Early Pregnancy Loss

INTRODUCTION

For both the doctor and the patient, early pregnancy loss is a frustrating and heart-wrenching situation. Unfortunately, early pregnancy loss is the most common complication of human gestation, occurring in at least 75% of all women trying to conceive. Most of these losses are unrecognized and occur before or with the expected next menses. Of those that are recognized, 15-20% are spontaneous abortions (SABs) or ectopic pregnancies diagnosed after clinical recognition of pregnancy. Approximately 5% of couples trying to conceive have 2 consecutive miscarriages, and approximately 1% of couples have 3 or more consecutive losses.

Early pregnancy loss is defined as the termination of pregnancy before 20 weeks' gestation or below a fetal weight of 500 grams. Most investigators agree that both ectopic and molar pregnancies should not be included in the definition. The following are more specific definitions:

• Chemical pregnancy loss - Loss of a biochemically evident pregnancy
• Early pregnancy loss/abortion of the first trimester - Loss of a pregnancy recognized histologically or by ultrasound
• Spontaneous abortion - Pregnancy loss before 20 weeks' gestation based on last menstrual period
• Stillbirth - Pregnancy loss after 20 weeks' gestation, and neonatal loss is the death of liveborn fetus
• Habitual/recurrent abortion - Three or more consecutive abortions

INCIDENCE

Early pregnancy loss occurs at a rate of 114 cases per hour. Most studies quote a spontaneous miscarriage rate of 10-15%. However, the true early pregnancy loss rate is closer to 50% because of the high number of “chemical pregnancies?that are not recognized in the 2-4 weeks after conception. Most of these pregnancy failures are due to gamete failure (eg, sperm, oocyte dysfunction). In a 1988 classic study by Wilcox et al, 221 women were followed during 707 total menstrual cycles. A total of 198 pregnancies were achieved, of which 43 (22%) were lost before the onset of menses, and another 20 (10%) were clinically recognized losses.

The recurrent miscarriage rate is 3-5%. The chance for a subsequent abortion increases with each successive abortion. Data from various studies indicate that after 1 SAB, the couple has approximately the baseline risk of having another one (15%). However, if 2 SABs occur, the subsequent risk increases to approximately 25%. Several studies have estimated that the risk of pregnancy loss after 3 successive abortions is 30-45%. Controversy exists as to how many (ie, 2 or 3) pregnancy losses a woman should experience before considering a diagnostic evaluation. One could argue that the diagnostic evaluation should take place after 2 losses because the diagnostic yield after 2 versus 3 miscarriages is identical. In addition, it appears that an increase in the prevalence of aneuploidy is noted when couples with 2 miscarriages are compared with normal controls.

ETIOLOGY

The etiology of early pregnancy loss is varied and often controversial. More than one etiologic factor is often present. The most common causes of recurrent miscarriages are as follows:

• Genetic balanced parental translocation
• Mendelian
• Multifactorial
• Other
• Robertsonian
• Reciprocal
• Uterine congenital
• Müllerian anomaly
• Diethylstilbestrol-linked
• Acquired defects
• Iatrogenic
• Uterine septum
• Hemiuterus
• Double uterus
• Incompetent cervix
• Leiomyomas
• Asherman syndrome
• Immune autoimmune
• Alloimmune
• Humoral mediated
• Cellular immunity mediated
• Endocrine luteal phase deficiency
• Other endocrine factors
• Antithyroid antibodies
• High luteinizing hormone synthesis
• Infection
• Hematologic
• Environmental

The gestational age at the time of the SAB can provide clues about the cause. For example, antiphospholipid syndrome (APS) and cervical incompetence losses tend to occur after the first trimester. It has been suggested that when a woman with a history of retroperitoneal lymphadenectomy (RPL) carries a pregnancy past mid gestation, her chances of having pregnancy complications, such as prematurity and low-birth weight infants, may be somewhat higher than controls.

GENETIC CAUSES

Single miscarriage

Most spontaneous miscarriages are caused by an abnormal karyotype of the embryo. At least 50% of all first trimester SABs are cytogenetically abnormal. However, this figure does not include abnormalities caused by single gene disorders (eg, Mendelian disorders) or mutations at several loci (eg, polygenic or multifactorial disorders) that are not detected by the evaluation of karyotypes. The highest rate of cytogenetically abnormal concepti occurs earliest in gestation, with rates declining after the embryonic period (>30 mm crown-rump length).

Cytogenetically abnormal embryos usually are aneuploid because of sporadic events, such as meiotic nondisjunction or polyploid from fertilization abnormalities. One half of the cytogenetically abnormal abortuses in the first trimester involve autosomal trisomy. Triploidy is found in 16% of abortions, with fertilization of a normal haploid ovum by 2 sperm (dispermy) as the primary pathogenic mechanism. Trisomies arise de novo because of meiotic nondisjunction during gametogenesis in parents with a normal karyotype. For most trisomies, maternal meiosis I errors have been implicated.

The incidence of trisomies increases with age. Trisomy 16, which accounts for 30% of all trisomies, is the most common. All chromosome trisomies have been reported in abortuses except for trisomy 1. Interestingly, trisomy 1 has been reported in embryos obtained with in vitro fertilization. This logically suggests that trisomy 1 is most likely lethal at the preimplantation stage. Autosomal monosomies are rarely, if ever, observed. In contrast, monosomy X (Turner syndrome) frequently is seen and is the most common chromosomal abnormality observed in SABs. Turner syndrome accounts for 20-25% of cytogenetically abnormal abortuses. Approximately one third of Down syndrome (trisomy 21) fetuses survive to term.

The standard of care is to offer genetic amniocentesis for all pregnant women of advanced maternal age, which is defined as women older than 35 years. The risk of a woman having an aneuploid fetus is 1 per 80 when she is older than 35 years, which is far greater than the inherent risk of fetal loss after amniocentesis, which is 1 per 200.

