Introduction

Group C and G Streptococcus (GCGS) are catalase-negative, gram-positive cocci that form large colonies with clear haemolysis on blood agar, similar to Streptococcus pyogenes (group A Streptococcus, GAS) [1,2,3]. GCGS human isolates are now grouped together as Streptococcus dysgalactiae subsp. equisimilis (SDSE) [1, 4]. They are normal commensal flora of the oral cavity and known colonisers of the skin, the gastrointestinal tract, and the female genitourinary tract [5, 6]. SDSEs are now widely acknowledged as significant pathogens that cause a myriad of infections, ranging from cellulitis to invasive infections, including bacteraemia and even toxic shock syndrome (TSS) [7, 8]. Large epidemiological studies have been hindered by taxonomic obscurity [1]. Several studies have reported the incidence of group C Streptococcus and group G Streptococcus individually, while others have reported it as another β-hemolytic Streptococcus (distinct from Streptococcus pyogenes and Streptococcus agalactiae) [9]. SDSEs express M proteins, which are highly genetically related to S. pyogenes, sharing several virulence factors and causing a similar range of clinical infections [2, 3, 8, 10]. The incidence of bacteraemia due to SDSE has been on the rise worldwide, but the reason for this is unclear; this may be related to the clonal expansion of a few strains carrying certain virulence factors [9, 11,12,13].

Several clinical case series have shown that invasive forms of this infection are more common in elderly polymorbid individuals, with reported mortality rates ranging from 15 to 18% [14]. They have also been reported to cause infective endocarditis (IE), although they were initially excluded as typical pathogens in the modified Duke criteria published in 2000 for IE [15]. However, the updated 2023 Duke-International Society for Cardiovascular Infectious Diseases (Duke-ISCVID) criteria now include SDSE as a typical microorganism causing IE [16]. Data on IE due to SDSE have been scarce, with quite a few studies dating back to the 1990s. In a few case series from the 1980s, the incidence of IE has been reported to be as high as 25 to 50% [6]. The incidence of IE may be underreported due to the relative infrequency of investigating SDSE bacteraemia for endocarditis. Lother et al. (2017) reported that transthoracic echocardiogram (TTE) was only performed in one-third of patients with SDSE infection, of whom 6% had endocarditis, with over half having embolic complications [11]. In a Danish study analysing 901 cases of SDSE bacteraemia over a 10-year period (2008–2017), IE was found in 6.4% of cases [17].

The aim of this study was to examine the epidemiology and disease manifestations of bacteraemia due to SDSE in Western Sydney over a 6-year period (2015–2020) by examining common presentations.

Materials and methods

Case definition

The criteria for inclusion of patients in the study were based on the isolation of group C, group G Streptococcus (GCGS) or Streptococcus dysgalactiae subsp. equisimilis (SDSE) via blood culture. The identification was as reported by the laboratory based on MALDI results, β -haemolysis, and large colony sizes (≥ 0.5 mm). Hence, the Streptococcus anginosus group, which can be grouped with Lancefield C and/or G on latex testing, was excluded.

Inclusion criteria

  • Age ≥ 18 years.

  • Presenting to four Western Sydney hospitals (Westmead, Blacktown, Mt Druitt, and Auburn) during the study period (January 2015 to December 2020 inclusive).

Exclusion criteria

  • Polymicrobial bacteraemia.

  • Mismatches between patient records in the initial database and electronic medical records (eMRs).

The Institute of Clinical Pathology and Medical Research (ICPMR) at Westmead Hospital is a public health reference laboratory providing pathology services to a large population in Western Sydney. The microbiology department at Westmead Hospital (Sydney, Australia) maintains a database of all positive blood cultures. We extracted samples from patients with GCGS or SDSE bacteraemia. The database included blood culture collection date, principal manifestation, antimicrobial therapy, and mortality date. Missing and additional data were collected by accessing eMRs, which have been in use in the study hospitals since 2015. These data included demographic information, admission details, comorbidities, number of positive blood cultures, and additional investigations, including echocardiography, use of antibiotics, their route, and duration. The 7-day and 30-day mortality rates were also reviewed. Chronic diseases were defined as per the treating physician and were grouped per organ system affected as recorded in the eMR. Obesity was only recorded as a chronic medical condition if it was documented on eMR and not based on individual BMI calculations. Healthcare-associated infections (HAIs) were defined as bacteraemia occurring at least 48 h after hospital admission. Postoperative skin and soft tissue infections within 30 days of surgery were included in HAI.

