Early and recurrent pregnancy loss: Etiology, Diagnosis, Treatment

Spontaneous early pregnancy loss

Pregnancy loss is common, affecting 10% to 30% of clinically recognized pregnancies ( ). Extremes of age increase the risk of pregnancy loss; loss is more common in women younger than 18 years and older than 35 years and rises both with increasing parity and number of prior losses ( Fig. 16.1 ). The loss rate is believed to approach 80% in women age 45 and older ( ).

Nearly 80% of all pregnancy losses occur in the first trimester, and the incidence declines with advancing gestational age ( ) ( Fig. 16.2 ). The loss rate when a gestational sac is visualized by ultrasound is 11.5%; it falls to 6% to 8% if embryonic cardiac activity is observed at 6 weeks and falls to 2% to 3% if cardiac activity persists at 8 to 12 weeks. Compared with their younger counterparts, women older than age 34 experience a loss rate twice as high, even after visualization of a fetal heartbeat ( ). Women with bleeding in the first trimester and confirmed presence of fetal cardiac activity have a 15% risk of suffering a pregnancy loss ( ).

Evaluation of the medical literature should take into account that the majority of patients with a prior pregnancy loss will eventually have a successful pregnancy; a study of more than 53,000 parous women reported that 43% experienced one or more prior first trimester pregnancy losses ( ). Remembering that 90% of pregnancy losses are not recurrent, watchful waiting may prove the best, albeit emotionally the hardest thing to do.

Recurrent pregnancy loss

Although sporadic pregnancy loss is common, fewer than 5% of women will experience two consecutive pregnancy losses, and only 1% experience three or more ( ). Women with two prior losses and no live births experience a loss rate of 25%, which increases to 45% with three consecutive losses ( ) ( Table 16.1 ).

Historically, RPL has been defined as three or more consecutive pregnancy losses. The American Society of Reproductive Medicine (ASRM) defines RPL as two or more failed clinical pregnancies, documented by ultrasonography or histopathologic examination, that are not necessarily consecutive ( ). For epidemiologic studies, however, ASRM recommends that ideally three or more losses be used as the threshold for RPL. The European Society of Human Reproduction Embryology (ESHRE) released a consensus statement in 2014 proposing that RPL describes repeated pregnancy loss, including nonvisualized pregnancy losses, but did not recommend a number of losses required to be defined as recurrent loss ( ). The rationale to include nonvisualized pregnancies (biochemical pregnancy losses and/or pregnancies of unknown location) 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 of gestation ( ). Nonvisualized pregnancy losses had the same negative impact on future live birth as clinical pregnancy losses.

It is important to emphasize to patients with RPL that their overall prognosis is good, with more than 80% of women with RPL younger than 30 years and 60% to 70% of women with RPL ages 31 to 40 achieving a successful pregnancy within 5 years of their first visit to a physician ( ).

Anomalies of uterine development

Women with uterine anomalies have higher rates of first and second-trimester pregnancy losses compared with women with a normal uterus . Uterine anomalies are present in approximately 12.6% of patients (range, 1.8% to 37.6%) with RPL compared with 4.3% of the general population (range, 2.7% to 16.7%) ( ). The wide range reflects the differences in diagnostic and imaging techniques used. Several imaging modalities may be used to evaluate the uterus, including hysterosalpingogram, transvaginal ultrasonography, and sonohysterography. Magnetic resonance imaging (MRI) and three-dimensional transvaginal ultrasonography may help better characterize the anomaly.

The septate uterus is the most common uterine anomaly in the general population and is also associated with poorest reproductive outcomes. Composed of fibromuscular tissue, the septum has a decreased blood supply, which may lead to poor implantation, resulting in early pregnancy loss. Later in gestation, the septum may compromise available space for growth, leading to pregnancy loss, malpresentation, or preterm term birth. Other congenital anomalies associated with pregnancy loss include unicornuate, didelphys, and bicornuate uteruses ( ).

Hysteroscopic septum incision, or metroplasty, is the treatment of choice for women with a septate uterus and a history of pregnancy loss. A large meta-analysis found that hysteroscopic resection of the uterine septum increased live birth rates from 6.1% before surgery to 83.2% after metroplasty. However, many of the retrospective studies included in this meta-analysis used patients as their own controls, which limits the interpretation of these results. The benefits of hysteroscopic metroplasty have not yet been assessed by a prospective randomized trial. In cases of bicornuate uterus, metroplasty may be performed transabdominally or laparoscopically but has not been shown to increase the live birth rate. Thus metroplasty is not recommended for women with bicornuate uterus . Uterine reconstruction is not feasible for the unicornuate uterus.

