Maternal and Perinatal Infection in Pregnancy: Bacterial

Epidemiology

Bacterial infections are the single most common medical complication encountered by the obstetrician. Some infections, such as puerperal endometritis and lower urinary tract infection, are of principal concern to the mother and pose little or no risk to the fetus or neonate. Others, such as listeriosis and group B streptococcal (GBS) infection, are of greatest concern to the fetus. Still others—such as pyelonephritis, chorioamnionitis, and syphilis—may cause serious morbidity, even life-threatening complications, for both the mother and baby. The purpose of this chapter is to review in detail the major bacterial infections and the key protozoan infection (toxoplasmosis) that the obstetrician confronts in daily clinical practice.

Pathogenesis

Chlamydiae are obligate intracellular bacteria separated into their own order, Chlamydiales, on the basis of a unique growth cycle that distinguishes them from all other microorganisms. This cycle involves infection of the susceptible host cell by a chlamydia-specific phagocytic process, so that these organisms are preferentially ingested. After attachment and ingestion, the chlamydiae remain in a phagosome throughout the growth cycle, but surface antigens of chlamydiae appear to inhibit phagolysosomal fusion. These two virulence factors (i.e., enhanced ingestion and inhibition of phagolysosomal fusion) attest to an exquisitely adapted parasite.

Once in the cell, the chlamydial elementary body , which is the infectious particle, changes to a metabolically active replicating form called the reticulate body , which synthesizes its own macromolecules and divides by binary fission. Chlamydiae are energy-dependent parasites; because they do not synthesize their own adenosine triphosphate, energy-rich compounds must be supplied to them by the host cell. By the end of the growth cycle, which lasts for approximately 48 hours, most reticulate bodies have reorganized into elementary bodies, which are released through mechanical disruption of the host cell to initiate a new infection cycle.

Controversy exists as to whether maternal cervical C. trachomatis infection is associated with adverse pregnancy outcome. Although some studies have demonstrated an association of maternal chlamydial infection with preterm birth, low birthweight, PPROM, and perinatal death, others have failed to confirm such an association . Harrison et al. and Sweet and associates demonstrated that a subgroup of infected women, in whom immunoglobulin M (IgM) antibody against C. trachomatis was present, were at significantly increased risk for PPROM, preterm birth, and delivery of a low-birthweight infant. These authors postulated that IgM seropositivity reflected recent acquisition and acute chlamydial infection, which may play a more important role than chronic infection.

In a historical control study, Ryan and colleagues reported that untreated chlamydia-infected pregnant women in a high-prevalence population (21% positive) had a significantly increased incidence of PPROM and of low-birthweight infants and decreased perinatal survival compared with treated women or women not infected with Chlamydia . Cohen and coworkers reported that treatment of chlamydial infection resulted in decreased rates of preterm birth, PPROM, preterm labor, and fetal growth restriction. There were experimental design flaws or limitations in both studies. However, because it is unethical to conduct a prospective, randomized, placebo-controlled trial in which some patients are not treated, these studies provide the best available evidence of the adverse perinatal effects of untreated chlamydial infection. A recent report from Western Australia by Reekie et al. showed that, when corrections are made for confounding variables, if women are adequately treated and do not develop recurrence, they are not at increased risk for spontaneous preterm birth, intrauterine growth restriction, or stillbirth.

Diagnosis

Before the introduction of nucleic acid amplification tests (NAATs), antigen-detection methods were widely used to diagnosis chlamydial infections. To a large extent, these tests have now been replaced by DNA/RNA-based methods, both nonamplified and amplified types. Nonamplified tests, such as the Gen-Probe PACE-2 assay, use DNA/TNA hybridization technology. In a large multicenter study, Black and coauthors reported that the sensitivity of PACE-2 ranged from 60.8% to 71.6%, and the specificity ranged from 99.5% to 99.6%. An important advantage of DNA probe–based testing is that it can be used in conjunction with a probe for the detection of Neisseria gonorrhoeae in a single swab.

