Preclinical abortions: incidence and significance in the Norfolk in vitro fertilization program*

Preclinical abortions: incidence and significance in the Norfolk in vitro fertilization program*

Vol. 53, No.4, April 1990 FERTILITY AND STERILITY Printed on acid·free paper in U.S.A. Copyright e 1990 The American Fertility Society Preclinical...

551KB Sizes 0 Downloads 25 Views

Vol. 53, No.4, April 1990

FERTILITY AND STERILITY

Printed on acid·free paper in U.S.A.

Copyright e 1990 The American Fertility Society

Preclinical abortions: incidence and significance in the Norfolk in vitro fertilization program*

Anibal A. Acosta, M.D.t Sergio Oehninger, M.D. Jeffrey Hammer, M.D.

Suheil J. Muasher, M.D. Hong Mei Liang, M.D. Debra L. Jones

The Howard and Georgeanna Jones Institute for Reproductive Medicine, Department of Obstetrics and Gynecology, Eastern Virginia Medical School, Norfolk, Virginia

Clinical and prognostic significance of preclinical abortions in assisted reproduction is ill defined. Strict diagnostic criteria include a transient and synchronous elevation of serum {3-human chorionic gonadotropin (hCG), estradiol, and progesterone levels 13 days after hCG administration, ending in a bleeding episode no more than 14 days after the missed period. The preclinical abortion study group (54 patients, 178 cycles) was compared with matched control groups A (54 patients, 132 cycles) and B (54 patients, 155 cycles), representing normal term pregnancies and all outcomes, respectively. Control group C included the overall population during the study period. The abortion rate per transfer (preclinical abortion and total miscarriage rates) and total pregnancy wastage in the study group were significantly higher; the ongoing pregnancy rate was significantly lower. Preclinical abortion should be considered as a true reproductive failure with similar implications. Fertil Steril53:673, 1990

Biochemical (preclinical) pregnancies in natural human reproduction and in programs of assisted reproduction represent a substantial number of very early conceptions that are detected only because of extremely careful monitoring of the theoretical periconceptional and postconceptional periods. These gestations are defined by a positive pregnancy test before an expected menstrual period that is not delayed; therefore the pregnancy terminates in a menstrual abortion.! (Because of the high-dose human chorionic gonadotropin [heG] used to trigger ovulation in the Norfolk in vitro fertilization [IVF] program, biochemical pregnancies cannot be diagnosed with the available

Received September 18, 1989; revised and accepted December 19, 1989. * Presented in part at the VIth Worid Congress of In Vitro Fertilization and Alternate Assisted. Reproduction, Jerusalem, Israel, April 2 to 7, 1989. ·t Reprint requests: Anibal A. Acosta, M.D., Jones Institute for Reproductive Medicine, 825 Fairfax Avenue, Norfolk, Virginia 23507. Vol. 53, No.4, April 1990

methodology.) Other early gestations terminate in a preclinical abortion. In the Norfolk program, preclinical abortion has been defined1 as a pregnancy identified by two consecutive heG determinations after day 11 of the luteal phase, which ends with the onset of spontaneous menstruation no more than 14 days after the missed period (28 days after oocyte aspiration and fertilization). The second heG determination should show a rising titer. To ascertain that only true early conceptions were included in this paper, the definition of preclinical abortion was based on stricter criteria than those originally used: transient and synchronous elevation of serum ,a-heG, estradiol (E 2 ), and progesterone (P) levels occurring 13 days after heG has been administered and resulting in a bleeding episode no more than 14 days after the missed menstrual period (preclinical abortion). As experience with this entity increased, its incidence was reported in the results of assisted reproduction programs, with significant increase in the actual pregnancy rates. It is a consensus now that these pregnancies should not be counted in the staAcosta et aI.

