INTRODUCTION — In the mid-nineteenth century, it was common for clinicians to perform autopsies on women who had died of postpartum infection. In the absence of handwashing between autopsies and deliveries, the clinicians could transmit group A Streptococcus (GAS) to laboring women, leading to postpartum infection [1]. During this time, Semmelweis deduced that physicians transmitted infection via their hands to pregnant women during labor and delivery and showed that handwashing could reduce transmission [2]. A century later, maternal mortality was reduced further with the introduction of antiseptics and penicillin [3].

Issues involving pregnancy-related GAS infections will be reviewed here. Issues related to nongynecological GAS infections such as necrotizing soft tissue infection and streptococcal toxic shock are discussed separately. (See "Necrotizing soft tissue infections" and "Surgical management of necrotizing soft tissue infections" and "Invasive group A streptococcal infection and toxic shock syndrome: Epidemiology, clinical manifestations, and diagnosis" and "Invasive group A streptococcal infection and toxic shock syndrome: Treatment and prevention".)

EPIDEMIOLOGY — Invasive GAS infections re-emerged in the mid-1980s [4], including those associated with pregnancy and childbirth [5]. In the United States between 2000 and 2004, 5400 cases of invasive GAS infection were identified (3.5 cases per 100,000 persons); the case fatality was rate 13.7 percent [6].

In the United States between 1995 and 2000, the annual incidence of GAS postpartum infection was 6 per 100,000 live births [7].

In London and the South East of England between 2010 and 2016, the incidence of invasive GAS infection among women within 28 days postpartum was 109 per 100,000 [8]. One review in Israel between 2008 and 2015 noted an incidence of 35 cases per 100,000 [9]. Globally, puerperal sepsis causes approximately 75,000 maternal deaths per year [10], with the highest maternal mortality in Asia (11 percent), Africa (9 percent), and Latin America and the Caribbean (7 percent) [11].

The attack rate of invasive GAS infection is 20-fold higher for pregnant and postpartum women compared with nonpregnant women [12,13]. Among patients with pregnancy-related GAS infection, approximately 85 to 93 percent of infections occur postpartum [9,14,15]. In one study including 67 patients with pregnancy-related GAS, 84 percent followed vaginal delivery, and the majority occurred within the first four days postpartum (73 percent) [14]. In another report of 24 maternal deaths due to sepsis in Japan between 2010 and 2016, 13 infections were attributable to GAS (54 percent); of these, 77 percent of cases occurred antepartum [16]. Death occurred within 24 hours after admission in 54 percent of cases (median time to death of 12 hours).

Most cases of pregnancy-associated GAS infection are community acquired; approximately 15 to 25 percent of GAS postpartum infections are nosocomially acquired [9,17]. Patients with intrapartum or late postpartum infection are more likely to have had upper respiratory tract GAS infection prior to development of intrauterine infection. Contact with young children during pregnancy represents an underappreciated risk factor for maternal GAS acquisition, since pharyngeal GAS colonization rate is higher in children than adults (25 versus 5 percent). In one study, primiparity was an independent protective factor against puerperal GAS infection, presumably due to diminished exposure to children carrying GAS [15]. Thus, pharyngeal screening for GAS is appropriate for pregnant women with pharyngitis or upper respiratory illness; this is particularly relevant for women who have close contact with young children.

Maternal mortality is highest when infection develops within four days after delivery. However, one case report described a patient with GAS bacteremia during the third trimester of pregnancy who received prompt and aggressive treatment, survived, and went on to have an uneventful delivery [18].

Premature rupture of membranes increases the risk for any postpartum infection [19]. In one study including 161 women, premature rupture of membranes for >12 hours was associated with an infection rate of 13 percent in mothers (chorioamnionitis and endometritis) and an infection rate of 6 percent in newborns. Prophylactic antibiotics reduced the infection rate to 2.6 percent in mothers and 3.8 percent in newborns [19]. (See 'Antibiotic prophylaxis' below.)

Cesarean section is also a risk factor for endometritis [20]. Cesarean during the second stage of labor (complete cervical dilation) has been associated with an increased risk of endometritis compared with cesarean during the first stage of labor (less than 10 cm cervical dilation) [20]. However, Cesarean delivery with perioperative antibiotic prophylaxis is a protective factor against puerperal GAS infection [15].

PATHOGENESIS

Factors related to pregnancy — During pregnancy, upper respiratory tract infection precedes hematogenous seeding of the placenta and uterus, suggesting an important tropism of GAS for these tissues during pregnancy.

