- Email: [email protected]
Estimates of human fertility and pregnancy loss*†
FERTILITY AND STERILITY Copyright
©
Vol. 65, No.3, March 1996
1996 American Society for Reproductive Medicine
Printed on acid·free paper in U. S. A
Estimates of human fertility and pregnancy loss*+ Michael J. Zinaman, M.D.:j:§ Eric D. Clegg, Ph.D.11 Charles C. Brown, Ph.D.~
John O'Connor, Ph.D.** Sherry G. Selevan, Ph.D.11
Georgetown University Medical Center, and U.S. Environmental Protection Agency, Washington, D.C.; Loyola University Medical Center, Maywood, Illinois; National Cancer Center, Bethesda, Maryland; and Columbia University, New York, New York
Objective: To examine the fertility and pregnancy wastage rates in a group of presumably fertile couples. Design: Prospective observational study of 200 couples desiring to achieve pregnancy over 12 menstrual cycles coupled with pregnancy outcome follow-up. Setting: A university-based obstetrics and gynecological center. Patients: Personal interviews and questionnaires were used to screen couples for entry into the study. Couples were counseled to have intercourse centered on predicted day of ovulation. Phase 1 included the first three cycles in which women collected daily morning urine samples, underwent midcycle postcoital tests, and, if late for their menses, presented for serum hCG testing. Phase 2 encompassed the next nine cycles in which women were contacted monthly by phone and underwent serum hCG testing if menses was delayed. Urine samples from cycles in which clinical (serum hCG) pregnancy did not occur underwent sensitive hCG testing to detect occult pregnancies. Pregnancies were followed until delivery to ascertain outcome. Results: Eighty-two percent ofthe 200 couples followed for the entire study period conceived. The maximal fertility rate was approximately 30% per cycle in the first two cycles. This rate quickly tapered over the remainder of the study. Pregnancy wastage during phase 1 accounted for 31% of the pregnancies detected. Forty-one percent (15/36) of these losses were seen only by urine hCG testing and were categorized as occult. Eleven of these same patients later achieved clinically recognized conceptions during the study. Conclusions: These results support the concept that the efficiency of human reproduction is maximum at approximately 30% per cycle. A very significant number of these pregnancies end in spontaneous abortion. In addition, pregnancy loss before missed menses occurs in a significant proportion of women. Fertil SterilI996;65:503-9 Key Words: Human fertility, pregnancy loss
The efficiency of human reproduction remains a subject of debate despite the published studies that deal with this topic. Fertility rates in presumably Received June 29, 1995; revised and accepted September 29, 1995. * Supported by grant CR-820787 from the United States Environmental Protection Agency, Washington, D.C. t The views expressed in this paper are those of the authors and do not necessarily reflect the views or policies of the U.S. Environmental Protection Agency. The U.S. Government has the right to retain a nonexclusive royalty-free license in and to any copyright covering this paper. :j: Department of Obstetrics and Gynecology, Georgetown University Medical Center. § Reprint requests and present address: Michael J. Zinaman, M.D., Department of Obstetrics and Gynecology, Loyola University Medical Center, 2160 South First Avenue, Maywood, Illinois 60153 (FAX: 708-216-5669). Vol. 65, No.3, March 1996
fertile populations are difficult to ascertain due to the complexity and cost of conducting such studies. Theoretical models of fecundity have suggested the possibility of conception rates as high as 68% per cycle (1). Wilcox et al. (2), following 221 couples without known impediments of fertility, observed a per cycle conception rate of 25% over the first three cycles. At the lower end of estimates of normal female fertility are donor insemination trials (3) that reveal a 10% per cycle rate. Reports of clinically recognized abortions are more
II Office of Research and Development, U.S. Environmental Protection Agency, Washington, D.C. ~ Division of Cancer Prevention and Control, National Cancer Institute. ** Irving Center for Clinical Research, Columbia University. Zinaman et at Fertility and pregnancy loss
503
reliable, suggesting loss rates of 15% to 20% (4, 5). In prospective studies, estimates for spontaneous abortions occurring between the first missed menses and 28 weeks of gestation have ranged from 13% to 22% (6-10). Since 1980, a series of studies (2, 1113) also have used hCG measurements as an indication of pregnancy. Despite differences in methods, clinical pregnancy loss rates were quite similar, 12% to 14%. When pregnancies detected solely by means ofthe hCG measurement (i.e., occult pregnancy) are included, the range of total pregnancy losses increased to between 20% (13) and 62% (12). This wide range is solely due to the differences in the number of occult pregnancies found (i.e., 8% to 50%). Assay methodology and the definition of occult pregnancy differed in each of the reports. As part of a broad-based fertility study of men and women, 210 couples without recognized risk factors for infertility were studied prospectively. In this paper, fertility and pregnancy wastage results are reported that incorporate the use of sensitive hCG monitoring to detect occult pregnancies. MATERIALS AND METHODS
Couples discontinuing contraception to become pregnant were sought for this study. Recruitment was primarily by means of radio and newspaper advertisements as well as physician and acquaintance referral. Initial screening was done by telephone to limit entry to women who were between the ages of 21 and 37 years and who had regular cycles (25 to 33 days). The men were required to be between the ages of21 and 60 years. At the time of the telephone screening, the couple could not have been without contraception for >3 months. Qualified couples were invited to make appointments in which each partner was interviewed separately and administered a questionnaire with the purpose of excluding the couple ifthey had known risk factors for infertility (e.g., previous tubal surgery) or recurrent pregnancy loss. Extensive sociodemographic information also was collected. Informed consent, as approved by the Georgetown University Institutional Review Board, was obtained from both members of the couple and participation in the study began with the next menses. Couples were offered compensation and free early pregnancy ultrasonography in return for completing the study. The study allowed approximately 12 months for each couple to attempt conception. If a clinically recognized pregnancy occurred, follow-up continued until delivery to ascertain pregnancy outcome. Couples were counseled on coital timing and frequency. They were told that coitus a minimum of every other day around the day of anticipated ovulation was optimal. 504
Zinaman et al. Fertility and pregnancy loss
