- Email: [email protected]
Epidemiology of invasive group A Streptococcus infections in Sydney, Australia
Pathology (June 2015) 47(4), pp. 365–371
MICROBIOLOGY
Epidemiology of invasive group A Streptococcus infections in Sydney, Australia SHOBINI SIVAGNANAM1,2, FEI ZHOU1, ANDIE S. LEE2
AND
MATTHEW V. N. O’SULLIVAN1
1Centre for Infectious Diseases and Microbiology, Marie Bashir Institute for Infectious Diseases and Biosecurity, Westmead Hospital, Sydney, and 2Department of Microbiology, Royal Prince Alfred Hospital, Sydney, Australia
Summary There is concern of global resurgence of invasive group A Streptococcus (iGAS) infections. We compared the clinical and molecular epidemiology of patients admitted with iGAS over two time periods, 2008 and 2010, in Western Sydney, Australia. The annual incidence was 19 cases per 100,000 admissions in 2008, compared to 33 per 100,000 in 2010. An increasing proportion of patients died (0% versus 13%), had an APACHE II score 30 (0% versus 19%), and had no known risk-factors (12% versus 25%). A potential skin source was identified as a trigger in fewer cases in 2010 (36% versus 11%). In total, there were 27 different emm types and 11 different emm clusters. There were some new emm types/ clusters in 2010 that were not present in 2008. However, the study was not adequately powered to detect statistically significant differences in the distribution of emm types ( p ¼ 0.06) and emm clusters ( p ¼ 0.16) between the two years. There were also no clear associations between emm types/clusters and severity and clinical manifestations of iGAS infections. Although the proposed 30-valent M protein vaccine encompasses only 47% of our isolates, it will likely provide coverage for at least 71% of iGAS infections due to cross-opsonisation. Key words: Clinical epidemiology, emm cluster, emm types, epidemiology, group A Streptococcus, molecular epidemiology, Streptococcus pyogenes, toxic shock, typing. Received 20 September, revised 22 December, accepted 5 January 2015
INTRODUCTION The incidence of invasive group A Streptococcus (iGAS) infection is estimated to be 2.45 cases per 100,000 person years in developed countries globally, with a case fatality rate of 15%.1 The incidence in less developed countries and in the indigenous communities is even greater. The minimum estimate of 500,000 deaths per year from serious GAS infections is comparable to infections due to rotavirus, measles and Hemophilus influenzae type b.1 Further, epidemiological studies have shown global resurgence of severe iGAS infections since the 1980s.2 In addition, outbreaks of iGAS infections in the community, institutions and within household contacts have been described.3,4 Yet, globally, GAS infection is not a public health notifiable disease except in a few countries. M protein, encoded by the emm gene, is a major virulence protein of Streptococcus pyogenes that is responsible for the production of type specific antibodies in humans.5 A portion of the hypervariable region of this emm gene can be typed using Print ISSN 0031-3025/Online ISSN 1465-3931 DOI: 10.1097/PAT.0000000000000247
#
conserved primers in a polymerase chain reaction (PCR) and sequence typing to determine serospecificities of various S. pyogenes isolates.6 Additionally, the recently proposed emmclusters, containing closely related M proteins that share binding and structural properties, provide a functional classification system to facilitate a better understanding of streptococcal virulence and vaccine efficacy.7 Global emm type distribution varies widely; in established market economies, the most common emm type is emm 1, followed by emm 12, emm 28 and emm 3. Within the indigenous Australian communities and Pacific Island countries, the most common emm type is emm 18, followed by emm 80, emm 114 and emm 1.8 There might be an association between emm types and various clinical manifestations although this remains controversial.9 Whether certain emm types are more transmissible is also unclear. Epidemiological data are necessary in order to determine the level of coverage provided by vaccination strategies, including the proposed 30-valent M protein vaccine,10 in various countries. This is particularly relevant with increasing globalisation and international travel. There is anecdotal evidence to suggest an increase in the frequency and severity of iGAS in Sydney, New South Wales (NSW), Australia. However, to date, there are no published reports of the clinical manifestations and molecular epidemiology of iGAS infections in NSW, despite having the largest population size in Australia. This study reviews the clinical and molecular epidemiology of iGAS infections in Western Sydney Local Health District (WSLHD), covering a catchment area of 876,500 people in NSW, to determine trends in iGAS infections over time and the association between emm types/clusters and types and severity of iGAS infections.
MATERIALS AND METHODS Design WSLHD covers four public hospitals which predominantly see adult patients: Westmead, Blacktown, Auburn and Mount Druitt hospitals. These hospitals refer their microbiology specimens to a single laboratory. We conducted a retrospective review of iGAS infections in WSLHD over two time periods, one recent year at the time of initiation of data collection with suspected high incidence of severe iGAS infections (year 2010) and another randomly selected year prior to 2010 (year 2008). There were no changes in the culture collection or testing practices between the two selected years to account for any differences observed. Data collection was commenced in year 2011. Ethics approval was obtained from the WSLHD human research ethics committee. Definitions Invasive GAS infection was defined as the isolation of S. pyogenes from a usually sterile site such as blood, joint, sterile tissue or cerebrospinal fluid. Those
2015 Royal College of Pathologists of Australasia
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
366
Pathology (2015), 47(4), June
SIVAGNANAM et al.
with invasive disease without positive cultures from sterile sites were not captured. Bacteraemia without a focus was defined as case patients with a positive blood culture without an identifiable clinical syndrome. Streptococcal toxic shock syndrome (TSS) was defined as proposed by the working group on severe streptococcal infections.11 Definite and probable TSS cases were combined to give a single category, TSS, for analysis in this study. Healthcare associated infection was defined as a positive culture in patients who had been hospitalised for at least 48 h at the time of onset of symptoms or the infection was a complication of a medical intervention such as surgery or indwelling catheters and devices. Data collection Cases were identified using the laboratory information system. Medical records were retrieved from all four hospitals and the following data were collected: patient demographics, baseline co-morbidities, clinical manifestations, laboratory results, predisposing factors including skin diseases, chronic medical conditions, immune status, pregnancy, smoking, intravenous drug abuse and alcoholism,4,12 treatment and outcome. Severity of the infection was determined based on clinical outcome (death versus cure) and the worst Acute Physiology and Chronic Health Evaluation (APACHE) II score within the first seven days of culture isolation (30 versus <30), collected systematically from patient records. An APACHE II cut-off value of 30 was used to indicate severity as patients admitted to the intensive care unit with septic shock with an APACHE II score of 30 and above have been shown to have a mortality rate of 82%.13 Overall baseline health was determined using Charlson’s Weighted Index of Co-morbidity (WIC).14 Microbiological methods The isolates were confirmed to be S. pyogenes prior to storage by beta-hemolysis on Horse Blood Agar (HBA), presence of Lancefield group A antigen, negative catalase test, positive reaction to pyrrolidonyl-beta-naphthylamide (PYR) and sensitivity to bacitracin. These stored isolates were subsequently retrieved and their emm types were determined using published methods by CDC11 with modifications. PCR was performed using the Qiagen HotStarTaq DNA polymerase kit (Qiagen, Australia) according to the manufacturer’s instructions with the following conditions: 958C for 15 min; 40 cycles of 948C for 30 s, 46.58C for 30 s and 728C for 3 min; 728C for 10 min; and a hold at 228C. The PCR mixture contained: 5 mL template DNA, 0.125 mL each of primer 1 (100 pmol mL1) and primer 2 (100 pmol mL1), 1 mL dNTPs (2.5 mM of each dNTP), 2.5 mL 10 PCR buffer, 0.1 mL Qiagen HotStarTaq DNA polymerase (5 units mL1), and water to 25 mL. PCR products (9 mL) were further purified using a PCR product pre-sequencing kit (USB Corporation, USA) according to the manufacturer’s procedures. These products were sequenced using the BigDye Terminator (version 3.1) cycle sequencing kit in the ABI Prism 3100 genetic analyser (Applied Biosystems, USA) Data analysis Using published data on emm types and clusters,7 we classified all emm types into emm clusters. Subsequent data analysis was performed for emm types as well as emm clusters. Firstly, the emm types and clusters were compared between years 2008 and 2010 to determine if there had been the emergence of a predominant strain or strains. Secondly, clinical data and the emm sequence typing and cluster data were assessed for associations between emm types/ clusters and the severity and type of infections. Thirdly, the factors associated with the most severe outcome, i.e. death, in iGAS infection were determined. Finally, geographic clustering was determined by comparing the postcodes and ethnicity of patients admitted during the two year groups. Statistical analysis was performed using SAS 9.4 (SAS Institute, USA). For continuous data, Wilcoxon sum rank test was used to compare medians; for non-continuous data, either Chisquared test or Fisher’s exact test was used as appropriate to compare proportions. Forward stepwise logistic regression was performed to determine risk factors associated with death in iGAS infection, including emm type. Other candidate variables for the multivariate model were those with p < 0.2 on univariate analysis which were considered a priori to be potential risk factors for mortality in iGAS infections. These variables were included in the model starting with the variable with the strongest association with mortality on univariate analysis. Likelihood ratio tests were used to guide inclusion of variables in the model and interaction terms were fitted. The multivariate analysis was performed using Stata V.11.0 (Stata Corp, USA).
