- Email: [email protected]
Bleeding and spontaneous abortion after therapy for infertility
FERTILITY AND STERILITY威 VOL. 74, NO. 3, SEPTEMBER 2000 Copyright ©2000 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A.
Bleeding and spontaneous abortion after therapy for infertility Kevin Pezeshki, M.D.,a,c Joseph Feldman, Dr. P.H.,b Daniel E. Stein, M.D.,a Susan M. Lobel, M.D.,a and Richard V. Grazi, M.D.a Maimonides Medical Center and State University of New York Health Science Center at Brooklyn, Brooklyn, New York
Received December 17, 1999; revised and accepted March 31, 2000. Reprint requests: Richard V. Grazi, M.D., Division of Reproductive Endocrinology and Infertility, Maimonides Medical Center, 4802 Tenth Avenue, Brooklyn, NY 11219 (FAX: 718-2836580; E-mail: drgrazi @aol.com). a Division of Reproductive Endocrinology and Infertility, Maimonides Medical Center, Brooklyn, New York. b Department of Preventive Medicine and Community Health, State University of New York Health Science Center at Brooklyn. c Present address: Northridge Hospital Medical Center, Roscoe Boulevard Campus, 18300 Roscoe Boulevard, Northridge, CA 91328. 0015-0282/00/$20.00 PII S0015-0282(00)00707-X
504
Objective: To determine the incidence of early-pregnancy bleeding and spontaneous abortion (SAB) after various treatments for infertility and to assess whether bleeding is predictive of SAB. Design: An historic cohort study of women who conceived after various treatments. Setting: Hospital-based private practice. Patient(s): We studied 418 patients in whom 500 consecutive clinical pregnancies occurred. Intervention(s): Patients were grouped according to the method of conception: ovulation induction, IVF, and other. The latter category included interventions not requiring ovulation induction, such as surgery and insemination. A fourth group of subjects who conceived independently of treatment was used as the control. Main Outcome Measure(s): Bleeding and pregnancy outcome (SAB, ectopic pregnancy, or ongoing pregnancy). Result(s): Rates of SAB did not differ among the treatment groups. SAB occurred significantly more often after bleeding than when bleeding did not occur (30.8% versus 19.8%, respectively). Bleeding was predictive of SAB only in patients ⬍35 years old (odds ratio 2.4). Conclusion(s): Infertile women who conceive after reproductive therapy are not at increased risk for SAB compared with women who conceive naturally. There appears to be no association between previous diagnosis or treatment and the occurrence of SAB in previously infertile women. Bleeding is associated with a twofold relative risk of SAB. (Fertil Steril威 2000;74:504 – 8. ©2000 by American Society for Reproductive Medicine.) Key Words: Bleeding, infertility, pregnancy, spontaneous abortion
Spontaneous abortion (SAB) occurs in 20%–25% of pregnancies (1, 2). In the setting of previous infertility, SAB is especially troubling to the patient, who typically questions whether the treatment leading to her pregnancy was related in some way to the adverse outcome. Any amount of bleeding that occurs during early pregnancy in these women evokes a similar question. Counseling of patients with a history of infertility who conceive is often required to explain the significance of bleeding and the risk of miscarriage.
the presence or absence of an association between fertility therapies and subsequent bleeding during pregnancy.
Although most reproductive specialists are aware that bleeding during early pregnancy is common and usually does not herald pregnancy loss, there are few data upon which to base this reassurance. Our review of the current literature yielded no statistics on the precise incidence of bleeding in pregnancies after fertility therapy. There are also no data describing
We performed a retrospective cohort analysis of women who conceived after infertility treatments at Brooklyn IVF from May 1996 through September 1998. We used the patients’ medical records to simulate the prospective experience. The cohorts were determined on the basis of exposure to a specific type of treatment for infertility and the presence or
This study was performed to determine the incidence of early-pregnancy bleeding after various treatments for infertility, to compare it with the incidence of bleeding in naturally occurring pregnancies, and to assess whether such bleeding is related to the occurrence of SAB.
MATERIALS AND METHODS
absence of bleeding. The predictor variables were cause of infertility, method of treatment, and bleeding. The outcome of primary interest was pregnancy outcome (SAB, ectopic pregnancy, or ongoing pregnancy).
Subjects All patients sought treatment for infertility at Brooklyn IVF, a hospital-based private practice. All patient records are kept in an electronic database (Labsystem; Kretz Consulting, Cabin John, MD) that runs on a Novell network of 30 linked PCs. Every encounter is recorded, including patient telephone calls, which are routed electronically from receptionists to the clinicians responsible for responding. These messages, as well as the responses from the clinicians, become part of the electronic record. Patients’ subjective complaints are also recorded at the time of each office visit with the use of an electronic version of the traditional subjective, objective, assessment, and plan protocol. A list of all patients who conceived during the study period was assembled with the use of data from serum hCG testing at the on-site laboratory. This hormone was assayed with an immunofluorescence technique (3). Levels of hCG of ⱖ10 mIU/mL were considered positive. All patient records corresponding to a positive test were reviewed. Patients who conceived before initiating any fertility evaluation or treatment as well as patients with “chemical pregnancies” (transiently positive hCG and no sac detected on ultrasonographic (US) examination) were excluded from the study. Patients who conceived after evaluation and treatment but independent of the treatment given were included in the study as controls. During the study period, all patients were given US examinations at 6 and 8 weeks of pregnancy and were referred out for obstetric care at 12 weeks.
