Clinical Chemistry 49:12 2045–2049 (2003) Endocrinology and Metabolism Human Chorionic Gonadotropin Isoforms in the Diagnosis of Ectopic Pregnancy Paola T.A. Borrelli,1,2 Stephen A. Butler,1,3 Suzanne M. Docherty,1,4 Edyta M. Staite,1 Antonio L. Borrelli,2 and Ray K. Iles1,4* Background: Early diagnosis of ectopic pregnancy uses ultrasound with serial measurements of total human chorionic gonadotropin (hCG). The objective of this study was to explore the possibility that an isolated measurement of hCG isoforms/subunits rather than total hCG could be used as a single test for ectopic pregnancy. Methods: Total and intact hCG, free hCG ␤- and ␣-subunits (hCG␤ and -␣), and hCG ␤-core fragment were measured by RIA and IRMA in the serum and urine of 76 women presenting at outpatient emergency departments with a positive pregnancy test, lower abdominal pain, and/or vaginal bleeding. Final diagnoses were based on outcomes of pregnancies and tissue histology. Results: Twenty-seven of the 76 women were subsequently diagnosed with viable pregnancies, 37 with spontaneous miscarriage, and 12 with ectopic pregnancy. Concentrations of all forms of hCG were lower in cases of ectopic pregnancy and spontaneous miscarriage than in viable pregnancies. Serum samples gave better results than urine samples. The free hCG␤ isoform (P <0.0001) had 100% sensitivity at a specificity of 79% at a 281 pmol/L (6.5 ␮g/L) cutoff. Total hCG (P ⴝ 0.005) had comparable ROC characteristics with a 100% sensitivity and 68% specificity at a cutoff value of 1053 pmol/L (375 IU/L). Neither hCG␤ (P ⴝ 0.7) nor total hCG (P ⴝ 0.4) 1 Williamson Laboratory, Department of Obstetrics and Gynaecology, Barts and The London Queen Mary School of Medicine, London EC1A 7BE, UK. 2 Centro di Diagnosi Prenatale, Dipartimento di Ginecologia, Ostetricia e Neonatologia, Seconda Università degli Studi di Napoli, Naples, Italy. 3 Diagnostica Medica, Centro di Fisiopatologia della Reproduzione, Mercogliano (AV), Italy. 4 School of Health and Social Science, Middlesex University, Trent Park, Bramley Road, London, UK. *Address correspondence to this author at: Williamson Laboratory, Department of Obstetrics and Gynaecology, East Wing, St. Bartholomew’s Hospital, West Smithfield, London EC1A 7BE, UK. Fax 44-0207-601-7050; e-mail r.k.iles@qmul.ac.uk. Received May 16, 2003; accepted September 17, 2003. DOI: 10.1373/clinchem.2003.022095 could distinguish ectopic pregnancies from spontaneous miscarriage. Conclusion: Measurement of serum free hCG␤ at the time of presentation can identify women with a high probability of ectopic pregnancy who may benefit from closer surveillance, reducing the risk of tubal rupture. © 2003 American Association for Clinical Chemistry Ectopic pregnancy is the major cause of morbidity and mortality in women of reproductive age, accounting for ⬃9% of all deaths associated with first-trimester pregnancies and affecting ⬃2% of all pregnancies (1, 2 ). Since the introduction of laparoscopy in the late 1960s (3 ), considerable effort has been made to define more sensitive and specific tests for its early diagnosis because ⬎80% of ectopic pregnancies had been diagnosed after tubal rupture. Several biochemical and imaging markers have been studied as single or combined (algorithm) tests for occult ectopic pregnancy (4 – 6 ). The most common of these involves the combination of transvaginal ultrasound and serial measurements of human chorionic gonadotropin (hCG)5 (7, 8 ). The rate of increase of serum hCG indicates the presence or absence of viable uterine pregnancies: hCG doubles every 1.5 days up to 5 weeks after the last menstrual period, and then every 3.5 days from the 7th week (or when hCG is ⬎10 000 IU/L) (9 ). Although hCG concentrations do not increase at this rate in most ectopic pregnancies, they do in approximately one-third of cases, and anomalies in the dynamics of the hCG increase are also related to spontaneous miscarriage (10 ). Despite these problems, the detection of hCG remains an indispensable tool in the diagnosis of ectopic pregnancy. In women with inconclusive ultrasonographic findings, a “discriminatory serum hCG zone” has been used, where a diagnosis of ectopic pregnancy is most likely in 5 Nonstandard abbreviations: hCG, human chorionic gonadotropin; hCG␤ and hCG␣, hCG ␤- and ␣-subunits, respectively; and hCG␤cf, hCG ␤-core fragment. 