RESULTS A total of 72 patients met the case definition for iGAS at WSLHD. There were 25 (35%) cases in 2008 and 47 (65%) in 2010. The sterile site from which GAS was isolated was blood in 60% of cases in 2008 and 64% in 2010. The characteristics of the two year groups are given in Table 1. In total, 58% of iGAS infections occurred in patients under the age of 50. Five (7%) patients were children under 15 years of age, including four patients under the age of 5. Fourteen (19%) patients were older adults, aged 65 years. Slightly more infections occurred in females (56%) compared to males. The dominant ethnic group was European (64%), followed by the Pacific Islander and Indigenous Australian/New Zealander group (21%). There were no significant differences between the baseline co-morbidities between the two year groups. While 48% of the patients had a chronic medical condition, most (82%) had a score of two or less in the Charlson’s WIC. Clinical presentations, predisposing factors and management Among the 72 iGAS infections, skin and soft tissue infection was the most common clinical presentation (57%), followed by bacteraemia without a focus (18%). Skin and soft tissue infection included two cases of necrotising fasciitis, which both occurred in 2010. Other manifestations of iGAS infections included pharyngitis/quinsy (13%), septic arthritis (7%), empyema (3%), endocarditis (1%) and pelvic inflammatory disease (1%). There were no cases of puerperal sepsis. More patients had TSS (21% versus 12%) and bacteraemia without a focus (21% versus 12%) in 2010 compared to 2008. Three (12%) patients in 2008 and 12 (26%) patients in 2010 had no identifiable risk factors for iGAS infection. Nine (36%) patients had skin diseases including eczema, burns, varicella zoster (VZV) infection or trauma in 2008 compared to only five (11%) cases in 2010. The one patient with recent VZV infection prior to iGAS infection was an adult. Treatment consisted of betalactam antibiotics alone in 67% and beta-lactam with clindamycin in 19% of cases. There was a small increase in the use of clindamycin in 2010 (27%) compared to 2008 (20%). Thirty percent of patients with GAS bacteraemia, 44% requiring management in the intensive care unit (ICU) and 60% with TSS received clindamycin in 2010 compared to 7%, 25% and 33% respectively in 2008. Of the 13 patients with toxic shock syndrome, three (23%) received intravenous immunoglobulin (IVIG) therapy, all in 2010. Severity of infection and outcome In 2008, all patients had an APACHE II score that was less than 30, compared to nine (19%) patients with an APACHE II score of 30 or more in 2010 ( p ¼ 0.02). A total of 13 (18%) patients required admission to ICU, without a significant difference between the two year groups. All patients survived in 2008; however, the case fatality rate was 13% in year 2010 ( p ¼ 0.06). Among the patients who died in 2010, 67% were female, 83% were at least 50 years of age and 67% had at least one chronic medical condition. Table 2 outlines the characteristics of patients with fatal and non-fatal outcomes. There was an associated skin disease in 22% of patients who survived compared to no patients with skin disease in those who died. In univariate analysis, the factors associated with death included APACHE II score 30 ( p < 0.001), admission to the intensive care unit ( p ¼ 0.07), toxic shock syndrome
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
EPIDEMIOLOGY OF INVASIVE GROUP A STREPTOCOCCUS INFECTIONS
Table 1
367
Differences in the clinical and demographic features of invasive group A Streptococcus infections between years 2008 and 2010 in Western Sydney
Demographics and clinical features Age, median years (IQR) Sex, male (%) Ethnicity European East and South East Asian Pacific and Indigenous Australian/New Zealander Middle Eastern South Central Asian Clinical Duration of illness prior to admission, median days (IQR) Bacteraemia Length of hospitalisation, median days (IQR) Type of infection Skin and soft tissue infections Pharyngitis/quinsy Endocarditis Empyema Septic arthritis Pelvic inflammatory disease Bacteraemia without a focus Possible predisposing factors Skin diseases (burns, varicella zoster virus infection, skin disease, wound) Smoking (n*¼59) Intravenous drug abuse Alcohol consumption Diabetes mellitus Obesity Lymphoedema Chronic medical condition Charlson’s weighted index of co-morbidity 0 1–2 3 or more Immunosuppression No known risk factors Healthcare associated infection Severity of infection Death APACHE II score 30 Admission to intensive care unit Toxic shock syndrome Acute renal failure (n*¼70) Coagulopathy (n*¼52) Acute respiratory distress syndrome (n*¼67) Acute liver injury (n*¼65) Treatment Surgical intervention required Antibiotics (n*¼69) b-lactam alone b-lactam þ clindamycin b-lactam þ other Non-b-lactam
2008 (n ¼ 25)
2010 (n ¼ 47)
P value
45 (38-50) 10 (40%)
40 (26-64) 22 (47%)
0.90 0.58
19 (76%) 1 (4%) 5 (20%) 0 0
27 4 10 5 1
0.34
3 (1–7) 15 (60%) 8 (2–16)
2 (1–7) 30 (64%) 7 (2–17)
0.36 0.75 0.89
16 3 1 1 1
25 (53%) 6 (13%) 0 1 (2%) 4 (9%) 1 (2%) 10 (21%)
0.36
(64%) (12%) (4%) (4%) (4%) 0 3 (12%)
(57%) (9%) (21%) (11%) (2%)
9 12 3 11 4 5 1 13
(36%) (52%) (12%) (44%) (17%) (21%) (4%) (54%)
5 10 1 8 8 6 3 21
(11%) (28%) (2%) (17%) (17%) (13%) (6%) (45%)
0.01 0.06 0.08 0.01 0.97 0.37 0.70 0.45
13 8 4 1 3 1
(52%) (32%) (16%) (4%) (12%) (4%)
19 19 9 4 12 3
(40%) (40%) (20%) (9%) (26%) (6%)
0.64
0.47 0.18 0.67
4 3 5 4
0 0 (16%) (12%) (21%) (27%) 0 (19%)
6 9 9 10 13 10 5 18
(13%) (19%) (20%) (21%) (28%) (27%) (11%) (42%)
0.06 0.02 0.71 0.33 0.50 0.98 0.10 0.07
11 (44%)
21 (46%)
0.89
19 (76%) 4 (16%) 0 2 (8%)
27 9 3 5
0.46
4
(61%) (20%) (7%) (11%)
IQR, interquartile range. *
Number of patients in whom this information was either documented or was able to be determined.
