Human chorionic gonadotrophin concentrations in early pregnancy after in-vitro fertilization

Sverre Bjercke1,3, Tom Tanbo1, Per Olav Dale1, Lars Mørkrid2 and Thomas Åbyholm1

1 Department of Obstetrics and Gynecology and 2 Department of Clinical Chemistry, Rikshospitalet,0860 Oslo, Norway


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
There is increased risk of early pregnancy loss after assisted reproduction. In this study the use of serum human chorionic gonadotrophin (HCG) concentrations on day 12 after in-vitro fertilization (IVF) and embryo transfer was evaluated to predict pregnancy outcome. A total of 417 IVF pregnancies were included. Early pregnancy loss was defined as biochemical pregnancies, ectopic pregnancies and first trimester abortions. Vital pregnancies were defined as delivered singletons, multiple pregnancies and second trimester abortions. On the post embryo transfer day 12, the mean HCG concentration of the vital pregnancy group was significantly higher than in early pregnancy loss outcomes (P < 0.00001). Receiver operating characteristic (ROC) curve analysis was performed to evaluate the cut-off value of HCG giving maximal sensitivity and specificity in order to discriminate early pregnancy losses from vital pregnancies. A patient with a HCG value higher than the calculated cut-off value (55 IU/l) had a 90% chance of having a vital pregnancy after IVF and embryo transfer. It can be concluded that a discriminatory HCG value on day 12 after IVF and embryo transfer cycles may be useful in predicting pregnancy outcome and may guide clinicians in identifying those pregnancies at risk for adverse outcomes and instituting more intensive surveillance in this population.

Key words: : clinical pregnancy/human chorionic gonadotrophin/in-vitro fertilization


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Pregnancies obtained after in-vitro fertilization (IVF) and embryo transfer are at increased risk of adverse outcome compared with natural conceptions (Edwards et al ., 1984 Go; Ben-Rafael et al ., 1988 Go; Correy et al ., 1988 Go). A reliable and inexpensive diagnostic test to differentiate between vital pregnancies and pregnancies with early adverse outcome might reduce the psychological tension and anxiety present in many of these patients, and also reduce the cost by making the treatment more efficient. On the other hand, in patients with a high risk of unfavourable outcome based on the test a more careful follow up might reduce the risks associated with these abnormal pregnancies.

Ultrasound (US) examination is effective for evaluation of ongoing pregnancies, but a gestational sac is not reliably visible until 33–37 days after the luteinizing hormone (LH) surge (Shapiro et al ., 1992 Go). As a result of this inability of US to identify very early pregnancy abnormalities, there is an ongoing effort to institute a method that can forecast pregnancy outcome. Several studies have investigated hormones like progesterone, oestradiol, relaxin, pregnancy-specific ß1-glycoprotein, inhibin, and CA-125 in predicting prognosis (Witt et al ., 1990 Go; Hahlin et al ., 1991 Go). The exact role of these markers and their clinical utility, however, has not yet been established. Researchers have reported the usefulness of early serum human chorionic gonadotrophin (HCG) (Confino et al ., 1986 Go; Dor et al ., 1988 Go; Heiner et al ., 1992 Go; Fridstrøm et al ., 1995 Go; Glatstein et al ., 1995 Go). One of these studies has reported clinically discriminatory HCG values on different days post embryo transfer (Glatstein et al ., 1995 Go), but none of them as early as on day 12.

The aim of this study was to investigate whether a single HCG value on day 12 after embryo transfer could be used as a marker to predict the outcome of a pregnancy after IVF.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Subjects
During the period of January 1994 to December 1996, 417 pregnancies obtained after IVF and embryo transfer treatment for infertility due to tubal factor (61%), polycystic ovarian syndrome (4%), endometriosis (12%), unexplained (14%), or male factor (9%) were studied. Conception was defined as a HCG serum concentration higher than 10 IU/l. All patients had their HCG assay performed inside our institution on day 12 after embryo transfer. The median age of the participant was 33 (21–44 years).