More recently, DNA/RNA amplification technology has been introduced into clinical practice. NAATs have excellent sensitivity and specificity for chlamydial testing. Clinically available NAATs include tests based on polymerase chain reaction (PCR) transcription-mediated amplification, and strand displacement amplification. NAATs have performed better than culture, antigen detection, or DNA probe techniques for detection of C. trachomatis . A major advantage of NAATs is their ability to identify patients with a low inoculum of C. trachomatis . However, the use of a vaginal swab has been shown to have equivalent or better sensitivity and specificity and is better accepted by patients, especially when patient-obtained specimens are used.

The CDC recommends that all pregnant women be routinely tested for C. trachomatis at their first prenatal visit. Women at increased risk for chlamydial infection should be retested during the third trimester to prevent maternal postnatal complications and chlamydial infection in the infant.

Treatment

The most appropriate regimen for treatment of chlamydia in pregnancy is a single 1 g oral dose of azithromycin. This regimen is inexpensive and is usually very well tolerated. An alternate regimen is 500 mg oral amoxicillin 3 times daily for 7 days.

In light of the sequelae that can occur in the mother and neonate if chlamydial infection persists, repeat testing, preferably by NAATs, is recommended for all pregnant women 3 weeks after completion of therapy to ensure cure. Sex partners should be referred for evaluation, testing, and treatment. The CDC suggests that if concerns exist that sex partners will not seek evaluation and treatment, consideration should be given for delivery of antibiotic therapy by female patients to their sex partners. This approach decreases the rate of persistent or recurrent Chlamydia compared with standard partner referral.

Epidemiology

Streptococcus agalactiae is a gram-positive encapsulated coccus that produces β-hemolysis when grown on blood agar. On average, 20% to 25% of pregnant women in the United States harbor this organism in their lower genital tract and rectum. GBS is one of the most important causes of early-onset neonatal infection . The prevalence of neonatal GBS infection currently is approximately 0.5 per 1000 live births, and approximately 10,000 cases of neonatal streptococcal septicemia occur each year in the United States.

Neonatal GBS infection can be divided into early-onset and late-onset infection, and Table 58.1 summarizes characteristics of both. Eighty to 85% of cases of neonatal GBS infection are early in onset and result almost exclusively from vertical transmission from a colonized mother . Early-onset infection presents primarily as a severe pneumonia and overwhelming septicemia. In preterm infants, the mortality rate from early-onset GBS infection may approach 25%. In term infants, the mortality rate is lower, averaging approximately 5%. Current nationwide prevention strategies have decreased the number of cases of early-onset GBS sepsis by approximately 3900 per year and the number of deaths by 200 per year.

Late-onset neonatal GBS infection occurs as a result of both vertical and horizontal transmission . Of note, 16% to 45% of nursery personnel are carriers of GBS, and nosocomial acquisition in newborns is common. Late-onset infection is typically manifested by bacteremia, meningitis, and pneumonia. The infection also may be localized in the middle ear, sinuses, conjunctiva, lungs, bones, breasts, joints, and skin. The mortality rate from late-onset infection is approximately 5% for both preterm and term infants. Unfortunately, obstetric interventions have proved ineffective in preventing late-onset neonatal infection . Therefore the remainder of this discussion focuses on early-onset infection.

Major risk factors for early-onset infection include preterm labor, especially when complicated by PPROM; intrapartum maternal fever (usually the result of chorioamnionitis); prolonged rupture of membranes, defined as greater than 18 hours; previous delivery of an infected infant; young age; and black or Hispanic ethnicity . Approximately 25% of pregnant women have at least one risk factor for GBS infection. The neonatal attack rate in colonized patients is 40% to 50% in the presence of a risk factor and less than 5% in the absence of a risk factor. In infected infants, neonatal mortality approaches 30% to 35% when a maternal risk factor is present but is less than 5% when a risk factor is absent.

Maternal Complications

Several obstetric complications occur with increased frequency in pregnant women who are colonized with GBS. The organism is one of the major causes of chorioamnionitis and postpartum endometritis . It may cause postcesarean delivery wound infection, usually in conjunction with bacteria from the skin and pelvic flora. The organism also is responsible for 2% to 3% of lower urinary tract infections in pregnant women. GBS urinary tract infection, in turn, is a risk factor for PPROM and preterm labor. For example, Thomsen and colleagues reported a study of 69 women at 27 to 31 weeks’ gestation who had streptococcal urinary tract infections. Women were assigned to treatment with either penicillin or placebo. Treated patients had a significant reduction in the frequency of both PPROM and preterm labor.

