Immunoprocedures for detecting human chorionic gonadotropin: clinical aspects and doping control

Transcription

1 Clinical Chemistry 43: (1997) Doping in Sports Symposium Immunoprocedures for detecting human chorionic gonadotropin: clinical aspects and doping control Ulf-Håkan Stenman, 1 * Leila Unkila-Kallio, 2 Juha Korhonen, 2 and Henrik Alfthan 1 The pregnancy hormone human chorionic gonadotropin (hcg) is also present at low concentrations in plasma and urine of men and nonpregnant women. hcg immunoreactivity occurs in various molecular forms: Besides the intact hcg heterodimer, considerable amounts of proteolytically cleaved forms, free subunits, and fragments are found in plasma and urine. Especially in urine, proteolytic fragments constitute a major part of the hcg immunoreactivity. The different forms of hcg cross-react to various degrees in immunoassays and constitute a problem for standardization of specific hcg determinations. After injection of hcg ( IU of Pregnyl ; Organon), above-normal concentrations of hcg can be detected in serum and urine for 7 11 days. Most immunoassays for hcg also measure hcg. Quantitative hcg determinations are mainly performed on serum samples, and very few commercial hcg determinations have been validated for determination of urine samples. Considerable care must therefore be exercised when utilizing such assays to analyze urines for doping control. INDEXING TERMS: standardization steroid hermones urine reference values pregnancy cancer luteinizing hormone estrogen The gonadotropins luteinizing hormone (LH) 3 and chorionic gonadotropin (hcg) are therapeutically used to stimulate gonadal steroid production, both in women and men. hcg is also used by some athletes to stimulate Departments of 1 Clinical Chemistry and 2 Obstetrics and Gynecology, Helsinki University Central Hospital, Haartmaninkatu 2, FIN Helsinki, Finland. *Author for correspondence. Fax ; ulf-hakan. stenman@hyks.mailnet.fi. 3 Nonstandard abbreviations: hcg, human chorionic gonadotropin; IOC, International Olympic Committee; hcg, free beta subunit of hcg; hcg, free alpha subunit of hcg; LH, luteinizing hormone; hcg cf, core fragment of hcg ; hcgn, nicked hcg; hcg n, nicked hcg ;LH cf, core fragment of LH. Received October 1, 1996; revised February 18, 1997; accepted February 26, testosterone production and to normalize the testosterone/epitestosterone ratio during or after administration of testosterone and other anabolic steroids. The International Olympic Committee (IOC) imposed a ban on hcg administration in 1987 [1], but no official decision limit for hcg in urine has been defined. IOC prescribes that hcg should be determined by two different assays, but neither the decision limit nor the specificity of the assays has been defined. Assay of hcg in serum and urine can now be performed with extremely sensitive and specific methods, which can determine the low basal concentrations found in nonpregnant women and men. From the hcg concentrations in urine from 1400 men, Laidler et al. have recommended that a cutoff value 10 IU/L should be used for doping control [1]. However, this recommendation may be considered valid only for the method used. Different assay methods for determining hcg in serum are poorly standardized e.g., the results obtained by various methods may vary by a factor of five [2] and different methods show large variation in reactivity with various forms of hcg [3]. Furthermore, given the more extensive heterogeneity of hcg in urine than in serum, larger differences will be observed in urine. Therefore, the various forms of hcg in circulation and urine and the design of the assays used will affect the results. Little is known about the extent of this effect, because quantitative assays for hcg in urine have very little clinical use, and kit manufacturers have therefore not validated their methods for assay of urine. Thorough knowledge of the forms of hcg detected in urine by various methods under normal conditions and after hcg administration are prerequisites for the use of hcg determinations for doping control. Molecular Forms of hcg and Its Subunits The following nomenclature for well-characterized molecular forms of hcg has been suggested by the Working Group for standardization of hcg determinations appointed by the IFCC [2]. Human chorionic gonadotropin (hcg) is the intact heterodimer, consisting of an - and a -subunit. hcg is 1293

