Spontaneous preterm birth: Pathogenesis

INTRODUCTION — Preterm birth (PTB) is a leading cause of infant morbidity and mortality in the United States and globally [1,2]. For this reason, the pathogenic processes leading to PTB and the development of preventive interventions are major targets of obstetric research.

PTB may be "spontaneous" (sPTB) or "indicated" because of concerns about maternal or fetal status. Compelling clinical and research evidence suggest that a number of pathogenic processes can lead to a final common pathway that results in sPTB (figure 1). The four major pathogenic processes are:

●Premature activation of the fetal hypothalamic-pituitary-adrenal axis in response to maternal and/or fetal stress

●Exaggerated inflammatory response/infection and/or an altered genital tract microbiome

●Decidual hemorrhage (placental abruption)

●Pathologic uterine distention

These four processes, and other less common processes, may be initiated long before preterm labor or preterm prelabor rupture of membranes is clinically evident, which is a common feature of cascade biologic pathways.

This topic will discuss the pathogenic processes leading to sPTB. Risk factors for PTB and risk reduction interventions, as well as diagnosis and treatment, are reviewed separately.

●(See "Spontaneous preterm birth: Overview of risk factors and prognosis".)

●(See "Preterm labor: Clinical findings, diagnostic evaluation, and initial treatment".)

●(See "Inhibition of acute preterm labor".)

THE FOUR MAJOR PATHOGENIC PATHWAYS TO sPTB

#1 Stress-induced premature activation of the HPA axis

Causes of stress — Fetal stress is caused by uteroplacental vascular insufficiency, which, if severe, may manifest clinically as preeclampsia or fetal growth restriction. Pregnancies manifesting these disorders are more likely to end in sPTB and/or indicated PTB [3,4]. In regression models, pregnancies that end in sPTB have a four- to sevenfold increase in histologic evidence of placental vascular damage, bleeding, fetal vascular disruption, or lack of normal physiologic conversion of maternal spiral arteries [3]. First pregnancies that end in sPTB are at increased risk of PTB, preeclampsia, and fetal growth restriction in the subsequent pregnancy compared with uncomplicated first pregnancies, and the earlier the sPTB, the higher the risk of one of these complications in the next pregnancy [5].

Major maternal psychosocial stress (eg, depression, posttraumatic stress disorder, anxiety) can activate the maternal hypothalamic-pituitary-adrenal (HPA) axis and has been associated with a small (generally less than twofold) increased risk of sPTB, but the association is weaker than that observed with fetal stress [6-19]. A prospective cohort study of individuals with depressive symptoms early in pregnancy found that they had almost twice the PTB risk of pregnant individuals without such symptoms, and the risk increased with increasing severity of depression, suggesting a dose-response effect [17].

Relationship between stress and hormonal changes leading to sPTB — Fetal or, less commonly, maternal stress can induce premature activation of the fetal HPA axis resulting in a cascade of hormonal changes leading to sPTB. A synopsis of these changes follows:

●Stress-induced increases in placental production and release of corticotropin-releasing hormone (CRH) appear to program a "placental clock" for earlier labor and delivery [20-22]. PTB occurs when this process is accelerated, which may be marked by elevated maternal CRH levels.

●In the mother, CRH released by the hypothalamus stimulates secretion of adrenocorticotropic hormone (ACTH) from the pituitary, promoting adrenal secretion of cortisol, which inhibits hypothalamic CRH and pituitary ACTH release via a negative feedback loop. During pregnancy, the trophoblast, amniochorion, and decidual cells also secrete CRH (called placental CRH). In a normal pregnancy, it is hypothesized that maturation of the fetal HPA axis and development of the fetal zone of the fetal adrenal gland beginning at 25 weeks of gestation cause a progressive physiologic increase in fetal cortisol secretion, which enhances placental CRH release [23] because, in contrast to the maternal HPA axis, placental CRH production is stimulated by cortisol in a positive feedback loop [24-27]. Placental CRH secretion stimulates the fetal pituitary and adrenal to release cortisol but also stimulates the amniochorion and decidua to release prostaglandins, which further stimulate placental CRH release [27] via a second positive feedback loop [28].