The most likely day of ovulation was approximated from each woman's previous menstrual cycle length history and adjusted for each cycle based on the previous cycle's length. The 12-month interval for attempting conception was divided into two phases. The first phase lasted three complete menstrual cycles. The duration of the second phase was approximately 9 months, ensuring that 12 complete menstrual cycles were available for each couple to conceive. During phase 1 (first three cycles), women were asked to maintain a detailed menstrual calendar for each cycle. They recorded days of bleeding, acts of intercourse, symptoms of pregnancy, and any illnesses or new medications. In addition, to verify active coitus, the women underwent a midcycle postcoital test. They also collected a daily aliquot of first morning urine for later hCG testing. These were kept in their home freezers and delivered to the laboratory after each cycle. In the laboratory, samples were maintained at -20°C until analysis. If a woman was> 1 day late for her anticipated menses, conventional serum pregnancy testing was done. All patients noted to be pregnant in this manner were followed monthly until delivery to ascertain pregnancy outcome. In phase 2 of the study, the couples continued to attempt pregnancy for nine complete menstrual cycles. The women continued to complete their menstrual diaries. Monthly phone contact took place over this time to ensure the couples were continuing to attempt conception and to determine the outcome of the pregnancies. Serum hCG testing was performed as in phase 1. Follow-up continued for the 12 cycles or, if pregnancy occurred, until its outcome was known. The daily urine samples for each woman were secured in duplicate 4.5-mL screwcap vials and maintained at -20°C. At the time of assay, one vial per cycle day was thawed. Urine samples were assayed from all cycles in which pregnancy was not known to have occurred as well as from 30 cycles in which the women were known to be pregnant already by conventional methods. Sixteen samples per cycle were assayed beginning with the first day of menstruation for the next cycle and counting back into the luteal phase. Each woman's urine samples for all cycles were assayed in a single batch and in quadruplicate. Intact urinary hCG was measured using the monoclonal antibodies B108 (14, 15) and B109 (16) for detection and capture, respectively. The assay was modified from the original immunoradiometric assay of O'Connor et al. (17) and subsequent immunoenzymometric assay of Taylor et al. (18). Briefly, Dynatech Immulon-2 microtiter plates (Dynatech Laboratories, Inc., Alexandria, VA) were coated (200 Fertility and Sterility
J
J.LL per well) with B109 (5 J.Lg/mL) in coating buffer (pH 9.5). Mter removal ofthe coating buffer solution, the plates were air dried for 48 hours at room temperature. Standards and samples were added in 100J.LL quadruplicate aliquots with 100-J.LL buffer for the last 16 days of each menstrual cycle. Mter overnight incubation, biotinylated B108 was added and the plates again were incubated for 24 hours. The alkaline phosphatase-streptavidin signal amplification using the substrate p-nitrophenylphosphate was detected at 410 and 570 nm for the test and reference dose, respectively, using a Dynatech Microelisa (Dynatech Laboratories, Inc.) reader and data reduction program. Standard curves covering the range of 0.048 to 25 ng/mL were constructed using purified intact hCG (17). The assay sensitivity with urine was consistently in the range of 0.05 to 0.06 ng/mL. The intra-assay and interassay coefficients of variation were 4.6% and 10.2%, respectively. An occult pregnancy was considered to have occurred when hCG titers exceeded 0.15 ng/mL urine for 2e:3 consecutive days. Serum hCG tests were done the day the specimen was drawn. Throughout the study period all serum hCG assays were done by the same laboratory using an 1125 Coat-A-Count hCG assay (Diagnostic Products Corporation, Los Angeles, CA.) The sensitivity of the assay was 5 mIU/mL. Results above this level were reported as positive and thereby confirmed pregnancy. Statistical methods used X2 tests to compare two proportions. In the case of comparing two means, ttests were used. Standard life-table methods were used to account for study dropouts when estimating the monthly cumulative proportion of couples producing a pregnancy.
RESULTS A total of 210 couples were recruited into the study. Ten were excluded from analysis: six for severe oligospermia, two for noncompliance with the study protocol, one for bilateral tubal occlusion detected after acceptance, and one couple who changed their minds about participating. One hundred ninety-two couples completed the entire study protocol while four left the study after one cycle (no pregnancies), one after two cycles (not pregnant), and three after three cycles (one occult pregnancy during cycle 2 and one clinical pregnancy during cycle 3). The demographics of the 200 couples in the final study group are shown in Table 1. The ages of the men ranged from 20 to 51 years with a mean of 32.3 years, whereas the women ranged from 23 to 37 years with a mean of 30.6 years. Volunteers for this study tended to be white/non-Hispanic (>90%) Vol. 65, No.3, March 1996
Table 1 Demographic Description of the Study Group*
Age* Ethnicity (%) White Non-Hispanic Hispanic African American AsianlPacific Islander Other Education (%) Less than High School High School Some College College Advanced Training Religion Protestant Catholic Jewish Other None Family income <$20,000 $20,000 to $49,000 $50,000 to $74,000 2>$75,000 Alcohol consumption 2>20 drinks per lifetime (%) Over last month (%) No. drinksllast month* Cigarette smoking Ever smoked 100 Smoked in last 3 months No. per day last 3 months*
Husband
Wife
32.3 :+: 4.8 (20 to 51)
30.6 :+: 3.3 (23 to 37)
92.0 3.5 3.0 1.5
91.5 2.0 4.0 2.0 0.5
0.5 3.5 12.5 37.5 46.0
5.0 14.0 45.7 35.2
35.5 31.5 13.5 6.0 11.5
32.0 32.5 15.0 10.6 8.6
1.5 27.5 35.0 34.5
2.0 26.5 39.0 32.5
95.0 81.5 16.7 :+: 16.2 (1 to 80)
90.5 77.5 7.5 :+: 7.5 (1 to 48)
32.0 9.0 9.7 :+: 6.9 (1 to 20)
28.1 6.5 7.8 :+: 6.6 (1 to 20)
* Values are means:+: SD with ranges in parentheses. No. of couples = 200.
and higher socioeconomic status (SES). The higher SES was demonstrated by the level of education (>80% had at least a college degree) and family income (approximately 70% of the couples reported a family income greater than $50,000/year). Although approximately 30% reported a lifetime history of cigarette smoking, < 10% reported smoking during the last 3 months. More than 90% reported a lifetime history of alcohol consumption, with approximately 80% reporting consumption over the last month. Men's reported level of consumption of all alcohol products was more than double that of their mates' (16.7 versus 7.5 per month). The reproductive histories of the women are given in Table 2. The majority of women were nulliparous (52.5%) and most couples (59%) had ceased contracepting :s; 1 month before enrolling in the study. Although subsequent pregnancy occurrence and outcome were unrelated to age (male or female), it was related to a woman's being nulliparous or parous. Seventy-two percent of nulliparous women and 88% of parous women became pregnant within the 1-year study period (P = 0.008). Zinaman et al. Fertility and pregnancy loss
505
35%
Table 2 Reproductive History of the 200 Women Participating in the Study*
30%
Current marriage
Lifetime history 25%
Total pregnancies (%) Live births (%) Spontaneous abortions (%)
Stillbirths (%) Induced abortions (%)
31.0 (1.4 :!: 0.7; 1 to 4) 24.0 (1.1 :!: 0.3; 1 to 2) 9.5 (1.1 :!: 0.3; 1 to 2) 0.0 7.0 (1.0 :!: 0; 1 to 1)
(1.6 :!: (1.1 :!:
(1.1 :!: (1.2 :!:
47.5 0.8; 1 28.0 0.4; 1 12.5 0.3; 1 0.0 23.5 0.4; 1
Q)
to 4)
1ii
a:
20%
>-
to 2)
g
ro
~15%
Q)
to 2)
0: 10% 4/52
to 3)
5%
2140
... --------1/35
* Values are percentages with means:!: SD and ranges in parentheses.
