Urinary hCG Screening in the Gynecologic Oncology Population

Abstract

Objective: Human chorionic gonadotropin (hCG) testing is well established as a diagnostic tool for gestational trophoblastic disease (GTD) but it may also be produced by other gynecologic malignancies. The primary goal of this study was to evaluate whether urinary hCG screening using a highly sensitive testing system might identify gynecologic and nongynecologic malignancies that produce hCG. Design: This study was conducted as a retrospective chart review. Methods: The records of 1003 consecutive patients who presented to the Yale University Gynecologic Oncology Clinic were retrospectively reviewed and clinical and histologic data were extracted. All of these women routinely underwent urinary hCG testing. Results: Forty-one percent of all patients diagnosed with a gynecologic malignancy were found to have uterine cancer, and 38% of all patients had ovarian cancer. Less than 15% of patients were diagnosed with cervical, fallopian tube, primary peritoneal, vulvar, or vaginal carcinoma. A total of 8 patients had a positive urine pregnancy test. Six of these patients were known to be pregnant. One patient was referred with a complete hydatidiform mole. One patient was diagnosed with stage IIC clear cell ovarian carcinoma. Conclusions: Less than 0.1% of the nonpregnant patients without gestational trophoblastic disease had a positive urine pregnancy test. Based on the results of this study, urinary hCG screening can not be recommended to screen for malignancies that produce hCG. J GYNECOL SURG 27:143)

Introduction

Low levels of human chorionic gonadotropin (hCG) or its metabolites have been reported to be produced by cancers arising from numerous sites, in addition to being produced by gestational trophoblastic disease (GTD) and germ cell malignancies.^1–11 Recent advances in testing for urinary hCG allow for the identification of urinary hCG at levels as low as 25 mIU/mL.

Germ cell tumors, especially embryonal carcinomas, polyembryomas, and dysgerminomas, frequently express hCG.² Gynecologic tumors such as ovarian, cervical, endometrial, and vulvovaginal cancers have been associated with hCG or its metabolite, the beta core fragment of hCG.^2,4–6 Elevated urine hCG levels have been linked to the prognosis of cervical and vulvovaginal cancers.^5,6 Serum hCG has been shown to be an independent prognostic factor in ovarian cancer.⁷

Ectopic hCG expression has also been described in nongynecologic malignancies such as bladder, renal, gastrointestinal, lung, and breast cancer.^2,8–11

Therefore, hCG might not only be an important test for GTD diagnosis and management, but its measurement might add some insight into the diagnosis and management of non-GTD gynecologic and nongynecologic malignancies.

Urinary hCG can be detected with urine pregnancy tests. First developed in the 1920s, urine pregnancy tests now use two monoclonal antibodies that selectively detect hCG's beta subunit (β-hCG).^12,13 These tests consist of one antibody, the capturing antibody, that binds to one site on the β-hCG and immobilizes it, and a second antibody labeled with an enzyme tracer. The tracer antibody binds a distant site on β-hCG and labels the immobilized complex.¹² This immobilized antibody-hCG dye antibody complex can then be detected using a qualitative immunoassay.¹²

Urine β-hCG testing is cheap and easy. It is well established as a diagnostic tool for GTD but may also be useful in screening for other gynecologic and nongynecologic malignancies. The primary purpose of this study was to identify non-GTD gynecologic malignancies that produced β-hCG, using a qualitative urinary assay technique sensitive to 25 mIU/mL. A secondary goal was to identify patients who might have an unrecognized pregnancy, prior to recommending surgical procedures, radiation therapy, or chemotherapeutic treatments.