Definitions Infertility was defined as the inability to conceive after 12 months of unprotected coitus. Causes of infertility were categorized as follows. Patients were considered to have male factor infertility if they were being treated for a known defect in their husbands’ sperm count, motility, or morphology as defined by World Health Organization (4) or the criteria of Kruger et al. (5). Patients were diagnosed with ovulatory dysfunction if they were being treated for anovulation, luteal phase defect, or oligoovulation. Anatomic infertility included women with a diagnosis of tubal occlusion on hysterosalpingogram or with laparoscopic confirmation of adhesive disease. Endometriosis was diagnosed by laparoscopic findings. Cervical factor infertility was diagnosed when repeated, well-timed postcoital testing showed no viable sperm but the partner’s semen analysis was normal. Miscellaneous diagnoses included known uterine anomalies, immunologic infertility, advanced maternal age, hyperprolactinemia, and history of infertility after recurrent pregnancy loss. Unexplained infertility was diagnosed when all of the aforementioned causes of infertility had been exFERTILITY & STERILITY威
cluded. Any patient with more than one known cause of infertility was categorized as having multifactorial infertility. Using primary modes of treatment exclusively, we identified and labeled three distinct treatment groups as follows. Group I (ovulation induction) consisted of women who conceived after ovulation induction with either clomiphene citrate (CC), gonadotropins, or gonadotropins in combination with a GnRH analog (GnRH-a). The use or nonuse of intrauterine insemination (IUI) did not alter assignment to this group. Group II (IVF) included women who conceived after IVF with or without intracytoplasmic sperm injection. All patients were pretreated with a GnRH-a and gonadotropins with either a long or short protocol (6). Group III (other) included patients who conceived after any type of reproductive surgery, IUI during an unstimulated cycle, or medical treatment, including bromocriptine and P. Group IV (untreated controls) was comprised of women who presented with infertility and were evaluated and/or treated, but who conceived independently of any treatment. The diagnosis of bleeding during early pregnancy was based upon the patients’ subjective complaints and/or physical examination. The outcome of every early pregnancy was known. Outcomes were coded as SAB, ectopic pregnancy, or ongoing pregnancy. Spontaneous abortions included complete, incomplete, and missed abortions. Ongoing pregnancies were defined as those progressing to ⱖ12 weeks.
Database Data were organized with Microsoft Access (Microsoft, Redmond, WA), a relational database program that can be customized for unlimited data fields. The age and parity of every patient were entered. The primary diagnosis as well as any secondary diagnosis and treatment were entered. We used the age breakdown of ⬍35, 35–39, and ⱖ40 years to be consistent with the IVF Registry (7). The presence or absence and the severity of bleeding were noted in the database. Mild, moderate, and severe bleeding were combined in our analysis because patients’ subjective experience of bleeding could easily affect this classification, and it was not always possible to examine patients at the time of the reported episode. Fisher’s exact test was used for assessing the association of two variables. An exact trend test was used to examine the associations by age group. Multiple logistic regression was performed to analyze the association among several variables. In the regression model, SAB was the outcome and bleeding was the predictor variable; age, treatment, and diagnosis were the covariates. Interactions among bleeding, age, treatment, and diagnosis were examined. Models were assessed for goodness of fit and outliers. Statistical significance was defined as P⬍.05, and confidence intervals (CIs) were at the 95% level. Computations were performed with the use of software from StatXact (Cytel Software, Cambridge, MA) and SPSS (Chicago, IL). 505
TABLE 1
TABLE 3
Treatments leading to pregnancy, grouped according to age.
Rate of SAB by maternal age and treatment. Treatment groupsa,b
Treatment groupa Age (y) Age (y) ⬍35 35–39 ⱖ40 Total
I
II
III
IV
Total
191 (50.1) 41 (46.6) 8 (25.8) 240 (48.0)
69 (18.1) 19 (21.6) 8 (25.8) 96 (19.2)
44 (11.5) 10 (11.4) 3 (9.7) 57 (11.4)
77 (20.2) 18 (20.5) 12 (38.7) 107 (21.4)
381 88 31 500
Note: Values are n (%). Numbers in parentheses indicate percentage of total age group. a Group key: I ⫽ ovulation induction; II ⫽ IVF; III ⫽ other; IV ⫽ spontaneous conception. Pezeshki. Bleeding and spontaneous abortion. Fertil Steril 2000.