( p ¼ 0.07), bacteraemia ( p ¼ 0.08), bacteraemia without a focus ( p ¼ 0.009), acute renal failure ( p ¼ 0.04) and acute respiratory distress syndrome ( p ¼ 0.03). However, after multivariate analysis, APACHE II score was the only independent risk factor associated with mortality [adjusted odds ratio (OR) 1.26 per 1-unit increment, 95% confidence interval (CI) 1.03–1.54, p ¼ 0.02]. emm types and clusters A total of 55 (76%) isolates [17 (68%) in 2008 and 38 (81%) in 2010] were available for emm sequence typing. Of the 17 unavailable isolates, 12 (71%) were from non-blood specimens. Overall, 27 different emm types and 11 different emm clusters were obtained. The distribution of emm types and emm clusters
between the two year groups is shown in Fig. 1. The most common emm type was emm 1 (13%), followed by emm 101 (11%), emm 12 (7%), emm 66 (7%) and emm 3 (5%). Predominance of a single clone did not account for the increased numbers of iGAS infections seen in 2010. The distribution of emm types in 2008 and 2010 were different ( p ¼ 0.06), with the emergence of some new emm types in 2010 that were not present in 2008. For example, there were 16 additional emm types in 2010 that were not present in 2008, accounting for 49% (n ¼ 27/55) of infections; similarly, there were six additional emm types in 2008 which did not occur in 2010, accounting for 16% (n ¼ 9/55) of infections. The most common emm cluster was D4 (20%), followed by A-C3, E3, E4 and E6, representing 13% each of the total typed isolates. There were some notable differences in the distribution of emm clusters between the two
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
368
Pathology (2015), 47(4), June
SIVAGNANAM et al.
Table 2
Factors associated with death in invasive group A Streptococcus infections in Western Sydney in both 2008 and 2010 (using univariate analysis) Death (n ¼ 6)
Demographics Male Age 50 Pacific and Indigenous Australian/New Zealander Possible risk factors Chronic medical conditions No risk factors Diabetes mellitus Obesity Lymphoedema Malignancy Immunosuppression Healthcare associated infection Skin diseases Microbiology emm1 type (n*¼54) Clinical features Duration of illness prior to presentation, median days (IQR) APACHE 30 APACHE, mean (SD) Admission to intensive care unit Toxic shock syndrome Requirement for surgery Bacteraemia Bacteraemia without a focus Acute renal failure (n*¼68) Coagulopathy (n*¼51) Acute respiratory distress syndrome (n*¼65) Acute liver impairment (n*¼62)
No death (n ¼ 64)
Crude OR (95% CI)
2 (33%) 5 (83%) 1 (17%)
28 (44%) 25 (39%) 14 (22%)
0.64 (0.11–3.77) 7.80 (0.86–70.75) 0.72 (0.08–6.63)
0.69 0.08 1.00
4 1 1 2 1 1
30 12 11 9 3 7 5 4 14
2.27 0.87 0.96 3.06 4.07 1.63
(0.39–13.27) (0.09–8.12) (0.10–9.08) (0.49–19.20) (0.35–46.65) (0.17–16.02) – – –
0.42 1.00 1.00 0.24 0.31 0.53 1.00 1.00 0.34
1.79 (0.17–18.82)
0.52
(67%) (17%) (17%) (33%) (17%) (17%) 0 0 0
1 (20%) 1 6 42.7 3 3 1 6 4 4 3 2 3
(47%) (19%) (17%) (14%) (5%) (11%) (8%) (6%) (22%)
6 (12%)
(1–1) (100%) (7.8) (50%) (50%) (17%) (100%) (67%) (67%) (50%) (33%) (50%)
19 3 13.0 10 10 31 37 9 14 11 3 19
(7–31) (5%) (11.0) (16%) (16%) (48%) (58%) (14%) (23%) (24%) (5%) (34%)
P value
0.81 per 1-day increment (0.58–1.15) – 1.20 per 1-point increment (1.07–1.34) 5.4 (0.95–30.67) 5.40 (0.95–30.67) 0.21 (0.02–1.93) – 12.22 (1.95–76.79) 6.86 (1.13–41.43) 3.09 (0.54–17.59) 9.33 (1.19–72.99) 1.95 (0.36–10.59)
0.15 <0.01 0.07 0.07 0.21 0.08 0.01 0.04 0.33 0.03 0.66
CI, confidence interval; IQR, interquartile range; SD, standard deviation. *
Number of patients in whom this information was either documented or was able to be determined.
year groups ( p ¼ 0.16); there were four additional emm clusters in 2010 that were not present in 2008, representing 24% of the typed isolates. In 2008, although there were three additional emm clusters, this represented only 5% of the typed isolates. There were no significant differences in the emm types/ clusters between those with fatal and non-fatal outcomes. Five of the six patients with fatal iGAS had isolates available for typing. These belonged to emm types 1, 3, 12, 114 and 232 and emm clusters E4 (2 isolates), A-C3, A-C4, and A-C5.
within different ethnic groups to suggest geographic clustering. The four healthcare associated cases, one in 2008 and three in 2010, were unrelated and had four different emm types.
DISCUSSION This present study, which analyses the clinical and molecular epidemiology of iGAS infections in WSLHD, is the first study looking at iGAS infections in New South Wales, Australia. Most data in Australia on iGAS infections come from Victoria,12,15,16 Queensland17 and the Northern Territory.18 The annual incidence of iGAS in WSLHD was 19 cases per 100,000 hospital admissions in 2008 compared to 33 cases per 100,000
Geographic clustering There were no significant differences in the pattern of emm types and clusters isolated within the different postcodes or
6
No. of cases
5
4
3
2
1
0 1
12
3
6
A-C3 A-C4 A-C5 Clade Y
71 D2
33
53
70 D4
101
4 E1
66
90
9
E2
87
103
113
E3
8
28
89 E4
114
232
11
63
75
81
85
177
E6
Emm types and clusters 2008
2010
Fig. 1 Emm type and emm cluster distribution between 2008 and 2010 in Western Sydney.
