Recurrent pregnancy loss: Definition and etiology

INTRODUCTION — Couples with pregnancy loss need empathy and understanding. Early pregnancy loss, especially when recurrent, is an emotionally traumatic experience, similar to that associated with stillbirth or neonatal death.

Recurrent pregnancy loss (RPL) is one of the most frustrating and difficult areas in reproductive medicine because the etiology is often unknown and there are few evidence-based diagnostic and treatment strategies. Studies on the etiology, evaluation, and management of RPL are often flawed. Common methodologic weaknesses include failure to adhere to generally accepted criteria for RPL, ascertainment bias, improper selection of controls, uneven monitoring of cohorts, no exclusion of aneuploid fetuses, lack of stratification for important factors such as number of previous losses, premature termination of study after interim analysis, and excessive postrandomization patient withdrawal [1].

In this topic, when discussing study results, we will use the terms "woman/en" or "patient(s)" as they are used in the studies presented. We encourage the reader to consider the specific counseling and treatment needs of transgender and gender diverse individuals.

DEFINITION — The definition of RPL varies, which makes studying the phenomenon, and determining which couples to counsel or treat, more challenging. As examples, varying definitions have included:

●Two or more failed clinical pregnancies as documented by ultrasonography or histopathologic examination [2].

●Three consecutive pregnancy losses, which are not required to be intrauterine [3,4].

In our practice, we start investigating after two failed clinical pregnancies, including biochemical pregnancies for women undergoing in vitro fertilization. The rationale for including biochemical pregnancies and nonvisualized pregnancies (ectopic pregnancies) in the definition of RPL comes from a retrospective cohort study of 587 women who had three or more consecutive pregnancy losses before 12 weeks gestation [5]. Nonvisualized pregnancy losses (biochemical pregnancy losses and/or pregnancies of unknown location) had the same negative impact on future live birth as an intrauterine pregnancy losses.

In response to these varied definitions, the European Society of Human Reproduction and Embryology released a 2017 consensus statement proposing that RPL describes two or more pregnancy losses, diagnosed by either serum or urine human chorionic gonadotropin [6]. This diagnosis includes biochemical pregnancies and treated pregnancies of unknown location but does not include confirmed ectopic or molar pregnancies.

RPL can be further divided into primary or secondary processes [6,7]. Primary RPL refers to pregnancy loss in women who have never carried to viability (24 weeks gestation or beyond). By contrast, secondary RPL refers to pregnancy loss in a woman who has had a previous live birth. The prognosis for successful pregnancy is better with secondary RPL [7].

There is no specific term for describing women who have had multiple spontaneous miscarriages interspersed with normal pregnancies (ie, nonconsecutive pregnancy losses).

INCIDENCE — Approximately 15 percent of pregnant women experience sporadic loss of a clinically recognized pregnancy. Just 2 percent of pregnant women experience two consecutive pregnancy losses and only 0.4 to 1 percent have three consecutive pregnancy losses [8]. The observed frequency of three consecutive pregnancy losses is slightly higher than that expected by chance alone: if the frequency of miscarriage is 15 percent, then the probability of three consecutive miscarriages would be (0.15)(0.15)(0.15), or 0.003 (0.3 percent). By comparison, the observed frequency of two consecutive pregnancy losses is that which would be expected by chance alone (0.15)(0.15) = 0.0225 (2.25 percent).

These mathematical relationships constitute one basis for defining the disorder, but do not take into account the effect of maternal age or the gestational age at which the miscarriage occurred (see 'Definition' above). The prevalence of miscarriage is higher with increasing maternal age (table 1) and at very early gestational ages (eg, at less than 6 weeks of gestation the risk of miscarriage is 22 to 57 percent versus 15 percent at 6 to 10 weeks and 2 to 3 percent after 10 weeks [9]).

RISK FACTORS AND ETIOLOGY — Couples who have had a pregnancy loss have two major concerns: the cause and the risk of recurrence. Although RPL is an important problem in women's health, there are many unsolved questions regarding etiology, evaluation, and management. Unfortunately, the cause of RPL can be determined in only 50 percent of patients [10]. General etiological categories of RPL include anatomic, immunological, genetic, endocrine, infectious, thrombophilic, and environmental factors.