Preclinical abortions in IVF

673

tistical evaluation of results, although it has been recommended! that they should be reported separately. Therefore, this very early type of conception ending in preclinical abortion, when properly identified, represents a true clinical entity; its incidence and significance should be clearly understood and defined. This study was undertaken to fulfill those objectives. MATERIALS AND METHODS

Series 14 to 29 of the Norfolk IVF program (April 1984 through December 1987) with a total of 1,081 patients, 1,887 cycles, and 1,699 transfers were reviewed. Fifty-four patients who showed evidence of preclinical abortion comprise the study group. These patients had a total of 178 cycles ending in 173 transfers; 63 of the cycles showed evidence of preclinical abortion ("index" cycles). Control group A (54 patients, 132 cyclfls, 128 transfers), matched by age, series number, primary etiology of infertility, and type of ovarian stimulation used, was patients with normal term pregnancies. Control group B (54 consecutive patients with 155 cycles, 139 transfers) was matched identically but irrespectively of IVF outcome. Control group C was the overall population attempting IVF during the study period. Ages (mean ± SD) of the study and control groups A and B were 34.1 ± 3.7, 33.5 ± 3.6, and 35.0 ± 3.4 years, respectively; the corresponding ages for group C were 34.7 ± 3.7 years. No significant differences were found among the groups. The primary etiologies in the study group and control groups A and B, respectively, were tubal factor (59.2%, 68.5%, 60.0%), endometriosis (31.4%, 7.4%, 15.1%), male factor (3.7%, 11.1%, 14.8%), unexplained (5.5%, 12.0%, 6.0%), and other (4.1% in control group B only). Except for the incidence of endometriosis in the study and control group A, no significant differences were found. The ovarian stimulation protocols that were used have been described. 2 The stimulations used were 2 ampules of follicle-stimulating hormone (FSH)/2 ampules of human menopausal gonadotropin (hMG) (60.3% of study group, 72.2% of group A, and 57.4 % of group B), 2 ampules of hMG (4.7% of study group, 1.8% each of groups A and B), 3 ampules of hMG (3.1% of study group, 1.8% of group A), 4 ampules of hMG (1.5% of study group, 1.8% of group A), 2 ampules of FSH (7.9% of stu4y group, 1.8% of group A, 7.4% of group B), 3 ampules of FSH (3.1% of study group, 3.7% of group A, 5.5% of group B), 4 ampules of FSH 674

Acosta et al.

Preclinical abortions in IVF

(17.4% of study group, 14.8% of group A, 25.9% of group B), and 8 ampules of FSH (1.5% of study group, 1.8% each of groups A and B). No'significant differences were observed among the groups; 10,000 IV of hCG were administered when leading follicles reached 16 mm in diameter. The cycle monitoring techniques, as well as oocyte retrieval, pre-embryo culture, and transfer procedures have been reported. 3 ,4 Serum E 2, P, and ,B-hCG levels were determined by radioimmunoassay after published techniques. 5 The luteal phase was supplemented, as has been the norm in the Norfolk program, using 25 mg of P in oil administered intramuscularly (1M), starting on the day of transfer and continuing for 5 days. The dose was then increased to 50 mg/d until the period began or until hCG results were obtained. If the hCG determination was positive, P was continued for 7 to 10 days; after that the patient began receiving 17-hydroxy P caproate (Carter-Glogau Laboratory, Inc., Glendale, AZ) 250 mg/wk 1M; this was continued until the 14th week of pregnancy. Evaluation of the luteal phase was done by comparing daily E2 and P levels and ratios and also by using the area under the curve (AVC). Statistical analysis was done by the unpaired Student's t-test, x2, and Duncan's multiple range test, as appropriate. RESULTS

Estradiol response, the terminal E2 pattern judged by published classifications,6 peak E2 levels in the periovulatory period, and the mean number of pre-embryos transferred are shown in Table 1. No significant differences were observed among the groups. The incidence of preclinical abortion in the total population was 4.9% per patient (54/1,081), 2.9% per cycle (54/1,887), 11.6% per pregnancy (54/465), 33.5% per abortion (54/161), and 16.6% recurrence rate per patient (9/54). Table 2 compares the reproductive performance of patients in the study group (excluding the index cycles) with control groups Band C. The overall pregnancy rate, abortion rate, preclinical abortion rate, and the ongoing pregnancy rate per transfer, also the abortion rate per pregnancy and overall pregnancy wastage in the study and control groups Band C are shown in Table 3, with and without consideration given to the index cycles. The abortion rate per transfer in the study group tended to be higher than in control groups B and C when the index cycles were excluded and Fertility and Sterility