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
EPIDEMIOLOGY OF INVASIVE GROUP A STREPTOCOCCUS INFECTIONS
Table 3
369
Association between various clinical manifestations of invasive group A Streptococcus infections, emm types and emm clusters
Clinical manifestation
emm cluster (frequency)
emm types (frequency)
Bacteraemia without primary focus
A-C3 (1) A-C5 (1) D4 (2) E2 (2) E3 (1) E4 (3) E6 (2) A-C3 (1) A-C4 (1) E1 (1) A-C4 (1) E2 (1) E4 (1) A-C3 (1) A-C5 (1) D2 (1) D4 (1) D4 (1) A-C3 (4) A-C4 (2) A-C5 (1) Clade Y (1) D4 (7) E1 (1) E2 (2) E3 (6) E4 (3) E6 (5) A-C3 (3) A-C4 (1) A-C5 (2) Clade Y (1) D4 (1) E4 (3) E6 (2) A-C4 (1) A-C5 (1) A-C3 (1) A-C4 (1) A-C4 (1) E4 (2)
emm1(1) emm3(1) emm101(1), emm70(1) emm66(1), emm90(1) emm103(1) emm232(2), emm114(1) emm11(1), emm75(1) emm1(1) emm12(1) emm4(1) emm12(1) emm66(1) emm89(1) emm1(1) emm3(1) emm71(1) emm33(1) emm101(1) emm1(1) emm12(2) emm3(1) emm6(1) emm101(4), emm33(1), emm53(1), emm70(1) emm4(1) emm66(2) emm103(1), emm113(2), emm87(1), emm9(2) emm28(1), emm8(1), emm89(1) emm177(1), emm11(1), emm63(1), emm81(1), emm85(1) emm1(3) emm12(1) emm3(2) emm6(1) emm101(1) emm232(1), emm114(1), emm28(1) emm11(1), emm63(1) emm3(1), emm12(1)
admissions in 2010. This is an underestimation as our data does not capture those with clinical iGAS infection without a positive sterile site culture. This equates to at least four per 100,000 adults in 2008 and seven per 100,000 adults in 2010. This is a minimum incidence estimate only as we did not have information on the number of patients admitted to the children’s or private hospitals with iGAS infections; these additional cases would likely increase the incidence further. These rates are significantly higher than the rates of infection observed in other parts of Australia12 and in some European countries and USA.4,19–22 A New Zealand study has shown an annual incidence rate of greater than 8.1 cases per 100,000, driven largely by the indigenous Maori and Pacific Islander population.23 Approximately 21% of subjects in our study belonged to Pacific Islander or Indigenous Australian/New Zealand ethnic groups; our high incidence rate may be reflective of this. This study also highlights the increase in severity of iGAS infections over time. In 2010, more patients died, had a high APACHE II score and had TSS. In addition, a larger proportion of patients had no identifiable risk factors for iGAS infection (26% versus 12%) and had bacteraemia without a focus (21% versus 12%) in 2010. A recent study showed that children with bacteraemia without a focus had a lower risk for severe disease.24 In our study, only one child had bacteraemia without a
focus, in year 2010, and survived. In addition, bacteraemia without a focus was associated with a greater odds of death (OR 12.22; 95% CI 1.95–76.79). Skin breaches have been considered portals of entry for iGAS infections.4 A potential skin source was identified in more patients in 2008 compared to 2010. These observations support the hypothesis that there was an increase in the frequency of more severe infections in previously healthy hosts without an identifiable trigger in more recent times. In some countries, the increasing frequency of iGAS infections was attributed to the emergence of a new strain of S. pyogenes.25 In our study, there was no evidence for the emergence of a single clone contributing to the observed increase in frequency and severity of infections. However, similar to what is seen in New Caledonia,26 there is significant variation in the emm type distribution over time in Western Sydney, with 49% of infections in 2010 being caused by emm types that were absent in 2008. Additionally, we did not observe an association between emm types/clusters and type of infection (Table 3) or severity of infections. Emm 1 has been shown to be associated with more severe infections in some studies.4 In our study, the odds ratio of death in those who had emm 1 type was 1.79 (95% CI 0.17–18.82). The small number of isolates of this type may have reduced our ability to detect a statistically significant association between emm type and mortality.
Empyema Infective endocarditis Pharyngitis and/or quinsy
Pelvic inflammatory disease Septic arthritis
Septic arthritis with skin and soft tissue infection Skin and soft tissue infection
Toxic shock syndrome
Necrotising fasciitis Death
emm1(1) emm12(1) emm3(1) emm232(1), emm114(1)
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
370 Table 4
Pathology (2015), 47(4), June
SIVAGNANAM et al.
Coverage provided by the 30-valent M protein vaccine for infections caused by group A Streptococcus in Western Sydney
30-valent M protein vaccine coverage
emm cluster (frequency)
emm types (frequency)
Vaccine serotypes
A-C3 (7) E6 (4) E4 (4) A-C4 (4) A-C5 (3) E1 (2) Clade Y (1) E3 (1) D4 (3) E6 (2) E2 (4) D2 (1) E4 (1) E3 (2) D4 (2) E6 (1) E4 (2) D4 (6) E3 (4) E2 (1)
emm1(7) emm11(2), emm75(1), emm81(1) emm114(1), emm28(1), emm89(2) emm12(4) emm3(3) emm4(2) emm6(1) emm87(1) emm33(2), emm53(1) emm63(1), emm85(1) emm66(4) emm71(1) emm8(1) emm9(2) emm70(2) emm177(1) emm232(2) emm101(6) emm103(2), emm113(2) emm90(1)
Non-vaccine serotypes, protected by cross-opsonisation
Non-vaccine serotypes, not protected by cross-opsonisation Non-vaccine serotypes with no available data on cross-opsonisation
*
Proportion of infections in this study (n ¼ 55*) 47%
24%
4% 25%
Number of isolates available for emm sequence typing.
Other factors associated with death in iGAS infections shown in prior studies include age 50, TSS, bacteraemia without a focus, chemotherapy, diabetes, and liver disease.4 In our study, whilst age 50, TSS, APACHE II 30, bacteraemia without a focus, bacteraemia, admission to intensive care unit, acute renal impairment and acute respiratory distress syndrome were associated with mortality on univariate analysis, only APACHE II 30 was significantly associated with death in the multivariate analysis. Diabetes mellitus, immunosuppression, and chronic medical conditions did not appear to influence outcome (Table 2). There is growing evidence to suggest a mortality benefit from adjunctive therapy with clindamycin and IVIG in iGAS infections.27,28 In our study, clindamycin and IVIG use were very low. However, there was an increase in usage of clindamycin in 2010 compared to 2008, particularly in patients with severe infections. In addition, all three patients who received IVIG were treated in 2010. Two of the three patients had severe disease requiring intubation and inotropes as soon as they presented to the emergency department. There may be a number of reasons for these low numbers. Firstly, there are no randomised controlled trials to suggest a clear benefit from clindamycin and IVIG use. Secondly, opinions vary on which subpopulations with iGAS are most likely to benefit from these treatments. However, increased use of clindamycin and IVIG in 2010 may be related to higher numbers of severe infections observed in 2010 as well as increased clinician awareness. In addition, infectious diseases physicians and intensivists are more likely to be involved in the care of patients with severe iGAS infections; they may be more likely to recommend clindamycin and IVIG. The fatality rate of iGAS is not negligible, particularly when the condition occurs in hosts without identifiable risk factors. Phase I/II studies among healthy adults have shown that a 26valent M protein vaccine, which was based on the predominant strains in USA, may be safe and immunogenic.29 A 30-valent M protein vaccine has been shown to evoke cross-opsonic bactericidal antibodies against non-vaccine serotypes resulting in greater vaccine efficacy10 The 30-valent M protein based