Materials and Methods

One thousand consecutive patients referred to the Yale Gynecologic Oncology Service from May 3, 2004 until February 26, 2007 participated in this study. Their charts were reviewed. As part of this study, a urine pregnancy test was routinely performed on every patient who presented to the Yale Gynecologic Oncology Service by two patient care associates (PCA) in the clinic who had been trained in performing and reading the test. The QuickVue^® urine pregnancy test was used. This pregnancy test is the urine pregnancy test used by the clinical facilities at Yale–New Haven Hospital. It is a qualitative immunoassay and detects β-hCG with a sensitivity of 25 mIU/mL in the urine (Quidel^® Corporation package insert). The testing kits for β-hCG were supplied by the manufacturer to the investigators. Urine was aspirated with a disposable pipette. Three drops of urine were added to the sample well on the test cassette with the pipette. The test was read 3 minutes after placement of the drops by a PCA. A blue line appeared in the control field and confirmed that the procedure was adequately performed. Presence of a pink-to-red test line confirmed the presence of urine β-hCG at 25 mIU/mL or more (Quidel^® Corporation package insert). All urine test results were recorded as “positive” or “negative”.

Medical charts, including admission and discharge notes, as well as surgical pathology reports and radiation records, were reviewed and epidemiological data (age, ethnicity, gravity, and parity), histologic data (stage, grade), as well as hormone status (last menstrual period, pre- or postmenopausal status) was extracted.

This research was approved by the Human Investigation Committee at Yale University School of Medicine (HIC #0905005220).

Results

A total of 1003 consecutive patients underwent urine β-hCG screening in the Yale Gynecologic Oncology Center. Their mean age was 53 years (range 14 to 95 years). The majority (78%) of the patients (n=785) were Caucasian (Table 1). Postmenopausal women comprised 52% of those who underwent urine pregnancy screening (n=525), whereas 47% of the patients were premenopausal (n=467) (Table 1).

Table 1.

Patient Demographics

Patient characteristics	Number of patients	Percentage
Ethnicity
Caucasian	785	78.3%
African American	56	5.6%
Hispanic	33	3.3%
Asian	14	1.4%
Missing	115	11.4%
Total	1003	100%
Menopause status
Premenopausal	467	46.6%
Postmenopausal	525	52.3%
Missing	11	1.1%
Total	1003	100%

Three hundred forty-three patients were diagnosed with a malignancy and 532 patients presented with benign disease (Table 2). Of all the patients diagnosed with a gynecologic malignancy, 41% (n=141) were found to have uterine cancer and 38% (n=129) of all patients had ovarian cancer. Less than 15% of all patients were diagnosed with cervical, fallopian tube, primary peritoneal, vulvar. or vaginal carcinoma.

Table 2.

Gynecologic Histology

Histology	Number of patients	Percentage
Benign	532	53.0%
Uterine cancer	141	14.0%
Ovarian cancer	129	12.9%
Intraepithelial neoplasia (cervical, vaginal, vulvar)	73	7.3%
Cervical cancer	42	4.2%
Vulvar cancer	10	1.0%
Primary peritoneal cancer	9	0.9%
Vaginal cancer	7	0.7%
Gestational trophoblastic disease	3	0.3%
Fallopian tube cancer	2	0.2%
None available	55	5.5%
Total	1003	100%

One hundred twenty-nine patients had been diagnosed with a nongynecologic malignancy in the past (12.9% of patients) (Table 3). The majority of those patients (63.6%, n=82) had a history of breast cancer (Table 3). Fifteen (11.5%) patients had been diagnosed with skin cancers in the past, including basal cell and squamous cell carcinomas, and melanomas (Table 3). For 55 patients, no histologic diagnosis was available as there were healthy women participating in the Yale Ovarian Cancer Early Detection Program.

Table 3.

Nongynecologic Malignancies in Referred Patients

Nongynecologic malignancy	Number of patients	Percentage
Breast cancer	82	63.6%
Skin cancer (melanoma, basal cell carcinoma, squamous cell carcinoma)	15	11.5%
Colon cancer	11	8.4%
Lymphoma	4	3.1%
Lung cancer (small cell, adenocarcinoma)	4	3.1%
Renal carcinoma	4	3.1%
Pancreatic cancer	2	1.6%
Leukemia	2	1.6%
Thyroid carcinoma	2	1.6%
Bladder carcinoma	1	0.8%
Gastric cancer	1	0.8%
Larynx carcinoma	1	0.8%
Total	129	100%

A total of 994 patients had a negative urine pregnancy test (Table 4). Eight patients had a positive pregnancy test. One patient was unable to urinate for the test. Of the patients with a positive test result, 6 were pregnant. All 6 were referred to the Gynecologic Oncology Clinic for the evaluation and management of adnexal masses in pregnancy or abnormal Papanicolaou smears. None of them were found to have malignant disease. No patients were first diagnosed to be pregnant based on the study test.