The net effect of these positive and negative feedback loops is increasing levels of CRH as gestation progresses, with high concentrations in the maternal and fetal blood and amniotic fluid by term, and increasing levels of glucocorticoids and prostaglandins [29]. Prostaglandins are an integral part of the common final pathway leading to parturition since they promote cervical maturation and uterine contractility [30]. In a normal pregnancy, the rise in prostaglandins ultimately results in labor occurring at term. The effects of CRH are augmented near term by a large reduction in maternal plasma CRH-binding protein [31], allowing saturation of CRH-binding protein and making free CRH available to act as a parturition trigger [22], in part because CRH also has some ability to stimulate uterine contractility directly [32].

●Stress-induced activation of the fetal HPA axis stimulates increased adrenal synthesis of dehydroepiandrosterone sulfate (DHEA-S), which is converted to 16-hydroxy-DHEA-S in the fetal liver. Placental CRH directly augments fetal adrenal DHEA production as well [33]. The placenta converts DHEA and DHEA-S to estrone (E1), estradiol (E2), and estriol (E3), which, in the presence of estrogen receptor-alpha (ER-alpha) and reduced progesterone receptor (PR) action, increases gap junction formation, oxytocin receptors, prostaglandin activity, and enzymes responsible for coordinated muscle contraction (myosin light chain kinase, calmodulin) [34-38].

Elevated circulating estrogens are a marker for sPTB. In a prospective study that obtained serial measurements of salivary E3, mean salivary E3 was higher from 24 to 34 weeks of gestation in individuals with singleton pregnancies giving birth preterm than in those giving birth at term, a surge in salivary E3 occurred three to four weeks before the onset of labor in both groups, and the predictive value of salivary E3 measurements was greatest for later PTBs (≥34 weeks) [39]. The precise mechanism for this limitation remains unknown; however, it likely reflects the requirement for development of the fetal zone of the fetal adrenal, which is responsible for the bulk of DHEA production and enlarges rapidly in the last six weeks of gestation [40].

●Glucocorticoids induce the immunophilin co-chaperone FK506-binding protein 51 (FKBP51), which binds to the ligand binding site on the PR and glucocorticoid receptor (GR) to inhibit ligand binding to PR and GR, nuclear translocation, and receptor-mediated transcription. While this negative feedback circuit prevents pathologic effects accruing from excess glucocorticoids, it also inhibits PR activity. Consistent with maturation of the fetal HPA axis and rising maternal cortisol levels at term, human labor at term is accompanied by increased expression of FKBP51 in decidual cells, resulting in functional progesterone withdrawal [41]. In situ FKBP51 protein expression is increased, as is nuclear FKBP51-progesterone (PR) binding in decidual cells of patients with idiopathic PTB compared with gestational age-matched controls. Moreover, in a murine model, Fkbp5-deficient (Fkbp5 -/-) mice are highly resistant to stress-induced depressive and anxiety-like behaviors [42]. Maternal restraint stress in wild type Fkbp5 +/+ mice increases Fkbp5, decreases PR, and elevates expression of the progesterone metabolizing enzyme AKR1C18 in uteri at E17.25 days of a 20-day murine gestation, followed by reduced uterine progesterone levels and increased oxytocin receptor expression at 18.25 days, resulting in PTB. Conversely, Fkbp5 -/- (knock out) mice have prolonged gestations and are completely resistant to stress-induced PTB and the proparturition biochemical changes found in the uteri of stressed Fkbp5 +/+ mice. These findings indicate that stress induces a functional intracellular progesterone withdrawal mediated by increased decidual FKBP51 expression and FKPB51-PR binding to promote sPTB.

#2 Inflammation — Intra-amniotic inflammation resulting from infection or damage-associated molecular patterns (DAMPs) or "alarmins" released because of cellular stress or damage is the best-established causal link to preterm labor and birth [43]. In the setting of infection-induced fetal stress, the fetal cortisol to DHEA ratio remains low, and there is no direct correlation between the fetal adrenal gland volume and either cortisol or DHEA levels [44]. This suggests that infection-mediated sPTB may act via a different pathway than the premature activation of the HPA axis described above.

