Pathogenesis to Treatment: Preventing Preterm Birth Mediated by Infection

Prevention of preterm birth and subsequent newborn immaturity is a primary goal of obstetrical care worldwide. Accumulated evidence shows that 1) as many as 25–50% of preterm births are caused by common genital tract infections and subsequent maternal/fetal inflammatory responses; 2) microbial and maternal host factors (phospholipases, proteases, etc.) play roles in preterm labor and preterm premature rupture of membranes (pPROM); 3) integrated aspects of maternal and fetal host responses (inflammation, altered immune adaptations, endocrine and paracrine mechanisms) play increasingly understood roles in premature activation of parturition; and 4) identification and systemic treatment of common genitourinary infections, most importantly bacterial vaginosis (BV), reduce the risks of preterm delivery and PROM.

".Infection in the female reproductive tract (especially in the cewix) can cause premature rupture of membranes and induce premature labor.... This process is responsible for many preventable infant deaths. 1,, estational age at birth and birth weight are the most important biologic determinants worldwide of an individual child's chances of survival and healthy growth, e Immediate sequelae of biologic immaturity at birth include respiratory distress, intraventricular hemorrhage, leukomalacia, necrotizing enterocolitis, prolonged hospitalization, and death. 3,4 Among survivors, life-long complications can include cerebral palsy, cognitive impairment, blindness, and deafness. 3,4 The direct and indirect costs of biologic immaturity at birth can be immense. Estimates of the excess direct medical costs attributable to preterm infants totaled $6 billion U.S. in 1988. s The total individual, family, and societal burdens imposed by biologic immaturity at birth are beyond measure.

BACTERIAL VAGINOSIS (BV) AND PRETERM BIRTH
The evidence for microbial causes of preterm birth is most extensive and convincing for BV. Nevertheless, there is continuing misapprehension that lower reproductive tract infections and BV are "mere markers" of upper tract intrauterine infection. 6 BV is not truly an infection, but rather a microecological condition in which there are dramatic alterations in the endogenous vaginal microflora. Hydrogen peroxide-producing Lactobadllus strains (including L.jensenii and L. crispatus) are reduced in number. :,s BV features multilog population increases in a characteristic set of microflora which includes Gardnerella vaginalis, genital anaerobes, and genital mycoplasmas.: These microbes, along with coliforms and streptococci, are the same as found in most cases of chorioamnionitis. 9'1 BV is associated with increased vaginal and cervical fluid concentrations of endotoxin, proteases, mucinases, sialidases, IgA proteases, and phospholipases A z and C. [11][12][13][14][15][16] Observational studies show that the presence of BV early in pregnancy is associated with second trimester labor and perinatal loss or so-called late miscarriage. 6,17 Multiple studies have been completed worldwide which demonstrate consistent associations of preterm birth with BV as well as our ability to reduce the risks with systemic (oral) treatmerit. 6,,1,16-9   PROM, and 3) among women who continue preterm labor to delivery compared with women in whom preterm labor is successfully interrupted. 9,zs-Recent investigations show that increased levels of interleukin-6 are detectable within cervical fluid of women with idiopathic preterm labor who have intraamniotic fluid infection and among women during antenatal care who subsequently deliver prior to 37 weeks gestation.
In vitro and animal model studies confirm and inform clinical studies showing causal relationship between infection and preterm birth. Interleukin-1, interleukin-6, and TNF-c production have been demonstrated by cultured human decidual cells stimulated by bacterial products. ,zs,3 Similarly, amnion or decidual explants have been shown to produce prostaglandins in response to these cyto-   36 These and other cytokines directly contribute to increased levels of uterotonic prostaglandins. 36 Such cytokines may also contribute to the onset of  preterm birth by induction of matrix metalloproteases, which in turn enhance cervical ripening and weakening of the amnionchorion. 37 Experiments performed with human fetal membranes show that substances produced by both microbes and host inflammatory cells have been shown to mediate prostaglandin release and membrane weakening in vitro. 21,22 Importantly, addition of appropriate antimicrobials including erythromycin and metronidazole during in vitro experiments of fetal membrane tensile strength prevents bacteria-induced fetal membrane weakening. 28,38 Choice of antibiotics to treat infectious agents in both in vitro and in vivo and clinical investigations regarding preterm birth is crucial: agents such as clindamycin, erythromycin, aminoglycosides, and metronidazole tend to shut down bacterial virulence factor production. Conversely, beta lactam antibiotics (penicillins, cephalosporins) act primarily by impairing cell wall synthesis, allowing for increased local tissue release of bacterial cell constituents, including endotoxins [lipopolysaccharides (LPS)]. Use of bactericidal antibiotic agents may "throw gasoline" on the "fire of inflammation" and thus worsen outcomes. Further understanding of the nature and timing of microbe-maternal and fetal interactions and how best to interfere with these processes will lead to development of improved regimens for both prevention and treatment of preterm birth mediated by reproductive tract infection and inflammation.
OBSTETRICAL STUDIES: CHLAMYDIA, GONORRHEA, AND TRICHOMONIASIS Past observation and anecdotal studies (Table 1) show that treatment for Neisseria gonorrhoeae and Chlamydia trachomatis provides benefits in terms of reduced rates of preterm birth and PROM as well as prevention of ophthalmia neonatorum ( Table  2). [39][40][41][42] Retrospective comparisons of women who received antenatal treatment for N. gonorrhoeae compared with untreated women demonstrated significant reductions of preterm birth and rupture of membranes. 39 Ryan and colleagues 41 reported significant reductions in the rate of low birth weight infants [odds ratio (OR) 0.56, 95% confidence interval (CI) 0.46-0.68] born to women who received antenatal treatment for C. trachomatis compared with retrospective, untreated control women. Similarly, Cohen and colleagues 4e described an 84% (OR 0.16, 95% CI 0.06-0.47) reduction in the rate of preterm birth among successfully treated, chlamydia-positive women compared with a retrospective control group of untreated chlamydia-positive women. Among results presented from the Vaginal Infections and Prematurity (VIP) study, Martin et al. 43 reported that among women enrolled at the New Orleans site, low birth weight (erythromycin treated 7.9% vs. placebo treated 21.2%, P 0.01) and PROM (erythromycin treated 1.2% vs. placebo treated 7.4%, P 0.04) were each significantly reduced among chlamydiapositive women who received erythromycin compared to women who received placebo. The VIP study also showed a 40% increased risk of low birth weight and prematurity if subjects suffered trichomoniasis. 44 Results of this large study confirm the role of Trichomonas vaginalis in preterm birth shown in prior smaller investigations. 4s,46