In-vitro fertilization protocol
The patients underwent IVF and embryo transfer according to described protocols (Åbyholm et al ., 1990 Go; Tanbo et al ., 1995 Go). Pituitary down regulation with a gonadotrophin releasing hormone agonist (GnRHa) (Suprefact ® ; Hoechest, Frankfurt am Main, Germany ) was used in all patients. Ovarian hyperstimulation was performed using human menopausal gonadotrophin (HMG) (Pergonal ® ; Serono, Aubonne, Switzerland ) and follicle stimulating hormone (FSH) (Fertinorm ® and Fertinorm HP ® ; Serono ). Transvaginal ultrasound (US) and serum oestradiol concentrations were used to monitor the IVF cycle. When three or more of the follicles had a diameter exceeding 18 mm, 10 000 IU units of HCG (Profasi ® ; Serono ) were administered s.c. to induce ovulation. Oocyte aspiration was performed approximately 34–36 h after HCG was given using a standard transvaginal US-guided approach. Embryo transfer was performed 2 days after oocyte aspiration and a maximum of two embryos was transferred. All patients included in this series received progesterone support for 4 weeks, either as vaginal suppositories (600 mg daily) (Progestan ® ; Organon, Oss, The Netherlands ) or i.m. injections (25 mg daily) starting on the day prior to oocyte aspiration.

Serum HCG assay
Samples of venous blood were collected routinely on the 12th day after embryo transfer. Serum HCG concentrations were measured by microparticle enzyme immunoassay for the intact HCG molecule (Imx ® ; Abbot Laboratories, Abbot Park, IL, USA ) at the Central Chemical Laboratory of our hospital. The assay was calibrated using the World Health Organization (WHO) Third International Standard (75/537) (formerly designated the WHO First International Reference Preparation). The intra-assay coefficient of variation (CV) was 4.1% for a mean HCG of 23 IU/l; 3.9% for a HCG mean of 143 IU/l, and 3.4% for a mean HCG of 707 IU/l. The inter-assay CV was 4.6%, 6.1% and 9.6% respectively, and the sensitivity was 2 IU/l. Serum samples that were not obtained by day 12 after embryo transfer were discarded. As the HCG analyses were only performed on Monday, Wednesday and Friday, and not at weekends, some of the samples were stored for 1 to 2 days before being tested. If stored for more than 24 h, the samples were frozen at –24°C. Otherwise they were maintained at 4°C.

Pregnancy outcome
Pregnancy outcomes were divided into two groups: early pregnancy losses and vital pregnancies. Early losses included pregnancies classified as chemical (subclinical), ectopic gestations and first trimester spontaneous abortions. Vital pregnancies were defined as delivered singletons, multiple pregnancies or second trimester abortions occurring with a documented fetal heart, before the abortion. Our series contained five patients who experienced ectopic pregnancy. No heterotopic pregnancies were detected among our patients during the study period. The same study was performed with ectopic pregnancies excluded from the failures, and only singleton pregnancies were included in the successful group.

Statistical analysis
The Mann–Whitney U test was used to compare the differences between the groups. A receiver operating characteristic (ROC) curve analysis (McNeil et al ., 1975 Go; Hanley and McNeil, 1982 Go) was used to estimate the predictive power of the measured variables (Swets, 1979 Go), and to determine the HCG cut-off that optimally discriminated early pregnancy losses from vital pregnancy cycles. ROC curve was obtained by plotting the sensitivity of the HCG value versus 1-specificity for a series of multiple cut-off points on the 12th post embryo transfer day.

The cut-off point was chosen to maximize sensitivity as well as specificity. As a measure for the diagnostic ability of the cut-off to predict pregnancy outcome, the positive and negative predictive values were determined at the calculated cut-off value. As described (Glatstein et al ., 1995 Go) the following definitions were used: sensitivity: the probability that a patient with an early pregnancy loss will have an HCG value less than the cut-off value, the true positive rate; specificity: the probability that a patient with a vital pregnancy will have HCG value greater than the cut-off value; positive predictive value: the probability that a patient with a HCG value less than the cut-off will have an early pregnancy loss; and negative predictive value: the probability that a patient with an HCG value greater than the cut-off will have a vital pregnancy. Then positive and negative post test probability were calculated and found to coincide with the positive and negative predictive value respectively. Unless otherwise stated, values are expressed as means ± SE of the mean with a probability value of P < 0.05.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Table I Go shows the pregnancy outcomes of the 417 studied patients. Vital pregnancies represented 74% and early pregnancy failure 26% of all pregnancies, respectively. There is a statistically significant difference between the HCG concentrations in the early pregnancy failure group versus the vital pregnancy group (P < 0.00001 ). The HCG concentration (mean + SE) for the three groups: spontaneous abortions, vital singletons and vital twins are illustrated in Figure 1 Go. Using the Mann–Whitney U test, there was a statistically significant difference in serum HCG values between singleton and multiple pregnancies (P < 0.0001 ) and between singletons and spontaneous abortions (P < 0.0001 ).