⬍35 35–39 ⱖ40 Total
I
II
III
IV
36 (18.8) 11 (26.8) 4 (50.0) 51 (21.3)
13 (18.8) 3 (15.8) 3 (37.5) 19 (19.8)
12 (27.3) 3 (30.0) 3 (100.0) 18 (31.6)
15 (19.5) 8 (44.4) 5 (41.7) 28 (26.2)
Note: Values are n (%). Numbers in parentheses indicate percentage of total age group who received the indicated treatment and had SAB. a Group key: I ⫽ ovulation induction; II ⫽ IVF; III ⫽ other; IV ⫽ spontaneous conception. b P ⫽ not significant for all treatment and age groups. Pezeshki. Bleeding and spontaneous abortion. Fertil Steril 2000.
The analysis focused on risk factors for SAB compared with ongoing pregnancies. Ectopic pregnancies, although shown in the tables for completeness, were not included in the analysis. The implication of failing to find a difference in the frequency of bleeding by treatment group was quantified by estimating the power to detect such a difference given the current sample sizes.
RESULTS The study group consisted of 418 patients, among whom 500 clinical pregnancies occurred. The age breakdown was ⬍35 years old (76.2%), 35–39 years (17.6%), and ⱖ40 years (6.2%). Approximately two thirds of the study population was nulliparous. The diagnostic groups of infertility were as follows: 11.8% male factor, 29.2% ovulatory dysfunction, 12.8% anatomic, 3.6% endometriosis, 4.2% cervical factor, 10.2% miscellaneous, 16.4% unexplained, and 11.8% multifactorial. Table 1 shows the age distribution of patients in treatment groups I–IV. There were 240 women in group I (ovulation induction), 96 in group II (IVF), 57 in group III (other), and 107 in group IV (untreated). Table 2 presents the outcomes of pregnancy in the study
population by age group. The overall rate of SAB was 23.2%. SAB occurred significantly less often in women ⬍35 years old than in women in the two older categories (odds ratio [OR] 1.6, 95% CI 0.9 –2.7 and OR 3.6, 95% CI 1.7–7.8; Ptrend⬍.01). There was no association between the diagnosis for which treatment was given and the occurrence of SAB (data not shown). Table 3 groups the patients who had SAB by maternal age and treatment. Rates of SAB were not significantly different across treatment groups (Pⱕ.26). Furthermore, the association of SAB with age was consistent among the treatment groups (Pinteractionⱕ.82). Patients in group I (ovulation induction) also were analyzed according to whether they were treated with CC or gonadotropins. There was no statistically significant difference in the rates of SAB between the groups (data not shown). In Table 4, pregnancy outcome is analyzed by age and the
TABLE 4 Early-pregnancy bleeding and pregnancy outcome. Age (y) ⬍35
TABLE 2
35–39
Age and pregnancy outcome. Age (y)
ⱖ40
Ongoing
SAB
Ectopic
Total
276 (72.4) 57 (64.8) 15 (48.4) 348 (69.6)
76 (19.9)a 25 (28.4) 15 (48.4) 116 (23.2)
29 (7.6) 6 (6.8) 1 (3.2) 36 (7.2)
381 88 31 500
Total ⬍35 35–39 ⱖ40 Total
Bleeding
Ongoing
SAB
Ectopic
No Yes No Yes No Yes No Yes
197 (75.8) 79 (65.3) 37 (63.8) 20 (66.7) 12 (46.2) 3 (60.0) 246 (71.5) 102 (65.4)
39 (15.0) 37 (30.6)a 16 (27.6) 9 (30.0) 13 (50.0) 2 (40.0) 68 (19.8) 48 (30.8)b
24 (9.2) 5 (4.1) 5 (8.6) 1 (3.3) 1 (3.8) —— 30 (8.7) 6 (3.8)
Note: Values are n (%). a P⬍.01.
Note: Values are n (%) a P⬍.01 compared with SAB rate among women in the same age group with no early-pregnancy bleeding (OR 2.4, 95% CI 1.4 –3.98). b P⬍.01 compared with SAB rate among all women with no early-pregnancy bleeding (OR 1.85, 95% CI 1.18 –2.89).
Pezeshki. Bleeding and spontaneous abortion. Fertil Steril 2000.
Pezeshki. Bleeding and spontaneous abortion. Fertil Steril 2000.
506
Pezeshki et al.
Spontaneous abortion after infertility
Vol. 74, No. 3, September 2000
presence of early bleeding. SAB was seen more commonly after bleeding than when bleeding did not occur (30.8% versus 19.8%, respectively; P⬍.01). Bleeding was associated with an 85% increased risk of subsequent SAB as compared with ongoing pregnancy (OR 1.85, 95% CI 1.18 – 2.89) after adjusting for age. In analyses stratifying by age, bleeding was predictive of miscarriage in patients ⬍35 years old (OR 2.4, 95% CI 1.4 –3.98; P⬍.01) but not in women ⬎35 years old (OR 0.81, 95% CI 0.34 –1.89). The attributable risk of SAB, which is the difference in occurrence between women who had bleeding and those who did not, was 15.6% in women ⬍35 years old. There was no association between the occurrence of bleeding and the diagnosis for which patients were treated (data not shown). By treatment group, bleeding varied from 26.2% among spontaneous pregnancies to 34.4% among patients who had ovulation induction (Pⱕ.52). The power to detect an 8% difference in bleeding, if one existed, by treatment in the population with the present sample size was only 24%.