Previous pregnancy loss — In a first pregnancy, the risk of miscarriage is 11 to 13 percent [11]. After one miscarriage, this rate rises slightly to 14 to 21 percent. After two or three miscarriages, the rate is 24 to 29 percent and 31 to 33 percent, respectively. However, several factors influence these rates:

●The cause of the pregnancy loss impacts the miscarriage risk. As an example, carriers of a 22:22 translocation will almost always miscarry, whereas women with 13:14 translocation have a 25 percent risk.

●The gestational age of prior pregnancy loss and interpregnancy interval (IPI) may impact the risk of repeat pregnancy loss. In a secondary analysis of 677 women with a prior pregnancy loss (mean gestational age 8.6±2 weeks), the IPI after a first-trimester loss did not impact the live birth rate in a subsequent pregnancy [12]. However, in a study of women with pregnancy loss between 14 and 19 weeks of gestation (that adjusted for maternal age), an IPI ≤3 months was associated with an increased rate of recurrent loss compared with an IPI of >9 to 12 months (22 versus 11 percent) [13].

●Advancing maternal age is associated with a higher rate of pregnancy loss of both normal and abnormal conceptuses [14]. This probably reflects poor oocyte quality in this age group. The increased risk of miscarriage with advancing age was illustrated in a review of over 1 million pregnancies with known outcome and with admission to a hospital [14]. The overall rate of spontaneous pregnancy loss was 11 percent. The approximate rates of clinically recognized miscarriage according to maternal age were: age 20 to 30 years (9 to 17 percent), age 35 (20 percent), age 40 (40 percent), and age 45 (80 percent). (See "Effects of advanced maternal age on pregnancy".)

●Increasing parity is also associated with an increased rate of miscarriage [15]. This may be partially explained by the correlation between increasing maternal age and greater parity and by "reproductive compensation" (ie, a behavior pattern in which couples make repetitive attempts to bear children after miscarriage) [16].

●Previous pregnancy outcome can positively or negatively affect the risk in the next pregnancy. It is generally agreed that the risk of miscarriage increases with each successive pregnancy loss [17], whereas some studies report a pregnancy ending in live birth reduces the risk of miscarriage in the subsequent gestation [17].

●Gestational age at the time of pregnancy loss should be considered in determining both the etiology and risk of recurrence. RPL typically occurs at a similar gestational age in consecutive pregnancies [18]. The recurrence risk increases as gestational age at the time of loss increases.

Uterine factors — Acquired and congenital uterine abnormalities are responsible for 10 to 50 percent of RPL [19].

Anomalies — Congenital uterine anomalies (figure 1 and figure 2) are present in 10 to 15 percent of women with RPL versus 7 percent of all women [20]. Pregnancy loss may be related to impaired uterine distention or abnormal implantation due to decreased vascularity in a septum, increased inflammation, or reduction in sensitivity to steroid hormones [21].

The septate uterus is the uterine anomaly associated with the poorest reproductive outcome and the most common uterine abnormality associated with RPL [22]. The fetal survival rate in women with untreated septate uterus is 6 to 28 percent and the miscarriage rate is greater than 60 percent. The longer the septum, the worse the prognosis [23]. However, data on the pregnancy effects of reproductive tract anomalies have been derived from relatively small observational studies that are subject to ascertainment biases.

The mechanism by which a septate uterus causes pregnancy loss is not clearly understood, but poor blood supply to the septum leading to poor implantation is one possibility [22]. (See "Congenital uterine anomalies: Clinical manifestations and diagnosis".)

Leiomyoma — Submucous leiomyomas that protrude into the endometrial cavity can impede normal implantation as a result of their position, poor endometrial receptivity of the decidua overlying the myoma, or degeneration with increasing cytokine production (figure 3) [24]. An association between pregnancy loss and intramural or subserous myomas is less clear, having been demonstrated in some, but not all, studies.

Adenomyosis — In a meta-analysis of 15 studies evaluating the effect of adenomyosis on fertility and in vitro fertilization clinical outcomes, the pregnancy loss rate was higher in individuals with adenomyosis than in those without (odds ratio 2.2, 95% CI 1.53-3.15) [25]. Inflammatory changes in the endometrium associated with adenomyosis and/or endometriosis have been postulated as contributing to RPL [26]. Whether one should perform laparoscopy to diagnose endometriosis in women with RPL is not known.