Other abnormalities include those related to abnormal fertilization (eg, tetraploidy, triploidy). These abnormalities are not compatible with life. Tetraploidy occurs in approximately 8% of chromosomally abnormal abortions, resulting from failure of a very early cleavage division in an otherwise normal diploid zygote.

Structural chromosomal problems are a third category of abnormalities. Structural rearrangements occur in approximately 3% of cytogenetically abnormal abortuses. It is thought that structural chromosomal abnormalities are inherited more commonly from the mother. Structural chromosomal problems that are found in men seem to lead to lower sperm concentrations, male infertility, and, thus, a lesser chance of pregnancy and miscarriage. The exception to this is the couple undergoing assisted reproductive technologies, whereby selected sperm can be injected into oocytes to force fertilization using potentially genetically abnormal sperm.

The incidence of translocations increases with the number of abortions. In addition, women were more frequently the carriers. Slightly more than one half of the unbalanced rearrangements result from the abnormal segregation of Robertsonian translocations (ie, fusion of 2 acrocentric chromosomes at the centromere). Approximately one half of all unbalanced translocations arise de novo during gametogenesis. Of the familial translocations, about two thirds are derived maternally, and one third are paternal in origin. In 2-3% of couples who have had 2 or more spontaneous miscarriages, one partner has a balanced translocation. In addition, the rate is slightly higher (1.7-4.6%) in couples with a history of both recurrent abortion and anomalous or stillborn infants. Other structural rearrangements, such as inversions or ring chromosomes, are more rare.

The final group is gene abnormalities. It is thought that there may be certain mutations of genes involved with implantation that may predispose a patient to either infertility or even miscarriage. An example of a single gene disorder associated with recurrent pregnancy loss is myotonic dystrophy, an autosomal dominant disorder with high penetrance.

This disorder is a progressive, degenerative neuromuscular disease with an extremely variable phenotype. The cause of the abortion is unknown but may be related to abnormal gene interactions combined with disordered uterine function. Other presumed autosomal dominant disorders that affect the fetus and are associated with pregnancy loss include lethal skeletal dysplasias, such as thanatophoric dysplasia and type II osteogenesis imperfecta. In these cases, the parents are phenotypically normal because the mutation presumably occurs during gametogenesis. The rare case of recurrence in these families is presumed to be due to gonadal mosaicism in the ovary or the testes.

Maternal disease associated with increased fetal wastage includes connective tissue disorders, such as Marfan syndrome, Ehlers-Danlos syndrome, homocystinuria, and pseudoxanthoma elasticum. Women with sickle cell anemia are at increased risk for fetal loss, possibly because of placental bed microinfarcts. Other hematologic abnormalities associated with recurrent pregnancy loss include dysfibrinogenemia, factor XIII deficiency, and congenital hypofibrinogenemia and afibrinogenemia.

Recurrent miscarriage may result from 2 different chromosomal abnormalities, a structural abnormality derived from one of the parents or the recurrence of a numerical abnormality, which usually is not inherited. Studies have analyzed whether the presence of karyotypic abnormalities in one abortus was predictive of a similar abnormality in the next pregnancy. Warburton et al found that when the effects of maternal age are taken into account, there is no increase in the risk of trisomy in a second abortion following a trisomic abortion. There is no increased risk of trisomy in a second abortion following a previous abortion with another karyotype. Conversely, there is a significant increase in the risk of a nontrisomic abnormal karyotype after a previous abortion with a similar karyotype.

Genetic counseling after a miscarriage

The study by Warburton et al indicates that a routine karyotype analysis after one miscarriage is not cost-effective or prognostic. However, after 2 miscarriages, analysis of the abortuses is useful, a theory supported in a 1990 study by Drugan. This study sampled 305 women with 2 or more miscarriages, with either chorionic villus sampling or amniocentesis. Drugan found a higher risk for fetal aneuploidy in couples with recurrent miscarriages. The risk was 1.6, similar to the aneuploidy risk in a woman older than 40 years. In a woman with a prior trisomic livebirth, an approximately 1% increased risk exists for subsequent trisomic birth. The recurrence risk probably is limited to only those trisomies compatible with life, such as trisomies 13, 16, 18, and 21 or to parental trisomy mosaicism. These couples should always have their karyotypes evaluated.

If the karyotype of both parents is normal and the next pregnancy ends as a spontaneous miscarriage, cytogenetic studies of the abortus should be preformed to provide prognostic information and to assess the efficacy of treatments for other potential causes of miscarriage. Because abnormalities caused by single gene mutations (eg, Mendelian disorders) or mutations at several loci (eg, polygenic or multifactorial disorders) are not detected by karyotype analysis, molecular techniques are being used more frequently to complement standard cytogenetics.

The analysis of very small structural deletions and rearrangements that are not detectable with standard cytogenetic techniques can be identified with specialized methods, such as fluorescence in situ hybridization (FISH). However, if a parental chromosome abnormality is found, then this should be the starting point for familial testing. If an inherited abnormality is found, then proper family counseling is recommended. If an increased risk for future pregnancies is identified, then each alternative should be discussed, including foregoing any attempts at further conception, adoption, trying to conceive again with early prenatal testing, sperm or oocyte donation, or preimplantation diagnosis (PGD).

PGD entails in vitro fertilization (IVF), removal of a blastomere from the developing embryo for genetic analysis, and then implantation into the uterus only those embryos that are genetically normal. However, difficulties remain with this procedure. The delivery rate for IVF is only 35% even if normal embryos are transferred. The cost is $10,000-15,000 per treatment cycle. Of note, it is possible that all the cells within the early embryo are not genetically similar and that a so-called normal embryo actually may be abnormal.

In reciprocal translocations, the chromosomes involved in the translocation together with their normal homologues, form 1 quadrivalent instead of 2 bivalents. Alternate segregation results in 1 gamete receiving both normal chromosomes and the other gamete receiving both translocation chromosomes. The children created from these gametes have normal and carrier karyotypes. Adjacent segregation results in unbalanced distribution of the chromosomes involved in the translocation, leading to partial trisomy for 1 chromosome and partial monosomy for the other chromosome. The severity of the phenotype depends on the chromosomes involved and the positions of their breakpoints. The risk is higher if the translocation is carried by the female partner.