The apparent increase in susceptibility to GAS infection among postpartum women could be due to compromised mucosal or cutaneous barriers (eg, an open cervix, vaginal mucosal tears, episiotomy, or cesarean section incision), a transiently more neutral vaginal pH after amniotic fluid release (which could favor growth of the organism) [21], or suppressed immunity during pregnancy. For example, alternatively activated M2 macrophages predominate in the postpartum uterus and are focused on uterine repair [22]; the relative absence of proinflammatory M1 macrophages in this setting could leave the host vulnerable to bacterial pathogens. In addition, a unique population of transitional cells present during uterine regeneration after parturition expresses the filament protein vimentin, a known GAS adhesin [23-25], which could mediate the increased risk of GAS infection in the postpartum setting.

Genetic susceptibility — Immunologic polymorphisms play a role in susceptibility to invasive GAS infection in general [26], though few studies have investigated this relative to postpartum infections specifically. An inherent resistance to GAS infection is suggested by the fact that a much larger proportion of women are colonized with GAS than actually develop infection. Specifically, one study noted that, during the third trimester of pregnancy, 0.03 percent (30 per 100,000) of women had GAS vaginal-rectal colonization [27]. However, GAS postpartum sepsis occurs with a frequency of only 0.006 percent (6 cases per 100,000 live births). These findings suggest that some level of innate or acquired mucosal immunity exists in most pregnant and postpartum women or that not all colonizing strains of GAS are capable of causing infection during pregnancy or after delivery.

Bacterial factors — Development of pregnancy-related GAS infection may depend on the specific M-type of GAS in the patient's environment. Only a handful of GAS strains are associated with invasive GAS infections in pregnant women. The importance of a specific M-type or its associated toxins and virulence factors is yet to be determined.

For example, a population-based study conducted from 1995 to 2000 by the United States Centers for Disease Control and Prevention showed that, among postpartum GAS blood isolates, the most common M-types were M28 (21 percent), M1 (14 percent), and types 4, 11, 12, and 13 (7 percent) [7]. A similar report of 18 cases from Utah demonstrated that severe GAS puerperal infection was solely associated with M-types 1 and 28 [28]. A report from Great Britain described an outbreak of GAS infections on a maternity ward involving several patients and health care workers associated with an M-type 1 strain [29]. A subsequent study from Prague found that M types 1 and 3 were commonly isolated from pregnant women with or without invasive disease; however, M types 12, 28, 75, and 89 were uniquely expressed among strains causing invasive infections [30]. In addition, the incidence of adherence molecule genes was significantly lower, while the iron-chelating protein gene perR was significantly higher, among isolates from invasive infection compared with those from the genital tracts of asymptomatic multiparous women [30]; no significant differences were observed in the superantigen profiles between groups [30].

CLINICAL MANIFESTATIONS — Patients with GAS puerperal sepsis typically present with fever, abdominal pain, with or without hypotension, tachycardia, or leukocytosis. Fever of greater than 38.5°C develops within the first 48 hours postpartum and persists for more than 4 hours. Abdominal tenderness is also a prominent feature. These manifestations can be difficult to evaluate since postpartum abdominal pain is common in the absence of infection, and leukocytosis can be significant even in a normal childbirth. Other significant clinical manifestations are summarized in the table (table 1).

GAS can cause invasive infections such as endometritis, necrotizing fasciitis, or streptococcal toxic shock syndrome (TSS) [31]. The involved areas may include the uterus, vagina, and external genitalia as well as other sites. In one study including 430 women with postpartum infection within 30 days of delivery, sites of infection included breast, wound, vagina, urinary tract, respiratory tract, and endometrium [32,33]. Patients who develop signs and symptoms of capillary leakage have a 3.5 times higher relative risk of an adverse outcome compared with those without such findings [34].

Clinical clues for severe invasive GAS infection include sudden onset of shock and organ dysfunction, including renal failure and acute respiratory distress syndrome. Hypotension, tachycardia, and leukocytosis are signs of developing streptococcal toxic shock syndrome (TSS) and are associated with higher mortality; in one study, TSS occurred in approximately 20 percent of patients [9]. (See "Invasive group A streptococcal infection and toxic shock syndrome: Epidemiology, clinical manifestations, and diagnosis".)

During pregnancy — GAS infection may be ascending (originating in the vagina) or descending (originating the respiratory tract). In one series, most patients had normal pregnancies until the sudden onset of symptoms in third trimester; one patient presented during the first trimester with streptococcal TSS and had a spontaneous abortion [14].