Other investigations have confirmed the association between GBS colonization and preterm labor and PPROM. Women with the latter complication who are colonized with GBS tend to have a shorter latent period and higher frequency of chorioamnionitis and puerperal endometritis compared with noncolonized women. Approximately 20% of patients with chorioamnionitis or endometritis will have GBS as part of the complex microbiological flora causing the infection.

Diagnosis

The gold standard for the diagnosis of GBS infection is bacteriologic culture. Todd-Hewitt broth, Lim broth, and Trans-Vag broth are the preferred media. These broths are used to enhance the growth of a low bacterial inoculum before the specimen is subcultured on sheep's blood agar. Specimens for culture should be obtained from the lower vagina, perineum, and perianal area, using a simple cotton swab.

In recent years, considerable research has been devoted to assessment of rapid diagnostic tests for the identification of colonized women. Table 58.2 summarizes the results of several investigations of rapid diagnostic tests. The information in this table is primarily based on the review by Yancey and associates and Ahmadzia et al. The former authors noted that, although the rapid diagnostic tests had reasonable sensitivity in identifying heavily colonized patients, they had poor sensitivity in identifying lightly and moderately colonized patients.

Although the first-generation rapid diagnostic tests were not as valuable as originally hoped, Bergeron and colleagues reported exceptionally favorable results with a new PCR assay for GBS. In a series of 112 patients, the authors documented a sensitivity of 97%, specificity of 100%, positive predictive value of 100%, and negative predictive value of 99%. This PCR assay is currently commercially available and offers clear promise as a rapid test for screening patients at the time of admission for labor. Ahmadzia and colleagues reviewed the diagnostic accuracy of intrapartum NAATs for GBS and reported that the sensitivity ranged from 91% to 94%. These newer tests are particularly accurate when the specimen is initially incubated in a selective medium. Of course, this incubation step delays the ultimate test result.

Prevention of Group B Streptococcal Infection

In the past 20 years, several strategies have been proposed for the prevention of neonatal GBS infection. Each strategy has had imperfections. However, in 1996 the CDC published a series of recommendations that incorporated the major advantages of previous protocols and minimized some of the more problematic aspects of selected strategies. The initial CDC guidelines recommended either universal culturing of all patients at 35 to 37 weeks’ gestation and intrapartum treatment of all colonized women or selective treatment on the basis of identified risk factors. Subsequently, in a large population-based survey, Rosenstein and Schuchat assessed the theoretic impact of the CDC recommendations and showed that a strategy of universal culturing and treatment of all colonized patients would prevent 78% of cases of neonatal infection. In contrast, only 41% of cases were prevented when patients were targeted for prophylaxis just on the basis of risk factors. In addition, Locksmith and colleagues confirmed that universal culturing also was of value in decreasing the rate of maternal infection compared with a strategy of only treating on the basis of risk factors.

In 2019, the ACOG issued its most recent guidelines for the prevention of early-onset GBS infection. The newest guidelines recommend universal cultures in all patients as the optimal method of prevention. Cultures should be performed at 36 to 37 weeks’ gestation. All patients who test positive should receive intrapartum antibiotic prophylaxis, as outlined in Fig. 58.1 . In brief, penicillin G is the drug of choice, 5 million units initially, then 2.5 to 3.0 million units every 4 hours intravenously until delivery. Ampicillin is an equally effective alternative, 2 g initially, then 1 g every 4 hours until delivery. For patients with a nonlife-threatening reaction to penicillin, the drug of choice is cefazolin, 2 g initially, then 1 g every 8 hours until delivery. If the patient is at high risk for anaphylaxis and the organism is known to be susceptible, 900 mg clindamycin intravenously every 8 hours may be used. However, 10% to 15% of strains of GBS are resistant to clindamycin. Therefore if the sensitivity is not known and the patient is at high risk of anaphylactic reaction to β-lactam antibiotics, she should receive vancomycin 15 mg/kg every 8 hours, not to exceed 2 g in a single dose, infused over 1 to 2 hours.