5
6
7
8
10
11
12
Cycle N umber in Study
Couples that had not taken precautions to avoid conception for up to 5 months before beginning phase 1 were accepted into the study. Numbers of couples not contracepting for 0 to 5 months were 83 (41.5%), 35 (17.5%), 37 (18.5%), 33 (16.5%), 8 (4), and 4 (2%), respectively. Thus, fertility results could have been influenced by the prior length of time in which couples were not contracepting. Over the 12-month study period, the proportion of couples conceiving was significantly higher for 0 months not contracepting (83 couples, 90%) than for :2: 1 months not contracepting (117 couples, 77%, P = 0.02). However, because there was no pregnancy rate trend across the number of months not contracepting (1 to 5 months), the effect is unclear. For that reason and because the conception "effort" was likely to have been quite different between the couple's prestudy and on-study periods, the results presented here were not adjusted for prior months not contracepting. Eighty-two percent of the couples followed for the entire study conceived, suggesting that the volunteer study group was highly fertile. The analysis of fertility and pregnancy outcome counted only the first pregnancy (if more than one), whether it was occult or clinical. The cycle-specific pregnancy rates
Figure 1
Cycle-specific pregnancy rates for entire study period.
over the phase 1 study period (first three cycles) are displayed in Table 3. Of the 200 couples entering the study, 59 (29.5%) achieved conception during the first cycle. Of the 137 remaining couples entering cycle 2, (200; 59 pregnant, 4 dropped out), 41 (29.9%) achieved conception during the second cycle and, of the 95 remaining couples entering cycle 3 (137; 41 pregnant, 1 dropped out), 16 (16.8%) achieved conception during the third cycle. Pregnancy rates tended to decline over the remainder of the study (Fig. 1). Cumulative pregnancy rates revealed that 50% of women had achieved pregnancy in two cycles and 65% by four cycles (Fig. 2). Thirteen percent (151 116) of the pregnancies were occult in nature, as they were pregnancies detected by the hCG assay only. Because of the strict follow-up protocol and immediate implementation of serum pregnancy testing, the cycle lengths during which these 15 occult
60
-----------------+------------------------------------------------------- 90% Phase One
[ Phase Two
-------------+--------------------------~--~~~'::":-:-::'-
I
50
--------------1------
------------------------------------------------
I ---
Table 3 Pregnancy Occurrence and Outcome During Phase 1 of the Study (Includes Only First Pregnancy During First Three Cycles)
I
-------------------- ------------
I
------r---------------------------------------------·----------
I
"2
-------r-------------------------------------- -----------------
020
_______1_________ - - -- --------------- ---------------- - - --
I
E
Cycle no.
-----------------------
z
70%
~
0%
0%
1
2
3
Total
506
Zinaman et al.
Fertility and pregnancy loss
~
0
3~
~
--------------------------------------------- 20% 10 -------------
Entered cycle not pregnant 200 137 95 Pregnant during cycle 16 59 116 41 Occult 15 7 7 1 Clinical Live birth 40 10 29 79 21 Spontaneous abortion 12 5 4 I Lost to follow-up 1 6 Dropped out not pregnant 4 1 1 31.3 pregnancy (%) Loss 29.3 33.3 _ _per __ ____ _ _ _ _32.2 ____ ___ ____
i
OO/o~
clinical pregnancies only
Cycle outcome
r
00/0 ~
10
11
10%
12
Cycle Number
Figure 2 Cycle-specific conception and pregnancy outcome resuIts. Phase 1 includes live births, spontaneous abortions, and occult pregnancies. Phase 2 includes live births and spontaneous abortions. only. D, live births; D, spontaneous abortions; ., occult pregnanCIes.
Fertility and Sterility
The pregnancy loss rates were 19/59 (32%), 12/41 (29%), and 5/15 (33%) for cycles 1 to 3, respectively. The contribution of occult pregnancy to these loss rates are 7/19 (37%), 7/12 (58%), and 115 (20%), respectively. Thus, occult losses accounted for 42% of the pregnancy losses detected during phase 1.
10
DISCUSSION
0.1
-+---,----,0--,---,-----,--,-------.------, 25
26
27 31 28 29 30 Length of Cycle (days) for occult pregnancy
32
33
Figure 3 Maximum urinary hCG concentrations found in occult pregnancies by length of cycle in which they occurred.
pregnancies were detected were limited to 27 to 33 days (Fig. 3). Eleven of 14 couples who had an occult pregnancy during phase 1 and completed the entire study achieved a clinically recognized pregnancy during a later cycle; 6 of them within the first phase and 5 within the second phase of the study. A clinically recognized pregnancy was detected in the immediate ensuing cycle in five of the cases. The subsequent pregnancy outcomes of all 15 women who had occult pregnancies are listed in Table 4. Of the 11 women who conceived clinically after the occult pregnancy, all but one went on to deliver a viable offspring. Table 4 Listing of all Occult Pregnancies and Subsequent Study Outcome
Occult pregnancy no.
Occult in cycle
Clinical in cycle
1 1 1 1 2 2
2 2 2 3 3 3
1 1 2 2 3
6 6 8 12 12
Phase 1 1 2 3 4 5 6
Pregnancy outcome Live birth Live birth Live birth Live birth Live birth Spontaneous abortion
Phase 2 7 8 9 10 11 Not clinically pregnant 12 13 14 15
Vol. 65, No.3, March 1996
1 2 2 2
Live Live Live Live Live
birth birth birth birth birth
Dropped out after three cycles
In this study we determined fertility rates in a selected group of volunteers. Couples were scrutinized closely throughout their participation to ensure active attempts at conception as well as early detection of pregnancy. Fertility of couples was influenced only marginally by the number of months not contracepting before beginning the study. No correction was made because the prestudy and onstudy periods were not comparable. Fertility rates of 30% per cycle were seen for the first two cycles, with a progressive tapering over the remainder of the study. Only the first three cycles (phase 1) actually are comparable as coitus was confirmed by postcoital testing, the couples were monitored closely, and urine samples were collected only in those cycles. During the next nine cycles, the women were contacted and visits were limited to serum hCG testing if menses was delayed. Occult pregnancies would not have been detected during phase 2. There is an apparent maximum rate of conception that represents the most fertile of the couples under the study. The rapid subsequent tapering of the fertility rates would be expected because the proportion of couples with reduced fertility would increase with increasing number of exposed cycles. In this study 18% of couples failed to conceive after 12 complete menstrual cycles. Using a similar time period of 1 year, the National Center for Health Statistics reported an infertility rate of only 7.9%.(19) The higher rate is despite the considerable counseling and certainty that couples actively were attempting to conceive. Possible explanations for this difference include under-reporting of infertility (19); admission of some couples into this study, despite screening, that had a prior history offertility related problems; or our study population consisting primarily of urban and suburban, middle and upper-middle class whites had a higher than usual incidence of infertility. In 1976, Bloch (20) defined occult pregnancy "as a pregnancy that terminates so soon after implantation that no clinical suspicion exists as to its having existed." Emphasizing "no clinical suspicion," the present study took the approach that any delay in menses was suspicious. This is often the case in current clinical practice with sensitive urine and serum pregnancy testing widely available. Although the Zinaman et al. Fertility and pregnancy loss