DISCUSSION The initial purpose of this study was to provide information with which to counsel infertile women who conceived after reproductive therapy about their risks of SAB, particularly in the setting of first-trimester bleeding. Women who bled had a higher risk of SAB than women who did not. The frequency of SAB was twice as high when bleeding was encountered. Because bleeding occurred in one third of the women ⬍35 years old, the attributable risk was high (15.6%). Nonetheless, the fact that two thirds of the patients who bled did not lose their pregnancies should provide some reassurance. The high frequency of bleeding reported by women in this cohort (31.2%) may be related to the composition of the study group (infertility patients) who, after pregnancy is diagnosed, tend to be sensitive to even the smallest amount of staining or spotting. Although bleeding often resolves before an office visit, we were able to track the occurrence of bleeding accurately because our institutional protocol requires that all patient encounters, including telephone calls, be recorded in our electronic patient database. An additional purpose of this study was to compare rates of SAB in a cohort of previously infertile women who conceived spontaneously or after various treatments for infertility. We found that whether pregnancy occurred spontaneously or through treatment, women who conceived after experiencing infertility had a similar risk of losing the pregnancy before 12 weeks. This risk did not vary by the diagnosis for which the women were treated or by the specific treatment given. Again, these findings should be reassuring to infertile women and their physicians. The rate of SAB detected in this study was similar to that reported in studies of women with normal fertility and subfertility (2, 8). As in FERTILITY & STERILITY威
other studies on SAB (9), the risk in our study population was directly related to maternal age. Although this study is limited by its retrospective design, its inclusion of a relatively large number of patients as well as an untreated control group makes the conclusions more compelling. There is also some internal consistency to the data, in that SAB rates were significantly higher in older women and conceptions in older women were achieved independently of treatment more often than in their younger infertile counterparts (Table 1). The makeup of the study population in terms of age, gravidity, and distribution of diagnoses appears typical of women seeking treatment for infertility. The finding that bleeding was predictive of miscarriage only in the younger group of women is likely a result of the nature of SAB in older women. Although almost all women with SAB will bleed eventually if untreated, older women with histories of infertility are more likely to have suspicious US examinations at their first prenatal visit. This occurs because of the relatively more common occurrence in this population of chromosomal anomalies as a cause of miscarriage. Earlier diagnosis of SAB and termination of the pregnancy in these women may have reduced the frequency with which bleeding was encountered. In conclusion, infertile women who conceive after fertility therapy are not at increased risk of SAB compared with those who conceive spontaneously. Overall, their risk of SAB appears to be similar to that of infertile women who conceive without treatment as well as to that of women with normal fertility. There appears to be no association between previous diagnosis or treatment and the occurrence of SAB in previously infertile women. Although bleeding does not herald SAB in most of these women, it does in many, and its occurrence is associated with approximately a twofold risk of SAB compared with women who do not experience bleeding.
Acknowledgment: The authors thank Howard Minkoff, M.D., for his thoughtful review and comments on this article.
References 1. Laferia JJ. Spontaneous abortion. Clin Obstet Gynecol 1986;13:105–14. 2. Wilcox AJ, Baird DD, Weinberg CR. Time of implantation of the conceptus and loss of pregnancy. N Engl J Med 1999;340:1796 –9. 3. Henry JB. Clinical diagnosis and management by laboratory methods. 19th ed. Philadelphia: WB Saunders, 1996:483–91. 4. World Health Organization. Laboratory manual for the examination of human semen and semen-cervical mucus interaction. 3rd ed. New York: Cambridge University Press, 1993:43– 4. 5. Kruger TF, Menkveld R, Stander FSH, et al. Sperm morphological features as a prognostic factor in in vitro fertilization. Fertil Steril 1986;46:1118 –23. 6. Hughes EG, Fedorkow DM, Daya S, Sagle MA, Van de Koppel P, Collins JA. The routine use of gonadotropin-releasing hormone agonists prior to in vitro fertilization and gamete intrafallopian transfer: a metaanalysis of randomized controlled trials. Fertil Steril 1992;58:888 –96. 7. Society for Assisted Reproductive Technology, The American Society
507
for Reproductive Medicine. Assisted reproductive technology in the United States: 1996 results generated from the American Society for Reproductive Medicine/Society for Assisted Reproductive Technology Registry. Fertil Steril 1999;71:798 – 807. 8. Guzick DS, Carson SA, Coutifaris C, et al. Efficacy of superovulation
508
Pezeshki et al.
Spontaneous abortion after infertility
and intrauterine insemination in the treatment of infertility. N Engl J Med 1999;340:177– 83. 9. Abdalla HI, Billett A, Kan AKS, Baig S, Wren M, Korea L, et al. Obstetric outcome in 232 ovum donation pregnancies. Br J Obstet Gynaecol 1998;105:332–7.
