disease Update, in press 2002.

Use of hCG Tests for Evaluating Trophoblastic Diseases: Choosing an Appropriate hCG Assay, False Detection of hCG, Unexplained Elevated hCG, and Quiescent Trophoblastic Disease.

Laurence A Cole, PhD. 

USA hCG Reference Service,
Department of Obstetrics and Gynecology, University of New Mexico, Albuquerque, New Mexico, 87131, USA

INTRODUCTION

          hCG is a glycoprotein hormone composed of 2 dissimilar subunits, a- and  b-subunit, held together by charge and hydrophobic interactions. hCG is an extraordinary glycoprotein with 8 sugar side chains (oligosaccharides). This accounts for as much as 35% of the composition or molecular weight (MW). The combination of multiple subunits and multiple N-linked and O-linked oligosaccharide side chains causes significant heterogeneity in hCG structure. hCG, free subunits, degraded molecules, molecules with irregular N- and O-linked oligosaccharide side chains, and fragments are present in serum, urine and other bodily fluids. They are detected in pregnant women, in women with trophoblastic diseases, in men with testicular and women with ovarian germ cell tumors, and in some cases of men and women with other non-trophoblastic non-germ cell neoplasms.

In addition to regular hCG, 5 major variants of hCG are present in serum samples: hyperglycosylated hCG; nicked hCG; hCG missing the b-subunit C-terminal peptide; free b-subunit. Many other variants (i.e. hyperglycosylated free b-subunit, nicked hCG missing the b-subunit C-terminal peptide) are also present in serum samples.  The same large mixture of molecule plus urine b-core fragment are detected in urine samples. Table 1 summarizes the structure of the key hCG-related molecules. These vary in size from a molecular weight of 9,500 (b-core fragment) to approximately 40,000 (hyperglycosylated hCG).

Table 1. Structure of hCG-related molecules detected, to different extents, by commercial hCG immunoassays in serum and urine samples (1-9).

¶ Combinations of modification are present in serum and urine, such as hyperglycosylated nicked hCG, hyperglycosylated hCG missing the b-subunit C-terminal peptide, nicked free b-subunit, hyperglycosylated free b-subunit, nicked-hyperglycosylated free b-subunit and free b-subunit missing the C-terminal peptide.

          Over 40 different professional laboratory serum hCG tests are sold in the USA for quantifying serum hCG. These and many other tests may also be used to quantitatively assess urine hCG (i.e. positive test when urine concentration >20 IU/L). Almost all of these tests work through the multi-antibody “sandwich assay” method using labeled-enzyme or chemilumiscense detection. Few if any assays are sold today using the older competitive radioimmunoassay (RIA) method which was developed in the nineteen fifties.

          All hCG tests use at least one antibody directed against the b-subunit. This has led to the commonly used term “bhCG test.” Some tests detect hCG only and are called intact hCG, or simply hCG tests. Other tests hCG, its free b-subunit and possibly other degradation product. These tests are called total hCG tests, bhCG tests, or again can just be called hCG tests. The names are confusing. For the purpose of this publication we will refer to them all as hCG tests.

          All modern professional laboratory hCG tests use a combination of antibody to different sites on hCG. Commonly tests use an antibody to one site on the core of the b-subunit. A second antibody is then directed to an alternate site on the core of b-subunit, on the b-subunit C-terminal peptide, on the hCG dimer or subunit interface, on free subunits, or on the a-subunit. Because of these variations in antibody combination, different commercial hCG tests may measure very different mixtures of hCG-related molecules. Some tests may detect hCG only, others may detect all major hCG-related molecules. This may not be a problem for monitoring pregnancy in serum samples, 7 weeks of gestation until term, since hCG is consistently the principal molecule present (Table 2). It may, however, be a major problem in monitoring patients with trophoblastic diseases and non-trophoblastic malignancies. In such cases, different hCG variants may be the principal molecule present (Table 2).

          This article considers all aspects of hCG detection in trophoblastic diseases and other malignancies. It advises on the use of hCG tests to ensure that your laboratory is using an appropriate hCG test for trophoblast disease and non-trophoblastic malignancy applications. It examines the problems of variable sensitivity and of false positive hCG results. It describes how to avoid false positive results. The experience of the USA hCG Reference service is also presented: problems found with the sensitivity of specific hCG tests; false positive hCG results; and the complications of “quiescent gestational trophoblastic disease” and “unexplained elevated hCG”, two recently defined hCG-producing disorders.
 