Endometrial polyps — There have been no data showing a relationship between endometrial polyps and RPL. Infertility and endometrial polyps is discussed in detail separately. (See "Endometrial polyps".)

Intrauterine adhesions — Intrauterine adhesions or synechiae lead to pregnancy loss because there is insufficient endometrium to support fetoplacental growth. The main cause of intrauterine adhesions is curettage for pregnancy complications. Curettage of the endometrium, especially within the first four weeks postpartum, traumatizes the basalis layer, which then heals via formation of granulation tissue. Granulating tissue on opposing surfaces of the uterus can eventually fuse and produce tissue bridges, ranging from filmy adhesions composed of endometrial tissue to dense adhesions consisting entirely of connective tissue. The resulting adherence of the uterine walls, with partial or complete obliteration of the uterine cavity, leads to menstrual irregularities (hypomenorrhea, amenorrhea), cyclic pelvic pain, infertility, and RPL. (See "Intrauterine adhesions: Clinical manifestation and diagnosis".)

Cervical insufficiency — Cervical insufficiency is a cause of recurrent midtrimester, but not early, pregnancy loss. (See "Cervical insufficiency".)

Defective endometrial receptivity — Estrogen and progesterone prepare the endometrium for pregnancy [27]. Normal endometrial receptivity allows embryo attachment, implantation, invasion, and development of the placenta. These processes are likely to be disturbed when endometrial receptivity is defective, resulting in unexplained infertility and RPL. Causes of defective endometrial receptivity and biomarkers for evaluation of endometrial receptivity are under investigation [28]. One example is a primary endometrial receptor defect [29]. Additional research suggests that RPL is associated with uterine stem cell deficiency and enhanced cellular senescence, which then results in abnormal endometrial preparation for pregnancy that leads to recurrent loss [30]. Clinical tests are not yet available.

Immunologic factors — Each step in the establishment of normal pregnancy has been implicated as a possible site of immune-mediated reproductive failure. Both autoimmune and alloimmune mechanisms have been proposed. Since the mechanisms that allow a mother to tolerate her semi-allogeneic conceptus are not well defined, it is difficult to assess the role of aberrant immunologic factors in reproductive failure [31]. (See "Immunology of the maternal-fetal interface".)

Antiphospholipid syndrome — Several autoimmune diseases have been linked to poor obstetric outcome, but antiphospholipid syndrome (APS) is the only immune condition in which pregnancy loss is a diagnostic criterion for the disease. Five to 15 percent of patients with RPL may have APS [32]. (See "Antiphospholipid syndrome: Obstetric implications and management in pregnancy".)

Other immunological factors — Allogeneic factors may cause RPL by a mechanism similar to that of graft rejection in transplant recipients. If the blastocyst is developmentally normal and intact, the embryo should be entirely protected by trophoblast cells. However, in some pregnancies, the blastocyst is genetically deformed and not fully intact. As a result, paternally-derived antigens are exposed to the maternal immune system, which leads to a graft rejection response. A secondary immune response would be expected to cause early rejection in cases of recurrent spontaneous abortion. (See "Immunology of the maternal-fetal interface".)

Alternatively, some mothers with RPL may lack essential components of the networks that provide immunological protection to the embryos, such as appropriate expression of complement regulatory proteins (eg, mannose-binding lectin, apoptosis-inducing TNF superfamily members, macrophage inhibitory cytokine 1, Th1/Th2/Th3-type cytokines, and HLA-DR, HLA-G or HLA-E) [33-35]. Dysregulation of the normal immune mechanism probably operates at the maternal-fetal interface and may involve increased activity of uterine natural killer (uNK) cells, which appear to regulate placental and trophoblast growth, local immunomodulation, and control of trophoblast invasion [36,37]. These and other aspects of immunologically-mediated infertility are not well defined.

A rare immunologically mediated cause of RPL is alloimmunization to blood group antigen P [38]. Cytotoxic IgM or IgG3 antibody directed against the P and PK antigens has been associated with miscarriage in more than 50 percent of affected pregnancies, as well as fetal growth restriction in ongoing pregnancies; the authors speculated that early plasmapheresis could be therapeutic.