2 1294 Stenman et al.: Immunodetection of hcg: clinically and for doping control the major form of hcg immunoreactivity during pregnancy. The free beta subunit (hcg ) consists of a 22-kDa polypeptide containing 145 amino acid residues, with two N-linked and four O-linked carbohydrate chains. hcg is secreted by the placenta during pregnancy, and the serum concentrations are 1% of those of hcg. The free alpha subunit (hcg ) consists of a glycosylated 14-kDa polypeptide containing 92 amino acid residues and two N-linked carbohydrate chains. This subunit is common to hcg, LH, follicle-stimulating hormone (FSH), and thyroid-stimulating hormone (TSH). hcg is produced in excess of the -subunits by both the pituitary and the placenta [4, 5]. The core fragment of hcg (hcg cf) is a 10-kDa two-chain polypeptide lacking the amino acid sequences 1 5, 41 54, and The two chains are held together by disulfide bonds. hcg cf is probably formed from hcg and hcg by proteolytic degradation in the kidneys. Nicked hcg (hcgn) and nicked hcg (hcg n) are partially degraded forms of hcg and hcg, in which the peptide chain of hcg is cleaved in the region (usually between amino acids 47 and 48 and less frequently between 44 and 45). Assay Methods Serum assays. Specific determination of hcg in serum from nonpregnant subjects was first performed by RIA with antiserum SB6 [6]. This assay also measures hcg. Specific measurement of hcg in serum became possible with the development of ultrasensitive immunometric assays that were based on monoclonal antibodies [7, 8]. This technology has also made it possible to develop sandwich -type assays for measuring hcg in serum [9] and hcg cf in urine [10, 11]. Only a few commercially available methods are capable of measuring the physiological concentrations of hcg in men and nonpregnant women, and the concentrations of hcg and hcg cf can only be measured by some research methods. Many hcg immunoassays are designed to measure both hcg and hcg. Such methods are based on the use of two or three antibodies specific for hcg or of a monoclonal antibody and polyclonal antiserum. A common problem with these assays is that they tend to overestimate hcg [2]. On the other hand, some assays based on a combination of anti-hcg and anti- antibodies may underestimate hcgn [3]. In pregnancy, urine hcg may be extensively nicked, but whether this form occurs in urine of nonpregnant subjects is not known. By utilizing one antibody to hcg and another one to hcg or the dimer, one can design highly specific assays reacting with hcg but not with hcg. Cross-reaction with LH, which used to be a problem with conventional RIA methods, is usually negligible. However, denaturation of LH during storage has been shown to increase its cross-reactivity, even in highly specific assays based on monoclonal antibodies [7]. Urine assays. Commercial hcg assays have usually not been validated for determination of hcg in urine. The reactivity with hcg cf may be a problem in assays designed to measure both hcg and hcg. Another potential problem is the cross-reaction with denatured LH. The core fragment of LH (LH cf) is the predominant form of LH immunoreactivity in urine and may interfere in hcg assays. Assays for total hcg that overestimate hcg cannot be used to specifically measure hcg in urine, which may contain much hcg and hcg cf. Given the close structural similarity between hcg cf and LH cf, some assays for hcg cf also measure LH cf [12], which is present in urine from nonpregnant subjects at much higher concentrations than hcg cf [13]. However, assays based on two monoclonal antibodies with negligible cross-reaction with LH cf ( 2%) have been described [11]. Because of the large differences in assay design and specificity for different forms of hcg, the concentrations measured are highly dependent on the assay used [2, 3]. The concentrations measured in quality-control samples may vary fivefold [2]. The relative amounts of various hcg forms also depend on the clinical situation of the subject tested. All these factors contribute to the large variation in reported hcg concentrations. hcg can be detected in urine by highly specific massspectrometric methods. However, there is still at least a 100-fold difference in detection limit between mass-spectrometric and immunological methods [14]. The concentrations of protein hormones in urine are usually corrected for variations in urine flow. Mostly the concentrations are normalized by dividing the hormone concentrations with the creatinine concentration, but correction to mean urine density is also used [15]. In healthy subjects the mean density is mg/l (Alfthan and Stenman, unpublished); however, normal urine may range from 10-fold more dilute to 3-fold more concentrated than average. Although creatinine correction is widely used, in our experience the concentrations of gonadotropins are not directly correlated to creatinine in very dilute or concentrated urines [16], and better ways of normalizing urine concentrations of protein hormones in urine need to be developed. This appears to be especially important in athletes, whose urine output after strenuous exercise may be very low, which may be expected to increase the concentrations of protein hormones. Concentrations of hcg in Serum and Urine in Health and Disease in nonpregnant women and men Serum. The concentrations of hcg in serum of nonpregnant women are measurable by sensitive assays. These concentrations are dependent on gonadal function, showing a clear increase around the menopause, when the upper reference limit increases from 3 to 5 IU/L (from 8.6 to 15.5 pmol/l) [17]. Concentrations up to 8 10 IU/L are occasionally observed in healthy women.