2045 2046 Borrelli et al.: hCG Isoforms and Ectopic Pregnancy the absence of demonstrable intrauterine pregnancy but with an adnexal mass or free fluid in the pouch of Douglas and serum hCG ⬎6500 IU/L (11 ). With the advent of transvaginal ultrasound, the hCG cutoff has decreased to 1000 –2000 IU/L (12 ). Although the introduction of this hCG discriminatory zone appears effective, a particular hCG concentration has yet to be defined as a good diagnostic indicator for ectopic pregnancy (13 ). hCG is a glycoprotein heterodimer (14, 15 ). Only intact hCG, i.e., the ␣, ␤ heterodimer, has biological activity, but other forms of hCG can be found in the placenta, urine, and serum (16 ). hCG isoforms and subunits are generally found as metabolic breakdown products of the intact hormone, but occasionally free subunits can be expressed and released directly into the circulation. The most commonly assayed hCG fragments are the free ␤- and ␣-subunits of hCG (hCG␤ and hCG␣, respectively) and the terminal degradation product of hCG, hCG ␤-core fragment (hCG␤cf), which is found in urine (17 ). hCG assays currently used in the diagnosis of ectopic pregnancy measure so-called total hCG, i.e., intact hCG plus free hCG␤. Diagnostic algorithms for ectopic pregnancies that require serial measurements of hCG over several days necessitate delays in diagnosis because hCG test are often done on an outpatient basis, and the risk of tubal rupture is increased. The aim of our study was to explore whether measurement of the concentrations of a specific hCG isoform and/or its subunits in serum and/or urine at a single point in time could be used to test for ectopic pregnancy. Materials and Methods women and samples Between September 1998 and September 2000, matched urine and serum samples were collected at the Homerton and Whipps Cross Hospitals, London, emergency outpatient departments. Samples were collected from 97 women who presented at the Accident and Emergency Department with abdominal pain, vaginal bleeding, and a previous positive hCG test. Twenty-one were excluded for having no or insufficient matched urine or serum sample for all tests. Of the remaining 76 women, 27 (36%) were subsequently diagnosed with viable pregnancies, 37 (48%) with spontaneous miscarriage (11 missed, 6 complete, and 20 incomplete), and 12 (16%) with ectopic pregnancies. The 12 cases of ectopic pregnancy were confirmed by surgical intervention and histology. Serum and urine samples were frozen at ⫺20 °C and then assayed at the Williamson Laboratory at St. Bartholomew’s Hospital. Samples were free of contamination and contained no hemolyzed blood. Sample concentrations were determined over several assays when run undiluted and at multiple dilutions of up to 1 in 10 000 (in phosphate-buffered saline containing 100 mL/L horse serum). The CV for all assays were ⱕ5% for the midrange quality control. At the extremes of the assay range, results for quality-control samples varied up to 20%. However, sample results were recorded as the mean analyte concentration from multiple, parallel dilution, determinations. A laboratory technician and a junior faculty member of staff in laboratory medicine conducted the index tests. Data interpretation and quality-control audits were conducted by the Laboratory Director. Clinical data were collected by a House officer blind to the index test results and audited by a specialist registrar in Obstetrics and Gynecology. The North East Thames Region Health Authority Ethic Committee approved the study. hormone assays In accordance with IFCC recommendations (18 ), we have calibrated all assays in terms of picomolar concentration of the immunoreactive analyte [for a review, see Iles and Chard (16 )]. For discussion purposes, concentrations in the commonly used units for the hCG analyte have been added in parentheses where appropriate. Total hCG was measured by an in-house RIA with sheep anti-hCG␤ antiserum (S424) that binds both free hCG␤ and the intact hormone. Standards were obtained from the National Institute of Biological Standards (Potters Bar, UK). Free