vaccine consists of 12 serotypes found in our population. As shown in Table 4, if vaccine serotypes alone were used to predict immunity, coverage would be 47% (26/55). However, the emm clusters that contain at least one representative included in the vaccine may demonstrate cross protection, except for the D4 emm cluster.7 Given this extension of coverage beyond the emm types contained in the vaccine, at least an additional 8 emm types representing 24% (13/55) of infections would be covered by the 30-valent M protein vaccine in our population. Data are not currently available on the bactericidal activity of vaccine induced antibodies against strains belonging to emm types 90, 101, 103, 113, 177 and 232 which accounted for 25% (14/55) of all our invasive infections. In addition, there were two isolates that belonged to emm 70; the vaccine has been shown to induce only low level antibodies against this type.10 Of the six patients with a fatal outcome, five isolates were available for emm typing. Four of the five were emm types with coverage in the 30-valent vaccine; data are unavailable regarding emm 232. Overall, the 30-valent M protein vaccine which covers 90% of invasive isolates in the US and 78% in Europe10 would have covered at least 71% of the emm types isolated in Western Sydney during this study. Limitations of our study include the small sample size, retrospective nature of data collection, limiting the study to the adult public hospital patients only, using laboratory information system as the sole surveillance tool to determine cases of iGAS infections and not determining the toxin profile of the strains as virulence may be related to toxin production rather than the emm types. Additionally, not all iGAS isolates from the study period were available for emm typing. However, the aim of our study was to describe the clinical and molecular epidemiology of iGAS infections in Western Sydney and to draw hypotheses which would require validation in future large, prospective studies. The study highlights the need to prospectively monitor GAS infections systematically. A recent study suggested potential benefit from penicillin prophylaxis in patients with recurrent leg cellulitis.30 Our data do not provide evidence of clustering to support the use of antibiotic prophylaxis in contacts of patients with iGAS infections. However,
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
EPIDEMIOLOGY OF INVASIVE GROUP A STREPTOCOCCUS INFECTIONS
postcodes and ethnicity alone do not adequately capture epidemiological links between patients. A secondary case in a household contact was observed in this cohort, both belonging to emm type 101; larger studies would be required to determine the incidence of iGAS infection among household contacts and the benefit of antibiotic prophylaxis in our population. In conclusion, there is evidence to suggest an increase in the frequency and severity of iGAS infections in metropolitan Sydney. There was no clear association between the increased frequency and severity and various emm types/clusters, although there were new emm types/clusters seen in 2010 that were not present in 2008. Comprehensive monitoring of iGAS infections through public health notification, contact tracing and emm sequence typing of these isolates will help determine the role of vaccination and antibiotic prophylaxis in our population. Acknowledgements: Mitchell Brown and Marion Nguyen for characterising a Lancefield group A positive, groups B, C, F and G negative, bacitracin susceptible Streptococcus dysgalactiae, emm type stG485, which was originally misidentified as GAS; this was excluded from the 2010 analysis. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Address for correspondence: Shobini Sivagnanam, 906 Dexter Ave N., Apt L539, Seattle, WA 98109, USA. E-mail: [email protected]
References 1. Carapetis JR, Steer AC, Mulholland EK, et al. The global burden of group A streptococcal diseases. Lancet Infect Dis 2005; 5: 685–94. 2. Aziz RK, Kotb M. Rise and persistence of global M1T1 clone of Streptococcus pyogenes. Emerg Infect Dis 2008; 14: 1511–7. 3. Dooling KL, Crist MB, Nguyen DB, et al. Investigation of a prolonged group a streptococcal outbreak among residents of a skilled nursing facility, Georgia, 2009-2012. Clin Infect Dis 2013; 57: 1562–7. 4. Lepoutre A, Doloy A, Bidet P, et al. Epidemiology of invasive Streptococcus pyogenes infections in France in 2007. J Clin Microbiol 2011; 49: 4094–100. 5. Lancefield RC. Current knowledge of type-specific M antigens of group A streptococci. J Immunol 1962; 89: 307–13. 6. Centers for Disease Control and Prevention. Emm types as proportions of total diseases isolates in six global regions. 28 Sep 2009; cited 19 May 2014. http://www.cdc.gov/ncidod/biotech/strep/emmtype_proportions.htm. 7. Sanderson-Smith M, De Oliveira DM, Guglielmini J, et al. A systematic and functional classification of Streptococcus pyogenes that serves as a new tool for molecular typing and vaccine development. J Infect Dis 2014; 210: 1325–38. 8. Centers for Disease Control and Prevention. Protocol for emm typing. 8 Feb 2008; cited 19 May 2014. http://www.cdc.gov/ncidod/biotech/strep/protocol_emm-type.htm. 9. Rogers S, Commons R, Danchin MH, et al. Strain prevalence, rather than innate virulence potential, is the major factor responsible for an increase in serious group A streptococcus infections. J Infect Dis 2007; 195: 1625–33.
371
10. Dale JB, Penfound TA, Chiang EY, et al. New 30-valent M protein-based vaccine evokes cross-opsonic antibodies against non-vaccine serotypes of group A streptococci. Vaccine 2011; 29: 8175–8. 11. The Working Group on Severe Streptococcal Infection. Defining the group A streptococcal toxic shock syndrome. Rationale and consensus definition. JAMA 1993; 269: 390–1. 12. O’Grady KA, Kelpie L, Andrews RM, et al. The epidemiology of invasive group A streptococcal disease in Victoria, Australia. Med J Aust 2007; 186: 565–9. 13. Knaus WA, Draper EA, Wagner DP, et al. APACHE II: a severity of disease classification system. Crit Care Med 1985; 13: 818–29. 14. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40: 373–83. 15. Carapetis J, Robins-Browne R, Martin D, et al. Increasing severity of invasive group A streptococcal disease in Australia: clinical and molecular epidemiological features and identification of a new virulent M-nontypeable clone. Clin Infect Dis 1995; 21: 1220–7. 16. Lithgow A, Duke T, Steer A, Smeesters PR. Severe group A streptococcal infections in a paediatric intensive care unit. J Paediatr Child Health 2014; 50: 687–92. 17. Norton R, Smith HV, Wood N, et al. Invasive group A streptococcal disease in North Queensland (1996-2001). Indian J Med Res 2004; 119 (Suppl): 148–51. 18. Carapetis JR, Walker AM, Hibble M, et al. Clinical and epidemiological features of group A streptococcal bacteremia in a region with hyperendemic superficial streptococcal infection. Epidemiol Infect 1999; 122: 59–65. 19. Montes M, Ardanuy C, Tamayo E, et al. Epidemiological and molecular analysis of Streptococcus pyogenes isolates causing invasive disease in Spain (1998-2009): comparison with non-invasive isolates. Eur J Clin Microbiol Infect Dis 2011; 30: 1295–302. 20. Darenberg J, Luca-Harari B, Jasir A, et al. Molecular and clinical characteristics of invasive group A streptococcal infection in Sweden. Clin Infect Dis 2007; 45: 450–8. 21. Lamagni TL, Darenberg J, Luca-Harari B, et al. Epidemiology of severe Streptococcus pyogenes disease in Europe. J Clin Microbiol 2008; 46: 2359–67. 22. O’Loughlin RE, Roberson A, Cieslak PR, et al. The epidemiology of invasive group A streptococcal infection and potential vaccine implications: United States, 2000-2004. Clin Infect Dis 2007; 45: 853–62. 23. Safar A, Lennon D, Stewart J, et al. Invasive group A streptococcal infection and vaccine implications, Auckland, New Zealand. Emerg Infect Dis 2011; 17: 983–9. 24. Gauguet S, Ahmed AA, Zhou J, et al. Group A streptococcal bacteremia without a source is associated with less severe disease in children. Pediatr Infect Dis J 2015; 34: 447–9. 25. Meehan M, Murchan S, Bergin S, et al. Increased incidence of invasive group A streptococcal disease in Ireland, 2012 to 2013. Euro Surveill 2013; 18: 20556. 26. Baroux N, D’Ortenzio E, Amedeo N, et al. The emm-cluster typing system for Group A Streptococcus identifies epidemiologic similarities across the pacific region. Clin Infect Dis 2014; 59: e84–92. 27. Carapetis JR, Jacoby P, Carville K, Ang SJ, Curtis N, Andrews R. Effectiveness of clindamycin and intravenous immunoglobulin, and risk of disease in contacts, in invasive group a streptococcal infections. Clin Infect Dis 2014; 59: 358–65. 28. Linner A, Darenberg J, Sjolin J, Henriques-Normark B, Norrby-Teglund A. Clinical efficacy of polyspecific intravenous immunoglobulin therapy in patients with streptococcal toxic shock syndrome: a comparative observational study. Clin Infect Dis 2014; 59: 851–7. 29. McNeil SA, Halperin SA, Langley JM, et al. Safety and immunogenicity of 26-valent group a streptococcus vaccine in healthy adult volunteers. Clin Infect Dis 2005; 41: 1114–22. 30. Thomas KS, Crook AM, Nunn AJ, et al. Penicillin to prevent recurrent leg cellulitis. N Engl J Med 2013; 368: 1695–703.