Chromosomes arising from Robertsonian translocations are composed of 2 more or less complete acrocentric chromosomes. Abnormal meiotic segregation results in either complete trisomies or monosomies. Viable trisomies have been observed for chromosomes 13, 16, and 21.

With inversions, a loop is formed during meiotic pairing. Crossing over within a loop yields 2 normal chromosomes and 2 chromosomes showing duplication/deficiency. If the inversion is paracentric, involving the centromere, 1 of the abnormal chromosomes is dicentric and the other is acentric. Both are incompatible with life. In pericentric inversion, not involving the centromere, abnormalities may lead to offspring with congenital abnormalities.

AUTOIMMUNE CAUSES

An association exists between recurrent pregnancy loss and autoimmune diseases. More specifically, systemic erythematosus (SLE) has been implicated with an increased miscarriage rate for many years, and pregnancy loss has been associated since 1954 with antiphospholipid antibodies (APLAs). APLAs are specific antibodies that put women with SLE at an increased risk for miscarriage. The median rate of spontaneous miscarriage among patients with SLE is 10%, compared with the general population. However, the median rate of late pregnancy loss (ie, second and third trimesters) of 8% is considerably higher than that observed in the healthy general population. Therefore, excess pregnancy loss in patients with SLE seems to be isolated 75% of the time to fetal death in the second and third trimesters. Most, if not all, fetal deaths in these women are associated with the presence of APLAs.

Three other factors that are predictive include disease before conception, onset of SLE during pregnancy, and underlying renal disease. APLAs are antibodies that bind to negatively charged phospholipids. At least 3 APLAs are well known as having important clinical relevance, including lupus anticoagulant (LAC), anticardiolipin antibodies (aCLs) and the biologically false-positive serologic test for syphilis (FP-STS). Although found in otherwise healthy people, APLAs are implicated in several other obstetric conditions, including preeclampsia, intrauterine growth restriction, abnormal fetal heart rate tracings, preterm deliveries, and pregnancy wastage.

Other medical conditions associated with APLAs are arterial and venous thrombosis, autoimmune thrombocytopenia, autoimmune hemolytic anemia, livedo reticularis, chorea, pulmonary hypertension, and chronic leg ulcers. The diagnosis of APS, also known as LAC and Hugh syndrome, is made when both clinical events (obstetric or medical) are present and when specific levels of APLAs are present.

The International Consensus Workshop in 1998 proposed preliminary classification criteria for APS that includes the following clinical criteria:

• Vascular thrombosis
• One or more episodes of arterial, venous, or small-vessel thrombosis in any tissue or organ that is confirmed by imaging or Doppler studies or histopathology.
• For histopathologic confirmation, thrombosis should be present without significant evidence of inflammation on the vessel wall.
• Pregnancy morbidity
• Three or more unexplained consecutive miscarriages with anatomic, genetic, or hormonal causes excluded
• One or more unexplained deaths of a morphologically normal fetus at or after the 10th week of gestation with fetal morphology documented by ultrasound or by direct examination of the fetus
• One or more premature births of a morphologically normal neonate at or before the 34th week of gestation associated with severe preeclampsia or severe placental insufficiency
• Laboratory criteria
• aCL: Immunoglobulin G (IgG) and/or immunoglobulin M (IgM) isotype is present in medium or high titer on 2 or more occasions, 6 or more weeks apart.
• The aCL is measured by a standardized enzyme-linked immunosorbent assay (ELISA) for beta2-glycoprotein I-dependent aCL LAC.
• The abnormality is present in plasma on 2 or more occasions, 6 or more weeks apart, and detected according to the guidelines of the Scientific and Standardization Committee on lupus.
• Anticoagulants/phospholipid-dependent antibodies
• Demonstration of a prolonged phospholipid-dependent coagulation screening test (eg, activated partial thromboplastin time [aPTT]) kaolin clotting time, dilute Russell viper venom time, dilute prothrombin time [PT], and Textarin time.
• Failure to correct the prolonged screening test by mixing with normal platelet-poor plasma
• Shortening or correction of the prolonged d-screening test by the addition of excess phospholipid
• Exclusion of other coagulopathies as clinically indicated (eg, factor VIII inhibitor) and heparin

The demonstration of these antibodies can be made with either ELISA or a coagulation test positive for LAC. Therefore, the presence of the antibodies alone in the absence of other clinical symptoms does not define the syndrome. APS is associated with systemic autoimmune diseases and with various other connective tissue disorders; 7-30% of women with SLE have APLAs. APLAs are found in less than 2% of apparently healthy pregnant females, in less than 20% of apparently healthy females with recurrent fetal loss, and in greater than 33% of women with SLE.

APLAs have not been shown definitively to be a risk factor for pregnancy loss because most studies to date implicating pregnancy loss with LACs and aCLs are case series. In contrast to recurrent pregnancy loss, isolated miscarriages have not been associated with APLA. Extensive placental infarction has been noted in some studies of patients with APLAs and recurrent pregnancy loss; however, an underlying pathophysiologic mechanism is still being sought for fetal loss and thrombosis. Placentas obtained from patients with APLAs show accelerated atherosis and vascular occlusion. In animal models, LAC and thrombocytopenia frequently accompany pregnancy loss.

Anticoagulant treatments, such as aspirin, heparin, intravenous immunoglobulin interleukin 3 (IL-3), and ciprofloxacin, have been shown to be effective therapies. Ciprofloxacin is thought to work through IL-3. IL-3 control animals have very large placentas and fetuses; therefore, it is hypothesized that IL-3 acts as a placental growth hormone and can make up for damaged placental tissue.

It is thought that the thrombosis of APLA is caused by an increase in the thromboxane to prostacyclin ratio, leading to thrombosis. Thromboxane production by the placenta could lead to thrombosis at the uteroplacental interface and rationalizes the use of low-dose aspirin therapy during pregnancies in women with APLA. Other studies have proposed that the thrombosis is secondary to enhanced platelet aggregation, decreased activation of protein C, increased expression of tissue factor, and enhanced platelet-activating factor synthesis.