hCG␤, labeled with 125I by the chloramine-T method, was used as a tracer (NIH preparation CR123). The detection limit of this assay is 42 pmol/L (15 IU/L) (19 ). Intact hCG was measured with an in-house IRMA with a polyclonal anti-hCG␣ capture antibody conjugated to 1,1⬘-carbonyldiimidazole-activated cellulose and a 125Iradiolabeled monoclonal antibody (1/07; Quantum Bioscience) to epitopes on the hCG␤ C-terminal peptide. The calibration curve ranged from 2.6 to 813 pmol/L (0.93–289 IU/L) (19 ). Free hCG␣ was measured by use of an in-house RIA with an antiserum (S781) raised against free hCG␣ and affinity-adsorbed on a column of intact hCG conjugated to CNBr-activated Sepharose. The tracer was 125I-radiolabeled purified free hCG␣ (NIH preparation CR123 ␣-subunit). The calibration curve ranged from 62.5 to 6250 pmol/L (0.9 –90.6 ␮g/L) (19 ). Free hCG␤ was measured with an in-house IRMA using antiserum (S752) raised in sheep against free hCG␤ and conjugated to 1,1⬘-carbonyldiimidazole-activated cellulose. Captured hCG␤ was detected by an 125I-radiolabeled monoclonal antibody, 1/07, which recognizes epitopes on the hCG␤ C-terminal peptide. The calibration curve used hCG␤ (NIH preparation C123) and ranged from 22 to 1110 pmol/L (0.5–24.6 ␮g/L) (19 ). The in-house hCG␤cf RIA used a polyclonal antibody to hCG␤cf (S504) in a late-addition competition assay using 125I-radiolabeled purified hCG␤cf. The calibration curve ranged from 9 to 500 pmol/L (90 –5000 ␮g/L) (20 ). Data were summarized using box-and-whisker plots and ROC curves. The Kruskal–Wallis test was used for comparisons of concentrations. A P value of 0.05 was 2047 Clinical Chemistry 49, No. 12, 2003 regarded as the upper limit of significance. Statistical analysis was performed with the Stats-Direct program, and graphs were plotted using Origin 6.0 Results The mean gestational ages (weeks) in women with ectopic pregnancy, viable pregnancy, and spontaneous miscarriage were 7.0, 8.0, and 8.3 weeks, respectively; thus, no significant difference in gestational age was found among the groups. Table 1 shows the descriptive statistics for the concentrations of hCG and its isoforms in serum and urine sample for each patient group. The corresponding figure can be seen in the Data Supplement that accompanies the online version of this article at http://www. clinchem.org/content/vol49/issue12/. The Kruskal–Wallis test for nonparametric comparison of the patient groups showed that, generally, concentrations of hCG and its subunits were lower in ectopic pregnancy than in normal pregnancies to a statistically significant margin (see Table 2 in the online Data Supplement). The greatest difference was found in serum free hCG␤ in ectopic pregnancy vs viable pregnancy (P ⬍0.0001). By way of contrast, the concentrations of serum and urinary free hCG␣ were not statistically different for ectopic pregnancies and viable pregnancies (P ⫽ 0.07 and 0.48, respectively; Table 1; also see Table 2 in the online Data Supplement). However, women with early pregnancy losses by spontaneous miscarriage also had dramatically lower concentrations of the hCG isoforms and consequently could not be distinguished from those with ectopic pregnancy (P ⫽ 0.45 for serum total hCG and P ⫽ 0.70 for serum free hCG␤; see Table 2 in the online Data Supplement). The ROC curves for urinary and serum intact hCG, free hCG␤, total hCG, free hCG␣, and hCG␤cf are shown in Fig. 1. These curves show the likelihood of an ectopic pregnancy determined by the different analytes below optimal cutoff limits compared with a viable pregnancy (Table 3 in the online Data Supplement). Serum free hCG␤ and serum total hCG are the best hCG-related analytes and biological fluid for this purpose. Estimated areas under the curves (95% confidence intervals) were 0.91 (0.72–1.0) and 0.86 (0.65–1.0) for serum free hCG␤ and total hCG, respectively. The reason for this was not simply that they showed the greatest mean decrease in concentrations: When we compared the decrease in group median values, the serum analytes intact hCG and free hCG␤ had the lowest values, 12% and 13%, respectively, of the values in women with viable pregnancies. However, it was the variability in the concentrations