The data were entered into Microsoft Excel. Statistical analyses, including statistical tests (chi-square test or Fisher’s exact test) and descriptive analysis (mean/percentage to characterise the data), were performed. A p value < 0.05 was regarded as statistically significant. The study was approved by the Western Sydney Local Health District Research Ethics Committee (HREA 2021/ETH00080).

Results

A total of 114 cases of SDSE bacteraemia were recorded in adults aged ≥ 18 years during the study period (2015–2020). We excluded four patients with polymicrobial bacteraemia (two with Staphylococcus aureus, one with Acinetobacter spp., one with Clostridium septicum) and two patients with no matching electronic records. The data of the remaining 108 patients were reviewed, of which 20 were reported as GCS (group C Streptococcus), 63 were GGS (group G Streptococcus) and 24 were reported to be SDSE. The baseline characteristics are described in Table 1.

60% of the patients were aged over 65 years, and over half were male (56%). The median age was 70 years (interquartile range [IQR] 58–85 years). Nearly a quarter of the patients were from a residential aged care facility (23%). The majority of the patients had at least one comorbidity (96%), with cardiovascular disease being the most common (46%), followed by chronic skin conditions (44%) and diabetes (37%). Almost one-third of patients (32%) had a history of minor skin trauma prior to hospital admission. The majority of infections were community acquired (91%), with only ten HAIs (p < 0.001).

Seven patients (6%) were documented as Persons Who Inject Drugs (PWID). Four patients were on dialysis. No blood cultures were secondary to complications of a central or peripheral cannula. Five patients had recently undergone surgery within seven days prior to hospital admission. All these patients underwent skin surgery (biopsy, debridement, or amputation).

Table 1 Baseline characteristics of 108 patients with SDSE bacteraemia
Full size table

Most patients had clearance blood cultures collected (93%), with only nine patients (8%) having more than one positive blood culture. Of the 9 patients with more than one positive blood cultures, two had infective endocarditis, one had septic arthritis, and three had skin and skin structure infections. The remaining patients had no identifiable focus. Less than half, 44% (48/108) of the patients had a TTE, with three patients having IE, 6.3% (3/48), using modified Dukes criteria. Eight out of these 48 patients had trans-oesophageal echocardiogram (TOE), while one patient refused TOE. Three patients with cardiac implantable electric devices (CIEDs) did not have TTE/TOE. Forty-six out of 108 patients underwent a computerised tomography (CT) scan to determine the source of infection or a deeper collection. Nuclear medicine scans were infrequently performed, with less than 6% of patients investigated with either of the available modalities (bone scan/gallium scan/white cell scan/positron-emission tomography scan).

Foci/site of infection

The most common sites of infection with SDSE bacteraemia were the skin and skin structure (59%), followed by bone and joint infection (13%). 11% of patients had no attributable source, while 3% had left-sided endocarditis. All patients with IE had prosthetic valves and cardiac comorbidities. The distribution is illustrated in Fig. 1.

Fig. 1
figure 1

SDSE bacteraemia cases according to principal diagnosis (108 cases)

Full size image

Mortality rate

During the first 30 days, the all-cause mortality rate was 13% (14/108). Patients with no clear focus had a high mortality rate, 40% (5/14), while one patient with IE died due to a perforated mitral valve. 16% of the total patients (17/108) required admission to the intensive care unit, of whom five (29%) died.

We reviewed mortality based on age (age ≥ 75 or < 75 years), with nine deaths in the ≥ 75 years age group and five in the < 75 years age group. There was no significant relationship between age and mortality (p = 0.055). The results are summarised in Table 2.