Epidemiologic studies using laboratory and clinical data show an association between sPTB and the presence of systemic and genitourinary tract pathogens [45-47], as well as an altered microbiome [47-49]:

●In a large retrospective population-based study of 199,093 deliveries, 2.5 percent of patients had asymptomatic bacteriuria, which was independently associated with PTB (adjusted odds ratio [OR] 1.6, 95% CI 1.5-1.7) [45]. Conversely, the diagnosis and treatment of asymptomatic bacteriuria appears to reduce the risk of PTB [46].

●Both clinical and subclinical chorioamnionitis are much more common in preterm than term deliveries and may account for 50 percent of sPTB before 30 weeks of gestation [50]. (See "Spontaneous preterm birth: Overview of risk factors and prognosis", section on 'Infection'.)

●Lactobacillus is the predominant flora of the microbial community in normal pregnancy, and the prevalence of a Lactobacillus-poor vaginal community state type 4 (CST 4) is inversely correlated with gestational age at delivery [51]. In addition, the risk for sPTB is more pronounced for individuals with CST 4 and elevated Gardnerella or Ureaplasma. However, treatment of bacterial vaginosis (BV) does not reduce sPTB rates in low-risk patients [48].

A nested case-control study of cervicovaginal microbiota among 107 patients with sPTB and 432 individuals with term deliveries found Mobiluncus curtisii/mulieris to be the only bacterial taxa significantly associated with increased risk of sPTB among all participants. However, among African American patients, five separate taxa were significantly associated with sPTB, and M. curtisii/mulieris and Mageeibacillus indolicus had the strongest association with sPTB with rates of over 60 and 55 percent, respectively, at their highest relative abundance [49]. This association disappeared among patients with these taxa but who also had abundant Lactobacillus.

●Metagenomic techniques, including surveillance of the 16S rRNA gene, have increased understanding of the spectrum of cultivated and uncultivated human microbial agents involved in the pathogenesis of sPTB [52]. By using metagenomic tools, several groups of investigators reported that, in pregnancies complicated by sPTB secondary to infection, approximately two-thirds of the amniotic fluid bacteria detected by culture-independent methods were not isolated in cultures [53,54]. These included both uncultivated and difficult-to-cultivate species, such as Fusobacterium nucleatum, Leptotrichia (Sneathia), Bergeyella, Peptostreptococcus, Bacteroides, and a species of the order Clostridiales.

In a cohort of approximately 12,000 samples as part of the integrative Human Microbiome Project, longitudinal analyses of 16S rRNA, metagenomic, metatranscriptomic, and cytokine profiles from 45 preterm and 90 term birth controls identified lower vaginal levels of Lactobacillus crispatus and higher levels of BV-associated bacterium (BVAB) 1 and TM7, Sneathia amnii, a group of Prevotella species, and nine additional taxa as harbingers of PTB [55].

Pathophysiology

●Bacterial degradation of membranes and direct uterotonic effects – Bacteria may have a direct role in the pathogenesis of sPTB.

•Bacteria produce phospholipase A2 (which leads to prostaglandin synthesis) and endotoxin, substances that stimulate uterine contractions and can cause preterm labor [56].

•Some organisms (eg, Pseudomonas, Staphylococcus, Streptococcus, Bacteroides, and Enterobacter) produce proteases, collagenases, and elastases that can degrade the fetal membranes, leading to preterm prelabor rupture of membranes (PPROM), with subsequent spontaneous or indicated PTB [57,58].

●Bacterial induction of maternal-fetal inflammatory responses, leading to uterotonic effects – Bacterial ligands bind to toll-like receptors (TLRs) expressed on decidual, amnion, chorion, cervical, placental, and local leukocyte cell membranes. This induces the transcription factor NF-kappaB, which then triggers a maternal and/or fetal inflammatory response in susceptible individuals that is linked to sPTB.

This TLR-mediated response is ultimately characterized by the presence of activated neutrophils and, to a lesser extent, activated macrophages and various proinflammatory mediators (eg, interleukin [IL] 1, 6, and 8; tumor necrosis factor [TNF], granulocyte colony-stimulating factor [G-CSF], colony-stimulating factor 2 [CSF-2], and matrix metalloproteinases [MMPs]). The key initial mediators of this response are IL-1beta and TNF, which enhance prostaglandin production by inducing cyclooxygenase 2 (COX-2) expression in the amnion and decidua while inhibiting the prostaglandin-metabolizing enzyme, 15-hydroxyprostaglandin dehydrogenase (PGDH), in the chorion [59,60]. Moreover, IL-1beta and/or TNF directly enhance the expression of various MMPs in the amnio-chorion, decidua, and cervix to degrade the extracellular matrix of the fetal membranes and cervix [61-63].