Table 4.

Patients with Positive Urine Pregnancy Test

Pregnancy test result	Number	Percentage
Negative	994	99.1%
Positive, pregnant	6	0.6%
Positive, hydatidiform mole	1	0.1%
Positive, ovarian carcinoma	1	0.1%
Unable to urinate	1	0.1%
Total	1003	100%

Two nonpregnant patients had a positive test. One of these 2 patients was a 27- year-old Caucasian who presented for follow-up of a complete molar pregnancy. The other patient was a 45-year-old African-American, gravida 3 para 1 woman, who had a pelvic mass and was subsequently diagnosed to have stage IIC clear cell ovarian carcinoma. The patient subsequently underwent surgical staging and chemotherapy. The work-up for other sources of hCG was negative. Her last menstrual period had been 3 weeks prior to her clinic visit. She was not pregnant.

Discussion

Human chorionic gonadotropin is a glycoprotein hormone composed of two noncovalently joined subunits, the alpha- and the beta-subunit.^12,14,15 It is a diverse molecule with three distinct biological variants expressing different functions: regular hCG, hyperglycosylated hCG and the free beta-subunit of hyperglycosylated hCG.¹² Regular hCG is produced by the villous syncytiotrophoblast throughout most of a normal pregnancy. It helps maintain efficient placentation via angiogenesis in the myometrial spiral arteries.¹² It is also the predominant hCG variant in cases of complete and partial moles.⁴

Hyperglycosylated hCG, an autocrine factor, is produced by extravillous invasive cytotrophoblast cells.¹² It not only plays a role in implantation during normal pregnancy, but is also detectable in GTD where it seems to promote invasion, growth, and malignancy.¹² It can help to differentiate invasive from noninvasive hydatidiform moles.^4,12

Hyperglycosylated free beta-subunit of hCG (hCG free beta) acts as an anti-apoptotic factor in non-trophoblastic malignancies.^4,12 It has been associated with a poor outcome in patients with cervical cancer.⁵ Muller and Cole also described the production of β-hCG by placental site trophoblastic tumors (PSTT), a characteristic that may help distinguish PSTT from choriocarcinoma.⁴

Although the short half-life of hCG and its subunits makes its use as a serum tumor marker difficult, there are higher concentrations in the urine, particularly of the free β subunit, making urine β-hCG a more interesting screening target for non-GTD malignancies.⁶

Papapetrou et al. studied 70 patients with malignant disease, using urine and serum hCG testing.¹⁶ Twenty-nine of the 70 patients were female. Ten of the 29 female patients were found to have positive urine hCG testing.¹⁶ Five of these women were diagnosed with breast cancer (3 premenopausal, 2 postmenopausal), 1 had ovarian cancer (premenopausal), 1 had cervical cancer (premenopausal), 1 had lung cancer (postmenopausal), 1 had gastric cancer (postmenopausal), and 1 had primary hepatic cancer (postmenopausal).¹⁶ Two breast cancer patients (1 pre- and 1 postmenopausal) with positive urine hCG levels had negative serum hCG levels.¹⁶ All premenopausal women had normal serum hCG levels. The normal serum levels for postmenopausal women were higher than for premenopausal women (upper limit of normal 4.4 ng hCG per mg creatinine, compared to 1.6 ng hCG per mg creatinine for premenopausal women). Using this cutoff, 5 postmenopausal control women were considered positive.¹⁶

Stenman et al. suggested the use of the beta core fragment of hCG in urine in combination with serum CA-125 in order to improve accuracy of biomarkers in the diagnosis of ovarian cancer.²

hCG production is associated with poor prognosis in bladder and cervical cancers.^5,6,17,18 It has been hypothesized that this poor prognostic effect of hCG might be linked to potential antiapoptotic activity of hCG via the transforming growth factor beta (TGF-β) pathway.^17,19 Having only one β-hCG positive patient in our study did not allow us to correlate the disease stage, grade, or prognosis and hCG positivity.