Vol. 74, No. 3, September 2000
Bleeding and spontaneous abortion after therapy for infertility Kevin Pezeshki, M.D.,a,c Joseph Feldman, Dr. P.H.,b Daniel E. Stein, M.D.,a Susan M. Lobel, M.D.,a and Richard V. Grazi, M.D.a Maimonides Medical Center and State University of New York Health Science Center at Brooklyn, Brooklyn, New York
Received December 17, 1999; revised and accepted March 31, 2000. Reprint requests: Richard V. Grazi, M.D., Division of Reproductive Endocrinology and Infertility, Maimonides Medical Center, 4802 Tenth Avenue, Brooklyn, NY 11219 (FAX: 718-2836580; E-mail: drgrazi @aol.com). a Division of Reproductive Endocrinology and Infertility, Maimonides Medical Center, Brooklyn, New York. b Department of Preventive Medicine and Community Health, State University of New York Health Science Center at Brooklyn. c Present address: Northridge Hospital Medical Center, Roscoe Boulevard Campus, 18300 Roscoe Boulevard, Northridge, CA 91328. 0015-0282/00/$20.00 PII S0015-0282(00)00707-X
504
Objective: To determine the incidence of early-pregnancy bleeding and spontaneous abortion (SAB) after various treatments for infertility and to assess whether bleeding is predictive of SAB. Design: An historic cohort study of women who conceived after various treatments. Setting: Hospital-based private practice. Patient(s): We studied 418 patients in whom 500 consecutive clinical pregnancies occurred. Intervention(s): Patients were grouped according to the method of conception: ovulation induction, IVF, and other. The latter category included interventions not requiring ovulation induction, such as surgery and insemination. A fourth group of subjects who conceived independently of treatment was used as the control. Main Outcome Measure(s): Bleeding and pregnancy outcome (SAB, ectopic pregnancy, or ongoing pregnancy). Result(s): Rates of SAB did not differ among the treatment groups. SAB occurred significantly more often after bleeding than when bleeding did not occur (30.8% versus 19.8%, respectively). Bleeding was predictive of SAB only in patients ⬍35 years old (odds ratio 2.4). Conclusion(s): Infertile women who conceive after reproductive therapy are not at increased risk for SAB compared with women who conceive naturally. There appears to be no association between previous diagnosis or treatment and the occurrence of SAB in previously infertile women. Bleeding is associated with a twofold relative risk of SAB. (Fertil Steril威 2000;74:504 – 8. ©2000 by American Society for Reproductive Medicine.) Key Words: Bleeding, infertility, pregnancy, spontaneous abortion
Spontaneous abortion (SAB) occurs in 20%–25% of pregnancies (1, 2). In the setting of previous infertility, SAB is especially troubling to the patient, who typically questions whether the treatment leading to her pregnancy was related in some way to the adverse outcome. Any amount of bleeding that occurs during early pregnancy in these women evokes a similar question. Counseling of patients with a history of infertility who conceive is often required to explain the significance of bleeding and the risk of miscarriage.
the presence or absence of an association between fertility therapies and subsequent bleeding during pregnancy.
Although most reproductive specialists are aware that bleeding during early pregnancy is common and usually does not herald pregnancy loss, there are few data upon which to base this reassurance. Our review of the current literature yielded no statistics on the precise incidence of bleeding in pregnancies after fertility therapy. There are also no data describing
We performed a retrospective cohort analysis of women who conceived after infertility treatments at Brooklyn IVF from May 1996 through September 1998. We used the patients’ medical records to simulate the prospective experience. The cohorts were determined on the basis of exposure to a specific type of treatment for infertility and the presence or
This study was performed to determine the incidence of early-pregnancy bleeding after various treatments for infertility, to compare it with the incidence of bleeding in naturally occurring pregnancies, and to assess whether such bleeding is related to the occurrence of SAB.
MATERIALS AND METHODS
absence of bleeding. The predictor variables were cause of infertility, method of treatment, and bleeding. The outcome of primary interest was pregnancy outcome (SAB, ectopic pregnancy, or ongoing pregnancy).
Subjects All patients sought treatment for infertility at Brooklyn IVF, a hospital-based private practice. All patient records are kept in an electronic database (Labsystem; Kretz Consulting, Cabin John, MD) that runs on a Novell network of 30 linked PCs. Every encounter is recorded, including patient telephone calls, which are routed electronically from receptionists to the clinicians responsible for responding. These messages, as well as the responses from the clinicians, become part of the electronic record. Patients’ subjective complaints are also recorded at the time of each office visit with the use of an electronic version of the traditional subjective, objective, assessment, and plan protocol. A list of all patients who conceived during the study period was assembled with the use of data from serum hCG testing at the on-site laboratory. This hormone was assayed with an immunofluorescence technique (3). Levels of hCG of ⱖ10 mIU/mL were considered positive. All patient records corresponding to a positive test were reviewed. Patients who conceived before initiating any fertility evaluation or treatment as well as patients with “chemical pregnancies” (transiently positive hCG and no sac detected on ultrasonographic (US) examination) were excluded from the study. Patients who conceived after evaluation and treatment but independent of the treatment given were included in the study as controls. During the study period, all patients were given US examinations at 6 and 8 weeks of pregnancy and were referred out for obstetric care at 12 weeks.
