Severe peripartum GAS infections, a scourge of the pre-antibiotic area, virtually disappeared during the mid 20th century, but have been reported sporadically since the early 1990s. While most of the reports have been 1-2 unrelated cases [3–14], some larger collections of patients have been reported. Nathan et al. [3] studied 18 GAS isolates from puerperal infections in the Salt Lake City region between 1991 and 1997; no outbreak strain was identified. By contrast, Raymond et al. [15] described a cluster of 5 cases of puerperal sepsis in which 3 of the 5 infections were clonal. Our 4 cases were separated by 1 year, and there was no epidemiologic list between them or any known GAS infections in their health care workers.

Some epidemiologic work has attempted to quantify the incidence of GAS puerperal infections in the USA. Chuang et al. [16] used CDC data (1995–2000) on 87 postpartum GAS infections from selected US counties to estimate that ~220 cases per year of invasive GAS occur in the USA annually. The mortality rate of the 87 documented cases was 3.5%. Interestingly, they found that cases caused by identical emm types clustered more than chance alone would suggest, raising the possibility that some of these cases may have a common source. Aronoff and Mulla [17] extended the earlier work of Chuang et al. [16] in a study limited to Florida. They found that 4 cases of hospitalized invasive postpartum GAS between 1996 and 2000, which represented 1.6% of all reported invasive GAS infections reported during that period.

The dramatic predisposition of pregnant or postpartum women for significant GAS infections was delineated by Deutscher et al. [18], who found that postpartum women have a 20-fold increased incidence of GAS infections as compared to nonpregnant women. The reasons for this predilection are not clear. GAS vaginal carriage in late pregnancy is only 0.03%, so GAS is clearly not “normal” vaginal flora [19]. Selected Streptococcal M types, especially 28, are overrepresented in GAS puerperal sepsis, suggesting that bacterial properties, in addition to host factors and peripartum changes, contribute to the predisposition to serious GAS infection [20].

Severe nonobstetrical GAS infections are most often individual and community acquired but can be hospital acquired in as many as 12% of cases. In some instances health care workers can serve as the means of transmission between patients [21], and nosocomal outbreaks may occur. Familial transmission of a severe GAS clone has also been reported [22] and explains some community acquired cases. Though rare, health care workers may acquire GAS infection from patients or even cadavers [23].

The clinical features of our 3 cases of severe peripartum sepsis are similar to most prior reports in that the onset was within a few days postpartum, occurred in previously health women, and progressed rapidly from fever and abdominal/pelvic pain to vasopressor dependent shock recurring ICU support for multiorgan failure [3–5, 7–14, 24]. Patients may present with puerperal sepsis due to GAS after discharge from the hospital [19], as one of our patients did; clinicians from emergency department or outpatient settings must be open to the possibility of a life-threatening infection when postpartum women present with unstable vital signs, high fever, or an unexpected “toxic” appearance. None of the cases had any symptoms prior to the second postpartum day, suggesting they were not hematogenous in origin but rather arose from vaginal colonization [25, 26]. Our fourth case, cervical necrosis due to GAS without sepsis, illustrates that not all serious peripartum GAS infections are life-threatening, but all require careful evaluation and aggressive care.

The treatment of severe peripartum sepsis requires antibiotics coverage aimed at the likely causes: GAS, S. aureus, anaerobes, and E. coli [19]. Given that GAS has special potential to be a “bad actor,” the use of clindamycin to block toxin production is recommended in addition to a broad spectrum beta-lactam agent with good gram negative and anaerobic activity (a carbapenem, piperacillin-tazobactam, etc.). MRSA coverage may be needed initially pending cultures [19]. When GAS infection is probable or proven, the addition of intravenous immunoglobulin may be indicated. Though its utility in GAS associated shock is not as firmly established as clindamycin, it may be a key adjunct to care in critical situations [27].

The prevention of puerperal sepsis is problematic. The infection is rare, occurs generally in the health host, and may be fulminant. Screening for vaginal carriage of GAS is unlikely to be helpful as such carriage is rare [19]; however, there are at least 2 cases where GAS vaginal carriage was documented (and ignored) prior to the onset of GAS puerperal sepsis [12]. A vaccine, the optimal prevention, is not yet available. Strict infection control practices will serve to limit the occasional case due to spread between patients or from staff to patients [19].

The CDC has issued guidelines to prevent invasive GAS infections among household contacts of severe GAS cases and to limit hospital outbreaks [28]. Chemoprophylaxis is usually not offered to household contacts. Single cases of postpartum or postsurgical GAS, as described in the report, should lead to enhanced surveillance, while ≥2 postpartum/postsurgical cases demand a full epidemiologic investigation with cultures of the relevant health care workers.

For sporadic single cases, it seems that prevention must take a back seat to prompt recognition and aggressive management of this rare, but sometimes terrifying, infection, of the postpartum period. We hope our cases and this review of the relevant literature serve as a reminder of this emerging infection.