Women who have delivered an infant between 16 and 36 weeks’ gestation have an increased risk of preterm birth in subsequent pregnancies. The risk increases with more than 1 preterm birth and is inversely proportional to the gestational age of the previous preterm birth. African American women have rates of recurrent preterm birth that are nearly twice that of women of other backgrounds. An approximate risk of recurrent preterm birth can be estimated by a comprehensive reproductive history, with emphasis on maternal race, the number and gestational age of prior births, and the sequence of events preceding the index preterm birth. Interventions including smoking cessation, eradication of asymptomatic bacteriuria, progestational agents, and cervical cerclage can reduce the risk of recurrent preterm birth when employed appropriately.
Efforts to reduce the incidence of preterm birth cannot yet be called successful but have produced sufficient information to justify management suggestions for clinicians, not only for strategies that do not work but also for those that do. The quality and number of clinical trials of interventions intended to reduce the incidence and/or morbidity of preterm birth has risen substantially in recent years, yielding data that can improve care for women at risk.
An important advance has been the revision of the traditional model of preterm birth in which painful contractions or preterm membrane rupture were understood as the most common initial steps preceding cervical change. This concept has been challenged by observations from several clinical studies:
- •
Drugs that arrest or inhibit uterine contractions can delay delivery but do not reliably reduce or prevent preterm birth.
- •
Assessment of contraction frequency is not a useful test to predict preterm birth, and its use to detect and arrest initial or recurrent episodes of preterm labor does not affect the rate of preterm birth.
- •
Antibiotics that eradicate microorganisms associated with preterm birth also do not reliably reduce, and may sometimes increase, the incidence of preterm birth, especially in women with intact membranes.
- •
Progestational agents such as 17α-hydroxy-progesterone caproate reduce recurrent preterm birth in some women with a prior preterm birth, especially those with early preterm birth and short cervix, but have no effect on preterm birth in multiple gestations.
- •
Cervical cerclage can reduce recurrent preterm birth in women with a prior preterm birth and short cervix but increases the risk of preterm birth in women with multiple gestation who also have a short cervix.
These findings suggest the following 4 substantive changes in the “contractions change the cervix” construct of preterm birth:
- 1
Contractions are not the primary initial driver of most preterm births. Cervical ripening and decidual activation are more common first steps in preterm parturition.
- 2
Progesterone affects the steps leading to preterm birth in women who deliver very preterm (<32 weeks), a group characterized by inflammation-driven preterm birth and short cervix without apparent contractions. The antiinflammatory effect of progestins is therefore a likely mechanism of their effect on preterm birth.
- 3
Cerclage helps some of this same group, especially those with a short cervical length, suggesting cerclage may help more by preventing membrane prolapse and bacterial invasion of the amniotic cavity from vaginal flora than by making the cervix stronger.
- 4
The parturitional process clearly begins in many women before 20 weeks’ gestation, making this historical boundary between miscarriage and birth obsolete.
These observations are the basis for an alternate model of preterm birth in which cervical ripening (short cervix), driven primarily by inflammation beginning in the early second trimester or even earlier, is the most common initial clinical manifestation of preterm parturition, especially in women with a previous preterm birth. Because preterm birth occurs in only 35-40% of all women with untreated short cervix and in 25-30% of women with a positive fetal fibronectin in the second trimester, many women apparently experience the initial phases of preterm parturition but do not progress to preterm birth. This is the model that underlies the recommendations discussed in the following text.
The contribution of prior preterm birth to preterm birth in general
Approximately 13 million of the more than 130 million babies born annually worldwide are born preterm, a global incidence of nearly 10%. There are more than 500,000 births before 37 weeks’ gestation in the United States each year, a rate of 12.7% in 2007. Of these, about 15% occur in women with a prior preterm birth. Current effective interventions could potentially eliminate as many as 35-50% of recurrent preterm births (n ∼ 37,000).
Identification of women with a prior preterm birth
A thorough obstetrical history is essential to identify women with prior preterm birth. Unfortunately, the nomenclature used to describe pregnancy outcomes, expressed as gravidity, parity (births after 20 menstrual weeks’ gestation), and abortions (births before 20 weeks’), lacks a “clear epidemiologic, biologic, or clinical basis” and is inconsistent with recent obstetrical epidemiology: Women whose prior pregnancy ended between 16 and 20 weeks have a risk of recurrent preterm birth that equals or exceeds the recurrence risk for women whose prior preterm birth occurred after 20 weeks ( Figure 1 from Edlow et al ). Women with a prior stillbirth are also often considered separately from those with a prior preterm birth, but their risk of subsequent spontaneous preterm birth is also increased. Thus, any woman with a prior birth between 16 0/7 and 36 6/7 weeks should be evaluated as possibly having had a preterm birth, whether the fetus was born alive or stillborn.