Clinically, pregnancy loss in patients with APS frequently occurs after 10 weeks' gestation (as opposed to the majority of SABs that tend to occur earlier). As mentioned previously, it is thought that placental insufficiency is the causal agent.

The combination of higher antibody titers and the IgG isotype has worse prognosis than does low titer and the IgM isotype. In addition, it does not make any difference whether the APLA is aCL, LAC, or anti–beta2-glycoprotein I. Treatment of patients with APS who have suffered prior fetal losses seems to improve pregnancy rates, but fetal loss may occur despite treatment. Preeclampsia, fetal distress, fetal growth impairment, and premature delivery are common.

Treatment data are difficult to analyze because most studies are not randomized and do not include appropriate controls. In addition, the serologic criteria for APLA, the clinical definitions of APS, and the dosing regimens for treatments vary greatly among studies. Overall, most studies report increases in pregnancy survival in women undergoing a treatment for APLA.

Treatment consists of subcutaneous heparin low-dose aspirin, prednisone, immunoglobulins, or combinations of the aforementioned drugs. Several well-controlled studies have shown that subcutaneous heparin (5,000 U bid) with low-dose aspirin (81 mg/d) increases fetal survival rates from 50% to 80% among women who have had at least 2 losses and who have unequivocally positive tests for APLA. Treatment started after pregnancy was confirmed and continued until the end of the pregnancy (just before delivery). This therapy (low-dose aspirin and subcutaneous heparin) has been shown to be equally effective and less toxic than prednisone (40 mg/d) plus aspirin.

Cowchock et al showed that women treated with prednisone plus aspirin had higher rates of hypertension, weight gain, diabetes, and premature rupture of membranes. Babies had higher incidences of amnionitis and prematurity. However, in women with secondary APS and SLE, consider the use of prednisone as a treatment modality. With long-term use of heparin, the physician must inform the patient about the risk of bone loss, bleeding, and thrombocytopenia. Loss of bone mineral density (as much as 10%) has been reported in women treated with heparin for an entire pregnancy; however, this loss of bone mineral density may be regained within 2 years.

In 1992, Branch et al reviewed 82 consecutive pregnancies in 54 women with APS who were treated during the pregnancy with the following: (1) prednisone and low-dose aspirin; (2) heparin and low-dose aspirin; (3) prednisone, heparin, and low-dose aspirin; and (4) other combinations of these medications or immunoglobulins. The overall neonatal survival rate was 73%, excluding SABs, but treatment failures (fetal and neonatal) occurred in all treatment groups. Patients with successfully treated pregnancies had fewer previous fetal deaths than those with unsuccessfully treated pregnancies. In addition, no significant differences occurred in outcome among the 4 treatment groups.

Intravenous immunoglobulin (IVIG) therapy has been shown to be effective with not only a decrease in fetal loss but also a decrease in preeclampsia and fetal growth restriction. However, to date, no properly controlled studies have been conducted. Intravenous treatment with immunoglobulins is very expensive and should not be used as first-line therapy until further data on its efficacy are available.

Antinuclear antibodies (ANAs) have been associated with recurrent pregnancy loss, even in patients without evidence of overt autoimmune disease. Elevated ANA titers (usually >1:40) were found in 7-53% of women with recurrent fetal loss compared with 0-8% of pregnant and nonpregnant controls. However, other studies have refuted the aforementioned study. In addition, the success rate in future pregnancies in women with increased ANA titers and previous pregnancy losses was similar to that in women with undetectable ANA. In most published studies, the ANA titers in women with recurrent miscarriages were only mildly elevated. These mild elevations are nonspecific and common in the general population (even in women with no history of pregnancy loss). Therefore, it is difficult to extrapolate this as a cause. Further studies are needed to prove or disprove ANA as a causal agent in recurrent miscarriages and is not recommended as part of an evaluation of recurrent miscarriage.

Unlike ANA, antithyroid antibodies are independent markers for an increased risk of miscarriage. Stagnaro-Green et al observed 500 consecutive women for the presence of thyroid specific autoantibodies (specifically, antithyroglobulin and/or antithyroid peroxidase) in the first trimester of pregnancy. Women with a positive test for thyroid autoantibodies had a 17% rate of pregnancy loss compared with 8.4% for women without evidence of thyroid autoantibodies. None of the women with thyroid autoantibodies had clinically evident thyroid disease, and the increase in pregnancy loss was not due to changes in thyroid hormone levels or APLA. The pathophysiology involved in this phenomenon is unclear and probably represents a generalized autoimmune defect rather than a thyroid induced abnormality. However, the available data thus far do not support the use of thyroid autoantibody testing in women with recurrent pregnancy loss.

ANATOMIC CAUSES

Anatomic uterine defects are known to cause obstetric complications including recurrent pregnancy loss, preterm labor and delivery, and malpresentation. A uterine malformation should be considered in any woman experiencing recurrent pregnancy loss. However, not all women with abnormal uteri have obstetric complications. Why some women have difficulties while others do not is not known. The incidence of uterine anomalies is estimated to be between 1 per 200 and 1 per 600, depending on the method used for diagnosis. When manual exploration is preformed at the time of delivery, uterine anomalies are found in approximately 3% of women. However, in women with a history of pregnancy loss, uterine abnormalities are present in approximately 27%.

The most common uterine defects include septate, bicornuate, and didelphic uteri. The unicornuate uterus is the least common.

When analyzing the different studies about uterine anomalies and their role in miscarriage, great disparity exists, caused by the lack of strict criteria in cataloging various types of uterine malformations. The largest study to date was conducted in 1996 by Acien. One hundred and seventy of the women had a prior pregnancy, of which only 32 (18.8%) had normal deliveries at term, compared with 70% in the control group and 0% in the hypoplastic uterus group. Sixty-two other patients with uterine malformations (36.5%) had abnormal deliveries (eg, premature, breech). Again, 59 (35%) had only reproductive losses (1 or more), including early abortions in 19.4%, late abortions in 4.7%, and immature deliveries in 10.6%. These loss rates were significantly higher than the control group.