507
term biochemical pregnancy often is used in those situations of delayed menses but without other signs of pregnancy, it also is used synonymously with occult (21) pregnancy. This often leads to confusion about whether the pregnancy was recognized by the patient. The 13% occult pregnancy rate determined in this study would be expected to disagree with studies that included patients with significant delays in menses in their occult category. In those studies of spontaneous abortion rates in which sensitive hCG assays were used to detect occult pregnancies, overall pregnancy loss rates ranged from 20% to 62% (2,11-13). The hCG assays used, minimum requirements for a positive result, and study populations have all differed. The most recent ofthese studies (2) cites an overall pregnancy loss rate of 31 %, which is not different from the present study. However, the percentage of occult and clinical pregnancies differs (22% versus 13%, 12% versus 18% for the Wilcox et al. (2) and present studies, respectively). The categories of pregnancy to which the losses were assigned appears to be important in this regard. Many of the occult losses in Wilcox et al. (2) occurred in women who were variably late for their menses. These women more likely would have been categorized as clinical losses in the present study. Differences in the participant population also exist. In particular, the present maximum age criterion was 37 versus 42 years for the Wilcox et al. study (2). In addition, the diminished sensitivity of the present hCG assay (B108-B109*; 0.05 ng/ mL) compared with Wilcox et al. (2) (B101-R525*; 0.01 ng/mL) might suggest that some occult pregnancies were missed in the present study. However, this is not likely, as no "borderline" cases were seen. The loss of sensitivity also would be expected to be at least partially offset by the fact that the (B108109*) configuration used in the present study does not require the presence ofthe hCG f3-carboxy-terminal peptide (i.e., the B108-109* assay has the potential to detect a larger family of hCG-related molecules). Although the overall number of occult pregnancies was limited, only three of the women with occult pregnancies who completed the study did not conceive subsequently. No women were identified who had two occult pregnancies, although there was only a three-cycle interval in which this could have occurred. Interestingly, five women who had an occult pregnancy were able to conceive in the next immediate cycle, suggesting little if any residual effect of the hCG on their subsequent ability to ovulate. This is not surprising, as the peak hCG concentration in this group was approximately 10 ng/mL (equivalent to 130 mIU/mL) with most peaks closer to 1 to 2 ng/ mL (equivalent to 13 to 26 mIU/mL). 508
Zinaman et al. Fertility and pregnancy loss
These results suggest that pregnancy loss accounts for approximately 30% of all detectable conceptions. The previously reported 20% to 62% range (2, 11-13) oflosses likely represents methodological differences. Improved specificity of antisera perhaps has been the greatest advance. At the same time, the demonstrated maximal fertility rates of approximately 30% per cycle in a tightly controlled setting suggests that there may be a practical upper limit to per cycle fertility rates in couples. Even with couples who are highly fertile, conception in a given cycle may be precluded by several factors, including a transient abnormal cycle, poor timing of coitus relative to ovulation, or loss of a conception before detection by hCG assay. Although the results from these study couples cannot be generalized to all couples attempting conception, upper limits appear to exist in natural cycles. This information should shed further light on the limitations and potential for conception. REFERENCES 1. Barrett JC, Marshall J. The risk of conception on different days of the menstrual cycle. Popul Stud 1969;23:455-61. 2. Wilcox AJ, Weinburg CR, O'Connor J, Baird DD, Schatterer JP, Canfield RE, et al. Incidence of early loss of pregnancy. N Engl J Med 1988;319:189-94. 3. Schwartz D, Mayaux MJ. Female fecundity as a function of age. N Engl J Med 1982;306:404-6. 4. Warburton D. Spontaneous abortion risks in man: data from reproductive histories collected in a medical genetics unit. Am J Hum Genet 1964;16:1-25. 5. Poland BJ, Miller JR, Jones DC, Trimble BK. Reproductive counseling in patients who have had a spontaneous abortion. Am J Obstet Gynecol 1977; 127:685-91. 6. French FE, Bierman J. Probabilities offetal mortality. Public Health Rep 1962;77:835-47. 7. Erhardt CL. Pregnancy losses in New York City, 1960. Am J Public Health 1963;53:1337-57. 8. Taylor WF. The probability of fetal death. In: Fraser FC, McKusick CA, editors. Congenital malformations. New York: Excerpta Medica, 1969:307-20. 9. Harlap S, Shiono P, Ramcharan S. A life table of spontaneous abortions and the effects of age parity and other variables. In: Porter IH, Hook EB, editors. Human embryonic and fetal death. New York: Academic Press, 1980:145-64. 10. Petterson F. Epidemiology of early pregnancy wastage: biological and social correlates of abortion. An investigation based on materials collected within Uppsala County, Stockholm, Sweden. Stockholm, Sweden: Svenska Bokforlaget, 1968. 11. Miller JF, Williams E, Glue J, Gordon YB, Grudzinskas JG, Sykes A. Fetal loss after implantation. Lancet 1980; 1:5546. 12. Edmonds KD, Lindsay KS, Miller JF, Williamson E, Wood PJ. Early embryonic mortality in women. Fertil Steril 1982; 38:447-53. 13. Whittaker PG, Taylor A, Lind T. Unsuspected pregnancy loss in healthy women. Lancet 1983;1:1126-7. 14. Ehrlich PH, Moustafa ZA, Krichevsky A, Birken S, Armstrong EG, Canfield RE. Characterization and relative orientation of epitopes for monoclonal antibodies and antisera to
Fertility and Sterility
15.
16.
17.
18.
human chorionic gonadotropin. Am J Reprod Immunol MicrobioI 1985;8:48-54. Ehrlich PH, Moyle WR, Canfield RE. Monoclonal antibodies to gonadotropin subunits. Methods Enzymol 1985; 109:63855. Ehrlich P, Moyle W, Moustafa ZA, Canfield RE. Mixing two monoclonal antibodies yields enhanced affinity for antigen. J Immunol 1982; 128:2709-13. O'Connor JF, Shlatterer JP, Birken S, Krichevsky E, Armstrong EG, McMahon D, et al. Development of highly sensitive immunoassays to measure human chorionic gonadotropin, its fJ-subunit and fJ core fragment in the urine: application to malignancies. Cancer Res 1988;48:1361-6. Taylor CA Jr, Overstreet JW, Samuels SJ, Boyers SP, Can-
Vol. 65, No.3, March 1996
field RE, O'Connor JF. Prospective assessment of early fetal loss using an immunoenzymometric screening assay for detection of urinary human chorionic gonadotropin. Fertil Steril 1992;57:1220-4. 19. Mosher WD, Pratt WF. Fecundity and infertility in the United States, 1965-88. Advance data from vital and health statistics, no. 192. Hyattsville (MD): National Center for Health Statistics, 1990. 20. Bloch SK. Occult pregnancy: a pilot study. Obstet Gynecol 1976;48:365-8. 21. Kline J, Stein Z, Susser M. Conception and reproductive loss: probabilities. In: Kline J, Stein Z, Susser M, editors. Conception to birth: epidemiology of prenatal development. New York: Oxford University Press, 1989:43-68.