Acquired uterine defects

Intrauterine adhesions (Asherman syndrome)

Intrauterine adhesions (IUAs), or synechiae, are scar tissue within the uterine cavity and are referred to as Asherman syndrome when associated with symptoms such as amenorrhea, infertility, or pregnancy loss. Pregnancy loss is thought to be a result of insufficient endometrium to support fetoplacental growth.

Adhesions are commonly caused by previous intrauterine surgery ( Table 16.4 ). Any intrauterine surgery may traumatize the basalis layer of the endometrium, which then heals by forming granulation tissue. Granulation tissue on opposing walls of the uterus can fuse, forming filmy adhesions composed of endometrial tissue and dense adhesions consisting entirely of connective tissue. These adhesions may result in partial or complete obliteration of the uterine cavity.

TABLE 16.4

Reported Incidence of Intrauterine Adhesion After Various Conditions and Procedures

Condition	Procedure	Incidence	Reference
G ravid
SAB	Suction D&C	15% 19%
First trimester TOP	Suction D&C	21%
Retained POC	Hysteroscopic resection Suction D&C	6% 13% 19% 30%	Smorgick, 2014
G ynecologic
Septum	Hysteroscopic septum resection (bipolar)	24%
Fibroids	Hysteroscopic myomectomy (bipolar) Abdominal myomectomy	8% 22%

D&C , Dilation and curettage; POC , product of conception; SAB , spontaneous abortion; TOP , termination of pregnancy.

Postpartum complications such as hemorrhage or retained placenta are major risk factors for adhesion formation. A greater number of intrauterine surgical procedures increases the risk of IUA formation. In a meta-analysis of 10 prospective studies, the risk of IUA development was 2.1-fold higher for two or more curettage procedures compared with one curettage procedure ( ). Adhesions can also form in a nongravid uterus as a result of procedures that injure the endometrium such as a myomectomy. In a study by Taskin and coworkers, hysteroscopic resection of multiple submucosal leiomyomas had the highest risk of IUA formation compared with other hysteroscopic procedures ( ).

Suspected IUAs may be seen in imaging studies such as ultrasound, saline infusion sonogram ( Fig. 16.4 ), and hysterosalpingogram ( Fig. 16.5 ). On hysterosalpingogram, the defects are typically irregular, with sharp contours and homogeneous opacity that persist in multiple views. The diagnosis is best confirmed and treated by hysteroscopy ( ).

The recommended treatment for symptomatic IUA is hysteroscopic lysis of the adhesions . The goal of surgery is to restore the size and shape of the uterine cavity. Unfortunately, adhesion recurrence rates have been reported to be as high as 30% to 66% after surgical lysis ( ). The prevention of adhesion recurrence is challenging, and no single method has shown superiority ( ). Suggested postlysis management approaches include placement of a mechanical barrier such as a pediatric balloon in the intrauterine cavity, postoperative treatment with estrogen and timed progestin therapy, and a repeat hysteroscopy ( ). In a meta-analysis of trials comparing multiple types of antiadhesion therapies with placebo or no intervention, treatment did not affect the subsequent live birth rate (odds ratio [OR], 0.99; 95% confidence interval [CI], 0.46 to 2.13; 3 studies, 150 women), although antiadhesion therapy was associated with fewer recurrent IUAs at the time of a second-look hysteroscopy (OR, 0.36; 95% CI, 0.20 to 0.64; 7 studies, 528 women) ( ). In women with severe IUAs, we suggest repeating an office hysteroscopy after the initial surgery . At that point, any new adhesions will be filmy and can be gently taken down. This process can then be repeated until the cavity has fully healed from surgery.

Gel barriers and regenerated cellulose adhesions barriers (Interceed) have also been investigated to reduce the recurrence of adhesions. Although studies have suggested that gel barriers (hyaluronic acid gel and polyethylene oxide-sodium carboxymethylcellulose) have a role in preventing IUA reformation ( ; ), a meta-analysis reported no definitive conclusion could be made because of the heterogeneity of study designs and lack of reproducibility of the study results ( ). Because the optimal approach to the prevention of IUAs is not yet known, the discussed treatment options are reasonable alternatives.