The highest rate of only reproductive losses was that of bicornuate uterus (47%), whereas the lowest was that of unicornuate uterus (17%). Women with unicornuate and didelphys had the highest rate of abnormal deliveries, while women with uterine septums had a 26% risk of reproductive loss.

Unicornuate uterus results from arrested or defective development of one of the müllerian ducts. The uterus often is deviated markedly to either side and is shaped like a banana. Although the studies on uterine anomalies are scarce, nearly all agree that the pregnancy outcome in women with unicornuate uteri is poor. Fetal survival rates for women with the unicornuate uterus average approximately 40%. Approximately 45% of pregnancies are lost within the first 2 trimesters.

In addition to fetal loss, prematurity occurs in 20% of all pregnancies and is thought to be due to the reduced capacity that does not allow for proper gestational growth and the possibility of cervical incompetence. Malpresentation and fetal growth restriction are other complications faced by women with unicornuate uteri. Fetal growth restriction is thought to be secondary to vascular anomalies in the distribution of the uterine artery.

The imaging studies of choice include hysterosalpingography and high-resolution ultrasonography. A bananalike cavity with a single fallopian tube is the most common finding. Approximately 65% of unicornuate uteri contain a rudimentary horn. Approximately one third contain endometrial tissue, and one half of these communicate with the main uterine cavity. A rudimentary horn without an endometrial cavity is present in approximately 34% of cases, and it is thought that pregnancies with rudimentary horns had a greater chance of delivering at or near full term. Fedele reported that the incidence of rudimentary horn pregnancy is approximately 12.5% because of the transmigration of sperm. Excision of the rudimentary horn is advised because of the high risk of morbidity secondary to intraperitoneal hemorrhage. However, removal of a solid rudimentary horn is not necessary because little evidence suggests that there is an adverse affect on pregnancy outcome in a solid (nonfunctioning) horn.

Prophylactic cervical cerclage should be considered in patients with a unicornuate uterus. Some authors support expectant management in these patients with serial assessments of cervical lengths by digital and ultrasonographic examinations. Uterine didelphys results from failed fusion of the paired müllerian ducts. A uterus didelphys consists of 2 endometrial cavities, each with a uterine cervix that is fused in the area of the lower uterine segment. A longitudinal vaginal septum running between the 2 cervices is present in most cases.

INFECTIOUS CAUSES

The theory that microbial infections can cause miscarriage has been present in the literature as early as 1917, when DeForest et al observed recurrent abortions in humans exposed to farm animals with brucellosis. Although infection has been cited as a cause of pregnancy loss, few studies exist, and results are inconsistent. Numerous organisms have been implicated in the sporadic cause of miscarriage, but common microbial causes generally have not been confirmed. In fact, infection is viewed as a rare cause of recurrent miscarriage. Those organisms implicated with SAB include the following:

• Bacteria
• Listeria monocytogenes
• Chlamydia trachomatis
• Ureaplasma urealyticum
• Mycoplasma hominis
• Bacterial vaginosis
• Viruses
• Cytomegalovirus
• Rubella
• Herpes simplex virus (HSV)
• Human immunodeficiency virus (HIV)
• Parvovirus
• Parasites
• Toxoplasmosis gondii
• Plasmodium falciparum
• Spirochetes - Treponema pallidum

Different theories have been postulated to explain exactly how an infectious agent leads to miscarriage and include the following:

• Toxic metabolic byproducts, endotoxin, exotoxin or cytokines could have a direct effect on the uterus or the fetoplacental unit.
• Fetal infection could cause fetal death or severe malformation incompatible with fetal viability.
• Placental infection could result in placental insufficiency, with subsequent fetal death.
• Chronic infection of the endometrium from ascending spread of organisms (eg, Mycoplasma hominis, Chlamydia, Ureaplasma urealyticum, HSV) from the lower genital tract could interfere with implantation.
• Amnionitis in the first trimester could play a role similar to chorioamnionitis in the third trimester, resulting in preterm labor (various common gram-positive and gram-negative bacteria, Listeria monocytogenes).
• Induction of a genetically and anatomically altered embryo or fetus by viral infection (eg, rubella, parvovirus B19, cytomegalovirus, coxsackievirus B, varicella-zoster, chronic cytomegalovirus [CMV], HSV, syphilis, Lyme disease) during early gestation.

Any patient undergoing infertility workup should be treated for any recognized vaginitis or cervicitis. In addition, chronic genital infection may be the most obvious initial manifestation of a general health problem. Chronic vulvovaginitis is known to be associated with both diabetes, other endocrinopathies, and, possibly, lupus erythematosus. In addition, elimination of both gonorrhea and chlamydia should take place before infertility workup (eg, hysterosalpingograms) for fear of spreading the infection to the upper genital tract. A recent review failed to find sufficient evidence for the notion that any type of infection could be identified as a causal factor for recurrent miscarriage. Most patients with a history of recurrent miscarriage will not benefit from an extensive infection workup. Exposure to a microbe that can establish chronic infection that can spread to the placenta in a patient who is immunocompromised is probably the most obvious risk situation in recurrent abortions.