3 Clinical Chemistry 43, No. 7, In men, the concentrations are lower but mostly are measurable by highly sensitive assays. In those 50 years old, the upper reference limit is 0.7 IU/L (2.1 pmol/l); in those 50, it is 2 IU/L (6.1 pmol/l) [17]. Concentrations as great as 3 4 IU/L are occasionally observed in healthy men. The concentrations of hcg are lower than those of hcg. Only a slight increase is observed with age and the sex difference is small. The upper reference limit in women 50 years is 1.6 pmol/l and in men 1.9 pmol/l. The corresponding values for those 50 years are 2.0 and 2.1 pmol/l, respectively. The concentrations of hcg cf in serum are below the detection limit of present assays [17]. Urine. The median concentrations of hcg and hcg are similar in serum and urine, whereas those of hcg cf are much higher in urine [17]. This suggests that, as in pregnancy, hcg cf is produced by proteolytic degradation of hcg and hcg in the kidneys. The concentrations of hcg and hcg in urine of women and men increase with age but not as sharply as those in serum. The upper reference limits for hcg in women 50 and 50 years are 3 and 4 IU/L (8.8 and 11.5 pmol/l), respectively. The corresponding values for men are 1 and 3 IU/L (2.9 and 8.4 pmol/l) [17]. The concentrations of hcg cf in urine are similar in women and men [17], increasing at age 50. By our method, the upper reference limits for pre- and postmenopausal women are 8.1 and 9.5 pmol/l, respectively, and those for men 50 and 50 are 6.7 and 8.5 pmol/l, respectively [17]. In nonpregnant subjects, the common glycoprotein hormone subunit (hcg ) is mainly derived from the pituitary, and its serum concentrations do not reflect hcg production. The concentrations in urine are three- to fivefold those in serum [18, 19]. in women with normal pregnancy The concentration of hcg in serum starts to increase 7 11 days after ovulation, corresponding to days after the last menstrual period [20]. The increase is nearly exponential during the first 5 weeks after implantation [21]. As measured by our method, the concentrations peak at IU/L during week 8 10 after the last menstrual period (i.e., 5 7 weeks after implantation), after which they decrease and reach a nadir of IU/L at the beginning of the second trimester [9]. (These values are method dependent [2, 3].) There is a small increase before delivery [22], and the concentrations of hcg and its subunits are higher in multiple pregnancies than in singleton ones [23]. Individual variation in hcg concentrations is large [9]. After normal delivery, the hcg concentrations decrease with a half-time of about h [24] and normalize within 1 3 weeks [25, 26]. Serum hcg concentrations after a first-trimester abortion may take 4 5 weeks to return to normal values [27]. Mean serum concentrations of hcg during pregnancy vary from 0.5% to 1.6% (range %) of the hcg concentrations [9, 28]. hcg cf can be detected in serum during pregnancy in quantities corresponding to 0.03% of the hcg content. The concentrations in urine are 4000-fold those in serum [29]. hcg cf is the major form of hcg immunoreactivity in pregnancy urine except during the first month of pregnancy, when hcg predominates [30]. The concentrations of hcg immunoreactivity in paired serum and urine samples are similar and correlate strongly [31]. Total hcg immunoreactivity is higher in urine [20, 32] apparently attributable to hcg cf. in pregnancy-related disorders Early fetal loss. About 20 25% of all conceptions end in an early spontaneous abortion [33, 34]. Often the only indication of pregnancy is an increased concentration of hcg in serum or urine; hence, the condition is also called subclinical abortion or biochemical pregnancy [35]. Spontaneous abortions occur most often during the first trimester. Initially, the concentration of hcg