Four patients required emergent cesarean section and, of these, two died; both had necrosis and/or inflammation of the uterus at autopsy. One of the two survivors required bilateral salpingo-oophorectomy. Three patients delivered vaginally and two of these mothers and their infants died. The method of delivery was not given in two patients; of these, one mother and baby died. Overall, fetal and maternal mortalities were 75 percent and 60 percent, respectively.

Initial clinical signs and symptoms include fever (78 percent), hypotension (56 percent), abdominal pain (44 percent), and tachycardia (44 percent) [14]. A prodrome of sore throat or upper respiratory tract infection was reported in 56 percent of patients. GAS was cultured primarily from the blood (78 percent) and/or respiratory tract (44 percent).

GAS isolated from the uterus in two fatal cases were M-types 1 and 3 [35]; these are the most common M-types associated with streptococcal TSS [36-39]. (See "Group A streptococcus: Virulence factors and pathogenic mechanisms", section on 'M and M-like proteins'.)

Postpartum

0 to 2 days — Among 26 patients with GAS infection 0 to 2 days postpartum, GAS was isolated from the vagina or urinary tract in 77 percent of cases, suggesting that infection resulted from prior or hospital-associated vaginal colonization [14]. The incidence of vertical transmission and infant mortality were low (11 percent and 5 percent, respectively).

A broad spectrum of GAS postpartum infections can develop rapidly, ranging from mild endomyometritis (with absence of tachycardia and leukocytosis) to fulminant endomyonecrosis and death. Clinical findings included abdominal pain (58 percent), purulent vaginal discharge (38 percent), and uterine tenderness (31 percent) [14]. Gastrointestinal symptoms (diarrhea, nausea, vomiting) occurred in 31 percent of cases. Systemic symptoms and signs included fever (73 percent), chills (35 percent), and hypotension (35 percent). Leukocytosis and tachycardia were not frequent findings. The highest postpartum maternal mortality was seen during the first two days postpartum (25 percent). Women who underwent abdominal surgery (31 percent) had necrosis of the ovaries, fallopian tubes, and/or uterus.

Maternal-to-infant transmission has been reported [39,40]; infections in infants ranged from necrotizing fasciitis of the scalp to respiratory distress.

3 to 4 days — Among 11 patients with GAS infection 3 to 4 days postpartum, ascending infection (from the vagina) likely followed urogenital acquisition of GAS from either the hospital or home environment. Clinical signs and symptoms as well as primary sites of bacterial isolation were consistent with ascending infection. Patients had hypotension (36 percent) and fever (73 percent). Leukocytosis (36 percent) and tachycardia (46 percent) were observed, though gastrointestinal symptoms were relatively uncommon. Exploratory surgery was performed in approximately one-third of cases. Surgical findings included necrosis, inflammation, and/or exudates of the ovaries, fallopian tubes, and/or uterus. One woman died.

5 to 8 days — Among seven patients with GAS infection 5 to 8 days postpartum, evidence of both ascending (from the vagina) and descending (from the respiratory tract) infections was observed [14]. These patients were more likely to have systemic involvement and some had hematogenous seeding of the extremities. Given the timeframe of infection onset, it is likely that patients in this group acquired GAS in the community.

Women presented with fever, hypotension, and leukocytosis. Diffuse erythema (scarlatina) or erythema associated with extremity pain were highest in this group. Three patients (43 percent) had debridement, drainage, or amputation of one or more extremities, suggesting spontaneous necrotizing fasciitis/myonecrosis and pyomyositis possibly related to antecedent soft tissue injury [24].

>8 days — Among seven patients with GAS infection 2 to 5 weeks after childbirth, the heterogeneity of signs and symptoms and sites of bacterial isolation suggest that these patients had infection not directly related to parturition [14]. Clinical findings included gastrointestinal symptoms (72 percent), fever (86 percent), chills (29 percent), and abdominal pain (57 percent). Leukocytosis was present in 57 percent, hypotension in 43 percent, and vaginal discharge in 43 percent.

The peritoneum was the most common site of GAS isolation (57 percent); this occurred more frequently than observed in GAS infection that developed earlier in the postpartum period. GAS was isolated from the respiratory tract (29 percent), blood (43 percent), and vagina or urinary tract (43 percent). One patient had GAS isolated from spinal fluid. Maternal survival was 100 percent; one patient required hysterectomy [36]. No infants died, suggesting again that maternal acquisition of GAS occurred well after delivery.

DIAGNOSIS — Timely diagnosis is critical; once shock develops, mortality approaches 60 percent [14,41].