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
MICROBIOLOGY
Epidemiology of invasive group A Streptococcus infections in Sydney, Australia SHOBINI SIVAGNANAM1,2, FEI ZHOU1, ANDIE S. LEE2
AND
MATTHEW V. N. O’SULLIVAN1
1Centre for Infectious Diseases and Microbiology, Marie Bashir Institute for Infectious Diseases and Biosecurity, Westmead Hospital, Sydney, and 2Department of Microbiology, Royal Prince Alfred Hospital, Sydney, Australia
Summary There is concern of global resurgence of invasive group A Streptococcus (iGAS) infections. We compared the clinical and molecular epidemiology of patients admitted with iGAS over two time periods, 2008 and 2010, in Western Sydney, Australia. The annual incidence was 19 cases per 100,000 admissions in 2008, compared to 33 per 100,000 in 2010. An increasing proportion of patients died (0% versus 13%), had an APACHE II score 30 (0% versus 19%), and had no known risk-factors (12% versus 25%). A potential skin source was identified as a trigger in fewer cases in 2010 (36% versus 11%). In total, there were 27 different emm types and 11 different emm clusters. There were some new emm types/ clusters in 2010 that were not present in 2008. However, the study was not adequately powered to detect statistically significant differences in the distribution of emm types ( p ¼ 0.06) and emm clusters ( p ¼ 0.16) between the two years. There were also no clear associations between emm types/clusters and severity and clinical manifestations of iGAS infections. Although the proposed 30-valent M protein vaccine encompasses only 47% of our isolates, it will likely provide coverage for at least 71% of iGAS infections due to cross-opsonisation. Key words: Clinical epidemiology, emm cluster, emm types, epidemiology, group A Streptococcus, molecular epidemiology, Streptococcus pyogenes, toxic shock, typing. Received 20 September, revised 22 December, accepted 5 January 2015
INTRODUCTION The incidence of invasive group A Streptococcus (iGAS) infection is estimated to be 2.45 cases per 100,000 person years in developed countries globally, with a case fatality rate of 15%.1 The incidence in less developed countries and in the indigenous communities is even greater. The minimum estimate of 500,000 deaths per year from serious GAS infections is comparable to infections due to rotavirus, measles and Hemophilus influenzae type b.1 Further, epidemiological studies have shown global resurgence of severe iGAS infections since the 1980s.2 In addition, outbreaks of iGAS infections in the community, institutions and within household contacts have been described.3,4 Yet, globally, GAS infection is not a public health notifiable disease except in a few countries. M protein, encoded by the emm gene, is a major virulence protein of Streptococcus pyogenes that is responsible for the production of type specific antibodies in humans.5 A portion of the hypervariable region of this emm gene can be typed using Print ISSN 0031-3025/Online ISSN 1465-3931 DOI: 10.1097/PAT.0000000000000247
#
conserved primers in a polymerase chain reaction (PCR) and sequence typing to determine serospecificities of various S. pyogenes isolates.6 Additionally, the recently proposed emmclusters, containing closely related M proteins that share binding and structural properties, provide a functional classification system to facilitate a better understanding of streptococcal virulence and vaccine efficacy.7 Global emm type distribution varies widely; in established market economies, the most common emm type is emm 1, followed by emm 12, emm 28 and emm 3. Within the indigenous Australian communities and Pacific Island countries, the most common emm type is emm 18, followed by emm 80, emm 114 and emm 1.8 There might be an association between emm types and various clinical manifestations although this remains controversial.9 Whether certain emm types are more transmissible is also unclear. Epidemiological data are necessary in order to determine the level of coverage provided by vaccination strategies, including the proposed 30-valent M protein vaccine,10 in various countries. This is particularly relevant with increasing globalisation and international travel. There is anecdotal evidence to suggest an increase in the frequency and severity of iGAS in Sydney, New South Wales (NSW), Australia. However, to date, there are no published reports of the clinical manifestations and molecular epidemiology of iGAS infections in NSW, despite having the largest population size in Australia. This study reviews the clinical and molecular epidemiology of iGAS infections in Western Sydney Local Health District (WSLHD), covering a catchment area of 876,500 people in NSW, to determine trends in iGAS infections over time and the association between emm types/clusters and types and severity of iGAS infections.
MATERIALS AND METHODS Design WSLHD covers four public hospitals which predominantly see adult patients: Westmead, Blacktown, Auburn and Mount Druitt hospitals. These hospitals refer their microbiology specimens to a single laboratory. We conducted a retrospective review of iGAS infections in WSLHD over two time periods, one recent year at the time of initiation of data collection with suspected high incidence of severe iGAS infections (year 2010) and another randomly selected year prior to 2010 (year 2008). There were no changes in the culture collection or testing practices between the two selected years to account for any differences observed. Data collection was commenced in year 2011. Ethics approval was obtained from the WSLHD human research ethics committee. Definitions Invasive GAS infection was defined as the isolation of S. pyogenes from a usually sterile site such as blood, joint, sterile tissue or cerebrospinal fluid. Those
2015 Royal College of Pathologists of Australasia
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
366
Pathology (2015), 47(4), June
SIVAGNANAM et al.
with invasive disease without positive cultures from sterile sites were not captured. Bacteraemia without a focus was defined as case patients with a positive blood culture without an identifiable clinical syndrome. Streptococcal toxic shock syndrome (TSS) was defined as proposed by the working group on severe streptococcal infections.11 Definite and probable TSS cases were combined to give a single category, TSS, for analysis in this study. Healthcare associated infection was defined as a positive culture in patients who had been hospitalised for at least 48 h at the time of onset of symptoms or the infection was a complication of a medical intervention such as surgery or indwelling catheters and devices. Data collection Cases were identified using the laboratory information system. Medical records were retrieved from all four hospitals and the following data were collected: patient demographics, baseline co-morbidities, clinical manifestations, laboratory results, predisposing factors including skin diseases, chronic medical conditions, immune status, pregnancy, smoking, intravenous drug abuse and alcoholism,4,12 treatment and outcome. Severity of the infection was determined based on clinical outcome (death versus cure) and the worst Acute Physiology and Chronic Health Evaluation (APACHE) II score within the first seven days of culture isolation (30 versus <30), collected systematically from patient records. An APACHE II cut-off value of 30 was used to indicate severity as patients admitted to the intensive care unit with septic shock with an APACHE II score of 30 and above have been shown to have a mortality rate of 82%.13 Overall baseline health was determined using Charlson’s Weighted Index of Co-morbidity (WIC).14 Microbiological methods The isolates were confirmed to be S. pyogenes prior to storage by beta-hemolysis on Horse Blood Agar (HBA), presence of Lancefield group A antigen, negative catalase test, positive reaction to pyrrolidonyl-beta-naphthylamide (PYR) and sensitivity to bacitracin. These stored isolates were subsequently retrieved and their emm types were determined using published methods by CDC11 with modifications. PCR was performed using the Qiagen HotStarTaq DNA polymerase kit (Qiagen, Australia) according to the manufacturer’s instructions with the following conditions: 958C for 15 min; 40 cycles of 948C for 30 s, 46.58C for 30 s and 728C for 3 min; 728C for 10 min; and a hold at 228C. The PCR mixture contained: 5 mL template DNA, 0.125 mL each of primer 1 (100 pmol mL1) and primer 2 (100 pmol mL1), 1 mL dNTPs (2.5 mM of each dNTP), 2.5 mL 10 PCR buffer, 0.1 mL Qiagen HotStarTaq DNA polymerase (5 units mL1), and water to 25 mL. PCR products (9 mL) were further purified using a PCR product pre-sequencing kit (USB Corporation, USA) according to the manufacturer’s procedures. These products were sequenced using the BigDye Terminator (version 3.1) cycle sequencing kit in the ABI Prism 3100 genetic analyser (Applied Biosystems, USA) Data analysis Using published data on emm types and clusters,7 we classified all emm types into emm clusters. Subsequent data analysis was performed for emm types as well as emm clusters. Firstly, the emm types and clusters were compared between years 2008 and 2010 to determine if there had been the emergence of a predominant strain or strains. Secondly, clinical data and the emm sequence typing and cluster data were assessed for associations between emm types/ clusters and the severity and type of infections. Thirdly, the factors associated with the most severe outcome, i.e. death, in iGAS infection were determined. Finally, geographic clustering was determined by comparing the postcodes and ethnicity of patients admitted during the two year groups. Statistical analysis was performed using SAS 9.4 (SAS Institute, USA). For continuous data, Wilcoxon sum rank test was used to compare medians; for non-continuous data, either Chisquared test or Fisher’s exact test was used as appropriate to compare proportions. Forward stepwise logistic regression was performed to determine risk factors associated with death in iGAS infection, including emm type. Other candidate variables for the multivariate model were those with p < 0.2 on univariate analysis which were considered a priori to be potential risk factors for mortality in iGAS infections. These variables were included in the model starting with the variable with the strongest association with mortality on univariate analysis. Likelihood ratio tests were used to guide inclusion of variables in the model and interaction terms were fitted. The multivariate analysis was performed using Stata V.11.0 (Stata Corp, USA).