Table 2. Occurrence of hCG-related molecules in serum and urine samples (1-9). Symbols refer to: principal component of hCG immunoreactivity (+++), major but secondary source hCG immunoreactivity (++), minor source of hCG immunoreactivity (+), significant source of hCG immunoreactivity in a small proportion  cases (±), and not normally present (- -).
 

¶ Combinations of modification are present in serum and urine, such as hyperglycosylated nicked hCG, hyperglycosylated hCG missing the b-subunit C-terminal peptide, nicked free b-subunit, hyperglycosylated free b-subunit, nicked-hyperglycosylated free b-subunit and free b-subunit missing the C-terminal peptide.

§ Non-trophoblastic malignancies such as reproductive tract, ovarian germ cell, gastrointestinal, bladder or lung cancer.
 

HCG IMMUNOCHEMISTRY

          Hyperglycosylated hCG is the predominant form of hCG produced in the three weeks following implantation in pregnancy (Table 2). During the first, second, third and subsequent weeks after implantation, hyperglycosylated hCG is gradually replaced with hCG; hyperglycosylated hCG accounting for >80%, 63%, 50% and 25 to <1% of hCG-forms, respectively (1-3). Hyperglycosylated  hCG is produced by invasive cytotrophoblast cells in early pregnancy (2). Hyperglycosylated hCG is produced by similar invasive cytotrophoblast cells in choriocarcinoma, and as such is also the principal hCG-related molecule produced in invasive gestational trophoblastic disease and in women with choriocarcinoma. In a blind test of 9 commercial hCG assays, all tests appropriately recognized hCG and hyperglycosylated hCG (Table 3). While the detection of hyperglycosylated hCG is probably not a problem in the choice of hCG immunoassay, measurement of hyperglycosylated hCG may potentially be useful in distinguishing invasive and non-invasive gestational trophoblastic disease (4).

          Circulating hCG (from hydatidiform mole) and hyperglycosylated hCG (from invasive trophoblast disease or choriocarcinoma) commonly becomes nicked as levels diminish after therapy (5,6) (Table 2). hCG and hyperglycosylated hCG are nicked or cleaved at b43-44, b44-45 or b47-48. This is a major a-subunit:b-subunit hydrophobic and charge interaction or linkage point.  As such, nicking leads to rapid dissociation of molecules (7), releasing the nicked free b-subunit (8). In these trophoblast disease cases, when hCG values fall below 100 mIU/ml, nicked hCG and free b-subunit often become the major or even sole sources of hCG immunoreactivity in serum (Table 2) (5,6). Cases have been reported in which a recurrence of invasive disease has been completely missed by use of an assay that does not detect nicked hCG (5). False negative results have also been observed using assays that do not detect free b-subunit or nicked hCG. It appears that measurement of these molecules is essential for accurately monitoring hCG levels until they become undetectable, or reach background (<2 IU/L) concentrations. It is also essential for demonstrating that background hCG and related molecule immunoreactivity remains undetectable and does not rise. As shown in Table 3, 2 of 9 common hCG tests in blind studies failed to detect or poorly detected nicked hCG. Four of 8 or half of common commercial immunoassays either poorly detected or failed to detect free b-subunit. Overall half of tests did not adequately detect nicked hCG and free b-subunit, so may not be appropriate for trophoblast disease applications. Similarly, as shown in Figure 2, nicked free b-subunit is commonly the sole hCG-related molecule present in serum samples of individuals with non-trophoblastic tumors, germs cell, bladder, gastrointestinal or lung malignancies. Five of the 9 assays evaluated are also probably inappropriate for monitoring hCG-related tumor markers.

          In two independent studies we have observed hCG missing the b-subunit C-terminal peptide in trophoblast disease patient serum (4.9). In the first study, we evaluate 10 serum samples from women with complete hydatidiform mole. In one case, the hCG result was 3,500 IU/L and 3,800 IU/L in two tests requiring the b-subunit C-terminal peptide to be present. The same sample gave results of 77,000 to 201,000 IU/L in 8 other immunoassays (9). In a more recent study, 5 of 76 cases of complete mole and choriocarcinoma were shown to have demonstrably lower hCG values in 3 of 9 assays requiring the b-subunit C-terminal peptide to be present (4). Taking the two independent studies together, 6 of 86 cases (about 1 in 14) required detection of hCG missing b-subunit C-terminal for meaningful management of trophoblastic disease. Tests using hCG b-subunit C-terminal antibodies do not detect hCG missing b-subunit C-terminal, and may yield misleading results or may miss persistence or recurrence of trophoblastic disease. Unfortunately, about one half of commercial laboratory serum hCG assay used in the USA today use an antibody against the b-subunit C-terminal peptide. As shown in Table 3, 6 of 9 tests, blindly evaluated, either poorly detected or failed to detect hCG missing the b-subunit C-terminal.