Specific pathogens are as follows:

• Gonorrhea: This is associated with premature rupture of membranes and chorioamnionitis.
• Chlamydia trachomatis: No association exists between a prior chlamydial infection and fetal loss in women with recurrent abortion. However, neonatal conjunctivitis and pneumonia are known sequelae.
• Women who are in high-risk groups are the only patients who should be screened. Serologic studies have suggested an association between C trachomatis and recurrent abortion, and routine C trachomatis screening has been recommended for all infertility patients. However, microbiologic testing for endocervical chlamydial infection during pregnancy has failed to confirm the association with recurrent abortion. In 1992, Witkin and Ledger reviewed the relationship between high-titer IgG antibodies to C trachomatis and recurrent SAB. They found that high-titer IgG antibodies to C trachomatis were associated with recurrent SABs. They proposed the mechanism to be reactivation of a latent chlamydial infection, endometrial damage from past chlamydial infection, or an immune response to an epitope shared by a chlamydial and a fetal antigen.
• Bacterial vaginosis: This is associated with preterm labor, intrauterine growth retardation chorioamnionitis, and late miscarriage; however, no studies have investigated its role in women with recurrent miscarriages. Most women are screened at their first prenatal visit and more frequently if there is a history of late miscarriages or preterm delivery.
• Genital mycoplasma: Mycoplasma hominis and Ureaplasma are isolated from the vagina in as many as 70% of pregnant women. Although these organisms are found more frequently in women with recurrent miscarriages, their elimination has not improved subsequent pregnancy outcome. Therefore, it is not recommended to screen for mycoplasma and ureaplasma in the typical patient with a history of recurrent miscarriage.
• L monocytogenes: Typically, this produces asymptomatic colonization of the maternal lower genital tract, although symptomatic maternal listeriosis characterized by bacteremia and influenzalike symptoms may occur. Symptomatic listerial infection typically is described as a complication of the third trimester, resulting from ingestion of unpasteurized milk or cheese. Asymptomatic genital Listeria colonization may result in high perinatal mortality and morbidity if the organism is spread to the fetus during labor and delivery. However, no evidence suggests that Listeria plays a role in patients with a history of recurrent pregnancy loss. Chronic genital infection with L monocytogenes, which could lead to recurrent abortion, would occur in patients who were immunocompromised, and, because of its low prevalence, screening for listeria during pregnancy or in routine cases of recurrent miscarriage is not recommended.
• Treponema pallidum: This is known to cause stillbirth and abortion in the second trimester. The timing of death is probably associated with the maturation of the fetal immune system at the 20th week of gestation. However, it is unlikely that syphilis contributes significantly to the general problem of recurrent miscarriage.
• Borrelia burgdorferi: Lyme disease can result in stillbirth and fetal infection. Obtain serology if the patient relates a history suspicious for infection with Lyme disease; however, it is unlikely that Lyme disease contributes significantly to the general problem of recurrent abortion.
• CMV: This is associated with random miscarriage but not recurrent miscarriage. A large study conducted by Stagno et al observed 3712 pregnant patients and documented only 21 per 3712 cases of primary maternal CMV during pregnancy. Only 11 of the 21 showed neonatal infection, and SABs did not occur in this group.
• HSV: Primary HSV has been associated with SAB, and chronic HSV is a possible cause of recurrent abortion (especially in a patient who is immunocompromised). The risk for in utero HSV transmission from chronic maternal disease is low (about 0-3% of pregnancies). Therefore, recurrent abortion secondary to chronic HSV is extremely low in the general population and does not warrant routine screening in patients with recurrent pregnancy loss.
• P falciparum: Malaria during pregnancy is associated with SAB, stillbirth, low birth weight, and prematurity. Screening is only important in those women where the disease is endemic or in symptomatic patients who have traveled to endemic countries.
• Toxoplasmosis: Primary infection with toxoplasmosis can lead to miscarriage and stillbirth. However, if the infection develops during the first trimester, the risk is less than 5%. In addition, repeated infections in subsequent pregnancies are unlikely, unless chronic infection develops in patients who are immunocompromised.
• HIV: Studies have failed to show an increase in miscarriage rates for asymptomatic patients with HIV.

ENVIRONMENTAL, HORMONAL, AND THROMBOPHILIC CAUSES

Environmental causes: Such causes of human malformation account for approximately 10% of malformations, and less than 1% of all human malformations are related to prescription drug exposure, chemicals, or radiation. It is important to recognize these preventable exposures. For example, the relationship between exposure to trace concentrations of waste anesthetic gases in the operating room and the possible development of adverse health effects has been a concern for many years. However, the studies that showed an increase incidence of miscarriage and congenital anomalies had many flaws.

Maternal exposure to tobacco and its effect on reproductive outcomes has been the subject of many studies. Cigarette smoke contains hundreds of toxic compounds. Nicotine is thought to have vasoactive actions, and thought to reduce placental and fetal circulation. Carbon monoxide depletes both fetal and maternal oxygen supplies, and lead is a known neurotoxin. Maternal smoking appears to only slightly increase the risk of SABs.

Endocrine factors: Ovulation, implantation, and the early stages of pregnancy are dependent on an integral maternal endocrine regulatory system. Historically, most attention has been directed on maternal systemic endocrine disorders, luteal phase abnormalities, and hormonal events that follow conception, particularly progesterone levels in early pregnancy.

Diabetes mellitus: Women with diabetes mellitus who have good metabolic control are no more likely to miscarry than are women without diabetes. However, women with diabetes with high glycosylated A1c levels in the first trimester are at a significantly higher risk of both miscarriage and fetal malformation. Insulin-dependent women with inadequate glucose control have a 2- to 3-fold higher rate of SAB than the general population of women. There is no value to screening for occult diabetes in asymptomatic women unless a random glucose is elevated. For a patient with an unexplained loss in the second trimester or with clinical signs of diabetes mellitus, investigation is needed.

Thyroid dysfunction: No direct evidence suggests that thyroid disease is associated with recurrent miscarriages. However, the presence of antithyroid antibodies is and may represent a generalized autoimmune abnormality rather than a specific thyroid dysfunction. Screening for thyroid disease is not useful unless the patient is symptomatic.

Low progesterone levels: Progesterone is the principal factor responsible for the conversion of a proliferative to a secretory endometrium, rendering the endometrium receptive for embryo implantation. In 1929, Allen and Corner published their classic results on physiologic properties of the corpus luteum, and, since then, it has been assumed that low progesterone levels are associated with miscarriage. Luteal support remains critical until about the seventh week of gestation at which time the trophoblast has acquired enough steroidogenic ability to support the pregnancy. In patients where the corpus luteum is removed before the seventh week, miscarriage results. If progesterone is given to these patients, the pregnancy is salvaged. Recent developments with RU486 (an antiprogestin) have shown that these can effectively terminate a pregnancy up to 56 days from the last menstrual period.