Cervical incompetence (cervical insufficiency)

Cervical incompetence is characterized by an asymptomatic dilation of the internal cervical os, leading to dilation of the cervical canal during the second trimester ( Figs. 16.6 and 16.7 ). Cervical abnormalities, either congenital or as a result of trauma, are risk factors for structural cervical weakness.

Cervical trauma may result from treatment of cervical intraepithelial neoplasia such as conization or loop electrosurgical excision procedures, rapid mechanical dilation before a gynecologic procedure, and obstetric lacerations. Congenital abnormalities that have been associated with cervical weakness include uterine anomalies, genetic disorders affecting collagen (e.g., Ehlers-Danlos syndrome), and in utero diethylstilbestrol exposure. Cervical insufficiency may be a cause of recurrent second-trimester loss but is not typically associated with early pregnancy loss.

Luteal phase deficiency

Progesterone is required for implantation and maintenance of pregnancy; therefore disorders related to impaired progesterone are likely to affect pregnancy success. For the first 7 weeks of gestation, progesterone is produced by the corpus luteum and maintains the endometrium.

After the corpus luteum regresses, progesterone is synthesized by the trophoblast and maintains the decidual tissue. If progesterone secretion from the corpus luteum is low or if the endometrium has inadequate response to normal circulating levels of progesterone, endometrial development may be unable to support the implanted blastocyst, leading to pregnancy loss.

There is no consensus on the best method of diagnosis or treatment for luteal phase deficiency ( ). The diagnosis of luteal insufficiency historically was made by performing a histologic examination of the endometrium; however, histologic dating of the endometrium has proven unreliable and not reproducible ( ). Similarly, the measurement of progesterone levels in the luteal phase is not considered a reliable diagnostic test. There is no standard progesterone value that defines normal luteal function because the corpus luteum varies from cycle to cycle in a normal fertile woman . In a 2015 committee opinion, the ASRM concluded “there is no reproducible, pathophysiologically relevant, and clinically practical standard to diagnose luteal phase deficiency and distinguish fertile from infertile women” ( 2015). Therefore we do not recommend performing luteal phase testing.

Abnormal luteal phase progesterone production may also occur as a result of medical conditions such as elevated prolactin or abnormal thyroid function and treatment of the underlying condition is recommended.

Thyroid disease

Risk of pregnancy loss is increased in women with poorly controlled thyroid disease, including hypo- and hyperthyroidism. Studies have also shown that subclinical hypothyroidism (SCH), defined as TSH greater than 4 mIU/L, is associated with pregnancy loss and treatment with levothyroxine improves pregnancy rates and decreases the rate of pregnancy loss. A meta-analysis of 18 cohort studies by Maraka and colleagues reported that SCH doubles the risk of pregnancy loss compared with euthyroid women ( ). The normal reference range for TSH changes in pregnancy. The upper limit of normal in most laboratories is 4 mIU/L for nonpregnant women and 2.5 mIU/L during the first trimester of pregnancy. The use of the first trimester pregnancy thresholds to diagnose and treat subclinical hypothyroidism in women attempting pregnancy remains controversial. According to the ASRM, if TSH levels before pregnancy are between 2.5 and 4 mIU/L, management options include either monitoring levels and treating when TSH is 4 mIU/L or greater or treating with levothyroxine to maintain TSH at 2.5 mIU/L or less ( ). The Endocrine Society recommends that patients with RPL be treated to keep a TSH level between 1 and 2.5 mIU/L. Once pregnant, women being treated for hypothyroidism often require an increase in thyroxine during pregnancy, beginning as early as the fifth week of pregnancy.

Studies regarding treating euthyroid patients with positive thyroid antibodies are conflicting. A large randomized controlled trial of 952 euthyroid women with positive antithyroid antibodies showed no benefit when treated with levothyroxine ( ). Routine screening for antithyroid antibodies in women with RPL is not currently recommended but may be considered if repeated TSH values are 2.5 mIU/L or more or when other risk factors for thyroid disease are present.

Hyperprolactinemia

Hyperprolactinemia has been associated with pregnancy loss. Elevated levels of prolactin can alter the hypothalamic-pituitary axis, resulting in impaired folliculogenesis, decreased oocyte maturation, and a short luteal phase. Normalization of prolactin levels with a dopamine agonist improves pregnancy outcomes in women with RPL . In a randomized controlled trial, 64 hyperprolactinemic women with RPL were randomly assigned bromocriptine therapy or no bromocriptine; treatment resulted in an 85.7% live-born rate compared with 52.4% in the untreated group ( ). Prolactin levels in early pregnancy were significantly greater in women who miscarried.