View this table:
[in this window]
[in a new window]
 
Table I. Pregnancy outcome in patients undergoing in-vitro fertilization with an initial human chorionicgonadotrophin (HCG) value >10 mIU/l
 


View larger version (10K):
[in this window]
[in a new window]
 
Figure 1. Day 12 post embryo transfer (mean ± SEM) of serum human chorionic gonadotrophin(HCG) in spontaneous abortions (sp. abort) (n = 107), viable singletons (n = 239) and viable twins (n = 67). The receiver operating characteristic (ROC) calculated cut-off is shown in a hatched line, dividingthe spontaneous abortions from viable pregnancies.

 
To determine the clinical significance of these statistical findings, a ROC curve was derived (Figure 2 Go). By analysing the ROC curve the cut-off point giving optimal sensitivity and specificity was found to be 55 IU/l. At the cut-off value of 55 IU/l the negative predictive value was 90%. Stated alternatively, this means that a patient who had a HCG concentration higher than the cut-off value had a 90% chance of having a vital pregnancy. The positive predictive value was 60%. Therefore, a patient in this study population who presented with a HCG concentration less than the cut-off (55 IU/l) had a probability of an early pregnancy failure equal to that percentage day 12 post embryo transfer. The negative and positive predictive values were found to correspond to the negative and positive post test probability figures which were calculated to be 91% and 61% respectively.



View larger version (20K):
[in this window]
[in a new window]
 
Figure 2. Receiver operating characteristic (ROC) curve plotting true positive versus false positive rate and theoptimal discriminatory human chorionic gonadotrophin (HCG) cut-off for day 12 post embryotransfer.

 
When the same studies were performed excluding ectopic pregnancies from the early pregnancy failure group, and multiple gestations from the successful pregnancy group, an identical cut-off value of 55 IU/l was found. Also the positive and negative predictive values were of the same order, 61% and 87% respectively. Thus the optimal cut-off value in order to discriminate between spontaneous abortion and singleton pregnancies is the same as the cut-off value that discriminates between the groups when the ectopic pregnancies and multiple gestations were included (55 IU/l).

Figure 3 Go illustrates the relative distribution of early pregnancy losses and vital pregnancies from the selected cut-off point of 55 IU/l to 100 IU/l. At the horizontal bar representing the chosen cut-off value at 55 IU/l 60% will have early pregnancy losses while 40% will have vital pregnancies. With increasing HCG cut-off values the figure shows that the relative distribution of vital pregnancies increases.



View larger version (13K):
[in this window]
[in a new window]
 
Figure 3. Horizontal bars illustrate the percentage (%) distribution of pregnancies classified as earlypregnancy losses (true positive) and vital pregnancies (false positive) with increasing cut-off valuesof human chorionic gonadotrophin (HCG). The vertical line in the middle of the figure represents thedivision between the percentage of early pregnancy losses (left) and the percentage of vitalpregnancies (right).

 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Various studies have addressed the question about early HCG values and their relationship to pregnancy outcome after IVF (Confino et al ., 1986 Go; Deutinger et al , 1986 Go; Dor et al ., 1988 Go; Yamashita et al. , 1989 Go; Lower and Yovich, 1993 Go; Fridstrøm et al ., 1995 Go; Glatstein et al ., 1995 Go). This study is unique in that the cut-off value of HCG was calculated on one single day and as early as day 12 post embryo transfer based on a generated ROC curve. The testing for HCG was performed at a single laboratory only, minimizing intra-assay and inter-assay variability (Confino et al ., 1986 Go). This report also included 417 IVF pregnancy cycles, the largest series to date in the literature of this question. In addition, the values reported were based on a commonly accepted HCG reference, the WHO Third International Standard.