Zinaman et al.
Fertility and pregnancy loss
509
©
Vol. 65, No.3, March 1996
1996 American Society for Reproductive Medicine
Printed on acid·free paper in U. S. A
Estimates of human fertility and pregnancy loss*+ Michael J. Zinaman, M.D.:j:§ Eric D. Clegg, Ph.D.11 Charles C. Brown, Ph.D.~
John O'Connor, Ph.D.** Sherry G. Selevan, Ph.D.11
Georgetown University Medical Center, and U.S. Environmental Protection Agency, Washington, D.C.; Loyola University Medical Center, Maywood, Illinois; National Cancer Center, Bethesda, Maryland; and Columbia University, New York, New York
Objective: To examine the fertility and pregnancy wastage rates in a group of presumably fertile couples. Design: Prospective observational study of 200 couples desiring to achieve pregnancy over 12 menstrual cycles coupled with pregnancy outcome follow-up. Setting: A university-based obstetrics and gynecological center. Patients: Personal interviews and questionnaires were used to screen couples for entry into the study. Couples were counseled to have intercourse centered on predicted day of ovulation. Phase 1 included the first three cycles in which women collected daily morning urine samples, underwent midcycle postcoital tests, and, if late for their menses, presented for serum hCG testing. Phase 2 encompassed the next nine cycles in which women were contacted monthly by phone and underwent serum hCG testing if menses was delayed. Urine samples from cycles in which clinical (serum hCG) pregnancy did not occur underwent sensitive hCG testing to detect occult pregnancies. Pregnancies were followed until delivery to ascertain outcome. Results: Eighty-two percent ofthe 200 couples followed for the entire study period conceived. The maximal fertility rate was approximately 30% per cycle in the first two cycles. This rate quickly tapered over the remainder of the study. Pregnancy wastage during phase 1 accounted for 31% of the pregnancies detected. Forty-one percent (15/36) of these losses were seen only by urine hCG testing and were categorized as occult. Eleven of these same patients later achieved clinically recognized conceptions during the study. Conclusions: These results support the concept that the efficiency of human reproduction is maximum at approximately 30% per cycle. A very significant number of these pregnancies end in spontaneous abortion. In addition, pregnancy loss before missed menses occurs in a significant proportion of women. Fertil SterilI996;65:503-9 Key Words: Human fertility, pregnancy loss
The efficiency of human reproduction remains a subject of debate despite the published studies that deal with this topic. Fertility rates in presumably Received June 29, 1995; revised and accepted September 29, 1995. * Supported by grant CR-820787 from the United States Environmental Protection Agency, Washington, D.C. t The views expressed in this paper are those of the authors and do not necessarily reflect the views or policies of the U.S. Environmental Protection Agency. The U.S. Government has the right to retain a nonexclusive royalty-free license in and to any copyright covering this paper. :j: Department of Obstetrics and Gynecology, Georgetown University Medical Center. § Reprint requests and present address: Michael J. Zinaman, M.D., Department of Obstetrics and Gynecology, Loyola University Medical Center, 2160 South First Avenue, Maywood, Illinois 60153 (FAX: 708-216-5669). Vol. 65, No.3, March 1996
fertile populations are difficult to ascertain due to the complexity and cost of conducting such studies. Theoretical models of fecundity have suggested the possibility of conception rates as high as 68% per cycle (1). Wilcox et al. (2), following 221 couples without known impediments of fertility, observed a per cycle conception rate of 25% over the first three cycles. At the lower end of estimates of normal female fertility are donor insemination trials (3) that reveal a 10% per cycle rate. Reports of clinically recognized abortions are more
II Office of Research and Development, U.S. Environmental Protection Agency, Washington, D.C. ~ Division of Cancer Prevention and Control, National Cancer Institute. ** Irving Center for Clinical Research, Columbia University. Zinaman et at Fertility and pregnancy loss
503
reliable, suggesting loss rates of 15% to 20% (4, 5). In prospective studies, estimates for spontaneous abortions occurring between the first missed menses and 28 weeks of gestation have ranged from 13% to 22% (6-10). Since 1980, a series of studies (2, 1113) also have used hCG measurements as an indication of pregnancy. Despite differences in methods, clinical pregnancy loss rates were quite similar, 12% to 14%. When pregnancies detected solely by means ofthe hCG measurement (i.e., occult pregnancy) are included, the range of total pregnancy losses increased to between 20% (13) and 62% (12). This wide range is solely due to the differences in the number of occult pregnancies found (i.e., 8% to 50%). Assay methodology and the definition of occult pregnancy differed in each of the reports. As part of a broad-based fertility study of men and women, 210 couples without recognized risk factors for infertility were studied prospectively. In this paper, fertility and pregnancy wastage results are reported that incorporate the use of sensitive hCG monitoring to detect occult pregnancies. MATERIALS AND METHODS
Couples discontinuing contraception to become pregnant were sought for this study. Recruitment was primarily by means of radio and newspaper advertisements as well as physician and acquaintance referral. Initial screening was done by telephone to limit entry to women who were between the ages of 21 and 37 years and who had regular cycles (25 to 33 days). The men were required to be between the ages of21 and 60 years. At the time of the telephone screening, the couple could not have been without contraception for >3 months. Qualified couples were invited to make appointments in which each partner was interviewed separately and administered a questionnaire with the purpose of excluding the couple ifthey had known risk factors for infertility (e.g., previous tubal surgery) or recurrent pregnancy loss. Extensive sociodemographic information also was collected. Informed consent, as approved by the Georgetown University Institutional Review Board, was obtained from both members of the couple and participation in the study began with the next menses. Couples were offered compensation and free early pregnancy ultrasonography in return for completing the study. The study allowed approximately 12 months for each couple to attempt conception. If a clinically recognized pregnancy occurred, follow-up continued until delivery to ascertain pregnancy outcome. Couples were counseled on coital timing and frequency. They were told that coitus a minimum of every other day around the day of anticipated ovulation was optimal. 504
Zinaman et al. Fertility and pregnancy loss