RESULTS A total of 72 patients met the case definition for iGAS at WSLHD. There were 25 (35%) cases in 2008 and 47 (65%) in 2010. The sterile site from which GAS was isolated was blood in 60% of cases in 2008 and 64% in 2010. The characteristics of the two year groups are given in Table 1. In total, 58% of iGAS infections occurred in patients under the age of 50. Five (7%) patients were children under 15 years of age, including four patients under the age of 5. Fourteen (19%) patients were older adults, aged 65 years. Slightly more infections occurred in females (56%) compared to males. The dominant ethnic group was European (64%), followed by the Pacific Islander and Indigenous Australian/New Zealander group (21%). There were no significant differences between the baseline co-morbidities between the two year groups. While 48% of the patients had a chronic medical condition, most (82%) had a score of two or less in the Charlson’s WIC. Clinical presentations, predisposing factors and management Among the 72 iGAS infections, skin and soft tissue infection was the most common clinical presentation (57%), followed by bacteraemia without a focus (18%). Skin and soft tissue infection included two cases of necrotising fasciitis, which both occurred in 2010. Other manifestations of iGAS infections included pharyngitis/quinsy (13%), septic arthritis (7%), empyema (3%), endocarditis (1%) and pelvic inflammatory disease (1%). There were no cases of puerperal sepsis. More patients had TSS (21% versus 12%) and bacteraemia without a focus (21% versus 12%) in 2010 compared to 2008. Three (12%) patients in 2008 and 12 (26%) patients in 2010 had no identifiable risk factors for iGAS infection. Nine (36%) patients had skin diseases including eczema, burns, varicella zoster (VZV) infection or trauma in 2008 compared to only five (11%) cases in 2010. The one patient with recent VZV infection prior to iGAS infection was an adult. Treatment consisted of betalactam antibiotics alone in 67% and beta-lactam with clindamycin in 19% of cases. There was a small increase in the use of clindamycin in 2010 (27%) compared to 2008 (20%). Thirty percent of patients with GAS bacteraemia, 44% requiring management in the intensive care unit (ICU) and 60% with TSS received clindamycin in 2010 compared to 7%, 25% and 33% respectively in 2008. Of the 13 patients with toxic shock syndrome, three (23%) received intravenous immunoglobulin (IVIG) therapy, all in 2010. Severity of infection and outcome In 2008, all patients had an APACHE II score that was less than 30, compared to nine (19%) patients with an APACHE II score of 30 or more in 2010 ( p ¼ 0.02). A total of 13 (18%) patients required admission to ICU, without a significant difference between the two year groups. All patients survived in 2008; however, the case fatality rate was 13% in year 2010 ( p ¼ 0.06). Among the patients who died in 2010, 67% were female, 83% were at least 50 years of age and 67% had at least one chronic medical condition. Table 2 outlines the characteristics of patients with fatal and non-fatal outcomes. There was an associated skin disease in 22% of patients who survived compared to no patients with skin disease in those who died. In univariate analysis, the factors associated with death included APACHE II score 30 ( p < 0.001), admission to the intensive care unit ( p ¼ 0.07), toxic shock syndrome
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
EPIDEMIOLOGY OF INVASIVE GROUP A STREPTOCOCCUS INFECTIONS
Table 1
367
Differences in the clinical and demographic features of invasive group A Streptococcus infections between years 2008 and 2010 in Western Sydney
Demographics and clinical features Age, median years (IQR) Sex, male (%) Ethnicity European East and South East Asian Pacific and Indigenous Australian/New Zealander Middle Eastern South Central Asian Clinical Duration of illness prior to admission, median days (IQR) Bacteraemia Length of hospitalisation, median days (IQR) Type of infection Skin and soft tissue infections Pharyngitis/quinsy Endocarditis Empyema Septic arthritis Pelvic inflammatory disease Bacteraemia without a focus Possible predisposing factors Skin diseases (burns, varicella zoster virus infection, skin disease, wound) Smoking (n*¼59) Intravenous drug abuse Alcohol consumption Diabetes mellitus Obesity Lymphoedema Chronic medical condition Charlson’s weighted index of co-morbidity 0 1–2 3 or more Immunosuppression No known risk factors Healthcare associated infection Severity of infection Death APACHE II score 30 Admission to intensive care unit Toxic shock syndrome Acute renal failure (n*¼70) Coagulopathy (n*¼52) Acute respiratory distress syndrome (n*¼67) Acute liver injury (n*¼65) Treatment Surgical intervention required Antibiotics (n*¼69) b-lactam alone b-lactam þ clindamycin b-lactam þ other Non-b-lactam
2008 (n ¼ 25)
2010 (n ¼ 47)
P value
45 (38-50) 10 (40%)
40 (26-64) 22 (47%)
0.90 0.58
19 (76%) 1 (4%) 5 (20%) 0 0
27 4 10 5 1
0.34
3 (1–7) 15 (60%) 8 (2–16)
2 (1–7) 30 (64%) 7 (2–17)
0.36 0.75 0.89
16 3 1 1 1
25 (53%) 6 (13%) 0 1 (2%) 4 (9%) 1 (2%) 10 (21%)
0.36
(64%) (12%) (4%) (4%) (4%) 0 3 (12%)
(57%) (9%) (21%) (11%) (2%)
9 12 3 11 4 5 1 13
(36%) (52%) (12%) (44%) (17%) (21%) (4%) (54%)
5 10 1 8 8 6 3 21
(11%) (28%) (2%) (17%) (17%) (13%) (6%) (45%)
0.01 0.06 0.08 0.01 0.97 0.37 0.70 0.45
13 8 4 1 3 1
(52%) (32%) (16%) (4%) (12%) (4%)
19 19 9 4 12 3
(40%) (40%) (20%) (9%) (26%) (6%)
0.64
0.47 0.18 0.67
4 3 5 4
0 0 (16%) (12%) (21%) (27%) 0 (19%)
6 9 9 10 13 10 5 18
(13%) (19%) (20%) (21%) (28%) (27%) (11%) (42%)
0.06 0.02 0.71 0.33 0.50 0.98 0.10 0.07
11 (44%)
21 (46%)
0.89
19 (76%) 4 (16%) 0 2 (8%)
27 9 3 5
0.46
4
(61%) (20%) (7%) (11%)
IQR, interquartile range. *
Number of patients in whom this information was either documented or was able to be determined.