that determined which analyte had superior diagnostic characteristic (Table 1; see also Tables 2 and 3 in the online Data Supplement). The range of values for each analyte was large, and most data sets did not follow a gaussian distribution. Serum free hCG␤ and total hCG had the Table 1. Descriptive statistics for concentrations of hCG and related fragment molecules in serum and urine samples from women with viable pregnancies, ectopic pregnancies, or spontaneous miscarriage. Statistics LMP,a weeks Urinary isoforms, pmol/mmol creatinine Total hCG Intact hCG Free hCG␤ Free hCG␣ hCG␤cf Serum isoforms, pmol/L Total hCG Intact hCG Free hCG␤ Free hCG␣ a Viable pregnancies Mean (SD) Range Mean (SD) Median (interquartile Mean (SD) Median (interquartile Mean (SD) Median (interquartile Mean (SD) Median (interquartile Mean (SD) Median (interquartile Mean (SD) Median (interquartile Mean (SD) Median (interquartile Mean (SD) Median (interquartile Mean (SD) Median (interquartile LMP, last menstrual period. b Interquartile range is 75th–25th centile. 8.2 (2.3) 6.8–10 range)b range) range) range) range) range) range) range) range) 8109 (10 824) 2874 (10 982–1599) 22 814 (35 875) 8613 (26 246–3408) 4007 (5958) 1481 (4233–563) 33 407 (144 020) 3519 (6632–947) 17 477 (27 201) 12 753 (20 336–1034) 8909 (23 299) 1548 (5891–688) 25 737 (200 813) 203 993 (480 245–39 231) 936 (939) 645 (989–357) 3171 (3714) 2664 (4329–1132) Ectopic pregnancies 7 (1.2) 6.7–8.0 1843 (3501) 716 (1591–178) 4499 (8041) 1649 (4395–660) 1698 (2020) 820 (3300–83) 26 054 (70 693) 1317 (5914–404) 4017 (4545) 2443 (7847–349) 360 (271) 284 (624–153) 59 664 (85 175) 24 719 (88 050–6134) 115 (90) 82 (195–35) 1231 (1177) 793 (2314–167) Spontaneous miscarriage 9 (2.8) 6.8–11 1320 (2091) 657 (1743–227) 4675 (8447) 890 (5651–188) 895 (1769) 299 (846–48) 3457 (5657) 850 (3110–251) 2440 (3255) 920 (3137–398) 7988 (26 813) 430 (1634–123) 38 900 (67 473) 16 977 (39 953–1153) 188 (474) 69 (129–34) 1027 (1105) 343 (1948–167) 2048 Borrelli et al.: hCG Isoforms and Ectopic Pregnancy Fig. 1. ROC curves for the hCG isoforms in urine (A) and serum (B). Solid light gray line, total hCG; solid black line, intact hCG, dashed black line, hCG␤; dotted light gray line, hCG␣; dot-dashed medium gray line, hCG␤cf. tightest distribution of values around the median (and mean) values and consequently gave the best results. Discussion The validity of the current diagnostic algorithm for ectopic pregnancy, i.e., ultrasound and serial measurements of hCG, has already been demonstrated by a decrease in the risk of tubal rupture and morbidity of women whose diagnosis was achieved by this protocol (21 ). Nevertheless, there are factors that reduce the sensitivity of ultrasound, such as lack of experience by the ultrasonographer, obesity of the patient, uterine myoma, or more commonly, “inconclusive ultrasound”, which then must be repeated several times (22 ). The subjective nature of scanning contributes to this uncertainty. When considering these drawbacks to a diagnostic algorithm, it is clear that a single blood test that would help to separate patients with high risk for an ectopic pregnancy from low-risk patients could possibly advance clinical practice and be less expensive. This study was designed to examine whether the answer lay within the isoforms of hCG produced by an ectopic pregnancy. We found that concentrations of hCG and its subunits were significantly lower in women with an ectopic pregnancy than in those with a normal pregnancy. There was, however, considerable variability in individual patient results, and serum results were far better discriminators than urinary concentrations, even when corrected for creatinine concentration. Nevertheless, certain hCG-related isoforms were clearly better than others. Serum free hCG␤ was ⬍500 pmol/L (⬍10 ng/mL) in all of the study women presenting with ectopic pregnancies but ⬎500 pmol/L in the vast majority of those women with viable pregnancies. This is consistent with the findings of Holman et al. (23 ), who found that all women with ectopic pregnancies in their study had hCG␤ concentrations ⬍25 ␮g/L (1250 pmol/L). ROC analysis demonstrated up to 100% sensitivity for ectopic pregnancy with a 