Diagnosis of GAS infections during pregnancy or early in the postpartum period is often difficult due to a low prevalence of infection and initial nonspecific symptoms. Because postpartum GAS infections are relatively uncommon, physicians often misconstrue the pain of developing infection as typical postpartum discomfort. The widespread use of pain-relieving agents after childbirth can further mask the signs of incubating infection, which can contribute to a delay in diagnosis with catastrophic outcomes [35]. Diagnosis is particularly problematic in pregnant women with premature rupture of membranes.

The occurrence of fever, chills, and abdominal pain in a pregnant or early postpartum woman should prompt consideration of GAS infection. Leukocytosis, hypotension, and tachycardia are signs of developing streptococcal toxic shock syndrome and are associated with higher mortality. A pelvic exam should be performed; vaginal discharge and/or infection at the site of an episiotomy may be detected via direct observation (or with a speculum to visualize the entire vagina and cervix).

Blood and urine cultures should be performed, and postpartum women should have endometrial aspiration for Gram stain and culture.

Marked leukocytosis or leukopenia can be seen. A marked bandemia (greater than 10 percent) may be seen even in the absence of a leukocytosis. Hemolysis or hemoconcentration may occur [1].

Computed tomography scans, magnetic resonance imaging, and ultrasonography may appear normal and should not delay aggressive management; typically, imaging demonstrates an edematous uterus that is larger than expected. GAS does not generally form abscesses or produce gas, whereas postpartum infections caused by Clostridium perfringens or mixed aerobic/anaerobic pathogens are clearly associated with gas in the tissues.

DIFFERENTIAL DIAGNOSIS — The differential diagnosis of pregnancy-related GAS infection includes infection due to C. perfringens and Clostridium sordellii. These clostridial species cause rapidly progressive infections of the uterus, external genitalia, and episiotomy site. Infections with these organisms are relatively uncommon but are associated with severe illness, high mortality, and devastating morbidity [42] (see "Clostridial myonecrosis" and "Toxic shock syndrome due to Clostridium sordellii").

Clinical clues for clostridial infection include lack of fever, hemoconcentration (hematocrit of 60 to 80 percent), leukemoid reaction (white blood cell count of 50,000 to 200,000 cells/microL), and diffuse capillary leak syndrome [34,43]. A strong clinical clue is the presence of gas in the tissue, detectable by clinical findings of crepitus and/or radiographically.

Other organisms associated with postpartum endometritis are discussed separately. (See "Postpartum endometritis".)

TREATMENT — Treatment of pregnancy-related GAS infection consists of aggressive fluid resuscitation, antibiotic therapy, and source control. Prompt intervention is critical; once shock develops, mortality approaches 60 percent [14,41].

Other issues related to management of streptococcal toxic shock syndrome are discussed separately. (See "Invasive group A streptococcal infection and toxic shock syndrome: Treatment and prevention".)

Antibiotics — Antibiotic therapy for treatment of severe GAS infection consists of penicillin G (4 million units intravenously every 4 hours) and clindamycin (900 mg intravenously every 8 hours) [44]. Vancomycin (table 2) or daptomycin (6 mg/kg intravenously every 24 hours) are acceptable alternatives to penicillin for patients with beta-lactam hypersensitivity. This approach is based on in vitro susceptibility data and results from animal studies; there are no trials of antibiotic therapy for pregnancy-related GAS infection in humans.

In vitro antimicrobial susceptibility testing must be performed on invasive strains of GAS, given emergence of clindamycin resistance [45,46] and reports of treatment failures in cases of clindamycin resistance [47]. Tedizolid or linezolid are acceptable alternative agents for treatment of infection due to clindamycin-resistant strains; like clindamycin, they are also capable of suppressing toxin production in GAS [48].

Combination therapy with penicillin and clindamycin should be continued until patients are clinically and hemodynamically stable for at least 48 to 72 hours; thereafter, penicillin monotherapy may be administered.

The duration of antibiotic therapy should be individualized. Patients with bacteremia should be treated for at least 14 days. In patients with complicated deep-seated infection, length of therapy depends on the clinical course and the adequacy of surgical debridement; therapy is usually continued for 14 days from the last positive culture obtained during surgical debridement.

In general, antibiotic regimens for empiric treatment for endometritis include activity against GAS if they include a beta-lactam (penicillin, ampicillin, or cephalosporin), clindamycin, or carbapenem. However, patients who are very ill initially or don’t respond to initial antibiotics should have antibiotic therapy adjusted as described above to target GAS as a possible pathogen. (See "Postpartum endometritis".)