in serum may increase normally before it starts falling [36]. Ectopic pregnancy. In ectopic pregnancy, the concentrations of hcg in serum tend to be lower than during normal pregnancy [36, 37], but they may also be normal for a long time. in malignant disease Gestational trophoblastic disease. The concentrations of hcg in serum and urine are strongly increased in choriocarcinoma and molar disease [38]. In benign trophoblastic disease, the concentrations of hcg are usually 5% of those of hcg, but in choriocarcinoma the proportion of hcg is higher [39, 40]. The proportion of hcgn is higher in serum of patients with choriocarcinoma than in pregnancy [41]. Testicular cancer. About 50% of the nonseminomatous testicular cancers cause an increase of hcg in serum and urine, often to very high concentrations. hcg is seldom increased in seminomas, but these tumors relatively often produce hcg [42]. Nontrophoblastic tumors. hcg is very rarely produced by nontrophoblastic tumors, but low-level expression of hcg is quite common. Because the serum concentrations of hcg in cancer patients are usually only moderately increased, hcg is a useful marker only if measured by an ultrasensitive assay [43]. The strong correlation between hcg in serum and hcg cf in urine suggests that, when excreted into urine, hcg is degraded to hcg cf. However, by immunohistochemistry staining with monoclonal hcg cf antibodies, positive staining has frequently been observed in nontrophoblastic cancer cells, suggesting that some hcg (and hcg) may be degraded within

4 1296 Stenman et al.: Immunodetection of hcg: clinically and for doping control Fig. 1. Concentrations of hcg in serum (S-hCG) and urine (U-hCG) and of hcg cf in urine after intramuscular injection of IU of hcg (Pregnyl) into (A) patients who did not conceive and (B) patients who conceived. The concentrations are given in pmol/l to facilitate comparison between the concentration of hcg and hcg cf (1 IU is equivalent to 2.93 pmol). The thin lines represent individual patients and the thick line mean values. The broken line indicates the upper reference limit. the cell before secretion [44]. The relative contribution of the various sources of hcg cf remains to be determined. Effect of Drugs gonadotropin-releasing hormone and estrogens In men and nonpregnant women, gonadotropin-releasing hormone causes a threefold increase in the serum concentrations of hcg, whereas treatment with estrogen and progestagen lowers the serum hcg by 50%. These results suggest that hcg in serum is derived from the pituitary [7]. hcg administration Intramuscular injection of partially purified urinary hcg is used to induce ovulation in hormone-stimulated menstrual cycles. We have measured the concentrations of hcg in serum and urine and of hcg cf in urine after injection of IU of Pregnyl to women during stimulated cycles. The assays used have been described in detail [9, 11, 17]. The procedures followed were in accordance with the Helsinki Declaration of 1975, revised in Peak serum hcg concentrations of IU/L are observed 1 day after injection, after which they decline with a mean ( SD) half-time of days (Fig. 1A). hcg concentrations in serum stay above the upper reference limit for 7 11 days after injection and reach baseline values within 8 13 days. Similar results have been obtained in another study [45]. The hcg concentrations in urine correlate fairly closely with those in serum, and there is very little delay in excretion. The concentrations remain increased for the same time as those in serum, and their elimination halftimes are similar, day (Fig. 1A). The pattern of hcg cf in urine is slightly different. Concomitant with the hcg peak in serum is a sharp peak of hcg cf in urine. The hcg cf concentration decreases rapidly within 1 day, after which it falls at the same rate as the concentration of hcg. The urine concentrations of hcg cf drop below the