Ideally, antibiotics should be administered at least 4 hours before delivery. DeCueto and coworkers demonstrated that the rate of neonatal GBS infection was reduced significantly when patients were treated for at least this long. In a subsequent investigation to assess timing of antibiotic prophylaxis, McNanley and associates showed that mean vaginal GBS counts decreased fivefold within 2 hours of antibiotic administration, 50-fold within 4 hours, and almost a thousand-fold within 6 hours.

The new CDC guidelines also addressed issues that previously had been imprecisely defined. Colonized patients scheduled for a planned cesarean delivery do not require intrapartum prophylaxis. Patients who tested positive for GBS in a previous pregnancy should not be assumed to be colonized and should be retested with each pregnancy . This recommendation is supported by a later report from Edwards and coworkers. These authors noted that only 59% of patients who were culture positive in a previous pregnancy were positive in the current pregnancy. Conversely, however, patients who have GBS bacteriuria in pregnancy, even if treated, should be considered heavily colonized and should be targeted for intrapartum prophylaxis. Moreover, patients who had a previous infant with GBS infection also should be considered colonized and should be treated during labor. Table 58.3 summarizes indications and nonindications for intrapartum prophylaxis against GBS.

Acute Urethritis

Acute urethritis, or acute urethral syndrome, is usually caused by one of three organisms: coliforms (principally Escherichia coli ), N. gonorrhoeae , and C. trachomatis . Coliform organisms are part of the normal vaginal and perineal flora and may be introduced into the urethra during intercourse or when wiping after defecation. N. gonorrhoeae and C. trachomatis are sexually transmitted pathogens. Patients with urethritis typically experience frequency, urgency, and dysuria. Hesitancy, dribbling, and a mucopurulent urethral discharge also may be present. On microscopic examination, the urine usually has white blood cells, but bacteria may not be consistently present. Urine cultures may have low colony counts of coliform organisms, and cultures of the urethral discharge may be positive for gonorrhea and chlamydia. A rapid diagnostic test, such as an NAAT, is now the preferred method for identification of gonorrhea and chlamydia.

Most patients with acute urethritis warrant empiric treatment before the results of laboratory tests are available. Infections caused by coliforms usually respond to the antibiotics described later for the treatment of asymptomatic bacteriuria and cystitis. If gonococcal infection is suspected, the patient should be treated with intramuscular (IM) ceftriaxone (250 mg in a single dose) plus 1000 mg oral azithromycin. If the patient is allergic to β-lactam antibiotics, an effective alternative is azithromycin 2000 mg orally in a single dose. This high dose of azithromycin is more likely to be associated with gastrointestinal side effects than more conventional lower doses. An alternative choice in the penicillin-allergic patient is ciprofloxacin 500 mg orally in a single dose. If chlamydial infection is suspected or confirmed, the patient should be treated with azithromycin 1000 mg in a single dose.

Asymptomatic Bacteriuria and Acute Cystitis

The prevalence of asymptomatic bacteriuria in pregnancy is 5% to 10%, and most cases antedate the onset of pregnancy . The frequency of acute cystitis in pregnancy is 1% to 3%. Some cases of cystitis arise de novo, whereas others develop as a result of failure to identify and treat asymptomatic bacteriuria.

E. coli is responsible for at least 80% of cases of initial infections and approximately 70% of recurrent cases . Klebsiella pneumoniae and Proteus species are also important pathogens, particularly in patients who have a history of recurrent infection. Up to 10% of infections are caused by gram-positive organisms, such as GBS, enterococci, and staphylococci.

All pregnant women should have a urine culture at their first prenatal appointment to detect preexisting asymptomatic bacteriuria. If the culture is negative, the likelihood of the patient subsequently developing an asymptomatic infection is less than 5% . If the culture is positive—defined as greater than 10 ⁵ colonies/mL of urine from a midstream, clean-catch specimen—prompt treatment is necessary to prevent ascending infection. In the absence of effective treatment, approximately one-third pregnant women with asymptomatic bacteriuria will develop acute pyelonephritis. In a recent report, treatment of asymptomatic bacteriuria was also associated with a reduction in the incidence of low-birthweight infants (relative risk [RR], 0.66; 95% confidence interval [CI], 0.49 to 0.89), but no difference in preterm delivery was seen.