21% falsepositive rate (viable pregnancy) for serum free hCG␤ at a cutoff of 281 pmol/L (6.5 ng/mL). Total hCG performed well when measured in both serum and urine samples, but again serum measurements were best, giving 100% sensitivity with a 32% false-positive rate (viable pregnancy) at a cutoff of 1053 pmol/L (375 IU/L). The data would suggest that free hCG␤ might have marginally better performance characteristics with respect to the diagnosis of ectopic pregnancy than total hCG. However, the confidence intervals of the areas under the ROC curves for these two measurements overlapped to such an extent that the difference in performance was not statistically significant. A much larger study is required to determine whether there is a significant difference in the diagnostic accuracy of free hCG␤ and total hCG for ectopic pregnancy. Measurement of intact hCG and free hCG␣ gave less significant results (Table 3 in the online Data Supplement), a finding inconsistent with that of Barnea et al. (24 ), who reported that in combination with total hCG, free hCG␣ could be used as a marker of ectopic pregnancy. Our results are also in disagreement with those of Reuter et al. (25 ), who found that free hCG␣ concentrations were unchanged in ectopic pregnancies and increased in spontaneous abortions. No urinary analyte apart from hCG␤cf was found to be a good predictor of ectopic pregnancy. Cole et al. (26 ) proposed that hCG␤cf could predict up to 84% of all ectopic pregnancies with only 1% false positives. In our study, although hCG␤cf measurements reached 100% sensitivity at a very high cutoff of 14 201 pmol/L, this was at the expense of a very poor specificity of 48% (see Table 3 in the online Data Supplement). Although serum free hCG␤ and total hCG concentrations resolved ectopic pregnancy from viable pregnancy with a high degree of diagnostic efficiency, spontaneous miscarriage could not be distinguished. Because ⬃50% of women presenting with suspected ectopic pregnancy will subsequently be diagnosed with spontaneous miscarriage, the clinical utility of a single hCG measurement is limited. Other markers, such as serum progesterone measurements, have been used to distinguish viable from Clinical Chemistry 49, No. 12, 2003 nonviable (spontaneous miscarriage) pregnancies but alone are unable to resolve ectopic from viable intrauterine pregnancies (22, 27 ). To date, no single-point cutoff value for hCG has been proposed as a discriminator of ectopic pregnancy. This pilot study suggests that quantitative measurement of free hCG␤ in serum may provide this single-point cutoff for resolving women at high risk from those with low risk of ectopic pregnancy. Serum total hCG gave diagnostic discriminator characteristics almost identical to those of serum free hCG␤. This may be because total hCG assays measure both free hCG␤ and intact hCG, with the free hCG␤ component being the discriminating factor. The study group in this report is too small to show any statistical difference between the ability of total hCG and free hCG␤ to discriminate high-risk from low-risk suspected ectopic pregnancies. In conclusion, based on our results, a single-point measure of total hCG or fee hCG␤ could potentially improve clinical practice in the diagnosis of ectopic pregnancy by defining low risk of ectopic from high risk of ectopic pregnancy for those women presenting at emergency departments with lower abdominal pain and/or vaginal bleeding and with a positive pregnancy test. A much larger study is underway in a clinical setting, using hospital-based automated assay platforms to compare the abilities of total hCG and free hCG␤ to differentiate ectopic from viable pregnancies. This, in combination with ultrasound and other serum markers that distinguish ectopic pregnancy from spontaneous miscarriage, may be a more useful test for ectopic pregnancy than current diagnostic algorithms. References 1. Ectopic pregnancy—United States, 1990 –1992. MMWR Morb Mortal Wkly Rep 1995;44:46 – 8. 2. Luciano AA, Roy G, Solima E. Ectopic pregnancy: from surgical emergency to medical management. Ann N Y Acad Sci 2001;943: 235–54. 3. Kadar N. Early recourse to laparoscopy in the management of suspected ectopic pregnancy. J Reprod Med 1990;35:1153– 6. 