Luteal phase defects (LPD): In 1943, Jones first discussed the concept of insufficient luteal progesterone resulting in either infertility or early pregnancy loss. This disorder was characterized by inadequate endometrial maturation resulting from a qualitative or quantitative disorder in corpus luteal function, which has been reported in 23-60% of women with recurrent miscarriage. Unfortunately, no reliable method is available to diagnose this disorder, and controversy exists due to the inconsistencies in the method of diagnosis.

Methods used to diagnose luteal phase defects include basal body temperature records, progesterone concentrations, and histologic dating of endometrial biopsy specimens. The criterion standard has been an endometrial biopsy taken in the luteal phase. However, significant interobservational and intraobservational discrepancies exist using the standard histologic criteria. This criteria uses development of stromal and glandular cells to determine how many days after ovulation the patient was at the time of the biopsy. A delay of more than 2 days in maturation compared to where the patient exactly was based on her luteinizing hormone (LH) surge is defined as LPD.

Most studies use the patient’s subsequent menses as a reference point, assuming the patient has a normal 28-day cycle. This accounts for many of the discrepancies in the literature. Consequently, as many as 31% of normally fertile women have a luteal phase defect by serial endometrial biopsies. In one of the few prospective studies evaluating women with 3 or more consecutive miscarriages, LPD was believed to be the cause in 17% of them. The biopsy samples were dated accurately by the pathologist using LH assays to pinpoint the time of ovulation. In this study, luteal phase serum progesterone levels were normal in the women with LPD. Luteal phase deficiency is most likely the result of an abnormal response of the endometrium to progesterone than a subnormal production of progesterone by the corpus luteum. This is evident in the fact that as many as 50% of women with histologic defined LPD have normal serum progesterone levels.

In treating LPD, it is important to realize that postimplantation failure or a very early nonviable pregnancy is associated with low serum progesterone levels. Only 1 randomized trial has shown treatment with progesterone supplementation to have a beneficial effect on pregnancy outcome. Most studies have opposite results, failing to show that any type of support (eg, progesterone, HCG) to have beneficial results. Therefore, the physician must be selective in deciding who should be screened for LPD. One approach is to screen patients with either a history of recurrent miscarriages or recurrent failures at infertility therapy. In addition, it would be most accurate if the histology is reviewed by the same pathologist, and the day of ovulation should be based on LH levels as opposed to subsequent menses. The dose of progesterone should be adequate enough to stimulate luteinization of the endometrium with the fewest adverse effects.

Endocrine modulation of decidual immunity: The transformation of endometrium to decidua affects all of the cell types present in the uterine mucosa. These morphologic and functional changes facilitate implantation but also help in controlling trophoblast migration and in preventing over invasion in maternal tissue. Attention focuses on the interaction between extravillous trophoblast and the leukocyte populations infiltrating the uterine mucosa. Most of these cells are large granular lymphocytes (LGL) and macrophages; very few T and B cells are present. The LGL population is unusual, staining strongly for natural killer (NK) cell marker CD56, but the cells do not express the CD16 and CD3 NK markers. NK cells with this distinct phenotype are found in high numbers, primarily in the progesterone–primed endometrium of the uterus. The CD56 cells are low in the proliferative phase endometrium, increase midluteal, and peak in the late secretory phase, suggesting that recruitment of LGLs is under hormonal control.

Progesterone is essential since LGLs are not found before menarche or after menopause or in conditions associated with unopposed estrogen, such as endometrial hyperplasia or carcinoma. In women who have been oophorectomized, LGLs appear only after treatment with both estrogen and progesterone. The increase in NK cells at the implantation site in the first trimester suggest its role in pregnancy maintenance. They preferentially kill target cells with little or no HLA expression. Extravillous trophoblast (which expresses modified forms of 1 HLA) is resistant to lysis by decidual NK cells under most circumstances, allowing invasion needed for normal placentation. These CD56 cells probably differentiate in utero from precursor cells, since serum levels are negligible.

The only cytokine that has been able to induce proliferation of these cells is interleukin 2 (IL-2). IL-2 also transforms NK cells into lymphokine-activated killer (LAK) cells, which are capable of lysing first trimester trophoblast cells in vitro. As expected, IL-2 has not been found in vivo at uterine implantation sites; otherwise, stimulation of decidual NK cells would cause widespread destruction of trophoblast. Trophoblast HLA expression is increased by interferon, a phenomenon that may offer protection from LAK cell lysis. Therefore, an equilibrium exists between the level of HLA expression on trophoblast and the amount of lymphokine activation of NK cells, leading to the concept of fine regulation of trophoblast invasion.

Thrombophilic defects: Many cases of recurrent miscarriages are characterized by defective placentation and the presence of microthrombi in the placental vasculature. Various components of the coagulation and fibrinolytic pathways are important in embryonic implantation, trophoblast invasion, and placentin. Because the association between APLA and recurrent miscarriage are now firmly established, interest has been fueled regarding the possible role of other hemostatic defects in pregnancy loss.

Pregnancy is a hypercoagulable state because of the following: (1) an increase in the levels of procoagulant factors, (2) a decrease in the levels of naturally occurring anticoagulants, and (3) a decrease in fibrinolysis. The levels of factors VII, VIII, X, and fibrinogen increase during a normal pregnancy, as early as 12 weeks' gestation.

This increase in factors is not balanced by an increase in anticoagulants, antithrombin III and proteins C and S. In fact, protein S decreases by 40-50%. Antithrombin III and protein C remain constant. Fibrinolytic activity also is altered, with levels of plasminogen activator inhibitors 1 and 2 increasing progressively during pregnancy. PAI-1 is produced by endothelial cells and inhibits release of plasminogen activator. PAI-2 is produced by the trophoblast and helps regulate placental growth. An increase occurs in platelet activation, which contributes to the prothrombic state of pregnancy reflected by an increase in platelet production of thromboxane and decreased platelet sensitivity to the antiaggregatory effects of prostacyclin. The hemostatic changes in pregnancy favor coagulation.