Patients with acute cystitis usually have symptoms of frequency, dysuria, urgency, suprapubic pain, hesitancy, and dribbling. Gross hematuria may be present, but high fever and systemic symptoms are uncommon. In symptomatic patients, the leukocyte esterase and nitrate tests are usually positive. When a urine culture is obtained, a catheterized sample is preferred because it minimizes the probability that urine will be contaminated by vaginal flora. With a catheterized specimen, a colony count greater than 10 ² colonies/mL is indicative of infection.

Asymptomatic bacteriuria and acute cystitis characteristically respond well to short courses of oral antibiotics. Single-dose therapy is not as effective in pregnant women as in nonpregnant patients. However, a 3-day course of treatment appears to be comparable to a 7- to 10-day regimen for an initial infection . Longer courses of therapy are more appropriate for patients with recurrent infections. Table 58.4 lists several antibiotics of value for treatment of asymptomatic bacteriuria and cystitis.

When sensitivity tests are available (e.g., for patients with asymptomatic bacteriuria), they may be used to guide antibiotic selection. When empiric treatment is indicated, the choice of antibiotics should be based on established patterns of susceptibility. In recent years, 20% to 30% of strains of E. coli and more than half the strains of Klebsiella have developed resistance to ampicillin . Thus this drug should not be used when the results of sensitivity tests are unknown unless the suspected pathogen is enterococci, in which case ampicillin or amoxicillin is the drug of choice.

When choosing among the drugs listed in Table 58.4 , the clinician should consider several factors. First, the sensitivity patterns of ampicillin, amoxicillin, and cephalexin are the most variable. Second, these drugs—along with amoxicillin-clavulanic acid—also have the most pronounced effect on normal bowel and vaginal flora and, thus, are the most likely to cause diarrhea or monilial vulvovaginitis. In contrast, nitrofurantoin monohydrate has only minimal effect on vaginal and bowel flora. Moreover, it is more uniformly effective against the common uropathogens, except for Proteus species, than trimethoprim-sulfamethoxazole. Third, amoxicillin-clavulanic acid and trimethoprim-sulfamethoxazole usually are the best empiric agents for treatment of patients with suspected drug-resistant pathogens. However, sulfonamides should be avoided in the first trimester of pregnancy because of possible teratogenicity, and they should be avoided immediately prior to delivery because of concern about displacement of bilirubin from protein binding sites, with resultant neonatal jaundice. The investigation by Crider et al. noted an association between sulfonamide use in the first trimester and anencephaly (adjusted odds ratio [aOR], 3.4, 95% CI, 1.3 to 8.8), coarctation of the aorta (aOR, 2.7, 95% CI, 1.3 to 5.6), hypoplastic left heart (aOR, 3.2, 95% CI, 1.3 to 7.6), choanal atresia (aOR, 8.0, 95% CI, 2.7 to 23.5), transverse limb deficiency (aOR, 2.4, 95% CI, 1.0 to 5.9), and diaphragmatic hernia (aOR, 2.4, 95% CI, 1.1 to 5.4). This same study also showed an association between the use of nitrofurantoin in the first trimester and anopthalmia or microophthalmos (aOR, 3.7, 95% CI, 1.1 to 12.2), hypoplastic left heart syndrome (aOR, 4.2, 95% CI, 1.9 to 9.1), atrial septal defect (aOR, 1.9, 95% CI, 1.1 to 3.4), and cleft lip with cleft palate (aOR, 2.1, 95% CI, 1.2 to 3.9).

For patients who have an initial infection and experience a prompt response to treatment, a urine culture for test of cure may not be clinically necessary or cost effective. Cultures during or immediately after treatment are indicated for patients who have a poor response to therapy or who have a history of recurrent infection. During subsequent clinic appointments, the patient's urine should be screened for nitrites and leukocyte esterase. If either of these tests is positive, repeat urine culture and retreatment are indicated.

Acute Pyelonephritis

The incidence of pyelonephritis in pregnancy is 1% to 2%. Most cases develop as a consequence of undiagnosed or inadequately treated lower urinary tract infection. Two major physiologic changes occur during pregnancy that predispose to ascending infection of the urinary tract. First, the high concentration of progesterone secreted by the placenta has an inhibitory effect on ureteral peristalsis. Second, the enlarging gravid uterus often compresses the ureters, particularly the right, at the pelvic brim, thereby creating additional stasis. Stasis, in turn, facilitates migration of bacteria from the bladder into the ureters and renal parenchyma ( Fig. 58.2 ).