Immunohistochemical staining of placentas from patients with chorioamnionitis-associated sPTB demonstrated significantly lower PR levels in decidual cells [64]. Moreover, treatment of cultured term decidual cells with IL-1beta decreased PR expression (both PR-B and A isoforms) as well as PR mRNA levels via ERK1/2 MAP kinase signaling. This reduction in PR expression may account for the increasing number of studies showing no benefit of progesterone supplementation in the prevention of PTB [65-67].

Elevated proinflammatory mediator levels have been demonstrated in the amniotic fluid of patients with preterm labor with intact membranes, and these levels correlated well with positive results from culture of the amniotic fluid and fetal membranes [68]. In a review of 17 primary studies including 6270 asymptomatic pregnant individuals, elevated cervicovaginal and amniotic fluid IL-6 levels at midgestation predicted PTB with ORs 3.05 (95% CI 2.00-4.67, number needed to treat = 7) and 4.52 (95% CI 2.67-7.65, number needed to treat = 7), respectively [69]. In addition, patients with intraamniotic inflammation destined to deliver preterm appear to present unique amniotic fluid, vaginal-cervical secretion, and serum proteomic profiles [70-74].

#3 Decidual hemorrhage — Decidual hemorrhage (placental abruption) originates in damaged decidual blood vessels and presents clinically as vaginal bleeding or retroplacental hematoma formation [75]. It is associated with a high risk of preterm labor and PPROM [76-78].

The development of PPROM in the setting of abruption may be related to the high decidual cell expression of tissue factor, the primary cellular mediator of hemostasis. Following intrauterine hemorrhage from placental abruption, decidual tissue factor combines with factor VIIa to activate factor Xa, which in turn complexes with its cofactor, Va, to generate thrombin. Hormonal factors, such as progesterone, play an important modulator role [79]. Thrombin activation (measured by serum thrombin-antithrombin [TAT] complex levels) has been noted in patients with preterm labor and in asymptomatic patients who subsequently delivered preterm [80,81].

In addition to its primary hemostatic functions, thrombin binds to protease-activated receptors (PAR) 1 and 3 to:

●Directly increase the frequency, intensity, and tone of myometrial contractions, an effect that is suppressed by blood containing thrombin inhibitors [82,83]. Myometrial thrombin-PAR-1 interactions activate myosin light chain kinase, promote Rho-induced phosphorylation of myosin light chain-2, and enhance COX-2 expression and prostaglandin (PG) E2 and F2 alpha synthesis [84].

●Upregulate the expression of proteases such as MMPs, which promote uterine contractility [85-87].

●Initiate functional progesterone withdrawal in the decidua by inhibiting expression of PRs in decidual cells [88]. As with IL-1beta-mediated inhibition of decidual cell PR expression, thrombin inhibition of PR protein and mRNA expression is mediated by ERK1/2 MAP kinase signaling. This may be a mechanism for contractions in pregnancies complicated by antepartum bleeding and, again, help account for the failure of progesterone supplementation to prevent PTB in some studies.

●Induce IL-8 in decidual cells, accounting for the dense neutrophil infiltrate observed in abruption-associated PPROM in the absence of infection [89]. The interactive effects of thrombin-enhanced MMPs with neutrophil-derived proteases promote degradation of the fetal membrane extracellular matrix, which can result in PPROM and, in turn, preterm labor and spontaneous and indicated PTB.

#4 Pathologic uterine distention — Multiple gestation, polyhydramnios, and other causes of excessive uterine distention are well-described risk factors for sPTB. Enhanced stretching of the myometrium induces formation of gap junctions, upregulation of oxytocin receptors, and production of inflammatory cytokines, prostaglandins, and myosin light chain kinase, which are critical events preceding uterine contractions and cervical dilation [90-92]. Myometrial distention also increases expression of genes with important roles in collagenolysis and inflammation [93]. (See "Physiology of parturition at term".)

Distention of the fetal membranes also contributes to myometrial activation, preterm cervical ripening, and PPROM, likely through the release of cytokines, prostaglandins, and collagenases [94-96].