INTERVENTION STUDIES VS. BV TO REDUCE RISKS OF PRETERM LABOR AND RUPTURE OF MEMBRANES
Recently published controlled trials demonstrate that important proportions of preterm births can be prevented in women considered to be at both "high" or "normal" risk for preterm birth by screening for and treating asymptomatic BV in pregnancy (Table 2). Large studies by Hillier et al. 19 and Hauth et al. 47 add conclusive weight to well over a dozen prior studies linking BV and prematurity. Hillier et al. 19 reported the most recently analyzed portion of the large, multicenter VIP study of over 10,000 U.S. women. There was a 40% increase in low birth weight in women with asymptomatic, untreated BV. 19 Treatment with agents effective against BV (metronidazole) eliminated this excess risk. 19 Meis et al. 48 noted a similar finding in an observational study in North Carolina. Hauth and colleagues 47 performed a placebocontrolled intervention trial using a 7 day course of oral metronidazole (500 mg twice daily) and enteric-coated erythromycin (300 mg twice daily) in women judged to be at increased risk of preterm birth owing to a prior history of short gestation or maternal low body mass (<50 kg). This combined regimen reduced risks of preterm birth in subjects with asymptomatic BV by approximately one-third in both risk groups. 47 In a subsequent analysis, treated women with no other identified causes of preterm birth had risks of PROM reduced by approximately 70% (Hauth, personal communication).
Such dramatic findings confirm those of a 1994 investigation by Morales et al. 49 in which women with both a history of prior preterm birth and findings of BV in the studied pregnancy were treated with either oral metronidazole or placebo. Morales et al. 49 similarly obtained approximately 70% reductions for prematurity, low birth weight, hospital admissions for preterm labor, and PROM.
Most recently, results from a randomized placebo-controlled treatment trial of two g oral doses of metronidazole vs. placebo among Australian women with heavy growth (3-4+) of Gardnerella vaginalis isolated from the vagina (as a surrogate for BV) demonstrated a two-thirds reduction in the rate of preterm birth among women who had had a prior preterm birth and 35% reduction among women without this history, s In each of these cited studies, the beneficial effects were demonstrated in asymptomatic women; symptomatic subjects were treated outside of these protocols and eliminated from analysis. 47'49's We conducted a large controlled, prospective evaluation of "screening and treating" prevalent infections in order to reduce preterm birth in Denver, CO. 17  among "high risk" women. Neither of these models takes into account costs incurred after neonatal discharge or increased liability costs owing to avoidable adverse pregnancy outcomes, s Other genitourinary tract infections such as chlamydia endocervicitis, trichomoniasis, and gonorrhea as well as asymptomatic bacteriuria are already sought out and treated as "standard of care" in most U.S. centers. Partners of pregnant women with sexually transmitted diseases should be treated and tests of cure should be performed as appropriate. Tests of cure for infected pregnant women and their partners with sexually transmitted disease are more urgent during pregnancy so as INFECTIOUS DISEASES IN OBSTETRICS AND GYNECOLOGY to 1) maximally reduce risks of preterm birth and 2) eliminate risks of neonatal ophthalmia and pneumonitis caused by "family pathogens." These costs are already provided for in contemporary obstetric care.

CONCLUSIONS
Preterm birth continues as an urgent international health priority as well as a widely accepted measure of effective health delivery. Children born with the biologic disadvantages of prematurity necessitate intensive care and expenditures of immense human and economic resources. Surviving children frequently lead lives of diminished personal and economic potential. Clearly, the morbidity and excess costs caused by preterm birth are better prevented or mitigated prior to birth than dealt with in intensive care settings, specialized schools, and sustaining social programs.
On the basis of epidemiologic, clinical, microbiologic, and biochemical evidence, there is now adequate evidence that 1) reproductive tract infection and subsequent inflammation cause significant numbers of women to suffer preterm labor, pPROM, and preterm birth; 2) these adverse effects are caused by infection and inflammation of upper reproductive tract organs and tissues (i.e., placenta, decidua, amnionchorion); 3) the majority of these infections arise from lower reproductive tract sources including BV, cervicitis, or abnormal colonization (enteropharyngeal pathogens, including coliforms); and 4) many of these preterm births are preventable with prompt diagnosis and systemic (oral) antibiotic treatment during pregnancy. Future studies will focus on more accurately defining women and babies at risk, elucidating pathogenic interactions between microbes and both maternal and fetal hosts, and refining diagnostic and treatment strategies so as to provide maximum benefits while incurring minimal adverse effects and costs.
Medical care providers now have new opportunities and obligations to prevent as many infectionmediated births as possible by identifying and treating prevalent reproductive tract infections in their patients. We wager that optimal approaches will be shown to include identification and treatment of susceptible women prior to pregnancy, as part of preconceptual counseling and care.