The most likely day of ovulation was approximated from each woman's previous menstrual cycle length history and adjusted for each cycle based on the previous cycle's length. The 12-month interval for attempting conception was divided into two phases. The first phase lasted three complete menstrual cycles. The duration of the second phase was approximately 9 months, ensuring that 12 complete menstrual cycles were available for each couple to conceive. During phase 1 (first three cycles), women were asked to maintain a detailed menstrual calendar for each cycle. They recorded days of bleeding, acts of intercourse, symptoms of pregnancy, and any illnesses or new medications. In addition, to verify active coitus, the women underwent a midcycle postcoital test. They also collected a daily aliquot of first morning urine for later hCG testing. These were kept in their home freezers and delivered to the laboratory after each cycle. In the laboratory, samples were maintained at -20°C until analysis. If a woman was> 1 day late for her anticipated menses, conventional serum pregnancy testing was done. All patients noted to be pregnant in this manner were followed monthly until delivery to ascertain pregnancy outcome. In phase 2 of the study, the couples continued to attempt pregnancy for nine complete menstrual cycles. The women continued to complete their menstrual diaries. Monthly phone contact took place over this time to ensure the couples were continuing to attempt conception and to determine the outcome of the pregnancies. Serum hCG testing was performed as in phase 1. Follow-up continued for the 12 cycles or, if pregnancy occurred, until its outcome was known. The daily urine samples for each woman were secured in duplicate 4.5-mL screwcap vials and maintained at -20°C. At the time of assay, one vial per cycle day was thawed. Urine samples were assayed from all cycles in which pregnancy was not known to have occurred as well as from 30 cycles in which the women were known to be pregnant already by conventional methods. Sixteen samples per cycle were assayed beginning with the first day of menstruation for the next cycle and counting back into the luteal phase. Each woman's urine samples for all cycles were assayed in a single batch and in quadruplicate. Intact urinary hCG was measured using the monoclonal antibodies B108 (14, 15) and B109 (16) for detection and capture, respectively. The assay was modified from the original immunoradiometric assay of O'Connor et al. (17) and subsequent immunoenzymometric assay of Taylor et al. (18). Briefly, Dynatech Immulon-2 microtiter plates (Dynatech Laboratories, Inc., Alexandria, VA) were coated (200 Fertility and Sterility
J
J.LL per well) with B109 (5 J.Lg/mL) in coating buffer (pH 9.5). Mter removal ofthe coating buffer solution, the plates were air dried for 48 hours at room temperature. Standards and samples were added in 100J.LL quadruplicate aliquots with 100-J.LL buffer for the last 16 days of each menstrual cycle. Mter overnight incubation, biotinylated B108 was added and the plates again were incubated for 24 hours. The alkaline phosphatase-streptavidin signal amplification using the substrate p-nitrophenylphosphate was detected at 410 and 570 nm for the test and reference dose, respectively, using a Dynatech Microelisa (Dynatech Laboratories, Inc.) reader and data reduction program. Standard curves covering the range of 0.048 to 25 ng/mL were constructed using purified intact hCG (17). The assay sensitivity with urine was consistently in the range of 0.05 to 0.06 ng/mL. The intra-assay and interassay coefficients of variation were 4.6% and 10.2%, respectively. An occult pregnancy was considered to have occurred when hCG titers exceeded 0.15 ng/mL urine for 2e:3 consecutive days. Serum hCG tests were done the day the specimen was drawn. Throughout the study period all serum hCG assays were done by the same laboratory using an 1125 Coat-A-Count hCG assay (Diagnostic Products Corporation, Los Angeles, CA.) The sensitivity of the assay was 5 mIU/mL. Results above this level were reported as positive and thereby confirmed pregnancy. Statistical methods used X2 tests to compare two proportions. In the case of comparing two means, ttests were used. Standard life-table methods were used to account for study dropouts when estimating the monthly cumulative proportion of couples producing a pregnancy.
RESULTS A total of 210 couples were recruited into the study. Ten were excluded from analysis: six for severe oligospermia, two for noncompliance with the study protocol, one for bilateral tubal occlusion detected after acceptance, and one couple who changed their minds about participating. One hundred ninety-two couples completed the entire study protocol while four left the study after one cycle (no pregnancies), one after two cycles (not pregnant), and three after three cycles (one occult pregnancy during cycle 2 and one clinical pregnancy during cycle 3). The demographics of the 200 couples in the final study group are shown in Table 1. The ages of the men ranged from 20 to 51 years with a mean of 32.3 years, whereas the women ranged from 23 to 37 years with a mean of 30.6 years. Volunteers for this study tended to be white/non-Hispanic (>90%) Vol. 65, No.3, March 1996
Table 1 Demographic Description of the Study Group*
Age* Ethnicity (%) White Non-Hispanic Hispanic African American AsianlPacific Islander Other Education (%) Less than High School High School Some College College Advanced Training Religion Protestant Catholic Jewish Other None Family income <$20,000 $20,000 to $49,000 $50,000 to $74,000 2>$75,000 Alcohol consumption 2>20 drinks per lifetime (%) Over last month (%) No. drinksllast month* Cigarette smoking Ever smoked 100 Smoked in last 3 months No. per day last 3 months*
Husband
Wife
32.3 :+: 4.8 (20 to 51)
30.6 :+: 3.3 (23 to 37)
92.0 3.5 3.0 1.5
91.5 2.0 4.0 2.0 0.5
0.5 3.5 12.5 37.5 46.0
5.0 14.0 45.7 35.2
35.5 31.5 13.5 6.0 11.5
32.0 32.5 15.0 10.6 8.6
1.5 27.5 35.0 34.5
2.0 26.5 39.0 32.5
95.0 81.5 16.7 :+: 16.2 (1 to 80)
90.5 77.5 7.5 :+: 7.5 (1 to 48)
32.0 9.0 9.7 :+: 6.9 (1 to 20)
28.1 6.5 7.8 :+: 6.6 (1 to 20)
* Values are means:+: SD with ranges in parentheses. No. of couples = 200.
and higher socioeconomic status (SES). The higher SES was demonstrated by the level of education (>80% had at least a college degree) and family income (approximately 70% of the couples reported a family income greater than $50,000/year). Although approximately 30% reported a lifetime history of cigarette smoking, < 10% reported smoking during the last 3 months. More than 90% reported a lifetime history of alcohol consumption, with approximately 80% reporting consumption over the last month. Men's reported level of consumption of all alcohol products was more than double that of their mates' (16.7 versus 7.5 per month). The reproductive histories of the women are given in Table 2. The majority of women were nulliparous (52.5%) and most couples (59%) had ceased contracepting :s; 1 month before enrolling in the study. Although subsequent pregnancy occurrence and outcome were unrelated to age (male or female), it was related to a woman's being nulliparous or parous. Seventy-two percent of nulliparous women and 88% of parous women became pregnant within the 1-year study period (P = 0.008). Zinaman et al. Fertility and pregnancy loss
505
35%
Table 2 Reproductive History of the 200 Women Participating in the Study*
30%
Current marriage
Lifetime history 25%
Total pregnancies (%) Live births (%) Spontaneous abortions (%)
Stillbirths (%) Induced abortions (%)
31.0 (1.4 :!: 0.7; 1 to 4) 24.0 (1.1 :!: 0.3; 1 to 2) 9.5 (1.1 :!: 0.3; 1 to 2) 0.0 7.0 (1.0 :!: 0; 1 to 1)
(1.6 :!: (1.1 :!:
(1.1 :!: (1.2 :!:
47.5 0.8; 1 28.0 0.4; 1 12.5 0.3; 1 0.0 23.5 0.4; 1
Q)
to 4)
1ii
a:
20%
>-
to 2)
g
ro
~15%
Q)
to 2)
0: 10% 4/52
to 3)
5%
2140
... --------1/35
* Values are percentages with means:!: SD and ranges in parentheses.