OTHER PATHOGENIC PATHWAYS

Cervical insufficiency — Cervical insufficiency refers to pathologic dilation and/or effacement of the uterine cervix in the absence of uterine contractions leading to previable pregnancy loss as well as sPTB. Cervical cerclage may be helpful in select instances [97]. (See "Cervical insufficiency".)

Cervical insufficiency due to intrinsic biochemical factors is probably a rare event (eg, Ehlers-Danlos syndrome). It is more likely that progressive cervical shortening and dilation prior to viability results from activation of the inflammatory and/or hemorrhagic pathways described above at a point in gestation when myometrial quiescence and amniotic fluid, fetal membrane, and decidual antiprotease activity are maximal.

Inherited pathogenic variants — sPTB demonstrates familial aggregation. Pregnant people who were themselves born preterm have a higher risk of sPTB, and the risk of sPTB increases by 80 percent in those whose sisters had sPTB. There are also significant ethnic differences in risk of sPTB. Taken together, these findings suggest that genetic factors may be involved. (See "Spontaneous preterm birth: Overview of risk factors and prognosis", section on 'Genetic variants'.)

A variety of single nucleotide polymorphisms, the majority associated with inflammation, have been linked to sPTB [98]. In a genome-wide association study among 43,568 females of European ancestry that used gestational duration as a continuous variable and either term or sPTB (<37 weeks) as a dichotomous outcome, four genetic loci (EBF1, EEFSEC, AGTR2, and WNT4) were significantly associated with gestational duration in both the discovery and replication datasets [99]. Functional analysis suggested that a variant in the WNT4 gene altered the binding of the estrogen receptor. Common variants in EBF1, EEFSEC, and AGTR2 showed association with PTB with genome-wide significance, and analysis of mother-infant dyads indicated that they act at the level of the maternal genome.

African American pregnant people have a 50 percent higher rate of sPTB than non-Hispanic White pregnant people [100,101]. This phenomenon may be partially explained by the observation that multiple maternal and fetal inflammatory pathway genetic polymorphisms linked to inflammation-associated sPTB are found in greater prevalence among African American mothers and/or fetuses than among other racial groups [102,103]. For example, African American mothers harboring both a polymorphism of the tumor necrosis factor-alpha gene and bacterial vaginosis are at significantly greater risk of PTB (odds ratio 6.1, 95% CI 1.9-21.0) [104]. Conversely, higher vaginal levels of the host-derived antimicrobial peptide, beta-defensin 2, are linked to a lowered risk of sPTB, while lower levels were linked to spontaneous PTB only among African American pregnant people [49].

SUMMARY AND RECOMMENDATIONS

●Mechanisms – There are four discrete mechanisms for the pathogenesis of preterm birth (PTB):

•Stress, which induces premature activation of the fetal hypothalamic-pituitary-adrenal axis leading to a cascade of hormonal changes that promote functional progesterone withdrawal and increase prostaglandin and protease production, resulting in spontaneous preterm birth (see '#1 Stress-induced premature activation of the HPA axis' above).

•Inflammation/infection/altered genital tract microbiome, in which bacterial degradation of membranes and/or bacterial induction of an inflammatory cytokine response has proparturition effects (see '#2 Inflammation' above).

•Decidual hemorrhage (placental abruption), which leads to high decidual cell expression of tissue factor and eventually generation of thrombin. In addition to its primary hemostatic functions, thrombin binds to its PAR-1 receptor in the reproductive tract to promote uterine contractility and weaken membranes, resulting in fetal membrane rupture and/or preterm birth (see '#3 Decidual hemorrhage' above).

•Pathologic uterine distention, which induces formation of gap junctions, upregulation of oxytocin receptors, and production of inflammatory cytokines, prostaglandins, and myosin light chain kinase, which are critical events preceding uterine contractions and cervical dilation. Myometrial distention also increases expression of genes with important roles in collagenolysis and inflammation (see '#4 Pathologic uterine distention' above).

●Final common pathway – Although each mechanism has distinct epidemiologic, genetic, and clinical characteristics, they are not mutually exclusive. They share a final common pathway involving the formation of uterotonic agents and proteases that weaken the fetal membranes and cervical stroma, eventually leading to preterm labor and birth (figure 1).