The present study shows that, by means of a single serum HCG determination on day 12 after embryo transfer, a group of patients can be identified that is at increased risk of carrying an abnormal pregnancy. At the cut-off value of 55 IU/l patients testing positively for HCG had a 61% risk of carrying an ectopic or non-viable intrauterine pregnancy. Therefore close monitoring of the course of the pregnancy is called for. Among those with a HCG value higher than 55 IU/l there was a 90% probability of a vital pregnancy. In total 26% of the patients in this study experienced early pregnancy loss while 74% had vital pregnancies. The latter group comprised one second trimester loss (16th week). Second trimester abortions are in this study regarded as successful implantation. As explained (Glatstein et al ., 1995 Go) early HCG concentrations can predict implantation outcome, but it cannot invariably predict pregnancy outcome. Therefore, it would be unlikely that a single early HCG value could reliably predict the occurrence of a live birth, an event influenced by multiple factors other than implantation quality like a second or third trimester abortion due to an incompetent cervix.

The calculated cut-off reported in this study was chosen to best discriminate between early pregnancy losses and vital pregnancies. There is a limit to what can be achieved with biochemical diagnosis in early pregnancy, since pathological pregnancies often have a normal growth rate early in their course (e.g. blighted ovum, initially beating heart). The earlier the test is performed, the higher is the rate of false negative tests. By definition, there is a trade off between sensitivity and specificity. If one wishes to increase the HCG cut-off value on day 12 after embryo transfer from 55 IU/l to 100 IU/l there is a decline in positive predictive value (true spontaneous abortion) from 60% to 32% (Figure 3 Go). At the same time the sensitivity increases from 73% to 90%, while the specificity falls from 83% to 35%. Among the 417 patients studied, there were only five (1.2%) who experienced ectopic pregnancy. Since there were very few ectopic pregnancies in the study, the main aim was to find a cut-off value that could discriminate between ordinary spontaneous and successful pregnancies. It is doubtful if any higher sensitivity (73%) than the one estimated at the cut-off value of 55 IU/l would be more suitable to detect ectopic pregnancies with enough certainty to be diagnostic. Mol et al . (1997) have reported on optimal HCG cut-off values for ectopic pregnancies in a group of 86 women where 24 (28%) experienced ectopic pregnancies (Mol et al ., 1997 Go). Their measurements were done on days 6, 9 and 15. The authors concluded that serum HCG measurement 9 days after embryo transfer could identify pregnancy failures with a 100% specificity at a cut-off value of 18 IU/l, but it could not identify patients with ectopic pregnancies with enough certainty to justify immediate treatment. Another recent study reported no difference in the measurable isoforms of HCG between normal intrauterine pregnancy and ectopic pregnancies (Mock et al ., 1998 Go). Abdominal pain or bleeding is usually an additional sign necessary to raise suspicion of ectopic pregnancy. Under these circumstances transvaginal sonography is of great diagnostic aid (Mol et al ., 1997 Go).

If one was to select one group for closer follow up on basis of the HCG measured on day 12 after embryo transfer it ought to be the patients with values equal to or below 55 IU/l. The lower the HCG value (lower sensitivity) the higher is the likelihood of non-viable pregnancy (specificity).

To set a higher cut-off point in order to increase the sensitivity does not seem reasonable in as much as this would decrease the specificity considerably and not contribute to a better discrimination between vital and the potentially life-threatening ectopic pregnancies.

The HCG concentrations were significantly higher on day 12 after embryo transfer for vital pregnancy to early pregnancy losses. This confirms previous studies (Heiner et al ., 1992 Go; Keith et al ., 1993 Go; Fridstrøm et al ., 1995 Go; Glatstein et al ., 1995 Go). By choosing day 12 as the test day, unnecessary delay of pregnancy diagnosis could be avoided, which is crucial both from a psychological and a medical point of view. Relief of tension might not only be important for the patient's well-being (Friedman, 1989 Go), but also for the outcome of pregnancy (Stray-Pedersen and Stray-Pedersen, 1984 Go). For patients using progesterone as luteal support, an early and reliable test for pregnancy is desirable in order to terminate luteal support in patients not achieving pregnancy.

It was shown in this series that multiple pregnancies had significantly higher HCG values compared with singletons on day 12 after embryo transfer. This finding is in agreement with previous reports (Heiner et al ., 1992 Go; Fridstrøm et al ., 1995 Go; Glatstein et al ., 1995 Go) but ultrasound confirmation of the number of fetuses and their viability still should be used to make clinical recommendations to patients in any case of multiple pregnancy.

Further it was calculated that this series of singletons had significantly higher HCG on day 12 after embryo transfer than spontaneous abortions. Only very small changes in predictive values were observed after excluding multiple pregnancies from the group of vital pregnancies.