Urokinase plasminogen (uPA) activator is active around the time of implantation. It triggers the localized production of plasmin, which catalyzes the destruction of the extracellular matrix, facilitating implantation. uPA also is found in the maternal venous sinuses and, therefore, plays a role in maintaining the patency of these channels. uPA receptors also are expressed on first trimester human trophoblast cells, primarily those that are not actively invasive, which serves to facilitate generation of plasmin at the interface of these cells with maternal plasma, thereby limiting deposition of fibrin within the intervillous spaces.

PAI-1 and PAI-2 have been localized to invasive trophoblast. Therefore, trophoblast implantation and invasion seemingly are regulated to some extent by the balance between plasminogen activators and inactivators. Indeed, defective trophoblast invasion of the spiral arteries has been a common finding in placental bed biopsies obtained from women who miscarry and from those patients with preeclampsia or intrauterine growth restriction.

In the normal placenta, important components of the hemostatic, fibrinolytic, and protein C anticoagulant (factor V Leiden) pathways are present and responsible for maintenance of hemostasis. Abnormal gestations are associated with an abnormal distribution of fibrin, and the production of certain factors (eg, cytokines) may convert a thrombo-resistant endothelium to one that is more thrombogenic. In support of this theory, fibrin deposition has been seen in chorionic villi that make allogenic contact with maternal tissue, which contains many factors and products of the hemostatic pathway. Endothelial cells in these areas appear to be deficient in the thrombin-thrombomodulin anticoagulant pathway and, therefore, are prone to clot formation. Normal villi have this pathway.

Compelling evidence suggests that women with a history of recurrent miscarriage are in a procoagulant state even when not pregnant. A large study of 116 nonpregnant women with recurrent miscarriages, who tested negative for LAC and aCLs, reported that 64% of these women had at least 1 abnormal fibrinolysis-related test, most commonly a high PAI-1. No abnormal defects were found in the control group, which consisted of 90 fertile women with no history of miscarriage. In 1994, in another study by Patrassi and colleagues, 67% of patients, regardless of whether they were aCL positive, had a defect in their fibrinolytic pathway.

Evidence also suggests that just before a miscarriage, defects in hemostatic variables are present. Tulpalla and coworkers revealed that in women with a history of recurrent miscarriages, an abundance of thromboxane production at 4-6 weeks' gestation and a decrease in prostacyclin production at 8-11 weeks exists, compared with women without such history. This shift in the thromboxane-to-prostacyclin ratio can lead to vasospasms and platelet aggregation, causing microthrombi and placental necrosis. A decrease seemingly occurs in the level of protein C and fibrinopeptide A just prior to miscarriage, suggesting an activation of the coagulation cascade.

Activated protein C (APC) resistance: Resistance to the anticoagulant effects of APC is inherited as an autosomal dominant trait and is the most important cause of thrombosis and familial thrombophilia. In more than 90% of cases, it is due to single-point mutation (glutamine for arginine) at nucleotide position 1691 in the factor V gene. This mutated gene is known as factor V Leiden. APC cleaves and inactivates coagulation factors Va and VIIIa in the presence of cofactor protein S. The mutated factor V is resistant to inactivation by APC, resulting in increased thrombin production and a hypercoagulable state. Its prevalence is 3-5%. In those with a prior venous thrombosis, the prevalence is as high as 40%.

In 1995, Rai and colleagues evaluated 120 women with a history of recurrent miscarriages. All women were negative for thrombosis history, LAC, and aCL. The prevalence of APC resistance was higher in those women who had had a second trimester miscarriage (20%) versus those with a first trimester loss (5.7%).

In normal pregnancies, a natural decrease in APC resistance occurs; however, it is those women with APC resistance prior to pregnancy who tend to have an even further drop in resistance. The best way to detect APC resistance is both a coagulation based assay and the DNA test to detect the actual mutation. They complement each other since one is a genetic test and one is a functional test.

Coagulation inhibitors: Little data exist evaluating deficiencies of antithrombin III, protein S, or protein C, and pregnancy loss.

Specific coagulation factor deficiencies: The deficiency is factor XII (Hageman) and is associated with both systemic and placental thrombosis and has been reported to be associated with recurrent miscarriage in up to 22% of patients evaluated. Again, data are limited.

Abnormal homocysteine metabolism: Homocysteine is an amino acid formed during the conversion of methionine to cysteine. Hyperhomocystinemia, which may be congenital or acquired, is associated with thrombosis and premature vascular disease. This condition also is associated with pregnancy loss; in one study, 21% of women were observed with recurrent pregnancy loss. The gene for the inherited form is transmitted in an autosomal recessive form. The most common acquired form is folate deficiency. In these patients, folic acid replacement achieves normal homocysteine levels within a few days.

Therapy for coagulation disorders: Low-dose aspirin (60-150 mg/d) irreversibly inhibits the enzyme cyclooxygenase in platelets and macrophages. This leads to a shift in arachidonic acid metabolism toward the lipoxygenase pathway, resulting in inhibition of thromboxane synthesis without affecting prostacyclin production. It also stimulates leukotrienes, which, in turn, stimulates production of IL-3 that is essential for implantation and placental growth. Heparin inhibits blood coagulation by 2 mechanisms. At conventional doses, it increases the inhibitory action of antithrombin III on activated coagulation factors XII, XI, IX, X, and thrombin. At high doses, it catalyzes the inactivation of thrombin by heparin cofactor 2. Heparin does not cross the placenta; therefore, no risk to the fetus is present.

The primary adverse effects are osteopenia if therapy is prolonged (usually therapeutic doses) and thrombocytopenia, which usually occurs within a few weeks of starting heparin (even low prophylactic doses). Osteopenia is reversed upon discontinuation of heparin, and platelet levels should be checked routinely.