Seventy-five to 80% of cases of pyelonephritis occur on the right side, 10% to 15% are left sided, and a slightly smaller percentage are bilateral . E. coli is again the principal pathogen. K. pneumoniae and Proteus species are also important causes of infection, particularly in women with recurrent episodes of pyelonephritis. Highly virulent gram-negative bacilli, such as Pseudomonas, Enterobacter, and Serratia, are unusual isolates except in immunocompromised patients. Gram-positive cocci do not frequently cause upper tract infection. Also, anaerobes are unlikely pathogens unless the patient is chronically obstructed or instrumented.

The usual clinical manifestations of acute pyelonephritis in pregnancy are fever, chills, flank pain and tenderness, urinary frequency or urgency, hematuria, and dysuria. Patients also may have signs of preterm labor, septic shock, and acute respiratory distress syndrome (ARDS). Urinalysis is usually positive for white blood cell casts, red blood cells, and bacteria. Urine colony counts greater than 10 ² colonies/mL in samples collected by catheterization confirm the diagnosis of infection.

Pregnant patients with pyelonephritis may be considered for outpatient therapy if their disease manifestations are mild, they are hemodynamically stable, and they have no evidence of preterm labor . Caution should be applied in treating diabetic women as outpatients as these women may be at risk for diabetic ketoacidosis. If an outpatient approach is adopted, the patient should be treated with agents that have a high level of activity against the common uropathogens. Acceptable oral agents include amoxicillin-clavulanic acid 875 mg twice daily or double-strength trimethoprim-sulfamethoxazole twice daily for 7 to 10 days. Alternatively, a visiting home nurse may be contracted to administer a parenteral agent, such as intravenous (IV) or IM ceftriaxone 2 g once daily. Although an excellent drug for lower tract infections, nitrofurantoin monohydrate does not consistently achieve the serum and renal parenchymal concentrations necessary for successful treatment of more serious infections.

Patients who appear to be moderately to severely ill or who show any signs of preterm labor should be hospitalized for IV antibiotic therapy. They should receive appropriate supportive treatment and should be monitored closely for complications, such as sepsis, ARDS, and preterm labor. One reasonable choice for empiric IV antibiotic therapy is ceftriaxone 2 g every 24 hours. Compared with a first-generation cephalosporin, such as cefazolin, ceftriaxone has expanded coverage against aerobic gram-negative bacilli and has the advantage of once-daily dosing. If the patient is critically ill or is at high risk for a resistant organism, a second antibiotic, such as gentamicin (7 mg/kg ideal body weight every 24 hours) or aztreonam (500 mg to 1 g every 8 to 12 hours), should be administered, along with ceftriaxone, until the results of susceptibility tests are available.

Once antibiotic therapy is initiated, approximately 75% of patients defervesce within 48 hours. By the end of 72 hours, almost 95% of patients are afebrile and asymptomatic. The two most likely causes of treatment failure are a resistant microorganism and obstruction . The latter condition is best diagnosed with computed tomography (CT) scan or renal ultrasonography and typically results from a stone or physical compression of the ureter by the gravid uterus.

Once the patient has begun to defervesce and her clinical condition has improved, she may be discharged from the hospital. Oral antibiotics should be prescribed to complete a total of 7 to 10 days of therapy. Selection of a specific oral agent should be based on considerations of efficacy, toxicity, and expense. A repeat urine culture should be obtained after therapy is completed to ensure that the infection has resolved.

Twenty to 30% of pregnant patients with acute pyelonephritis develop a recurrent urinary tract infection later in pregnancy. The most cost-effective way to reduce the frequency of recurrence is to administer a daily prophylactic dose of an antibiotic, such as 100 mg nitrofurantoin monohydrate. Patients receiving prophylaxis should have their urine screened for bacteria at each subsequent clinic appointment. They also should be questioned about recurrence of symptoms. If symptoms recur, or the dipstick test for nitrite or leukocyte esterase is positive, a urine culture should be obtained to determine whether retreatment is necessary.

Chorioamnionitis