As explained (Fridstrøm et al ., 1995 Go), several studies have shown that serial HCG determinations for differential diagnosis of early pregnancy are superior to single measurements with regard to diagnostic power (Marshall et al .,1968; Braunstein et al ., 1978 Go; Kadar et al , 1981 Go; Ankum et al ., 1993 Go). However, these studies concern patients presenting with actual suspicion of pathological pregnancy, in which active diagnostic measures are warranted. In women who experience recurrent abortions the most useful measurements in predicting the outcome in early pregnancy was recently reported to be ßHCG and gestation-sac diameter (Li et al ., 1998 Go). In contrast, more than 75% of the pregnant cycles in the present study led to successful implantation. It therefore seems reasonable to spare these patients the inconvenience of repeated blood sampling at short intervals.

As pointed out by Glatstein et al. (1995), ideally, each institution should analyse their own data to determine HCG cut-offs based on their own experience. Although quality control methods such as HCG reference preparations are standardized between clinical laboratories, minor differences in protocol may influence values. Glatstein et al . suggest a prospective study in multiple centres to confirm the validity of the cut-off values.

In summary, it can be concluded that HCG cut-off values determined by a ROC curve analysis is useful for discriminating between implantation successes and early pregnancy losses on day 12 after embryo transfer. The cut-off value might aid in the prognosis, clinical management and counselling of the patients. The lower the HCG values found, the greater is the need for repeated measurement and a closer follow up of the patients. Pregnant patients with HCG values greater than the cut-off value are followed with transvaginal US at 4–5 weeks post embryo transfer. Improved techniques for identification of pathological pregnancies after IVF will be of benefit for the patients and might also have cost saving consequences.


    Notes
 
3 To whom correspondence should be addressed Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Åbyholm, T., Tanbo, T., Dale, P.O.et al. (1990) The first attempt at IVF treatment.Results and requirements for satisfactory success rate.Eur. J. Obstet. Gynecol. Reprod. Biol.,38,125–132.[ISI]

Ankum, W.M., Van der Veen, F., Hamarlynck, J.V. and Lammes, F.B. (1993) Transvaginal sonography and human chorionic gonadotrophin measurements in suspected ectopicpregnancy: a detailed analysis of a diagnostic approach.Hum. Reprod.,8,1307–1311.[Abstract]

Ben-Rafael, Z., Fateh, M., Flickinger, G.L.et al. (1988) Incidence of abortion in pregnancies after in-vitro fertilization and embryo transfer.Obstet. Gynecol.,71,297–300.[Abstract]

Braunstein, G.D., Karow, W.G., Gentry, W.C.et al. (1978).First trimester chorionic gonadotrophin diagnosis of early human pregnancy disorder.Am. J. Obstet. Gynecol.,131,25–32.[ISI][Medline]

Confino, E., Demir, R., Friberg, J. and Gleicher, N. (1986) The predictive value of HCG-ß subunit levels in pregnancies achieved by in-vitro fertilizationand embryo transfer: an international collaborative study.Fertil. Steril.,45,526–531.[ISI][Medline]

Correy, J.F., Watkins, R.A., Bradfield. G.F.et al. (1988) Spontaneous pregnancies and pregnancies as a result of treatment on an in-vitro fertilizationprogram terminating in ectopic pregnancies or spontaneous abortions.Fertil. Steril.,50,85–88.[ISI][Medline]

Deutinger, J., Neumark, J., Reinthaller, A.et al. (1986) Pregnancy-specific parameters in early pregnancies after in-vitro fertilization: a prediction of earlypregnancy loss.Fertil. Steril.,46,77–88.[ISI][Medline]

Dor, J., Rudak, E., Rotmench, S.et al. (1988) The role of early post-implantation ßHCG concentrations in the outcome of pregnanciesfollowing in-vitro fertilization.Hum. Reprod.,3,663–667.[Abstract]

Edwards, R.G., Fishel, S.B., Cohen, J.et al. (1984) Factors influencing the success of in-vitro fertilization for alleviating human infertility.J. In Vitro Fert. Embryo Transf.,1,3–23.[Medline]

Friedman, T. (1989) Infertility and assisted reproduction.Baillière's Clinical Obstetrics and Gynecology 1989,3,751–766.