A comprehensive reproductive history should record prior spontaneous abortions and elective terminations, including the gestational age and methods of termination. Contrary to common belief, population-based studies have found that elective pregnancy terminations in the first and second trimesters are associated with a very small but apparently real increase in the risk of subsequent spontaneous preterm birth (PTB). The mechanism is unknown. The gynecological history is also important. Cervical surgery (cold knife, laser conization, and loop-electrosurgical excision procedures [LEEP]) is also associated with increased risk of preterm birth.
Because assisted reproduction techniques, including in vitro fertilization and all levels of ovulation promotion (clomiphene) and stimulation (gonadotropins), have been associated with a 2-fold increased risk of PTB, the circumstances of conception of a prior pregnancy ending in preterm delivery should be reviewed. Superovulation followed by conception from fresh eggs in the same cycle has a greater risk of preterm birth than does use of frozen eggs.
When the prior preterm birth occurred in a twin pregnancy, the risk of preterm birth in a subsequent singleton gestation varies according to the gestational age at delivery of the twins. The earlier the gestational age, the greater the risk of PTB in a subsequent singleton pregnancy, ranging from minimal if any increased risk for a twin birth after 34 weeks’ to as much as 40% when the prior twin birth occurred before 30 weeks.
Pregnancies complicated by increased levels of maternal serum alpha-fetoprotein (MSAFP) unexplained by incorrect dates, multifetal gestation, or birth defects have an increased risk of adverse pregnancy outcomes, including preterm birth associated with abruption and preterm ruptured membranes. An elevated MSAFP may also occur with the demise of an unrecognized or vanishing twin, another risk factor for preterm birth.
Document the clinical presentation of the prior preterm birth(s)
Preterm parturition is characterized by cervical ripening, decidual-membrane activation, and uterine contractions, any of which may predominate. The order in which these steps occurred can often be discerned from the history, offering clues to the etiology that can improve the prediction of recurrence risk. Obstetricians have traditionally been taught to ask about painful contractions as evidence that true labor has begun at term, but focusing on the presence and pain of contractions is inappropriate when taking a history of preterm parturition, in which cervical ripening and decidual-membrane activation are much more likely causes of the earliest signs and symptoms.
- •
The most common sequence preceding preterm birth is cervical ripening (shortening of the cervix), followed by decidual-membrane activation and then contractions. This sequence is characterized by mild symptoms of pelvic pressure, premenstrual-like cramping, and increased vaginal discharge that occur over several days or weeks. The initial clinical examination reveals a soft, effaced cervix, with minimal dilation. Contractions are perceived as absent or mild because the cervix is already effaced and slightly dilated. Spotting or ruptured membranes may occur a few days later. The ultimate clinical presentation in these women may be pelvic pressure with advanced dilation, ruptured membranes, or contractions that seem inadequate to explain the advanced effacement and dilation. The risk of recurrent PTB for women who describe this sequence is high.
- •
When preterm membrane rupture is the presenting complaint, a careful history may distinguish rupture that follows prolonged cervical ripening as described in the preceding text, from rupture after persistent spotting related to a vanishing twin or to a diagnostic amniocentesis, in which the recurrence risk is likely to be lower.
- •
A history of painful contractions without any premenstrual symptoms suggests pressure against a closed, unripe cervix, as might occur when the stimulus for labor is acute (eg, an occult abruption). The recurrence risk in this setting will vary with the cause of abruption.
- •
A confusing history (eg, of preterm birth in the first and third pregnancies with a birth at term in the second) raises the possibility of a uterine anomaly, in which preterm birth may be related to preterm separation of the placenta implanted on a uterine septum. We examine uterine anatomy with endovaginal ultrasound in these women. The recurrence risk will vary with the site of placental attachment.
- •
Although preterm births are commonly labeled as being spontaneous or indicated, these categories often overlap, can be difficult to distinguish, and may be misleading when caring for a subsequent pregnancy. Review of the records of any prior preterm birth, regardless of the clinical circumstances, is therefore useful. This is important for births between 16 and 26 weeks, in which fetal anomalies, placental abnormalities, uterine anomalies, and spontaneous preterm parturition are often conflated.