Table 2 Outcomes of hospital admission for 108 patients with SDSE bacteraemia
Full size table

Selection and duration of antimicrobial treatment

Of the 108 patients, six were allergic to penicillin. Of the patients who survived on day 30, details regarding choice and duration were not available for two patients, while one patient with IE was scheduled for lifelong suppressive therapy with suspected endovascular graft infection. These three patients were not included. Ninety-one of the remaining 108 patients were analysed. β-lactams were the preferred principal therapy in the majority of patients (85/91, 93%). Benzylpenicillin was the most common β-lactam and was used in 47 patients (52%), followed by cefazolin (15%) and ceftriaxone (12%). Only twelve patients (13%) were empirically treated with benzylpenicillin. 15% of patients (14/91) were also prescribed clindamycin, although the indications were not clearly documented. Penicillin allergy had no association with 30-day mortality.

Of the 14 patients who died, two patients had no antimicrobial therapy, as they were transitioned to end-of-life care early in their admission due to age, comorbidities and prior advanced care directives. Rest of the patients had empirical β-lactams.

The average length of hospital stay for patients who survived was 15 days, and the average duration of IV therapy was 12 days. We analysed the duration of antibiotic treatment for SSTIs, which represented close to 60% of our sample. The mean duration of IV therapy was 9 days, while the total duration of therapy was 16 days. The box plot in Fig. 2 shows the spread and skewness, with a total duration between 7 and 28 days.

Fig. 2
figure 2

Distribution of intravenous and total duration of therapy for SDSE bacteraemia due to skin/soft tissue infection (n = 64)

Full size image

Discussion

This study provides an overview of SDSE blood stream infections over a 6-year period (2015–2020). SDSE bacteraemia is typically a community-onset infection in older individuals with underlying chronic skin conditions or diabetes. The median age of individuals with SDSE bacteraemia was 70 years (interquartile range IQR, 58–85 years). SSTIs were the most common presentations (59%), while bone and joint infections (BJIs) represented 13% and IE represented 3% of our patients. The 30-day mortality rate was 13%, but it was greater in those aged > 75 years (21%). Our results are consistent with previous findings that, unlike patients with group A Streptococcus, patients with SDSE bacteraemia are relatively older with comorbidities [18,19,20,21]. Wright et al. (2022) examined epidemiologic data on invasive group C and G Streptococcus in Western Australia and reported similar findings [21]. Age, male sex, chronic skin conditions, intravenous drug use, cardiovascular disease, alcohol consumption, diabetes, obesity, and immunosuppression have been found to be risk factors for sepsis and bacteraemia caused by pyogenic streptococci [5, 10, 22, 23].

The clinical manifestations of SDSE infections and affected risk groups in our study are similar to those described in previous studies [18,19,20, 22,23,24]. Over 96% of the patients had one or more chronic medical conditions. SSTIs represented 59% of the patients in our study, while cardiovascular disease (46%), followed by chronic skin conditions (44%) and diabetes (37%) were the most common comorbidities. Cardiovascular disease and diabetes have been previously implicated as potential risk factors for infection with pyogenic streptococci strains, including SDSE [24, 25]. Any medical comorbidity leading to skin breakdown increases the risk of invasive infection due to SDSE, and chronic skin conditions have been commonly reported in patients with invasive streptococcal infections [9, 24]. A hyperglycaemic environment and immune dysfunction as well as microvascular and macrovascular complications lead to an increase in the frequency of infections among patients with diabetes [26].

In a population-based study from Finland by Rantala et al. (2009), the case fatality rates due to group C and G Streptococcus were reported to be 22% and 15%, respectively [27]. The reported mortality rate due to invasive infections ranges from 15 to 18% [14]. From an Australian perspective, Harris et al. (2011) from Queensland reported a 28-day all-cause death rate of 5.5% in a younger cohort with a mean age of 43 years, while more recently, Wright et al. (2022) reported a 90-day all-cause death rate of 9% in Western Australia for invasive SDSE infections, not limited to bacteraemia. [12, 21].

An association between obesity and an increased risk of SSTIs has been reported in previous studies [24, 28]. A Danish population-based study by Nevanlinna et al. (2023) reported obesity to be a common comorbidity and a risk factor for SDSE infection (53% of patients with a BMI > 30) [24]. Obesity was documented as a comorbidity in less than a quarter (23%) of our patients. This is lower than in previously reported studies, but individual BMIs were not calculated for our patients.