Definitions Infertility was defined as the inability to conceive after 12 months of unprotected coitus. Causes of infertility were categorized as follows. Patients were considered to have male factor infertility if they were being treated for a known defect in their husbands’ sperm count, motility, or morphology as defined by World Health Organization (4) or the criteria of Kruger et al. (5). Patients were diagnosed with ovulatory dysfunction if they were being treated for anovulation, luteal phase defect, or oligoovulation. Anatomic infertility included women with a diagnosis of tubal occlusion on hysterosalpingogram or with laparoscopic confirmation of adhesive disease. Endometriosis was diagnosed by laparoscopic findings. Cervical factor infertility was diagnosed when repeated, well-timed postcoital testing showed no viable sperm but the partner’s semen analysis was normal. Miscellaneous diagnoses included known uterine anomalies, immunologic infertility, advanced maternal age, hyperprolactinemia, and history of infertility after recurrent pregnancy loss. Unexplained infertility was diagnosed when all of the aforementioned causes of infertility had been exFERTILITY & STERILITY威
cluded. Any patient with more than one known cause of infertility was categorized as having multifactorial infertility. Using primary modes of treatment exclusively, we identified and labeled three distinct treatment groups as follows. Group I (ovulation induction) consisted of women who conceived after ovulation induction with either clomiphene citrate (CC), gonadotropins, or gonadotropins in combination with a GnRH analog (GnRH-a). The use or nonuse of intrauterine insemination (IUI) did not alter assignment to this group. Group II (IVF) included women who conceived after IVF with or without intracytoplasmic sperm injection. All patients were pretreated with a GnRH-a and gonadotropins with either a long or short protocol (6). Group III (other) included patients who conceived after any type of reproductive surgery, IUI during an unstimulated cycle, or medical treatment, including bromocriptine and P. Group IV (untreated controls) was comprised of women who presented with infertility and were evaluated and/or treated, but who conceived independently of any treatment. The diagnosis of bleeding during early pregnancy was based upon the patients’ subjective complaints and/or physical examination. The outcome of every early pregnancy was known. Outcomes were coded as SAB, ectopic pregnancy, or ongoing pregnancy. Spontaneous abortions included complete, incomplete, and missed abortions. Ongoing pregnancies were defined as those progressing to ⱖ12 weeks.
Database Data were organized with Microsoft Access (Microsoft, Redmond, WA), a relational database program that can be customized for unlimited data fields. The age and parity of every patient were entered. The primary diagnosis as well as any secondary diagnosis and treatment were entered. We used the age breakdown of ⬍35, 35–39, and ⱖ40 years to be consistent with the IVF Registry (7). The presence or absence and the severity of bleeding were noted in the database. Mild, moderate, and severe bleeding were combined in our analysis because patients’ subjective experience of bleeding could easily affect this classification, and it was not always possible to examine patients at the time of the reported episode. Fisher’s exact test was used for assessing the association of two variables. An exact trend test was used to examine the associations by age group. Multiple logistic regression was performed to analyze the association among several variables. In the regression model, SAB was the outcome and bleeding was the predictor variable; age, treatment, and diagnosis were the covariates. Interactions among bleeding, age, treatment, and diagnosis were examined. Models were assessed for goodness of fit and outliers. Statistical significance was defined as P⬍.05, and confidence intervals (CIs) were at the 95% level. Computations were performed with the use of software from StatXact (Cytel Software, Cambridge, MA) and SPSS (Chicago, IL). 505
TABLE 1
TABLE 3
Treatments leading to pregnancy, grouped according to age.
Rate of SAB by maternal age and treatment. Treatment groupsa,b
Treatment groupa Age (y) Age (y) ⬍35 35–39 ⱖ40 Total
I
II
III
IV
Total
191 (50.1) 41 (46.6) 8 (25.8) 240 (48.0)
69 (18.1) 19 (21.6) 8 (25.8) 96 (19.2)
44 (11.5) 10 (11.4) 3 (9.7) 57 (11.4)
77 (20.2) 18 (20.5) 12 (38.7) 107 (21.4)
381 88 31 500
Note: Values are n (%). Numbers in parentheses indicate percentage of total age group. a Group key: I ⫽ ovulation induction; II ⫽ IVF; III ⫽ other; IV ⫽ spontaneous conception. Pezeshki. Bleeding and spontaneous abortion. Fertil Steril 2000.