Table 3. Use of common brands of hCG immunoassays to detect hCG metabolic products commonly found in individuals with trophoblast disease or choriocarcinoma. All standards were calibrated by amino acid analysis, and converted into the international units, based upon molecular weight and molar equivalents of hCG. Samples were coded and tested blindly. Antigens appropriately detected are indicated  as “++”, those weakly detected (<150 IU/L or <75% of calibrated concentration) are indicated with as "<” and those extremely weakly detected or not detected at all (<50 IU/L or <25% of calibrate concentration) are indicted as “- -“. NE means not evaluated. Immunoassays are sorted in order of recognition of antigens.

          Urine hCG measurements are very useful in the management and diagnosis of trophoblastic diseases. They are useful in management, in that patients can readily collect and ship samples to clinical laboratories for long term monitoring of trophoblastic diseases.  Such a system is used by the Trophoblast Disease Center at Charing Cross Hospital in England. As discussed later in this article, some women are erroneously treated for gestational trophoblastic diseases because of false positive serum hCG assays (4,10-12). The interfering substance that causes false positive tests is only present in serum (4,10,12). Publications from our laboratory (10,11), and literature provided by certain hCG test manufacturers recommend the use of a urine tests to confirm the reality of serum hCG results. As such, urine hCG measurements have a clear role in diagnosis of trophoblastic diseases. Hyperglycosylated hCG and nicked hCG rapidly dissociate in serum releasing free b-subunit or nicked free b-subunit. This is further degraded in the kidney to urine b-core fragment. In most cases, urine b-core fragment is the principal hCG immunoreactive molecules in trophoblastic disease patient urine. For this reason use of a quantitative urine test that detects b-core fragment may be important. Only one of the assays evaluated here detected urine b-core fragment (Table 3).

          Quantitative urine hCG measurements are not widely available in the USA. This is not because of the lack of commercial tests to accurately measure hCG in urine, but because professional hCG tests sold in the USA have only been government licensed for qualitative applications or for Yes/No answers with urine samples. This does not mean that most professional laboratory hCG tests cannot be used for quantitative urine measurements, but rather that they are not a certified for this purpose. In fact, many common performed hCG test applications are not approved, including monitoring patients with trophoblastic diseases, use as a tumor marker, or use in Down syndrome screening. The only approved quantitative application is serum pregnancy testing. Quantitative urine measurements of hCG in cases with trophoblastic diseases are in many ways like quantitative serum measurement of hCG in cases with trophoblastic diseases, classified as experimental applications, yet are still carried out.

          It is concluded that hCG-related molecules may vary greatly in structure and size. In cases of trophoblast disease or non-trophoblastic malignancies, any one of the variants may constitute the sole form of hCG in serum or urine samples. It is important to talk with your laboratory and make sure that they are using an appropriate test which detects all the pertinent hCG-related molecules, before submitting samples for monitoring patients with trophoblastic diseases or other malignancies, or before making important decisions from the results.


THE EXPERIENCE OF THE USA HCG REFERENCE SERVICE

          The USA hCG Reference Service was started in January 1998 in response to repeated requests from physicians for help with confusing or inconsistent hCG results (12). It is a consulting service that investigates patient medical history, laboratory hCG records, and brands and versions of hCG tests used. It also independently measures in a single test concentrations of all common hCG-related molecules listed in Table 1 (DPC Immulite hCG test), and regular hCG only, nicked hCG only, hyperglycosylated hCG only, free b-subunit only, and b-core fragment only in parallel serum and urine samples provided by patients. Dilution parallelism is investigated (1X value in undiluted sample and 1/3^rd and 1/10^th of the value in 3-fold and 10-fold diluted sample), as is the affect of HBT (Scantibodies Inc.), a heterophilic antibody/interfering substance blocking agent on all assay results. A report is prepared for the physician on the nature of the hCG detected by the clinical laboratory, and on the most likely source of the immunoreactivity (normal pregnancy or invasive cell hCG, pituitary hCG, non-trophoblastic neoplasm hCG immunoreactivity, or false positive hCG).