In a large cohort study of 1,513,544 Danish women including 7667 women with multiples sclerosis, pregnancy loss was not significantly related to this autoimmune disorder [39].

Endocrine factors — Endocrine factors may account for 15 to 60 percent of RPL.

Diabetes mellitus — Although rare, poorly controlled diabetes mellitus is associated with early (and late) pregnancy loss. Several studies have linked high hemoglobin A1C values early in pregnancy (particularly values above 8 percent) to increased frequencies of miscarriage and congenital malformations (figure 4) [40]. The increased risk in poorly controlled diabetic women is believed to be secondary to hyperglycemia, maternal vascular disease, and possibly immunologic factors. There is no increased risk of miscarriage in women with well-controlled diabetes mellitus [41]. (See "Pregestational (preexisting) diabetes: Preconception counseling, evaluation, and management".)

Insulin resistance, as seen in women with polycystic ovaries, may also be a factor in pregnancy loss. (See 'Polycystic ovary syndrome' below.)

Polycystic ovary syndrome — The miscarriage rate in women with polycystic ovary syndrome (PCOS) may be as high as 20 to 40 percent, which is higher than the baseline rate in the general obstetric population (10 to 20 percent) [42]. The mechanism for excess pregnancy loss in these patients is unknown, but may be related to elevated serum luteinizing hormone (LH) levels, high testosterone and androstenedione concentrations (which may adversely affect the endometrium), or insulin resistance.

The sex hormone abnormalities in women with PCOS may cause premature or delayed ovulation, poor endometrial receptivity, and disturbances in synthesis/secretion/action of prostaglandins and ovarian growth factors/cytokines. In one study, a menstrual cycle longer than 34 days, which is common in women with PCOS, was the most important predictor for a RPL [43].

Women with RPL have a higher prevalence of insulin resistance than fertile controls, whether or not PCOS is present [44]. The role of insulin resistance was suggested in a study showing that treatment of insulin resistance with metformin in women with PCOS was associated with a reduction in pregnancy loss [45]. However, another well-designed trial showed no benefit of metformin compared with clomiphene in reducing the incidence of miscarriage, thereby casting doubt about the role of insulin resistance on pregnancy loss [46]. (See "Clinical manifestations of polycystic ovary syndrome in adults".)

Thyroid antibodies and disease — Some studies have reported an increased rate of fetal loss in women with high serum thyroid antibody concentrations (thyroid peroxidase or thyroglobulin), including those who are euthyroid [47]. Thyroid autoimmunity has also been related to unexplained infertility and implantation failure [48]. Direct evidence of causality, however, is still lacking [49] and conflicting data have also been reported [50]. However, in a randomized trial of 92 euthyroid women with recurrent pregnancy loss and positive for TPO-Ab, levothyroxine treatment did not increase the live birth rate [51].

Poorly controlled thyroid disease (hypo- or hyper-thyroidism) is associated with infertility and pregnancy loss. Excess thyroid hormone increases the risk of miscarriage independent of maternal metabolic dysfunction [52]. (see "Overview of thyroid disease and pregnancy").

Hyperprolactinemia — Normal circulating levels of prolactin may play an important role in maintaining early pregnancy. This hypothesis was supported by a study of 64 hyperprolactinemic women with RPL randomly assigned to bromocriptine therapy or no bromocriptine [53]. Treatment to lower prolactin concentration was associated with a higher rate of successful pregnancy (86 versus 52 percent). Prolactin levels during early pregnancy were significantly greater in women who miscarried. (See "Management of hyperprolactinemia".)

Luteal phase defect — A 2015 Committee Opinion released by the American Society for Reproductive Medicine concluded that "there is no reproducible, pathophysiologically relevant, and clinically practical standard to diagnose luteal phase deficiency and distinguish fertile from infertile women" [54]. The committee concluded that luteal phase deficiency as an independent cause of infertility has not been proven. We do not perform luteal phase testing.

Progesterone is required for successful implantation and maintenance of pregnancy; therefore, disorders related to impaired progesterone production or action are likely to affect pregnancy success. A defect in corpus luteum function (ie, luteal phase defect) has been hypothesized to be a potential cause of impaired progesterone production and resultant infertility or pregnancy failure. However, it is controversial as to whether such a defect really exists and is related to miscarriage, and there is no consensus on the best method of diagnosis or treatment [54].

