Laurence A. Cole, Ph.D., Sarah A. Khanlian B.S and Stephen Butler Ph.D., USA hCG Reference Service, University of New Mexico, Albuquerque, NM, USA
Brief description of authors:
Laurence A. Cole, Ph.D., is a Professor of Obstetrics and Gynecology, and of Biochemistry and Molecular Biology at the University of New Mexico, He is also chief of the new division of Womens Health Research at the UNM medical school. Dr. Cole maintains a formal affiliation with Yale University where he initially started the USA hCG Reference Service in 1998.
Stephen Butler Ph.D. joined Laurence A. Cole, Ph.D. as a post-doctoral fellow in 1999 at Yale University, moving with him to University of New Mexico. Sarah Khanlian is the technical associate that currently runs immunoassays for the Service.
The USA hCG Reference Service has been functioning for 3 years as a clinical consulting and testing service. The service and laboratory testing has been approved by the USA Department of Health and Human Services (CLIA #32D0972561). The service aids physicians with patient having irregular or contradictory (no symptoms) hCG results. During the first 2 years of operation we identified 12 patients that had been erroneously diagnosed with cancer (choriocarcinoma or gestational trophoblastic disease) based solely on a continuous false positive or phantom hCG result.
Typically, a women would have an incidental serum hCG pregnancy test prior to an X-ray or minor surgical procedure, or because of irregular vaginal bleeding. A positive result was recorded (10 to >900 IU/L). However, neither intra-uterine pregnancy could be detected (ultrasound, and dilation and curettage), nor an ectopic pregnancy could not be found (ultrasound and laparoscopy). In such cases, textbooks describe only one inference; choriocarcinoma or gestational trophoblastic disease. This would derive from placental cells, remaining in the uterus from a past pregnancy, miscarriage or other gestational event, which have turned malignant. Such cancers can be extremely invasive, textbooks recommend commencement of therapy even though MRI and CT-scanning reveal no tumor. This was the status in all the described cases. Patients would then receive chemotherapy. When this failed to reduce hCG levels a hysterectomy would be carried out, and when this failed to reduce hCG levels a highly cytotoxic combination chemotherapy would be considered. In several cases therapy or surgery went beyond point before the USA hCG Reference Service was contacted and false positive hCG was identified.
In all cases false positive hCG was demonstrated by the presence of hCG in serum and complete absence of hCG in urine, widely varying results (>5-fold) in different hCG immunoassays. It was confirmed by demonstrating other false positive immunoassays, and by successfully blocking false positive immunoreactivity with Scantibodies Inc (San Diego, USA) HBT blocking agent. In all cases the false positivity was seemingly due to interfering heterophilic or anti-animal antibodies. The clinical description of the first 12 patients was published recently in the Lancet (1). The immunoassay results, and blind testing of false positive serum samples in multiple immunometric assay were published in Clinical Chemistry (2).
It was hoped that these two publication and others (1-4), would lead to an awareness of false positive hCG results, a realization of their devastating consequences and a decline in occurrences. This is not what happened. In just the past one year we have observed a further 12 similarly devastating cases, bringing our total to 24. Here we briefly describe the 24 cases together with the false positive results of physician-ordered and several other USA hCG Reference Service requested commercial hCG immunoassays.
Sixty patients were referred to the hCG Reference Service because of potential false positive hCG result. Of these, 24 were shown to have false positive hCG results. The sixty samples came from all major area of the USA, and from Canada.
Results and discussion
A total of 60 cases have been referred to the hCG Reference Service for investigation of false positive hCG results. Of these 24 actually had false positive hCG results, the other 36 had true positive or real hCG value. Twenty three out of 24 false positive patients had minor surgery, dilation and curettage and/or laparoscopy. Sixteen of the 24 (67%) had chemotherapy for presumed cancer, hysterectomy, cytotoxic combination chemotherapy and/or other major surgery.
Most commonly, the assay used by a physicians that gave false positive results was found by phone calls to the clinic, and to the hospital laboratory or external laboratory, that had monitored the patient. This was done by the USA hCG Reference Service after completing testing, and while preparing a report for the physician. Of the 24 patients identified, 23 were being monitored (up to 45 consecutive measurement) by the Abbott AxSym hCGß test; one was being monitored using the Bayer Immuno-1 hCGß test. It is concluded, taking into consideration considering both the high usage of the Abbott AxSym hCGß test (approximately 50% of the market) and the more that 30 serum hCG/hCGß tests sold in North America, that the Abbott AxSym hCGß test has an exceptional false positive problem (considering 50% market share, odds are <0.001% of 23 of 24 cases being AxSym).