4. Braunstein GD, Ash RH. Predictive value analysis of measurements of hCG, pregnancy specific ␤-1-glycoprotein, placental lactogen and cysteine aminopeptidase for the diagnosis of ectopic pregnancy. Fertil Steril 1983;39:62–7. 5. Tornehave D, Chemnitz J, Westergaard JG, Teisner B, Poulsen HK, Bolton AE, et al. Placental proteins in peripheral blood and tissues in ectopic pregnancies. Gynecol Obstet Invest 1986;23:97–102. 6. Stabile I, Olajide F, Chard T, Grudzinskas JG. Maternal serum ␣-fetoprotein levels in ectopic pregnancy. Hum Reprod 1989;4: 835– 6. 7. Cacciatore B, Ylostalo P, Stenman UH, Widholm O. Suspected ectopic pregnancy: ultrasound findings and hCG levels assessed by an immunofluorometric assay. Br J Obstet Gynaecol 1988;95: 497–502. 8. Ankum WM. Diagnosing suspected ectopic pregnancy. BMJ 2000; 321:1235– 6. 9. Frits MA, Guo SM. Doubling time of hCG in early normal pregnancy: View publication stats 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 2049 relationship to hCG concentration and gestational age. Fertil Steril 1987;47:584 –9. Shepherd RW, Patton PE, Novy MJ, Burry KA. Serial data hCG measurements in the early detection of ectopic pregnancy. Obstet Gynecol 1990;75:417–20. Kadar N, De Vore G, Romero R. The discriminatory hCG zone. Its use in the sonographic evaluation of ectopic pregnancy. Obstet Gynecol 1981;58:156 – 61. Aleem FA, De Fazio M, Gintautas J. Endovaginal sonography for the early diagnosis of intrauterine and ectopic pregnancies. Hum Reprod 1990;5:755– 8. Di Marchi JM, Kosasa TS, Hale RW. What is the significance of human chorionic gonadotropin value in ectopic pregnancy? Obstet Gynecol 1989;74:851–5. Bellisario R, Carlsen RB, Bahl OP. Human chorionic gonadotropin. Linear amino acid sequence of the ␣-subunit. J Biol Chem 1973;248:6796 – 809. Carlsen RB, Bahl OP, Swaminathan N. Human chorionic gonadotropin. Linear amino acid sequence of the ␤-subunit. J Biol Chem 1973;248:6810 –27. Iles RK, Chard T. Molecular insights into the structure and function of human chorionic gonadotrophin. J Mol Endocrinol 1993;10:217–34. Cole LA. Immunoassay of human chorionic gonadotropin, its free subunits and metabolites. Clin Chem 1997;43:2233– 43. Birken S, Berger P, Bidart JM, Weber M, Bristow A, Norman R, et al. Preparation and characterization of new WHO reference reagents for human chorionic gonadotropin and metabolites. Clin Chem 2003;49:144 –54. Iles RK, Javid MK, Gunn LK, Chard T. Cross-reaction with luteinizing hormone ␤-core is responsible for the age-dependent increase of immunoreactive ␤-core fragment of human chorionic gonadotropin in women with nonmalignant conditions. Clin Chem 1999;45:532– 8. Lee CL, Iles RK, Shepherd JH, Hudson CN, Chard T. The purification and development of a radioimmunoassay for ␤-core fragment of human chorionic gonadotrophin in urine: application as a marker of gynaecological cancer in premenopausal and postmenopausal women. J Endocrinol 1991;130:481–9. Koh GH, Yeo GS. Diagnosis of ectopic pregnancy. Why we need a protocol. Singapore Med J 1997;38:369 –74. Gracia CR, Barnhart KT. Diagnosing ectopic pregnancy: decision analysis comparing six strategies. Obstet Gynecol 2001;93:464 – 70. Holman JF, Tyrey EL, Hammond CB. A contemporary approach to suspected ectopic pregnancy with use of quantitative and qualitative assays for the ␤-subunit of human chorionic gonadotropin and sonography. Am J Obstet Gynecol 1984;150:151–7. Barnea ER, Oelsner G, Benveniste R, Romero R, De Cherney AH. Progesterone, estradiol and ␣-human chorionic gonadotropin secretion in women with ectopic pregnancy. J Clin Endocrinol Metab 1986;62:529 –31. Reuter AM, Gaspard U, Deville JC, Vrindts-Gevaertv Y, Franchimont P. Evolution of serum levels of hCG and its subunits in normal pregnancy and in threatened and spontaneous abortion. Acta Endocrinol 1977;(Suppl):212–32A. Cole LA, Kardana A, Seifer DB, Bohler HC Jr. Urine hCG ␤-subunit core fragment, a sensitive test for ectopic pregnancy. J Clin Endocrinol Metab 1994;78:497–9. Mol BWJ, Lijimer JG, Ankum WM, van der Veen F, Bossuyt PMM. The accuracy of single serum progesterone measurement in the diagnosis of ectopic pregnancy: a meta-analysis. Hum Reprod 1998;13:3220 –7.