Several studies have shown that luteal phase defect diagnosed by endometrial biopsy is not predictive of infertility; luteal phase defect has been observed in as many as 25 percent of sequential endometrial biopsies in fertile women. Serum progesterone concentrations are not predictive of pregnancy outcome [55], and there is no high-quality evidence to support the use of exogenous progesterone supplementation to prevent early miscarriage. (See "Female infertility: Evaluation", section on 'Endometrial biopsy'.)

Abnormal luteal-phase progesterone production may occur as the result of medical conditions such as elevated prolactin or abnormal thyroid function; women suspected to have one of these disorders are evaluated and treated for the underlying condition [54]. (See 'Thyroid antibodies and disease' above and 'Hyperprolactinemia' above.)

Genetic factors — Abnormalities of chromosome number or structure are the most common cause of sporadic early pregnancy loss, accounting for at least 50 percent of such losses in multiple studies [56]. A significant proportion of RPL may also be associated with structural or numerical chromosomal abnormalities (eg, aneuploidy, mosaicism, translocation, inversion, deletion, fragile sites) [57]. Single-gene, X-linked, or polygenic/multifactorial disorders can also result in sporadic or recurrent miscarriage.

There appears to be an increased risk of RPL in first degree relatives of women with unexplained RPL [58]. This may be related to shared HLA types, coagulation defects, immune dysfunction, or other undefined heritable factors.

Aneuploidy — The risk of aneuploidy increases as the number of previous miscarriages increases [59]. However, a 2020 meta-analysis reported that individuals with sporadic pregnancy loss and RPL had similar percentages of anomalies detected in pregnancy tissue (46 percent) [60]. The relationship between the karyotype of the abortus and risk of RPL requires further study to better define which abnormalities are likely to be recurrent.

In some series, having one chromosomally abnormal spontaneous abortion appeared to increase the risk of a subsequent loss associated with a chromosomal abnormality [61,62]. As an example, one study reported that the frequency of an abnormal karyotype in a second abortus after a first aneuploid or euploid abortus was 70 and 20 percent, respectively [61]. Another study involving preimplantation genetic testing reported 532 of 764 embryos (70 percent) were abnormal in couples with RPL versus 97 of 215 embryos (45 percent) among controls [62]. In addition, an increased rate of aneuploid embryos has been demonstrated in young women with previous aneuploid conceptions [63].

However, these findings were not validated in other series [56,64,65]. As an example, in one study, there was no increased risk of chromosome abnormality in the next pregnancy if the prior abortus had a trisomy that was always lethal in utero or when the parental karyotypes were normal [64]. The aneuploid losses in this study, and others, may have been associated, in part, with the older age of the mothers, rather than solely a nonrandom risk in a predisposed couple.

Chromosomal rearrangements — Three to 5 percent of couples with RPL have a major chromosomal rearrangement (versus 0.7 percent of the general population); usually a balanced translocation (60 percent reciprocal, 40 percent Robertsonian) or, less commonly, an inversion [32,66]. One or both partners may harbor lethal genes in a heterozygous or balanced combination that does not affect them, but causes pregnancy loss when inherited by the embryo in a homozygous or unbalanced state. Balanced translocations are more common in the female than the male and more likely to result in pregnancy loss if the translocation is of maternal origin. As an example, an X-linked dominant condition may not cause disease or may result in mild disease in a heterozygous female, but can be lethal in males because of the lack of a normal compensatory gene [32]. (See "Chromosomal translocations, deletions, and inversions".)

The likelihood that RPL is related to parental karyotypic abnormality appears to be higher when one or more of the following characteristics are present: young maternal age at second miscarriage, a history of three or more miscarriages, or a history of two or more miscarriages in a sibling or the parents of either partner [66]. A family history of stillbirth or an abnormal liveborn also appear to be risk factors [24]. However, it is important to note that an abnormal parental karyotype, although present, may not be the cause of the RPL; therefore, a complete evaluation of RPL is indicated [67].

Other — A meta-analysis of studies evaluating whether there is an association between cytokine polymorphisms and RPL concluded there was either no association or no more than a mild effect [68]. Progesterone receptor gene polymorphisms, as well as other gene polymorphisms, may play a role in RPL [69]. This is an active area of investigation.