Furthermore, of the 23 patients monitored by their physicians with the Abbott AxSym test, at the time of hCG Reference Service analysis, 23 of 23 had hCG results greater than 10IU/L (100%) and 9 of 23 had results greater than 100 IU/L (41%) in the Abbott AxSym test. In the other commercial hCG/hCGß test results carried out by the physicians laboratory or by the USA hCG Reference Service on these 23 patients, 11 of 57 (<20%) test results were greater that 10 IU/L and 1 of 57 were greater than 100 IU/L (<2%). It is inferred, that the majority of these patients falsely monitored with the Abbott AxSym hCGß test would not have necessarily been diagnosed with choriocarcinoma or gestational trophoblastic disease and are much less likely to had hCG results necessitating chemotherapy or surgery if monitored with an alternative test.
The median hCG result at the time of hCG Reference Service analysis in the Abbott AxSym assay was 80 IU/L. The median in one other commercial assays was 8.3 IU, and in the 7 others (3 or more measurements) was <4 IU/L. Results clearly show that the Abbott test gives much higher, or much more elevated, false positive hCG results than other assays. Using a 2 sided Students t test, a statistical difference (P<0.05) was observed between the Abbott AxSym results and the 7 other test results which 3 or values (negative results were considered as 0). This indicated an error in Abbott AxSym hCGß assay function. The USA hCG reference service hCG test uses a commercial enzymelabeled anti-hCGß antibody tracer. We recently tested 9 "false positive" serum samples in this test, and in this test with one modification, replacement of our enzyme-labeled anti-ß tracer with that provided in the AxSym hCGß test pack. The modified test gave a respectable standard curve. The median "false positive" serum values increased from 3.3 IU/L to 301 IU/L, to values very close to those recorded in the AxSym test (Butler and Cole, unpublished data). It was inferred that a problem with the tracer antibody may be at the root of the high false positive results observed with the Abbott AxSym hCGß immunoassay.
While it is clear that the Abbott AxSym gave a preponderance of false positive results, it is also clear that other assays will also give false positive value. These other false positive values, while clearly lower, may lead to less aggressive therapy. hCG measurement are unique in that they are wholly positive, in pregnancies, trophoblastic diseases or cancers, and wholly negative in other individuals. As such a false positive result in an hCG test can be particularly devastating, compared to other assays. Clearly, care is needed in avoiding false positives in all hCG assays. Physicians need to be informed by laboratories either by letter or directly on hCG result forms of potential false positive problems with hCG assays in the diagnosis of ectopic pregnancy, cancer and trophoblastic diseases. It should be noted, "if hCG results do not concur with clinical observations, contact the laboratory for help". Laboratories can simply confirm the reality of hCG measurements by either showing the presence of hCG in serum and urine, showing parallel results in undiluted samples and samples diluted 1:10, or showing similar results in two different hCG immunoassays.
The USA hCG Reference Service has identified 24 women erroneously diagnosed with cancer because of false positive hCG results in North America. An unknown, possibly very large number of women are misdiagnosed with ectopic pregnancies because of false positive hCG results. Also an unknown, possibly very large number of women have been erroneously diagnosed with cancer throughout North America and throughout the World, and not recognized as having false positive hCG by a reference laboratory. A particular problem is indicated with the Abbott AxSym hCGß assay in both preponderance of false positive results and in extent of positivity. Data indicate that this assay should be avoided in ectopic pregnancy, trophoblast disease and cancer hCG determinations. All hCG assays should be considered as having the potential to give false positive results. Laboratories need to take precaution to avoid the devastating consequences.
1. Rotmensch S, Cole LA. False diagnosis and needless therapy of presumed malignant disease in women with false-positive human chorionic gonadotropin concentrations. The Lancet 2000;355:712-15.
2. Cole LA, Shahabi S, Butler S, Mitchell H, Newlands ES, Behrman HR, Verrill HL. Utility of commonly used commercial hCG immunoassays in the diagnosis and management of trophoblastic diseases. Clin Chem 2001;47:308-315.
3. Cole LA, Rinne KM, Shahabi S, Omrani A. False positive hCG levels leading to unnecessary surgery and chemotherapy, and needless occurrences of diabetes and coma. Clin Chem, 1999;45:313-14.
4. Cole LA. Phantom hCG and phantom choriocarcinoma. Gynecol Oncol, 1998;71:325 329.