Fridstrøm, M., Garoff, L., Sjøblom, P.et al. (1995) Human chorionic gonadotrophin patterns in early pregnancy after assisted reproduction.Acta Obstet. Gynecol. Scand.,74,534–538.[ISI][Medline]

Glatstein, I.Z., Hornstein, M.D., Kahana, M.J.et al. (1995) The predictive value of discriminatory human chorionic gonadotrophin levels in the diagnosis ofimplantation outcome in-vitro fertilization cycles.Fertil. Steril.,63,350–356.[ISI][Medline]

Hahlin, M., Sjøblom, P. and Lindblom, B. (1991) Combined use of progesterone and human chorionic gonadotrophin determination for differentialdiagnosis of very early pregnancy.Fertil. Steril.,55,492–496.[ISI][Medline]

Hanley, J.A. and McNeil, B.J. (1982) The meaning and use of the area under a receiver operating characteristic curve.Radiology,143,29–36.[Abstract]

Heiner, J.S., Kerin, J.F. and Schmidt, L.L. (1992) Can a single, early quantitative human chorionic gonadotrophin measurement in an in-vitrofertilization–gamete intraFallopian transfer program predict pregnancy outcome?Fertil. Steril.,58,373–377.[ISI][Medline]

Kadar, N., Caldwell, B.V. and Romero, R. (1981) A method of screening for ectopic pregnancy and its indications.Obstet. Gynecol.,58,162–165.[Abstract]

Keith, S.C., London, S.N., Weizman, G.A.et al. (1993) Serial transvaginal ultrasound scans and ß-human chorionic gonadotrophin levels in earlysingleton, multiple and abnormal pregnancies.Fertil. Steril.,59,1007–1010.[ISI][Medline]

Li, T.C., Spring, P.G., Bygrave, C.et al. (1998) The value of biochemical and ultrasound measurements in predicting pregnancy outcome inwomen with a history of recurrent miscarriage.Hum. Reprod.,13,3525–3529.[Abstract]

Lower, A.M. and Yovich, J.L. (1993) The value of serum levels of oestradiol, progesteron, and beta-human chorionic gonadotrophin inthe prediction of early pregnancy loss.Hum. Reprod.,7,711–717.[Abstract]

Marshall, J.R., Hammond, C.B., Griff, W.C.et al. (1968) Plasma and urinary chorionic gonadotrophin during early pregnancy.Obstet. Gynecol.,32,760–764.[ISI][Medline]

McNeil, B.J., Keeler, E., Edelstein, S.J.et al. (1975) Primer on certain elements of medical decision making.N. Engl. J. Med.,293,211–215.[Abstract]

Mock, P., Bischof, P., Rivest, R.et al. (1998) Modulation of human chorionic gonadotrophin bioactivity during the first trimester of pregnancy.Hum. Reprod.,13,2629–2632.[Abstract]

Mol, B.W.J., van der Veen, F., Hajenius, P.J.et al. (1997).Diagnosis of ectopic pregnancy after in-vitro fertilization and embryo transfer.Fertil. Steril.,68,1027–1032.[ISI][Medline]

Shapiro, B.S., Eschbar, M., Makuech, R.et al. (1992) A model-based prediction for transvaginal ultrasonographic identification of early intrauterinepregnancy.Am. J. Obstet. Gynecol.,166,1495–1500.[ISI][Medline]

Stray-Pedersen, B. and Stray-Pedersen, S. (1984) Etiologic factors and subsequent reproductive performance in 195 couples with prior history ofhabitual abortion.Am. J. Obstet. Gynecol.,148,140–146.[ISI][Medline]

Swets, J. (1979) ROC-analysis applied to the evaluation of medical image techniques.Invest. Radiol.,14,109–121.[ISI][Medline]

Tanbo, T., Omland, A., Dale, P.O.et al. (1995) In vitro fertilization/embryo transfer in unexplained infertility and minimal peritoneal endometriosis.Acta Obstet. Gynecol. Scand.,74,539–543.[ISI][Medline]

Witt, B.R., Wolf, G.C., Wainwright, C.J.et al. (1990) Relaxin, Ca, 125, progesteron, estradiol, Schwangerschaft protein and human chorionicgonadotrophin as predictors of outcome in threatened and not-threatened pregnancy.Fertil. Steril.,53,1029–1036.[ISI][Medline]

Yamashita, T., Okamoto, S., Thomas, A.et al. (1989) Predicting pregnancy outcome after in-vitro fertilization and embryo transfer using estradiol,progesterone, and human chorionic gonadotrophin ß-subunit.Fertil. Steril.,59,1081–1089.

Submitted on November 27, 1998; accepted on March 10, 1999.