Skewed X-inactivation is associated with increasing maternal age, but does not appear to be associated with RPL [70].

Thrombophilia and fibrinolytic factors — Thrombosis of spiral arteries and the intervillous space on the maternal side of the placenta can impair adequate placental perfusion. The resulting abnormalities of the uteroplacental circulation may cause late fetal loss, intrauterine growth restriction, placental abruption, or preeclampsia. A relationship to early pregnancy loss is less clear and may be restricted to specific thrombophilic defects that have not been completely defined, or the presence of multiple defects. There is a large and contradictory literature on the association between maternal inherited thrombophilia and RPL occurring in the first trimester; this is discussed in detail separately. (See "Inherited thrombophilias in pregnancy".)

●Impact of common thrombophilias – A meta-analysis of 89 observational studies reported the following thrombophilias were associated with increased risk of developing RPL when compared with individuals attempting pregnancy who did not have thrombophilia:

•Factor V Leiden G1691A mutation – Odds ratio (OR) 2.44, 95% CI 1.96-3.03

•Prothrombin G20210A mutation – OR 2.08, 95% CI 1.61-2.68

•Protein S deficiency – OR 3.45, 95% CI 1.15-10.35

Compared with the reference population, antithrombin and protein C deficiencies were not associated with increased risk of RPL. Sensitivity analysis supported these findings. Study imitations included reliance on observational data, inclusion of multiple definitions of RPL, and high heterogeneity of risk estimates [71].

●Factor XII deficiency – A systematic review of the association between fibrinolytic defects and RPL found a significant association for factor XII deficiency (odds ratio 18.11, 95% CI 5.52-59.4, five studies, 1096 women) [72]. There was no significant association with plasminogen activator inhibitor-I 4G/5G polymorphism, increased plasminogen activator inhibitor activity, factor XII C46T polymorphism, or factor XIII polymorphisms. Data on other fibrinolytic defects were sparse. (See "Overview of the causes of venous thrombosis", section on 'Factor XII deficiency'.)

Procoagulant microparticles can also contribute to the hypercoagulable state and thus may interfere with successful implantation and fetal growth. These microparticles were shown to be associated with early and late unexplained pregnancy loss in one pilot study [73].

●Role of placental factors – Placental mechanisms may play a role in RPL. Some data suggest that expression within the annexin A5 gene (annexin A5 is a placental anticoagulant protein) of the common haplotype M2 is associated with RPL [74,75]. As an example, one study reported that paternal M2 carriage and maternal M2 carriage play a role in RPL [74].

Environmental chemicals and stress — Although a frequent concern of patients, there is no high-quality evidence showing a relationship between RPL and occupational factors, stress, or low level exposure to most environmental chemicals [24,33]. Chemicals that have been associated with sporadic spontaneous pregnancy loss include anesthetic gases (nitrous oxide), arsenic, aniline dyes, benzene, ethylene oxide, formaldehyde, pesticides, lead, mercury, and cadmium [76]. (See "Overview of occupational and environmental risks to reproduction in females" and "Cigarette and tobacco products in pregnancy: Impact on pregnancy and the neonate".)

Other

Personal habits — The association between RPL and obesity, smoking, alcohol use, and caffeine consumption is unclear [1,77]. These factors may act in a dose-dependent fashion or synergistically to increase the rate of sporadic pregnancy loss. (See "Pregnancy loss (miscarriage): Terminology, risk factors, and etiology".)

Exercise does not appear to increase the rate of sporadic pregnancy loss or RPL [33].

Male factor — There is a trend toward repeated miscarriages in women whose male partner has abnormal sperm (eg, fewer than 4 percent normal forms, sperm chromosome aneuploidy) [78,79]. Paternal HLA sharing is not a risk factor for RPL, both maternal HLA allele type may play a role [1]. Advanced paternal age may be a risk factor for miscarriage. (See "Effect of advanced paternal age on fertility and pregnancy".)