⬍35 35–39 ⱖ40 Total
I
II
III
IV
36 (18.8) 11 (26.8) 4 (50.0) 51 (21.3)
13 (18.8) 3 (15.8) 3 (37.5) 19 (19.8)
12 (27.3) 3 (30.0) 3 (100.0) 18 (31.6)
15 (19.5) 8 (44.4) 5 (41.7) 28 (26.2)
Note: Values are n (%). Numbers in parentheses indicate percentage of total age group who received the indicated treatment and had SAB. a Group key: I ⫽ ovulation induction; II ⫽ IVF; III ⫽ other; IV ⫽ spontaneous conception. b P ⫽ not significant for all treatment and age groups. Pezeshki. Bleeding and spontaneous abortion. Fertil Steril 2000.
The analysis focused on risk factors for SAB compared with ongoing pregnancies. Ectopic pregnancies, although shown in the tables for completeness, were not included in the analysis. The implication of failing to find a difference in the frequency of bleeding by treatment group was quantified by estimating the power to detect such a difference given the current sample sizes.
RESULTS The study group consisted of 418 patients, among whom 500 clinical pregnancies occurred. The age breakdown was ⬍35 years old (76.2%), 35–39 years (17.6%), and ⱖ40 years (6.2%). Approximately two thirds of the study population was nulliparous. The diagnostic groups of infertility were as follows: 11.8% male factor, 29.2% ovulatory dysfunction, 12.8% anatomic, 3.6% endometriosis, 4.2% cervical factor, 10.2% miscellaneous, 16.4% unexplained, and 11.8% multifactorial. Table 1 shows the age distribution of patients in treatment groups I–IV. There were 240 women in group I (ovulation induction), 96 in group II (IVF), 57 in group III (other), and 107 in group IV (untreated). Table 2 presents the outcomes of pregnancy in the study
population by age group. The overall rate of SAB was 23.2%. SAB occurred significantly less often in women ⬍35 years old than in women in the two older categories (odds ratio [OR] 1.6, 95% CI 0.9 –2.7 and OR 3.6, 95% CI 1.7–7.8; Ptrend⬍.01). There was no association between the diagnosis for which treatment was given and the occurrence of SAB (data not shown). Table 3 groups the patients who had SAB by maternal age and treatment. Rates of SAB were not significantly different across treatment groups (Pⱕ.26). Furthermore, the association of SAB with age was consistent among the treatment groups (Pinteractionⱕ.82). Patients in group I (ovulation induction) also were analyzed according to whether they were treated with CC or gonadotropins. There was no statistically significant difference in the rates of SAB between the groups (data not shown). In Table 4, pregnancy outcome is analyzed by age and the
TABLE 4 Early-pregnancy bleeding and pregnancy outcome. Age (y) ⬍35
TABLE 2
35–39
Age and pregnancy outcome. Age (y)
ⱖ40
Ongoing
SAB
Ectopic
Total
276 (72.4) 57 (64.8) 15 (48.4) 348 (69.6)
76 (19.9)a 25 (28.4) 15 (48.4) 116 (23.2)
29 (7.6) 6 (6.8) 1 (3.2) 36 (7.2)
381 88 31 500
Total ⬍35 35–39 ⱖ40 Total
Bleeding
Ongoing
SAB
Ectopic
No Yes No Yes No Yes No Yes
197 (75.8) 79 (65.3) 37 (63.8) 20 (66.7) 12 (46.2) 3 (60.0) 246 (71.5) 102 (65.4)
39 (15.0) 37 (30.6)a 16 (27.6) 9 (30.0) 13 (50.0) 2 (40.0) 68 (19.8) 48 (30.8)b
24 (9.2) 5 (4.1) 5 (8.6) 1 (3.3) 1 (3.8) —— 30 (8.7) 6 (3.8)
Note: Values are n (%). a P⬍.01.
Note: Values are n (%) a P⬍.01 compared with SAB rate among women in the same age group with no early-pregnancy bleeding (OR 2.4, 95% CI 1.4 –3.98). b P⬍.01 compared with SAB rate among all women with no early-pregnancy bleeding (OR 1.85, 95% CI 1.18 –2.89).
Pezeshki. Bleeding and spontaneous abortion. Fertil Steril 2000.
Pezeshki. Bleeding and spontaneous abortion. Fertil Steril 2000.
506
Pezeshki et al.
Spontaneous abortion after infertility
Vol. 74, No. 3, September 2000
presence of early bleeding. SAB was seen more commonly after bleeding than when bleeding did not occur (30.8% versus 19.8%, respectively; P⬍.01). Bleeding was associated with an 85% increased risk of subsequent SAB as compared with ongoing pregnancy (OR 1.85, 95% CI 1.18 – 2.89) after adjusting for age. In analyses stratifying by age, bleeding was predictive of miscarriage in patients ⬍35 years old (OR 2.4, 95% CI 1.4 –3.98; P⬍.01) but not in women ⬎35 years old (OR 0.81, 95% CI 0.34 –1.89). The attributable risk of SAB, which is the difference in occurrence between women who had bleeding and those who did not, was 15.6% in women ⬍35 years old. There was no association between the occurrence of bleeding and the diagnosis for which patients were treated (data not shown). By treatment group, bleeding varied from 26.2% among spontaneous pregnancies to 34.4% among patients who had ovulation induction (Pⱕ.52). The power to detect an 8% difference in bleeding, if one existed, by treatment in the population with the present sample size was only 24%.