Table 1 E2 Response, Terminal E 2, and Follicular Phase Peak E2 Levels and Number of Pre-embryos Transferred in the Different Groups

E2 response (%) High Intermediate Low Terminal E2 pattern (%) A G B Peak E2 (pg/mL)a No. of embryos transferred/transfer a

Study group

Control A

Control B

57.8 41.2 1.5

44.4 55.6

51.8 46.3

87.3 11.1 1.5 750.0 ± 340.0 3.5 ± 1.5

83.0 11.1 5.5 736.2 ± 361.6 3.8 ± 1.4

81.0 12.0 7.0 807.2 ± 335.7 2.6 ± 2.0

a Values are expressed as means ± SD.

was significantly higher when the index cycles were included. The preclinical abortion and total miscarriage rates per transfer were also significantly higher. Total pregnancy wastage, including ectopic gestations, was also significantly higher. Consequently, the ongoing pregnancy rate per transfer in the study group was significantly lower. The study and control groups A and B showed no differences in the luteal phase, neither in terms of daily E 2 , P, and E 2 :P values nor with the same calculations performed using AVCs. The only exception was slightly lower values in the AVC in control group B (230.89) than in the study group (309.16) and control group A (307.36). This finding is of doubtful biological significance, since patients with normal term pregnancies showed values almost identical to those of the study group.

timation of hCG excretion. 7 In this investigation the possibility of misdiagnosing these very early pregnancies has been minimized, since a highly sensitive serum iJ-hCG assay5 was always used in our program, starting on day 12 or 13 of the luteal phase (day 0 was the day after hCG administration). Previous experience5 has demonstrated that exogenous hCG at that time has been cleared from the system. An increase in iJ-hCG levels in two or more consecutive samples was required for diagnosis of very early pregnancy, accompanied by a concomitant elevation of serum E2 and P, which signals early rescue of the corpus luteum. By using these strict criteria, we may have omitted other early pregnancies, decreasing the true incidence of Table 3 Reproductive Performance of Study and Control Groups Band C Study group

DISCUSSION

The very existence of biochemical pregnancies has been questioned on the basis of study design and the use of antisera of low specificity for the esTable 2 Reproductive Performance of Patients in the Study Group, Excluding the 54 Index Cycles, Compared with Overall Results of Control Groups Band C

No. of cycles No. of transfers a Pregnancies Ongoing Preclinical abortions Clinical abortions Ectopic Totals

Study group

Control B

Control C

124 119 (95.9)

155 139 (89.6)

1,187 1,699 (90.0)

12 9 8 2

11 2

288 54 107 16

31

36

465

a Values in parentheses are percents. Vol. 53, No.4, April 1990

23

Control B

Control C

%

Excluding "index" cycle Pregnancy rate/transfer Abortion rate/transfer Preclinical abortion rate/ transfer Ongoing pregnancy rate/ transfer Abortion rate Pregnancy wastage "Overall" performance (including "index" cycle) Ongoing pregnancy rate/ transfer Abortion rate

26.0 14.2

25.8 7.9

7.5

Od

10.0 54.8 61.2

16.5 30.5 d 36.1 d

6.9 83.5

27.3 9.4 3.1 b

16.9' 34.6 d 38.0 e

16.9 a 34.6 1

a P < 0.001 compared with study group. b P < 0.01. 'P = 0.05. d P<0.05. e P< 0.02. 1 P < .0001.

Acosta et al.