As for any diagnostic or screening test, there are limitations to hCG testing. Although the urine pregnancy test used for the present study (QuickVue^® urine pregnancy test) has a sensitivity and specificity of>99%, false-positive results are possible. False-positive serum hCG results have been described and are frequently attributed to heterophilic antibodies.^20–22

Muller and Cole reported that 68% of ovarian, 51% of endometrial, and 46% of cervical malignancies produce β-hCG.⁴ In the current study, these results could not be reproduced. Only 1 nonpregnant patient with ovarian cancer (<0.1%) presented with a positive urine pregnancy test. None of the patients with endometrial, cervical, vulvar, primary peritoneal, vaginal, or fallopian tube cancer included in the study had a positive urine pregnancy test.

The etiology of the positive urine pregnancy test in the patient with ovarian cancer is unclear. Low levels of hCG production are physiologic around the time of the midcycle preovulatory luteinizing hormone (LH) surge.²³ In the peri- and postmenopause, increasing hCG levels are detected even more frequently and seem to be a result of increasing pituitary hCG production.^20,23,24 The ovarian cancer patient in this present study was still premenopausal, and she was not midcycle at the time of her presentation to clinic. In turn, very high serum and urine β-hCG concentrations can lead to false-negative urine pregnancy test results.^25–27 It is possible that the positive urine pregnancy test was a false positive test result.

This current study does not suggest that routine urine hCG screening should be performed in gynecologic oncology clinics. In the 1003 patients evaluated from 2004-2007, only 1 patient (<0.1%) who was not pregnant was found to have a positive urine pregnancy test. Although this positive result may be explained by ectopic hCG production, a false-positive test result must also be considered. This patient showed disease progression after six cycles of chemotherapy. Although it has been suggested in the literature that positive hCG expression might correlate with poor prognosis in ovarian cancer, the numbers in this study are certainly too small to assess this question.⁷

One limitation of this study is the fact that hCG was routinely tested in the urine only. Serum hCG assays were not routinely performed. A comparison between the urine and serum hCG levels is therefore not possible.

This is the largest review of urine hCG screening, including 1003 patients, who were referred to a gynecologic oncology service. Although there is support in the literature to perform urine pregnancy tests as a screening test for malignancy, routine urine hCG screening cannot be recommended based on the results of this study.^2,4,16

Footnotes

Disclosure Statement

The QuickVue^® test kits for β-hCG were supplied by the manufacturer to the investigators. No competing financial conflicts exist.

References

Berkowitz

, Goldstein

. Current managemen of gestational trophoblastic diseases. Gynecol Oncol, 2009; 112:654.

Stenman

U-H

, Alfthan

, Hotakainen

. Human chorionic gonadotropin in cancer. Clin Biochem, 2004; 37:549.

Ajufo

, Lindow

, Canty

. Choriocarcinoma with markedly elevated serum hCG levels and negative urine hCG levels. J Obstet Gynaecol, 2006; 26:83.

Muller

, Cole

. The quagmire of hCG and hCG testing in gynecologic oncology. Gynecol Oncol, 2009; 112:663.

Crawford

, Iles

, Carter

, Caldwell

, Shepherd

, Chard

. The prognostic significance of beta human chorionic gonadrotrophin and its metabolites in women with cervical carcinoma. J Clin Pathol, 1998; 51:685.

Carter

, Iles

, Neven

, Ind

, Shepherd

, Chard

. The prognostic significance of urinary beta core fragment in premenopausal women with carcinoma of the cervix. Gynecol Oncol, 1994; 55:271.

Vartiainen

, Alfthan

, Lehtovirta

, Stenman

. Identification of choriocarcinoma be the hCG beta-to-hCG proportion in patients with delayed diagnosis caused by contraceptive use. Contraception, 1998; 57:257.

Dirnhofer

, Koessler

, Ensinger

, Feichtinger

, Madersbacher

, Berger

. Production of trophoblastic hormones by transitional cell carcinoma of the bladder: Association to tumor stage and grade. Hum Pathol, 1998; 29:377.