5 Clinical Chemistry 43, No. 7, upper reference limit 2 3 days earlier than does hcg (Fig. 1A). Pregnyl contains equal amounts of hcg and hcg cf, and the first hcg cf peak is explained by rapid excretion of hcg cf into urine [46]. In women who conceive after hcg stimulation, the hcg concentrations in serum and urine start increasing again before the injected hcg has reached baseline values in serum. The hcg concentrations in urine are similar to those in serum, but the hcg cf concentrations start increasing after a delay of 3 4 days (Fig. 1B). This delay is similar to the delay in excretion of LH cf after the midcycle LH surge [13]. On the basis of this, we expected to also see a delayed excretion of hcg cf into urine, but this was not the case. The metabolism of purified hcg may therefore differ from that of endogenous LH. Interestingly, after conception, the excretion of hcg cf is delayed (Fig. 1) in a way similar to that of LH cf [13]. Conclusions hcg is a normally occurring hormone, and its concentrations in serum and urine can be measured in men and in nonpregnant women by sensitive methods. The determination is complicated by the occurrence of many different forms of hcg immunoreactivity, which exert a variable response in different assays. The urine concentrations are also affected by variations in urinary flow rate. To be useful for doping control, the reference values for hcg need to be determined on a sufficiently large number of samples from women and men of the proper age group. Furthermore, the reference values need to be determined by the same method as used for doping control. The results obtained by various hcg assays vary by as much as fivefold because of differences in specificity and calibration. Therefore, some methods may give false-positive results whereas others may miss doping cases. It is also essential that the effect of exercise be determined. The disappearance curves for hcg in serum and urine after injection of urinary hcg in combination with appropriately established reference values provide the basis for detection of hcg administration. Intramuscular administration of hcg causes a prolonged increase of serum hcg and of the urinary excretion of hcg and hcg cf. The serum and urine concentrations decrease to less than the upper reference limit in 6 9 days. Generally, however, decision limits higher than the upper reference limit are used for doping control. If appropriately standardized and validated methods are used, investigators should be able to detect selfadministration of hcg in men as reliably as anabolic steroids and testosterone are now being detected by mass-spectrometric methods. However, although the methods for such determinations are available, very few have been appropriately validated. Therefore, the present recommendation by the IOC that hcg should be determined by two different immunoassays needs to be more specifically defined before an assay is used to sentence athletes. References 1. Laidler P, Cowan DA, Hider RC, Kicman AT. New decision limits and quality-control material for detecting human chorionic gonadotropin misuse in sports. Clin Chem 1994;40: Stenman U-H, Bidart J-M, Birken S, Mann K, Nisula B, O Connor J. Standardization of protein immunoprocedures. Choriogonadotropin (CG). Scand J Clin Lab Invest 1993;53(Suppl 216): Cole LA, Kardana A. Discordant results in human chorionic gonadotropin assays. Clin Chem 1992;38: Fein JL, Rosen SW, Weintraub BD. Increased glycosylation of serum human chorionic gonadotropin and subunits from eutopic and ectopic sources: comparison with urinary and placental forms. J Clin Endocrinol Metab 1980;50: Hoermann R, Spoettl G, Moncayo R, Mann K. Evidence for the presence of human chorionic gonadotropin (hcg) and free betasubunit of hcg in the human pituitary. J Clin Endocrinol Metab 1990;71: Vaitukaitis JL, Braunstein GD, Ross GT. A radioimmunoassay which specifically measures human chorionic gonadotropin in the presence of human luteinizing hormone. Am J Obstet Gynecol 1972;113: Stenman U-H, Alfthan H, Ranta T, Vartiainen E, Jalkanen J, Seppälä M. Serum levels of human chorionic gonadotropin in nonpregnant women and men are modulated by gonadotropin releasing hormone and sex steroids. J Clin Endocrinol Metab 1987;64: Odell WD, Griffin J. Pulsatile secretion of human chorionic gonadotropin in normal adults. N Engl J Med 1987;317: Alfthan H, Schröder J, Fraser R, Koskimies A, Halila H, Stenman U-H. Choriogonadotropin and its subunit separated by hydrophobic-interaction chromatography and quantified in serum during pregnancy by time-resolved immunofluorometric assays. Clin Chem 1988;34: O Connor JF, Schlatterer JP, Birken S, Krichevsky A, Armstrong EG, McMahon D, Canfield RE. Development of highly sensitive immunoassays to measure human chorionic gonadotropin, its -subunit, and core fragment in the urine: application to malignancies. Cancer Res 1988;48: Alfthan H, Stenman U-H. Pregnancy serum contains the -core fragment of human choriogonadotropin. J Clin Endocrinol Metab 1990;70: Neven P, Iles RK, Howes I, Sharma K, Shepherd JH, Edwards R, et al. Substantial urinary concentrations of material resembling -core fragment of chorionic gonadotropin -subunit in mid-menstrual cycle. Clin Chem 1993;39: Stenman U-H, Pettersson K, Koskimies A, Alfthan H. Changing ratio of free LH to LH in urine during the menstrual cycle [Abstract 22]. VII World Congress on Human Reproduction, Helsinki, Laidler P, Cowan D, Hider R, Keane A, Kicman A. Tryptic mapping of human chorionic gonadotropin by matrix-assisted laser desorption/ionization mass spectrometry. Rapid Commun Mass Spectrom 1995;9: Stenman U-H, Alfthan H, Koskimies A, Seppälä M, Pettersson K, Lövgren T. Monitoring the LH surge by ultrarapid and highly sensitive immunofluorometric assay. Ann N Y Acad Sci 1985;442: Demir A, Alfthan H, Stenman U-H, Voutilainen R. A clinically useful method for detecting gonadotropins in children: assessment of luteinizing hormone and follicle-stimulating hormone from urine as an alternative to serum by ultrasensitive time-resolved immunofluorometric assays. Pediatr Res 1994;36: Alfthan H, Haglund C, Dabek J, Stenman U-H. Concentrations of hcg, hcg and c hcg in serum and urine of nonpregnant women and men. Clin Chem 1992;38:

6 1298 Stenman et al.: Immunodetection of hcg: clinically and for doping control 18. Landy H, Schneyer AL, Whitcomb RW, Crowley WF. Validation of highly specific and sensitive radioimmunoassays for lutropin, follitropin, and free alpha subunit in unextracted urine. Clin Chem 1990;36: Iles RK, Lee CL, Howes I, Davies S, Edwards R, Chard T. Immunoreactive -core-like material in normal postmenopausal urine: human chorionic gonadotrophin or LH origin? Evidence for the existence of LH core. J Endocrinol 1991;133: Marshall JR, Hammond CB, Ross GT, Jacobson A, Rayford P, Odell WD. Plasma and urinary chorionic gonadotropin during early human pregnancy. Obstet Gynecol 1968;32: Pittaway DE, Reish RL, Wentz AC. Doubling times of human chorionic gonadotropin increase in early viable intrauterine pregnancies. Am J Obstet Gynecol 1986;152: Braunstein GD, Rasor J, Adler D, Danzer H, Wade ME. Serum human chorionic gonadotropin levels throughout normal pregnancy. Am J Obstet Gynecol 1976;15: Reuter AM, Gaspard UJ, Deville J-L, Vrindts-Gevaert Y, Franchimont P. Serum concentrations of human chorionic gonadotrophin and its alpha and beta subunits. 