Preclinical abortions in IVF

675

their occurrence, but we ascertained that the significance of preclinical abortion, in terms of subsequent reproductive performance, referred only to the true clinical entity. The incidence of biochemical pregnancies and menstrual abortions cannot be calculated in our program because of the use of high-dose hCG at midcycle to trigger ovulation. But pre-embryo loss, even under ideal circumstances, is very high. 8 The incidence of preclinical abortion in our program, with only a transient elevation of hCG as the diagnostic criterion, has been estimated as 3.8% per pregnancy5 in a group of 106 pregnancies. In the total population reviewed in this paper, the incidence was much higher (11.6%), even when strict criteria for diagnosis were used. Age, cause of infertility, type of stimulation used, and number of pre-embryos transferred did not seem to make any significant difference in preclinical abortion. The response to stimulation in terms of the E2 curve, peak E2 level, and terminal E2 pattern was not significantly different in study and control groups A and B. Therefore, these factors also did not seem to be significant. The cause of these very early miscarriages is difficult to establish, since no pregnancy material is available for pathological and chromosomal study. Therefore, it is as yet unknown whether they are euploidic or aneuploidic abortions. The fact that an important number of chromosomal errors have been described in IVF 9 after apparently normal fertilization may indicate that some or most of these could be aneuploidic reproductive failures. This possibility would seem to be supported by the normal luteal phase endocrine values in the study group. The reproductive performance of patients with a history of preclinical abortion in some IVF cycles is significantly impaired, regardless of whether the index cycle used to select the study group is considered or not in the statistical evaluation. The question of whether the index cycle should be included is highly debatable. On one hand, inclusion seems to skew the results by increasing the number of preclinical abortions in the study population. On the other hand, since all other cycles performed in those patients are considered, it seems legitimate to include the index cycle. Either way, the study population showed a significantly higher number of reproductive failures, both preclinical and clinical,

676

Acosta et aI. Preclinical abortions in IVF

leading to a significant decrease in the term pregnancy rate. The transfer rate per cycle in this group was excellent and no different from control groups Band C. Therefore, it is reasonable to consider reproductive performance in terms of the rate per transfer. These patients tend to have a recurrent abortion rate (16.6%) higher than patients with a single miscarriage and similar to patients with repeated (>3) clinical miscarriages (19% to 55% in retrospective studies; 14% to 68% in prospective studies).l0 Thus, when preclinical abortion is diagnosed repeatedly or when clinical and preclinical miscarriages tend to recur in an IVF patient, a complete work-up should be performed to rule out the most common causes of repeated miscarriages. REFERENCES 1. Jones HW, Jr, Acosta AA, Andrews MC, Garcia JE, Jones

2.

3.

4.

5.

6.

7.

8.

9.

10.

GS, Mantzavinos T, McDowell J, Sandow BA, Veeck L, Whibley TW, Wilkes CA, Wright GL, Jr: What is a pregnancy? A question for programs of in vitro fertilization. Fertil Steril40:728, 1983 Acosta AA, Bernardus RE, Jones GES, Garcia J, Rosenwaks Z, Simonetti S, Veeck LL, Jones D: The use of pure FSH alone or in combination for ovulation stimulation in in vitro fertilization. Acta Eur FertiI16:81, 1985 Jones HW, Jr, Acosta AA, Andrews MC, Garcia JE, Jones GS, Mayer J, McDowell JS, Rosenwaks Z, Sandow BA, Veeck LL, Wilkes CA: Three years of in vitro fertilization at Norfolk. Fertil Steril42:826, 1984 Jones GS, Garcia JE, Rosenwaks Z: The role of pituitary gonadotropins in follicular stimulation and oocyte maturation in the human. J Clin Endocrinol Metab 59:178,1984 Liu H-C, Kreiner D, Muasher SJ, Jones G, Jones H, Jr, Rosenwaks Z: ~-human chorionic gonadotropin as a monitor of pregnancy outcome in in vitro fertilization-embryo transfer patients. Fertil Steril 50:89, 1988 Jones HW, Jr, Acosta A, Andrews MC, Garcia JE, Jones GS, Mantzavinos T, McDowell J, Sandow B, Veeck L, Whibley T, Wilkes C, Wright G: The importance of the follicular phase to success and failure in in vitro fertilization. Fertil Steril40:317, 1983 Walker EM, Lewis M, Cooper W, Marnie M, Howie PN: Occult biochemical pregnancy: fact or fiction? Br J Obstet Gynaecol 95:659, 1988 Acosta AA, Moon SY, Oehninger S, Muasher SJ, Rosenwaks Z, Matta JF: Implantation potential of each pre-embryo in multiple pregnancies obtained by in vitro fertilization seems to be different. Fertil Steril 50:906, 1988 Wimmers MSE, van der Merwe M: Chromosome studies on early human embryos fertilized in vitro. Hum Reprod 3:894, 1988 Howert-de Jong MH, Eskes TKAB, Termigtelen A, Bruinse HW: Habitual abortion: a review. Eur J Obstet Gynecol Reprod Bioi 30:39, 1989

Fertility and Sterility