DISCUSSION The initial purpose of this study was to provide information with which to counsel infertile women who conceived after reproductive therapy about their risks of SAB, particularly in the setting of first-trimester bleeding. Women who bled had a higher risk of SAB than women who did not. The frequency of SAB was twice as high when bleeding was encountered. Because bleeding occurred in one third of the women ⬍35 years old, the attributable risk was high (15.6%). Nonetheless, the fact that two thirds of the patients who bled did not lose their pregnancies should provide some reassurance. The high frequency of bleeding reported by women in this cohort (31.2%) may be related to the composition of the study group (infertility patients) who, after pregnancy is diagnosed, tend to be sensitive to even the smallest amount of staining or spotting. Although bleeding often resolves before an office visit, we were able to track the occurrence of bleeding accurately because our institutional protocol requires that all patient encounters, including telephone calls, be recorded in our electronic patient database. An additional purpose of this study was to compare rates of SAB in a cohort of previously infertile women who conceived spontaneously or after various treatments for infertility. We found that whether pregnancy occurred spontaneously or through treatment, women who conceived after experiencing infertility had a similar risk of losing the pregnancy before 12 weeks. This risk did not vary by the diagnosis for which the women were treated or by the specific treatment given. Again, these findings should be reassuring to infertile women and their physicians. The rate of SAB detected in this study was similar to that reported in studies of women with normal fertility and subfertility (2, 8). As in FERTILITY & STERILITY威
other studies on SAB (9), the risk in our study population was directly related to maternal age. Although this study is limited by its retrospective design, its inclusion of a relatively large number of patients as well as an untreated control group makes the conclusions more compelling. There is also some internal consistency to the data, in that SAB rates were significantly higher in older women and conceptions in older women were achieved independently of treatment more often than in their younger infertile counterparts (Table 1). The makeup of the study population in terms of age, gravidity, and distribution of diagnoses appears typical of women seeking treatment for infertility. The finding that bleeding was predictive of miscarriage only in the younger group of women is likely a result of the nature of SAB in older women. Although almost all women with SAB will bleed eventually if untreated, older women with histories of infertility are more likely to have suspicious US examinations at their first prenatal visit. This occurs because of the relatively more common occurrence in this population of chromosomal anomalies as a cause of miscarriage. Earlier diagnosis of SAB and termination of the pregnancy in these women may have reduced the frequency with which bleeding was encountered. In conclusion, infertile women who conceive after fertility therapy are not at increased risk of SAB compared with those who conceive spontaneously. Overall, their risk of SAB appears to be similar to that of infertile women who conceive without treatment as well as to that of women with normal fertility. There appears to be no association between previous diagnosis or treatment and the occurrence of SAB in previously infertile women. Although bleeding does not herald SAB in most of these women, it does in many, and its occurrence is associated with approximately a twofold risk of SAB compared with women who do not experience bleeding.
Acknowledgment: The authors thank Howard Minkoff, M.D., for his thoughtful review and comments on this article.
References 1. Laferia JJ. Spontaneous abortion. Clin Obstet Gynecol 1986;13:105–14. 2. Wilcox AJ, Baird DD, Weinberg CR. Time of implantation of the conceptus and loss of pregnancy. N Engl J Med 1999;340:1796 –9. 3. Henry JB. Clinical diagnosis and management by laboratory methods. 19th ed. Philadelphia: WB Saunders, 1996:483–91. 4. World Health Organization. Laboratory manual for the examination of human semen and semen-cervical mucus interaction. 3rd ed. New York: Cambridge University Press, 1993:43– 4. 5. Kruger TF, Menkveld R, Stander FSH, et al. Sperm morphological features as a prognostic factor in in vitro fertilization. Fertil Steril 1986;46:1118 –23. 6. Hughes EG, Fedorkow DM, Daya S, Sagle MA, Van de Koppel P, Collins JA. The routine use of gonadotropin-releasing hormone agonists prior to in vitro fertilization and gamete intrafallopian transfer: a metaanalysis of randomized controlled trials. Fertil Steril 1992;58:888 –96. 7. Society for Assisted Reproductive Technology, The American Society
507
for Reproductive Medicine. Assisted reproductive technology in the United States: 1996 results generated from the American Society for Reproductive Medicine/Society for Assisted Reproductive Technology Registry. Fertil Steril 1999;71:798 – 807. 8. Guzick DS, Carson SA, Coutifaris C, et al. Efficacy of superovulation
508
Pezeshki et al.
Spontaneous abortion after infertility
and intrauterine insemination in the treatment of infertility. N Engl J Med 1999;340:177– 83. 9. Abdalla HI, Billett A, Kan AKS, Baig S, Wren M, Korea L, et al. Obstetric outcome in 232 ovum donation pregnancies. Br J Obstet Gynaecol 1998;105:332–7.
Vol. 74, No. 3, September 2000