Some investigators have suggested that bacteraemia with SDSE may be associated with underlying malignancy, but this has not been established [23]. In our cohort, approximately one in five patients (18%) had malignancies listed as a comorbidity.

In a recent study from Manitoba, Canada, Lother et al. (2017) reported a 6% incidence of IE with a mortality of 17%, despite only 33% of patients undergoing TTE [29]. Similar rates were observed in a large prospective cohort study by Murdoch et al. (2009) and a retrospective review by Loubinoux et al. (2013) [30, 31]. Our cohort had a lower incidence of IE than did those in previous studies (3%). In our population, fewer than half of the patients had a TTE in the setting of SDSE bacteraemia despite the potential risk of IE. Overall, 44% of patients underwent any form of investigation for IE (TTE/TOE). In our cohort, all 3 patients with IE had prosthetic valves and cardiac comorbidities The true incidence of IE due to SDSE is unclear, with reported rates < 1.5% to > 6%, depending on risk factors [16, 29,30,31,32]. Patients with multiple positive blood cultures and cardiac risk factors like prosthetic valve or CIED should be investigated with an echocardiogram to rule out IE. The recently updated 2023 Duke-ISCVID criteria will improve sensitivity for diagnosis of IE, although it is crucial that clinical judgement be the cornerstone for considering echocardiogram in patients with SDSE bacteraemia.

The treating physician guided the duration and choice of antimicrobial therapy. A formal Infectious Diseases consult was not requested for all patients. The data, except for bacteraemia secondary to skin and skin structure, were limited due to the small sample size. For SSTIs, the mean durations of IV and total therapy were 9 and 16 days, respectively. For many infections, including SDSE bacteraemia, no data have established the superiority of IV therapy or that a longer duration is better. Unnecessary hospital stays increase the risk of nosocomial infections, while shorter courses also help reduce the antibiotic footprint. Many modern RCTs have shown that short-course regimens are effective, and early transition to oral antibiotics is at least as effective as IV for a select group of patients [33, 34]. Prospective studies could examine the benefits of an Infectious Diseases consult and potentially shorter duration of therapy for uncomplicated bacteraemia and early transition to orals.

The strengths of our study include robust demographic and comorbidity data from four major public hospitals in the district. This is one of the largest reported series on SDSE bacteraemia. This study provides an overview of common presentations, including a good sample size for SSTIs. The same clinician reviewed all records. Clinical information was available regarding the need for hospital admission, length of stay, admission to the ICU, antimicrobial therapy, and mortality.

Limitations of this review include its retrospective study design and the use of eMRs for data collection. For example, obesity rates in our population were far lower than expected, as this was based on eMR documentation. Individual BMI calculations were not feasible due to incomplete or inaccurate data. Our sample size for infections other than skin/soft tissue infection was limited. Most other manifestations represented < 10% of the total cases. We did not review surgical interventions, and hence, we cannot infer whether they played any role in the duration or choice of antimicrobial therapy. Our review did not include blood cultures submitted to private laboratories, although the ICPMR laboratory at Westmead Hospital receives almost all blood culture specimens from hospitalised patients in the area. Alcohol use has been reported in previous studies as a risk factor, but we did not collect this information because it was not clearly documented on eMR. Antimicrobial susceptibility testing is not routinely reported given predictable susceptibility, so we cannot comment on any antimicrobial resistance.

In conclusion, this review describes the epidemiology and clinical features of SDSE bacteraemia. It is typically a community-onset infection in older individuals with underlying chronic skin conditions or diabetes. Skin and skin structure infections are common, with a low rate of complications such as BJI (13%) and IE (3%). The rates of invasive infections are expected to increase, and this review highlights areas that will benefit from future studies addressing current gaps in knowledge. A prospective study looking at the incidence and risk factors for IE for SDSE bacteraemia, including its incidence in PWID, would help address some of these gaps. Future studies could also address the impact of an Infectious Diseases review and its effect on the choice and duration of antimicrobial agents, hospital stay, and mortality in invasive SDSE bacteraemia patients.