Surgery — There are no trials comparing outcomes between conservative and surgical approaches, but most agree that a confirmed GAS infection in the presence of organ dysfunction should be managed surgically [14,41]. Source control may require wound or vulvar debridement, hysterectomy, or a combination of these. Hysterectomy may be life-saving in puerperal GAS sepsis and should be the default management given the high mortality rate. Patients without sepsis may be managed with antibiotic therapy alone with a plan to move to operative intervention if the condition worsens.

Intravenous immune globulin — We favor administration of intravenous immune globulin for patients with streptococcal toxic shock syndrome. This issue is discussed further separately. (See "Invasive group A streptococcal infection and toxic shock syndrome: Treatment and prevention", section on 'Intravenous immune globulin'.)

PREVENTION

Antibiotic prophylaxis

Cesarean section — Cesarean section is a risk factor for endometritis and administration of prophylactic antibiotics is important to mitigate this risk; this is discussed further separately. (See "Cesarean birth: Preoperative planning and patient preparation", section on 'Antibiotic prophylaxis'.)

Premature rupture of membranes — Premature rupture of membranes is a risk factor for endometritis, and administration of prophylactic antibiotics is important to mitigate this risk; this is discussed further separately. (See "Prelabor rupture of membranes at term: Management", section on 'Antibiotic prophylaxis'.)

Infection control — Approximately 14 percent of GAS postpartum infections are nosocomially acquired [17]. Hospital-associated epidemics of postpartum GAS infection can occur due to transmission from colonized health care workers [49,50]. Epidemiologic investigation with screening of health care personnel is warranted if two or more cases of invasive GAS infection occur at an institution in a six-month period [51].

Patients with endometritis in the absence of toxic shock may be managed with standard precautions. Patients with streptococcal toxic shock warrant droplet precautions (in addition to standard precautions); droplet precautions may be discontinued after the first 24 hours of antimicrobial therapy [52].

GAS screening in pregnancy — The value of screening for GAS during pregnancy is uncertain. GAS colonization of the genital tract during pregnancy is rare (0.03 percent), and most women who develop intrauterine infection due to GAS do not have vaginal GAS colonization [27]. In addition, there is no clear role for pre-emptive treatment of asymptomatic GAS colonization if identified in the course of routine prenatal care. (See "Early-onset neonatal group B streptococcal disease: Prevention", section on 'Management of incidental findings noted on culture report'.)

SUMMARY AND RECOMMENDATIONS

●Pregnant and postpartum women have a 20-fold increase in attack rate for invasive group A streptococcal (GAS) infection compared with nonpregnant women. (See 'Introduction' above.)

●Risk factors for pregnancy-related GAS infection include upper respiratory tract GAS infection and premature rupture of membranes. Contact with young children during pregnancy represents an underappreciated risk factor for maternal GAS acquisition. (See 'Epidemiology' above.)

●Pathogenesis may include factors related to pregnancy (compromised mucosal barriers and/or altered vaginal pH), host genetic susceptibility, and bacterial factors (M-types). (See 'Pathogenesis' above.)

●Clinical manifestations include fever, abdominal pain, and hypotension with or without tachycardia or leukocytosis. Hypotension, tachycardia, and leukocytosis are signs of developing streptococcal toxic shock syndrome (TSS) and are associated with higher mortality. Development of signs and symptoms of capillary leakage is strongly associated with adverse outcome. (See 'Clinical manifestations' above.)

●A bimanual pelvic exam should be performed to detect vaginal discharge and/or infection at the site of an episiotomy. Endometrial aspiration for Gram stain and culture, blood cultures, and urine culture should be performed. (See 'Diagnosis' above.)

●Treatment of pregnancy-related GAS infection consists of aggressive fluid resuscitation, antibiotic therapy, and source control. Prompt intervention is critical. (See 'Treatment' above.)

●We suggest that patients with pregnancy-related GAS infection be treated with combination antibiotic therapy including penicillin G and an antibiotic that suppresses protein synthesis such as clindamycin (rather than penicillin G alone) (Grade 2C). Antibiotic sensitivity testing is critically important in light of emerging clindamycin and erythromycin resistance. (See 'Antibiotics' above.)

●Source control may require wound or vulvar debridement, hysterectomy, or a combination of these. Patients without sepsis may be managed with antibiotic therapy alone with plan to move to operative intervention if the condition worsens. (See 'Surgery' above.)

●For patients with streptococcal TSS, we suggest administering intravenous immune globulin (Grade 2C). This is discussed further separately. (See "Invasive group A streptococcal infection and toxic shock syndrome: Treatment and prevention", section on 'Intravenous immune globulin'.)