          The USA hCG Reference Service is United States Department of Health and Human Services CLIA certified for preparing data/reports for inclusion in patient records (CLIA ID# 32D0972561). In three and a half years there have been approximately 160 patient referrals to the USA hCG Reference Service for various reasons. Commonly the USA hCG Reference Service data confirms physician’s diagnoses of gestational trophoblastic diseases, persistent mole, ectopic pregnancy, placental site trophoblastic disease, testicular choriocarcinoma, or ovarian germ cell or other non-trophoblastic neoplasms. Sometimes, the USA hCG Reference Service suggests an alternative assay for monitoring patients (free b-subunit, urine total hCG, hPL), or an alternative commercial laboratory. The USA hCG Reference Service also discusses and advises on a large number of cases on the telephone (over 50 each year), making recommendations or suggesting tests that can be carried out in the patient’s locality. In addition the web site (www.hcglab.com) attracts approximately 200 e-mails from patients each year. Questions about hCG results, pregnancy, trophoblast disease and the soundness of the hCG test are addressed. Often patient inquiries by e-mail led to home urine hCG testing, and sometimes to further outside hCG testing, physician involvement and USA hCG Reference Service testing to determine the validity of hCG results.

          Of particular interest has been the observation of a large number of false positive or phantom hCG cases, in which the patient received unnecessary therapy for the diagnosis of gestational trophoblastic disease (invasive trophoblastic disease/choriocarcinoma, presence not shown my imaging methods or pathology). Also of interest has been the observation of a significant number of women producing low levels of normal pregnancy hCG, treated for gestational trophoblastic disease ,with no imaging or other evidence of tumor (unexplained elevated hCG). Similarly, women with previous history of trophoblastic diseases, with persistent low levels of hCG, treated for gestational trophoblastic disease, with no imaging or other evidence of tumor  (quiescent gestational trophoblastic disease).

A) FALSE POSITIVE hCG

          In the first few months of operation of the USA hCG Reference Service, 3 unusual cases were investigated for gestational trophoblastic disease or choriocarcinoma (12). In all 3 cases the woman had an incidental pregnancy test which was positive. The positive hCG value persisted with small rises and reductions in values. Ultrasound, dilation and curettage and laparoscopy ruled out pregnancy or ectopic pregnancy. The diagnosis of gestational trophoblastic disease or choriocarcinoma was made, even though there was no previous history of trophoblastic disease or physical evidence of tumor. In two of the 3 cases chemotherapy was started, and in 1 case a hysterectomy was carried out. All 3 cases were then referred to the USA hCG Reference Service. At that time the reported hCG concentrations were 17, 53, 110 IU/L, respectively. It was a surprise when false positive hCG results were demonstrated in these individuals. The finding indicated that the hCG test used by the physicians was detecting unrelated or interfering molecules rather than hCG (12).

          Now, after nearly four years of operation, and multiple publications on the false positive hCG problem (4,10-13), 106 women have been referred to the USA hCG Reference Service for investigating potential false positive hCG results. Forty four women were shown to have had false positive results (Table 4). False positive results were identified by the following criteria (10-12):

1. The finding of more than 5-fold differences in serum hCG results with alternative immunoassays (critical criterion).

2. Presence of hCG in serum, absence of hCG or related molecule immunoreactivity in a urine sample (critical criterion).

3. False positive results in other tests for molecules not present in serum, such as urine b-core fragment (confirmatory criterion)

4. A heterophilic antibody blocking agent (Scantibodies Inc. HBR) prevents or limits false detection (confirmatory criterion).

Table 4. False positive hCG results in 9 common commercial immunoassays. Data from 44 women and 1 man, ages 20-44, without history of trophoblastic disease (or testicular cancer), mistakenly diagnosed with choriocarcinoma or gestational trophoblastic disease because of false positive hCG immunoassay data. These represent 45 out of 112 cases submitted to the center from physicians and centers throughout North America for evaluation for false positive hCG. All false positive cases were identified by the USA hCG Reference Service. The test used by the physician to manage treatment of patient is indicated as “initial”. The most recent serum hCG result prior to, or at the same drawing as that submitted to the USA hCG Reference Service, is shown. We do not show earlier or initial hCG values. Other hCG test results shown are those requested by the hCG Reference Service or clinician’s laboratory with a serum sample collected at or around the time of Reference Service testing. Blank spaces correspond to tests not performed. Histories are coded in order of event, D&C is dilation-curettage; Lap is laparoscopy; Mtx is methotrexate and AcD is actinomycin D chemotherapy; EMACO is etoposide-based 5 agent combination chemotherapy, HYS is hysterectomy, BSO is salpingo-oophorectomy; THO is thoracotomy; Coma is diabetic coma from therapy. Cases are sorted by extremity of therapy, commercial immunoassay are listed in order of occurrence of false positive results.