- •
Pathology reports should be reviewed for evidence of abruption, funisitis, and chorioamnionitis. Inflammation of the fetal membranes is found on the pathology report for most early preterm births, regardless of the process that led to preterm birth. When no inflammation is noted on pathological examination, we interpret that to suggest a relatively brief exposure of membranes to the vaginal flora, as might occur in cervical insufficiency.
Recurrent vs nonrecurrent preterm birth
Women with recurrent preterm birth are more likely to have a low prepregnancy weight (below 100 pounds, with a body mass index <19.8 kg/m 2 ) and to be African American than women with a single prior preterm birth. The risk of recurrent preterm birth increases as the gestational age of the previous preterm birth declines and as the number of previous preterm births increases, and thus is highest in women with more than 1 early preterm birth. Women with more than 1 preterm birth are also more likely to have prior early (<32 weeks) preterm birth, to demonstrate clinical and ultrasound evidence of short cervix (sonographic length <25 mm, and Bishop score >3), and to have a positive fibronectin at 22-24 weeks.
A study of women with a prior preterm birth found that the likelihood of recurrent preterm delivery varied greatly, from less than 10% in women with a cervical length greater than 35 mm and a negative fibronectin at 22-24 weeks to more than 60% when the cervical length was less than 25 mm and the fibronectin test was positive at 22-24 weeks.
Estimation of individual risk of recurrent preterm birth
The risk of recurrent preterm birth is commonly reported to be increased by 1.5- to 2-fold to 4-fold or more, depending on the population. The risk increases with the number of prior preterm births and as the gestational age decreases ( Figure 2 from McManemy et al ).
There are 3 historical factors that have a substantial influence on the likelihood of recurrent preterm birth. All are immediately available without cost: maternal race (African American vs all others), the gestational age of the index preterm birth, and the number of previous preterm births. Each confers a 1.5- to 2-fold increase in the risk of recurrence, over and above the 1.5- to 2-fold increased risk in women with any prior preterm birth. Comparison of conservatively estimated personal recurrence risks for 2 hypothetical women, each with 2 prior preterm births, illustrates this approach as used in our clinics:
- •
Patient A is a non-Hispanic white (NHW) woman whose first and second preterm births were both at 34 weeks. Her baseline risk is approximately 11% (overall rate of preterm birth for NHW women) and increases by approximately 1.5 for each prior PTB: 11% × 1.5 for the first PTB = 17.25% × 1.5 for the second PTB = an estimated risk of 26% in her third pregnancy.
- •
Patient B is a non-Hispanic black (NHB) woman with 2 prior preterm births at 28 and 26 weeks’ gestation. Her baseline risk is 17% (the rate for NHB women in the United States) and increases by 1.5-fold because of the first preterm birth and then by another 1.5-fold because the birth was early, before 32 weeks. Her risk entering her second pregnancy was therefore approximately 35% (17% × 1.5 × 1.5). After a second PTB before 32 weeks, this same calculation yields a risk for her third pregnancy of more than 75 % (35% × 1.5 X 1.5 = 79%).
These estimated risks are just approximations generated from published population risks and do not account for the sequence of term vs preterm births, in which the gestational age of the most recent birth predominates. More accurate estimates could be obtained from logistic regression analyses, but our simplistic method generates rate estimates that are consistent with reports in the literature. More specific information has been reported from studies of various tests for PTB risk, most usefully by Berghella et al ( Table ). Antenatal care decisions, such as the proper time to give antenatal corticosteroids, might be informed by knowledge of the risk of recurrent PTB according to the length of the cervix at various gestational ages ( Table ).