Louhimo

, Nordling

, Alfthan

, von Bogulaawski

, Stenman

, Haglund

. Specific staining of hCGβ in benign and malignant gastrointestinal tissue with monoclonal antibodies. Histopathology, 2001; 38:418.

10.

Lundin

, Nordling

, Lundin

, Alfthan

, Stenman

, Haglund

. Tissue expression of human chorionic gonadotropin beta predicts outcome in colorectal cancer: A comparison with serum expression. Int J Cancer, 2002; 95:18.

11.

Span

, Manders

, Heuvel

, Thomas

, Bosch

, Beex

, Sweep

. Molecular beacon reverse transcription-PCR of human chorionic gonadotropin-beta-3, -5, and -8 mRNAs has prognostic value in breast cancer. Clin Chem, 2003; 49:1074.

12.

Cole

. New discoveries on the biology and detection of human chorionic gonadotropin. Reprod Biol Endocrinol, 2009; 7:1.

13.

Hiroshe

. Exogenous stimulation of corpus luteum formation in the rabbit: Influence of extracts of human placenta, decidua, fetus, hydatid mole, and corpus luteum on the rabbit gonad. J Jpn Obstet Gynecol Soc, 1920; 16:1055.

14.

Cole

, Sutton

. HCG test in the management of gestational trophoblastic diseases. Clin Obstet Gynecol, 2003; 46:523.

15.

Swaminathan

, Bahl

. Dissociation and recombination of the subunits of human chorionic gonadotropin. Biochem Biophys Res Comm, 1970; 40:422.

16.

Papapetrou

, Sakarelou

, Braouzi

, Fessas

. Ectopic production of human chorionic gonadotropin (hCG) by neoplasms: The value of measurements of immunoreactive hCG in the urine as a screening procedure. Cancer, 1980; 45:2583.

17.

Iles

. Ectopic hCGβ expression by epithelial cancer: Malignant behaviour, metastasis and inhibition of tumor cell apoptosis. Mol Cell Endocrinol, 2007; 260–262:264.

18.

Martin

, Jenkins

, Zuk

, Oliver

, Baithun

. Human chorionic gonadotropin expression and histological findings as predictors of response to radiotherapy in carcinoma of the bladder. Virchows Arch A Pathol Anat Histopathol, 1989; 414:273.

19.

Butler

, Ikram

, Mathieu

, Iles

. The increase in bladder carcinoma cell population induced by the free beta subunit of human chorionic gonadotropin is a result of an anti-apoptosis effect and not cell proliferation. Br J Cancer, 2000; 82:1553.

20.

Stenman

U-H

, Tiitinen

, Alfthan

, Valmu

. The classification, functions and clinical use of different isoforms of HCG. Human Reprod Update, 2006; 12:769.

21.

McCready

, Braunstein

, Helm

, Wade

. Modification of the choriogonadotropin beta-subunit radioimmunoassay for determination of urinary choriogonadotropin. Clin Chem, 1978; 24:1958.

22.

Braunstein

. False-positive serum human chorionic gonadotropin results: Causes, characteristics, and recognition. Am J Obstet Gynecol, 2002; 187:217.

23.

Cole

, Khanlian

, Muller

. Detection of perimenopause and postmenopause human chorionic gonadotropin: An unnecessary source of alarm. Am J Obstet Gynecol, 2008; 198:275.

24.

Snyder

, Haymond

, Parvin

, Gronowski

, Grenache

. Diagnostic considerations in the measurement of human chorionic gonadotropin in aging women. Clin Chem, 2005; 51:1830.

25.

Ofinran

, Papaioannou

, Kandavel

, Shrivastava

, Hall

, Tzafettas . Negative pregnancy test: Could it be a molar pregnancy? J Obstet Gynaecol, 2007; 27:857.

26.

Tabas

, Strehlow

, Isaacs

. A false negative pregnancy test in a patient with a hydatidiform molar pregnancy. N Engl J Med, 2003; 349:2172.

27.

Wisdom

, Irvine

, Gray

, Dryburgh

, Greer

. Moles and sandwiches. Br J Obstet Gynaecol, 1994; 101:455.