Infection — Some infections, such as Listeria monocytogenes, Toxoplasma gondii, cytomegalovirus, and primary genital herpes, are known to cause sporadic pregnancy loss, but no infectious agent has been proven to cause RPL [33]. In one study that attempted to determine the role of infection in miscarriage, chorionic villi from aborted material were analyzed for the presence of Chlamydia trachomatis, Ureaplasma urealyticum, Mycoplasma hominis, cytomegalovirus, adeno-associated virus, and human papillomaviruses (HPV) [80]. The only infections detected were C. trachomatis DNA (detected in one sample) and HPV DNA (detected in eight samples). Karyotype was normal in 54 of 108 samples, abnormal in 38, and could not be determined in 16.

Diminished ovarian reserve — Diminished ovarian reserve appears associated with RPL but causation has not been established. A meta-analysis of 15 observational studies including over 3000 patients with RPL reported that more individuals with RPL had evidence of diminished ovarian reserve, specifically low anti-müllerian hormone (AMH) level and low antral follicle count (AFC) compared with those without RPL (AMH: odds ratio [OR] 2.77, 95% CI 1.41-5.46; AFC: OR 2.45, 95% CI 1.16-5.19, respectively) [81].

Celiac disease — Untreated celiac disease, even when subclinical, has been associated with pregnancy loss, menstrual disorders, and infertility. Treatment of celiac disease appears to prevent these problems [82,83]. (See "Epidemiology, pathogenesis, and clinical manifestations of celiac disease in adults".)

The relationship between untreated celiac disease and pregnancy loss is not well defined. Although some controlled studies have observed a significant association between poor reproductive performance, including miscarriage, and celiac disease [84], several others have not confirmed this association [85,86]. In addition, no study has proven that celiac disease causes repeated pregnancy loss. Nevertheless, due to the effects of the disease to health in general, and the possible positive effects of treatment of the disease, we suggest screening women with RPL for celiac disease.

EVALUATION AND TREATMENT — (See "Recurrent pregnancy loss: Evaluation" and "Recurrent pregnancy loss: Management".)

RESOURCES FOR PATIENTS — A Frequently Asked Questions page for patients is available online through the American College of Obstetricians and Gynecologists.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Recurrent pregnancy loss".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topic (see "Patient education: Repeat pregnancy loss (The Basics)")

PATIENT PERSPECTIVE TOPIC — Patient perspectives are provided for selected disorders to help clinicians better understand the patient experience and patient concerns. These narratives may offer insights into patient values and preferences not included in other UpToDate topics. (See "Patient perspective: Pregnancy loss".)

SUMMARY AND RECOMMENDATIONS

●Definition – A commonly used definition of recurrent pregnancy loss (RPL) is two or more pregnancy losses, diagnosed by either serum or urine human chorionic gonadotropin, although there are others. This includes pregnancies diagnosed by either serum or urine human chorionic gonadotrophin but excludes confirmed ectopic and molar pregnancies. It may be primary or secondary. (See 'Definition' above.)

●Incidence – 0.4 to 1 percent of women have three consecutive pregnancy losses. (See 'Incidence' above.)

●Common etiologies

•Chromosomal abnormalities – Three to 5 percent of couples with RPL have a major chromosomal rearrangement (versus 0.7 percent of the general population); usually a balanced translocation. Recurrent aneuploidy also appears to play a role in RPL. Chromosomal abnormalities are the most common cause of sporadic early pregnancy loss, accounting for at least 50 percent of such losses. (See 'Genetic factors' above.)

•Uterine abnormalities – Uterine abnormalities, both acquired (eg, submucosal leiomyoma, intrauterine adhesions) and congenital (eg, septum), have been reported to be responsible for 10 to 50 percent of RPL in small observational studies. (See 'Uterine factors' above.)

•Antiphospholipid antibody syndrome – Pregnancy loss is one of the diagnostic criteria for antiphospholipid syndrome. (See 'Other immunological factors' above.)

•Endocrine factors – Endocrine factors (eg, poorly controlled diabetes mellitus, elevated thyroid peroxidase or thyroglobulin antibodies, polycystic ovary syndrome, hyperprolactinemia) may account for some cases of RPL. (See 'Endocrine factors' above.)

●No causality – There is no high-quality evidence showing a relationship between RPL and occupational factors, stress, or low level exposure to most environmental chemicals. (See 'Environmental chemicals and stress' above.)

●Recurrence risk – RPL typically occurs at a similar gestational age in consecutive pregnancies. The recurrence risk increases as gestational age at the time of loss increases. (See 'Previous pregnancy loss' above.)