5
6
7
8
10
11
12
Cycle N umber in Study
Couples that had not taken precautions to avoid conception for up to 5 months before beginning phase 1 were accepted into the study. Numbers of couples not contracepting for 0 to 5 months were 83 (41.5%), 35 (17.5%), 37 (18.5%), 33 (16.5%), 8 (4), and 4 (2%), respectively. Thus, fertility results could have been influenced by the prior length of time in which couples were not contracepting. Over the 12-month study period, the proportion of couples conceiving was significantly higher for 0 months not contracepting (83 couples, 90%) than for :2: 1 months not contracepting (117 couples, 77%, P = 0.02). However, because there was no pregnancy rate trend across the number of months not contracepting (1 to 5 months), the effect is unclear. For that reason and because the conception "effort" was likely to have been quite different between the couple's prestudy and on-study periods, the results presented here were not adjusted for prior months not contracepting. Eighty-two percent of the couples followed for the entire study conceived, suggesting that the volunteer study group was highly fertile. The analysis of fertility and pregnancy outcome counted only the first pregnancy (if more than one), whether it was occult or clinical. The cycle-specific pregnancy rates
Figure 1
Cycle-specific pregnancy rates for entire study period.
over the phase 1 study period (first three cycles) are displayed in Table 3. Of the 200 couples entering the study, 59 (29.5%) achieved conception during the first cycle. Of the 137 remaining couples entering cycle 2, (200; 59 pregnant, 4 dropped out), 41 (29.9%) achieved conception during the second cycle and, of the 95 remaining couples entering cycle 3 (137; 41 pregnant, 1 dropped out), 16 (16.8%) achieved conception during the third cycle. Pregnancy rates tended to decline over the remainder of the study (Fig. 1). Cumulative pregnancy rates revealed that 50% of women had achieved pregnancy in two cycles and 65% by four cycles (Fig. 2). Thirteen percent (151 116) of the pregnancies were occult in nature, as they were pregnancies detected by the hCG assay only. Because of the strict follow-up protocol and immediate implementation of serum pregnancy testing, the cycle lengths during which these 15 occult
60
-----------------+------------------------------------------------------- 90% Phase One
[ Phase Two
-------------+--------------------------~--~~~'::":-:-::'-
I
50
--------------1------
------------------------------------------------
I ---
Table 3 Pregnancy Occurrence and Outcome During Phase 1 of the Study (Includes Only First Pregnancy During First Three Cycles)
I
-------------------- ------------
I
------r---------------------------------------------·----------
I
"2
-------r-------------------------------------- -----------------
020
_______1_________ - - -- --------------- ---------------- - - --
I
E
Cycle no.
-----------------------
z
70%
~
0%
0%
1
2
3
Total
506
Zinaman et al.
Fertility and pregnancy loss
~
0
3~
~
--------------------------------------------- 20% 10 -------------
Entered cycle not pregnant 200 137 95 Pregnant during cycle 16 59 116 41 Occult 15 7 7 1 Clinical Live birth 40 10 29 79 21 Spontaneous abortion 12 5 4 I Lost to follow-up 1 6 Dropped out not pregnant 4 1 1 31.3 pregnancy (%) Loss 29.3 33.3 _ _per __ ____ _ _ _ _32.2 ____ ___ ____
i
OO/o~
clinical pregnancies only
Cycle outcome
r
00/0 ~
10
11
10%
12
Cycle Number
Figure 2 Cycle-specific conception and pregnancy outcome resuIts. Phase 1 includes live births, spontaneous abortions, and occult pregnancies. Phase 2 includes live births and spontaneous abortions. only. D, live births; D, spontaneous abortions; ., occult pregnanCIes.
Fertility and Sterility
The pregnancy loss rates were 19/59 (32%), 12/41 (29%), and 5/15 (33%) for cycles 1 to 3, respectively. The contribution of occult pregnancy to these loss rates are 7/19 (37%), 7/12 (58%), and 115 (20%), respectively. Thus, occult losses accounted for 42% of the pregnancy losses detected during phase 1.
10
DISCUSSION
0.1
-+---,----,0--,---,-----,--,-------.------, 25
26
27 31 28 29 30 Length of Cycle (days) for occult pregnancy
32
33
Figure 3 Maximum urinary hCG concentrations found in occult pregnancies by length of cycle in which they occurred.
pregnancies were detected were limited to 27 to 33 days (Fig. 3). Eleven of 14 couples who had an occult pregnancy during phase 1 and completed the entire study achieved a clinically recognized pregnancy during a later cycle; 6 of them within the first phase and 5 within the second phase of the study. A clinically recognized pregnancy was detected in the immediate ensuing cycle in five of the cases. The subsequent pregnancy outcomes of all 15 women who had occult pregnancies are listed in Table 4. Of the 11 women who conceived clinically after the occult pregnancy, all but one went on to deliver a viable offspring. Table 4 Listing of all Occult Pregnancies and Subsequent Study Outcome
Occult pregnancy no.
Occult in cycle
Clinical in cycle
1 1 1 1 2 2
2 2 2 3 3 3
1 1 2 2 3
6 6 8 12 12
Phase 1 1 2 3 4 5 6
Pregnancy outcome Live birth Live birth Live birth Live birth Live birth Spontaneous abortion
Phase 2 7 8 9 10 11 Not clinically pregnant 12 13 14 15
Vol. 65, No.3, March 1996
1 2 2 2
Live Live Live Live Live
birth birth birth birth birth
Dropped out after three cycles
In this study we determined fertility rates in a selected group of volunteers. Couples were scrutinized closely throughout their participation to ensure active attempts at conception as well as early detection of pregnancy. Fertility of couples was influenced only marginally by the number of months not contracepting before beginning the study. No correction was made because the prestudy and onstudy periods were not comparable. Fertility rates of 30% per cycle were seen for the first two cycles, with a progressive tapering over the remainder of the study. Only the first three cycles (phase 1) actually are comparable as coitus was confirmed by postcoital testing, the couples were monitored closely, and urine samples were collected only in those cycles. During the next nine cycles, the women were contacted and visits were limited to serum hCG testing if menses was delayed. Occult pregnancies would not have been detected during phase 2. There is an apparent maximum rate of conception that represents the most fertile of the couples under the study. The rapid subsequent tapering of the fertility rates would be expected because the proportion of couples with reduced fertility would increase with increasing number of exposed cycles. In this study 18% of couples failed to conceive after 12 complete menstrual cycles. Using a similar time period of 1 year, the National Center for Health Statistics reported an infertility rate of only 7.9%.(19) The higher rate is despite the considerable counseling and certainty that couples actively were attempting to conceive. Possible explanations for this difference include under-reporting of infertility (19); admission of some couples into this study, despite screening, that had a prior history offertility related problems; or our study population consisting primarily of urban and suburban, middle and upper-middle class whites had a higher than usual incidence of infertility. In 1976, Bloch (20) defined occult pregnancy "as a pregnancy that terminates so soon after implantation that no clinical suspicion exists as to its having existed." Emphasizing "no clinical suspicion," the present study took the approach that any delay in menses was suspicious. This is often the case in current clinical practice with sensitive urine and serum pregnancy testing widely available. Although the Zinaman et al. Fertility and pregnancy loss