| Cervical length, mm | Week of pregnancy | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | |
| 0 | 76.3 | 73.7 | 70.9 | 67.9 | 64.7 | 61.4 | 58.0 | 54.5 | 51.0 | 47.5 | 44.0 | 40.5 | 37.2 | 33.9 |
| 5 | 67.9 | 64.8 | 61.5 | 58.1 | 54.6 | 51.1 | 47.6 | 44.0 | 40.6 | 37.2 | 34.0 | 30.9 | 28.0 | 25.2 |
| 10 | 58.1 | 54.7 | 51.2 | 47.6 | 44.1 | 40.7 | 37.3 | 34.1 | 31.0 | 28.0 | 25.3 | 22.7 | 20.3 | 18.1 |
| 15 | 47.7 | 44.2 | 40.7 | 37.4 | 34.1 | 31.0 | 28.1 | 25.3 | 22.7 | 20.4 | 18.2 | 16.2 | 14.3 | 12.7 |
| 20 | 37.4 | 34.2 | 31.1 | 28.1 | 25.4 | 22.8 | 20.4 | 18.2 | 16.2 | 14.4 | 12.7 | 11.2 | 9.9 | 8.7 |
| 25 | 28.2 | 25.4 | 22.8 | 20.4 | 18.2 | 16.2 | 14.4 | 12.7 | 11.3 | 9.9 | 8.7 | 7.7 | 6.7 | 5.9 |
| 30 | 20.5 | 18.3 | 16.3 | 14.4 | 12.8 | 11.3 | 9.9 | 8.7 | 7.7 | 6.7 | 5.9 | 5.2 | 4.5 | 3.9 |
| 35 | 14.5 | 12.8 | 11.3 | 10.0 | 8.8 | 7.7 | 6.8 | 5.9 | 5.2 | 4.5 | 4.0 | 3.5 | 3.0 | 2.6 |
| 40 | 10.0 | 8.8 | 7.7 | 6.8 | 5.9 | 5.2 | 4.5 | 4.0 | 3.5 | 3.0 | 2.6 | 2.3 | 2.0 | 1.7 |
| 45 | 6.8 | 5.9 | 5.2 | 4.5 | 3.9 | 3.4 | 3.0 | 2.6 | 2.3 | 2.0 | 1.7 | 1.5 | 1.3 | 1.1 |
| 50 | 4.6 | 4.0 | 3.5 | 3.0 | 2.6 | 2.3 | 2.0 | 1.7 | 1.5 | 1.3 | 1.2 | 1.0 | 0.9 | 0.8 |
| 55 | 3.0 | 2.7 | 2.3 | 2.0 | 1.8 | 1.5 | 1.3 | 1.2 | 1.0 | 0.9 | 0.8 | 0.7 | 0.6 | 0.5 |
| 60 | 2.0 | 1.8 | 1.5 | 1.3 | 1.2 | 1.0 | 0.9 | 0.8 | 0.7 | 0.6 | 0.5 | 0.4 | 0.4 | 0.3 |
Interventions to reduce the risk of recurrent preterm birth
Numerous strategies and treatments have been proposed to reduce the risk of recurrent preterm birth, but few have been found effective when tested in clinical trials. The interventions in the following text are grouped according to the strength of evidence supporting their use:
Interventions supported by firm evidence
- •
Smoking-cessation programs: pregnant women are uniquely receptive to smoking-cessation programs, especially when physicians participate directly and repeatedly. Smoking-cessation programs in pregnancy have been reported to reduce the rate of preterm birth by 16% (relative risk, 0.84; 95% confidence interval [CI], 0.72–0.98 ) to 31% (adjusted odds ratio [aOR], 0.69; 95% CI, 0.65–0.74).
- •
Screening and treatment for asymptomatic bacteriuria: screening for asymptomatic bacteriuria reduces recurrent urinary tract infections and may also reduce the risk of recurrent preterm birth.
- •
Prophylactic administration of progestational agents: there is now sufficient evidence from at least 6 trials to support the prophylactic use of 17 alpha-hydroxyprogesterone caproate (17P), 250 mg given intramuscularly, weekly between 16 and 36 weeks, to women with prior spontaneous preterm birth between 20 and 36 6/7 weeks. This therapy reduces the risk of recurrent preterm birth by approximately 35% and is especially effective for women with a prior early preterm birth. 17P does not reduce preterm birth risk in women with twin or triplet pregnancies. Progesterone supplementation appears to be safe for mothers and infants.
There are many questions about clinical use of progestational agents to reduce recurrent preterm birth for which the literature is incomplete (eg, the selection of the appropriate candidates for treatment and the optimal pharmacological preparation). Current trials may answer some of these clinical dilemmas. Common questions are listed in the following text followed by our comments:
- ○
Can progesterone supplementation be provided with oral progestational compounds? No. There is 1 randomized controlled trial in 150 women with prior preterm birth. This trial suggested efficacy, but larger studies are needed.
- ○
Should vaginal progestational compounds be used to prevent preterm birth? Not at this time. Although some studies indicate a reduction in recurrent preterm birth in women treated with vaginal progestins, the largest trial did not show benefit. Until more definitive evidence is available, we prefer injectable 17P because a larger number of studied subjects support its efficacy and safety, because the benefit was clear, and, because 17P is given as an injection, there is no question about whether the study subjects actually received the medication.
- ○
Should 17 P be offered to women with prior early (16-20 weeks) spontaneous birth? Because the clinical presentation of cervical insufficiency is indistinguishable from recurrent spontaneous preterm birth, 17P prophylaxis may be offered to all women with prior early (16-20 weeks) spontaneous preterm birth ( Figure 3 ). The woman should understand that there are few data for this population.
- ○