1. During normal singleton and twin pregnancies. Clin Endocrinol 1980;13: Rizkallah T, Gurpide E, Wiele RLV. Metabolism of hcg in man. J Clin Endocrinol 1969;29: Pastorfide GB, Goldstein DP, Kosasa TS, Levesque L. Serum chorionic gonadotropin activity after molar pregnancy, therapeutic abortion, and term delivery. Am J Obstet Gynecol 1974;118: Reyes FI, Winter JSD, Faiman C. Postpartum disappearance of chorionic gonadotropin from the maternal and neonatal circulations. Am J Obstet Gynecol 1985;153: Lähteenmäki P. The disappearance of HCG and return of pituitary function after abortion. Clin Endocrinol 1978;9: Ozturk M, Bellet D, Manil L, Hennen G, Frydman R, Wands J. Physiological studies of human chorionic gonadotropin (hcg), hcg and hcg as measured by specific monoclonal immunoradiometric assays. Endocrinology 1987;120: Wehmann RE, Blithe DL, Akar AH, Nisula BC. Disparity between -core levels in pregnancy urine and serum: implications for the origin of urinary -core. J Clin Endocrinol Metab 1990;70: Schroeder HR, Halter CM. Specificity of human -choriogonadotropin assays for the hormone and for an immunoreactive fragment present in urine during normal pregnancy. Clin Chem 1983; 29: Norman RJ, Buck RH, Rom L, Joubert SM. Blood or urine measurement of human chorionic gonadotropin for detection of ectopic pregnancy? A comparative study of quantitative and qualitative methods in both fluids. Obstet Gynecol 1988;71: McCready J, Braunstein GD, Helm D, Wade ME. Modification of the choriogonadotropin beta-subunit radioimmunoassay for determination of urinary choriogonadotropin. Clin Chem 1978;24: Jouppila P, Tapanainen J, Huhtaniemi I. Plasma hcg levels in patients with bleeding in the first and second trimesters of pregnancy. Br J Obstet Gynaecol 1979;86: Wilcox AJ, Weinberg CR, O Connor JF, Baird DD, Schlatterer JP, Canfield RE, et al. Incidence of early pregnancy loss. N Engl J Med 1988;319: Edmonds DK, Lindsay KS, Miller JF, Williamson E, Wood PJ. Early embryonic mortality in women. Fertil Steril 1982;38: Braunstein GD, Karow WG, Gentry WC, Rasor J, Wade ME. First-trimester chorionic gonadotropin measurements as an aid in the diagnosis of early pregnancy disorders. Am J Obstet Gynecol 1978;131: Rayford PL, Vaitukaitis JL, Ross GT, Morgan FJ, Canfield RE. Use of specific antisera to characterize biologic activity of hcg subunit preparations. Endocrinology 1972;91: Bagshawe KD. Choriocarcinoma. A model for tumor markers. Acta Oncol 1992;31: Stenman U-H, Alfthan H, Halila H. Determination of chorionic gonadotropin in serum of nonpregnant subjects and patients with trophoblastic cancer by time-resolved immunofluorometric assay. Tumor Biol 1985;5: Ozturk M, Berkowitz R, Goldstein D, Bellet D, Wands JR. Differential production of human chorionic gonadotropin and free subunits in gestational trophoblastic disease. Am J Obstet Gynecol 1988; 158: Bidart J-M, Troalen F, Lazar V, Berger P, Marcillac I, Lhomme C, et al. Monoclonal antibodies to the free -subunit of human chorionic gonadotropin define three distinct antigenic domains and distinguish between intact and nicked molecules. Endocrinology 1992;131: Mann K, Karl H-J. Molecular heterogeneity of human chorionic gonadotropin and its subunits in testicular cancer. Cancer 1983; 52: Alfthan H, Haglund C, Roberts P, Stenman U-H. Elevation of free subunit of human choriogonadotropin and core fragment of human choriogonadotropin in serum and urine of patients with malignant pancreatic and biliary disease. Cancer Res 1992;52: Kardana A, Taylor ME, Southall PJ, Boxer GM, Rowan AJ, Bagshawe KD. Urinary gonadotrophin peptide isolation and purification, and its immunohistological distribution in normal and neoplastic tissues. Br J Cancer 1988;58: Kicman AT, Brooks RV, Cowan DA. Human chorionic gonadotropin and sport. Br J Sports Med 1991;25: Wehmann RE, Blithe DL, Flack MR, Nisula BC. Metabolic clearance rate and urinary clearance of purified -core. J Clin Endocrinol Metab 1989;69:510 7.