507
term biochemical pregnancy often is used in those situations of delayed menses but without other signs of pregnancy, it also is used synonymously with occult (21) pregnancy. This often leads to confusion about whether the pregnancy was recognized by the patient. The 13% occult pregnancy rate determined in this study would be expected to disagree with studies that included patients with significant delays in menses in their occult category. In those studies of spontaneous abortion rates in which sensitive hCG assays were used to detect occult pregnancies, overall pregnancy loss rates ranged from 20% to 62% (2,11-13). The hCG assays used, minimum requirements for a positive result, and study populations have all differed. The most recent ofthese studies (2) cites an overall pregnancy loss rate of 31 %, which is not different from the present study. However, the percentage of occult and clinical pregnancies differs (22% versus 13%, 12% versus 18% for the Wilcox et al. (2) and present studies, respectively). The categories of pregnancy to which the losses were assigned appears to be important in this regard. Many of the occult losses in Wilcox et al. (2) occurred in women who were variably late for their menses. These women more likely would have been categorized as clinical losses in the present study. Differences in the participant population also exist. In particular, the present maximum age criterion was 37 versus 42 years for the Wilcox et al. study (2). In addition, the diminished sensitivity of the present hCG assay (B108-B109*; 0.05 ng/ mL) compared with Wilcox et al. (2) (B101-R525*; 0.01 ng/mL) might suggest that some occult pregnancies were missed in the present study. However, this is not likely, as no "borderline" cases were seen. The loss of sensitivity also would be expected to be at least partially offset by the fact that the (B108109*) configuration used in the present study does not require the presence ofthe hCG f3-carboxy-terminal peptide (i.e., the B108-109* assay has the potential to detect a larger family of hCG-related molecules). Although the overall number of occult pregnancies was limited, only three of the women with occult pregnancies who completed the study did not conceive subsequently. No women were identified who had two occult pregnancies, although there was only a three-cycle interval in which this could have occurred. Interestingly, five women who had an occult pregnancy were able to conceive in the next immediate cycle, suggesting little if any residual effect of the hCG on their subsequent ability to ovulate. This is not surprising, as the peak hCG concentration in this group was approximately 10 ng/mL (equivalent to 130 mIU/mL) with most peaks closer to 1 to 2 ng/ mL (equivalent to 13 to 26 mIU/mL). 508
Zinaman et al. Fertility and pregnancy loss
These results suggest that pregnancy loss accounts for approximately 30% of all detectable conceptions. The previously reported 20% to 62% range (2, 11-13) oflosses likely represents methodological differences. Improved specificity of antisera perhaps has been the greatest advance. At the same time, the demonstrated maximal fertility rates of approximately 30% per cycle in a tightly controlled setting suggests that there may be a practical upper limit to per cycle fertility rates in couples. Even with couples who are highly fertile, conception in a given cycle may be precluded by several factors, including a transient abnormal cycle, poor timing of coitus relative to ovulation, or loss of a conception before detection by hCG assay. Although the results from these study couples cannot be generalized to all couples attempting conception, upper limits appear to exist in natural cycles. This information should shed further light on the limitations and potential for conception. REFERENCES 1. Barrett JC, Marshall J. The risk of conception on different days of the menstrual cycle. Popul Stud 1969;23:455-61. 2. Wilcox AJ, Weinburg CR, O'Connor J, Baird DD, Schatterer JP, Canfield RE, et al. Incidence of early loss of pregnancy. N Engl J Med 1988;319:189-94. 3. Schwartz D, Mayaux MJ. Female fecundity as a function of age. N Engl J Med 1982;306:404-6. 4. Warburton D. Spontaneous abortion risks in man: data from reproductive histories collected in a medical genetics unit. Am J Hum Genet 1964;16:1-25. 5. Poland BJ, Miller JR, Jones DC, Trimble BK. Reproductive counseling in patients who have had a spontaneous abortion. Am J Obstet Gynecol 1977; 127:685-91. 6. French FE, Bierman J. Probabilities offetal mortality. Public Health Rep 1962;77:835-47. 7. Erhardt CL. Pregnancy losses in New York City, 1960. Am J Public Health 1963;53:1337-57. 8. Taylor WF. The probability of fetal death. In: Fraser FC, McKusick CA, editors. Congenital malformations. New York: Excerpta Medica, 1969:307-20. 9. Harlap S, Shiono P, Ramcharan S. A life table of spontaneous abortions and the effects of age parity and other variables. In: Porter IH, Hook EB, editors. Human embryonic and fetal death. New York: Academic Press, 1980:145-64. 10. Petterson F. Epidemiology of early pregnancy wastage: biological and social correlates of abortion. An investigation based on materials collected within Uppsala County, Stockholm, Sweden. Stockholm, Sweden: Svenska Bokforlaget, 1968. 11. Miller JF, Williams E, Glue J, Gordon YB, Grudzinskas JG, Sykes A. Fetal loss after implantation. Lancet 1980; 1:5546. 12. Edmonds KD, Lindsay KS, Miller JF, Williamson E, Wood PJ. Early embryonic mortality in women. Fertil Steril 1982; 38:447-53. 13. Whittaker PG, Taylor A, Lind T. Unsuspected pregnancy loss in healthy women. Lancet 1983;1:1126-7. 14. Ehrlich PH, Moustafa ZA, Krichevsky A, Birken S, Armstrong EG, Canfield RE. Characterization and relative orientation of epitopes for monoclonal antibodies and antisera to
Fertility and Sterility
15.
16.
17.
18.
human chorionic gonadotropin. Am J Reprod Immunol MicrobioI 1985;8:48-54. Ehrlich PH, Moyle WR, Canfield RE. Monoclonal antibodies to gonadotropin subunits. Methods Enzymol 1985; 109:63855. Ehrlich P, Moyle W, Moustafa ZA, Canfield RE. Mixing two monoclonal antibodies yields enhanced affinity for antigen. J Immunol 1982; 128:2709-13. O'Connor JF, Shlatterer JP, Birken S, Krichevsky E, Armstrong EG, McMahon D, et al. Development of highly sensitive immunoassays to measure human chorionic gonadotropin, its fJ-subunit and fJ core fragment in the urine: application to malignancies. Cancer Res 1988;48:1361-6. Taylor CA Jr, Overstreet JW, Samuels SJ, Boyers SP, Can-
Vol. 65, No.3, March 1996
field RE, O'Connor JF. Prospective assessment of early fetal loss using an immunoenzymometric screening assay for detection of urinary human chorionic gonadotropin. Fertil Steril 1992;57:1220-4. 19. Mosher WD, Pratt WF. Fecundity and infertility in the United States, 1965-88. Advance data from vital and health statistics, no. 192. Hyattsville (MD): National Center for Health Statistics, 1990. 20. Bloch SK. Occult pregnancy: a pilot study. Obstet Gynecol 1976;48:365-8. 21. Kline J, Stein Z, Susser M. Conception and reproductive loss: probabilities. In: Kline J, Stein Z, Susser M, editors. Conception to birth: epidemiology of prenatal development. New York: Oxford University Press, 1989:43-68.
Zinaman et al.
Fertility and pregnancy loss
509