( p ¼ 0.07), bacteraemia ( p ¼ 0.08), bacteraemia without a focus ( p ¼ 0.009), acute renal failure ( p ¼ 0.04) and acute respiratory distress syndrome ( p ¼ 0.03). However, after multivariate analysis, APACHE II score was the only independent risk factor associated with mortality [adjusted odds ratio (OR) 1.26 per 1-unit increment, 95% confidence interval (CI) 1.03–1.54, p ¼ 0.02]. emm types and clusters A total of 55 (76%) isolates [17 (68%) in 2008 and 38 (81%) in 2010] were available for emm sequence typing. Of the 17 unavailable isolates, 12 (71%) were from non-blood specimens. Overall, 27 different emm types and 11 different emm clusters were obtained. The distribution of emm types and emm clusters
between the two year groups is shown in Fig. 1. The most common emm type was emm 1 (13%), followed by emm 101 (11%), emm 12 (7%), emm 66 (7%) and emm 3 (5%). Predominance of a single clone did not account for the increased numbers of iGAS infections seen in 2010. The distribution of emm types in 2008 and 2010 were different ( p ¼ 0.06), with the emergence of some new emm types in 2010 that were not present in 2008. For example, there were 16 additional emm types in 2010 that were not present in 2008, accounting for 49% (n ¼ 27/55) of infections; similarly, there were six additional emm types in 2008 which did not occur in 2010, accounting for 16% (n ¼ 9/55) of infections. The most common emm cluster was D4 (20%), followed by A-C3, E3, E4 and E6, representing 13% each of the total typed isolates. There were some notable differences in the distribution of emm clusters between the two
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
368
Pathology (2015), 47(4), June
SIVAGNANAM et al.
Table 2
Factors associated with death in invasive group A Streptococcus infections in Western Sydney in both 2008 and 2010 (using univariate analysis) Death (n ¼ 6)
Demographics Male Age 50 Pacific and Indigenous Australian/New Zealander Possible risk factors Chronic medical conditions No risk factors Diabetes mellitus Obesity Lymphoedema Malignancy Immunosuppression Healthcare associated infection Skin diseases Microbiology emm1 type (n*¼54) Clinical features Duration of illness prior to presentation, median days (IQR) APACHE 30 APACHE, mean (SD) Admission to intensive care unit Toxic shock syndrome Requirement for surgery Bacteraemia Bacteraemia without a focus Acute renal failure (n*¼68) Coagulopathy (n*¼51) Acute respiratory distress syndrome (n*¼65) Acute liver impairment (n*¼62)
No death (n ¼ 64)
Crude OR (95% CI)
2 (33%) 5 (83%) 1 (17%)
28 (44%) 25 (39%) 14 (22%)
0.64 (0.11–3.77) 7.80 (0.86–70.75) 0.72 (0.08–6.63)
0.69 0.08 1.00
4 1 1 2 1 1
30 12 11 9 3 7 5 4 14
2.27 0.87 0.96 3.06 4.07 1.63
(0.39–13.27) (0.09–8.12) (0.10–9.08) (0.49–19.20) (0.35–46.65) (0.17–16.02) – – –
0.42 1.00 1.00 0.24 0.31 0.53 1.00 1.00 0.34
1.79 (0.17–18.82)
0.52
(67%) (17%) (17%) (33%) (17%) (17%) 0 0 0
1 (20%) 1 6 42.7 3 3 1 6 4 4 3 2 3
(47%) (19%) (17%) (14%) (5%) (11%) (8%) (6%) (22%)
6 (12%)
(1–1) (100%) (7.8) (50%) (50%) (17%) (100%) (67%) (67%) (50%) (33%) (50%)
19 3 13.0 10 10 31 37 9 14 11 3 19
(7–31) (5%) (11.0) (16%) (16%) (48%) (58%) (14%) (23%) (24%) (5%) (34%)
P value
0.81 per 1-day increment (0.58–1.15) – 1.20 per 1-point increment (1.07–1.34) 5.4 (0.95–30.67) 5.40 (0.95–30.67) 0.21 (0.02–1.93) – 12.22 (1.95–76.79) 6.86 (1.13–41.43) 3.09 (0.54–17.59) 9.33 (1.19–72.99) 1.95 (0.36–10.59)
0.15 <0.01 0.07 0.07 0.21 0.08 0.01 0.04 0.33 0.03 0.66
CI, confidence interval; IQR, interquartile range; SD, standard deviation. *
Number of patients in whom this information was either documented or was able to be determined.
year groups ( p ¼ 0.16); there were four additional emm clusters in 2010 that were not present in 2008, representing 24% of the typed isolates. In 2008, although there were three additional emm clusters, this represented only 5% of the typed isolates. There were no significant differences in the emm types/ clusters between those with fatal and non-fatal outcomes. Five of the six patients with fatal iGAS had isolates available for typing. These belonged to emm types 1, 3, 12, 114 and 232 and emm clusters E4 (2 isolates), A-C3, A-C4, and A-C5.
within different ethnic groups to suggest geographic clustering. The four healthcare associated cases, one in 2008 and three in 2010, were unrelated and had four different emm types.
DISCUSSION This present study, which analyses the clinical and molecular epidemiology of iGAS infections in WSLHD, is the first study looking at iGAS infections in New South Wales, Australia. Most data in Australia on iGAS infections come from Victoria,12,15,16 Queensland17 and the Northern Territory.18 The annual incidence of iGAS in WSLHD was 19 cases per 100,000 hospital admissions in 2008 compared to 33 cases per 100,000
Geographic clustering There were no significant differences in the pattern of emm types and clusters isolated within the different postcodes or
6
No. of cases
5
4
3
2
1
0 1
12
3
6
A-C3 A-C4 A-C5 Clade Y
71 D2
33
53
70 D4
101
4 E1
66
90
9
E2
87
103
113
E3
8
28
89 E4
114
232
11
63
75
81
85
177
E6
Emm types and clusters 2008
2010
Fig. 1 Emm type and emm cluster distribution between 2008 and 2010 in Western Sydney.
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
EPIDEMIOLOGY OF INVASIVE GROUP A STREPTOCOCCUS INFECTIONS
Table 3
369
Association between various clinical manifestations of invasive group A Streptococcus infections, emm types and emm clusters
Clinical manifestation
emm cluster (frequency)
emm types (frequency)
Bacteraemia without primary focus
A-C3 (1) A-C5 (1) D4 (2) E2 (2) E3 (1) E4 (3) E6 (2) A-C3 (1) A-C4 (1) E1 (1) A-C4 (1) E2 (1) E4 (1) A-C3 (1) A-C5 (1) D2 (1) D4 (1) D4 (1) A-C3 (4) A-C4 (2) A-C5 (1) Clade Y (1) D4 (7) E1 (1) E2 (2) E3 (6) E4 (3) E6 (5) A-C3 (3) A-C4 (1) A-C5 (2) Clade Y (1) D4 (1) E4 (3) E6 (2) A-C4 (1) A-C5 (1) A-C3 (1) A-C4 (1) A-C4 (1) E4 (2)
emm1(1) emm3(1) emm101(1), emm70(1) emm66(1), emm90(1) emm103(1) emm232(2), emm114(1) emm11(1), emm75(1) emm1(1) emm12(1) emm4(1) emm12(1) emm66(1) emm89(1) emm1(1) emm3(1) emm71(1) emm33(1) emm101(1) emm1(1) emm12(2) emm3(1) emm6(1) emm101(4), emm33(1), emm53(1), emm70(1) emm4(1) emm66(2) emm103(1), emm113(2), emm87(1), emm9(2) emm28(1), emm8(1), emm89(1) emm177(1), emm11(1), emm63(1), emm81(1), emm85(1) emm1(3) emm12(1) emm3(2) emm6(1) emm101(1) emm232(1), emm114(1), emm28(1) emm11(1), emm63(1) emm3(1), emm12(1)
admissions in 2010. This is an underestimation as our data does not capture those with clinical iGAS infection without a positive sterile site culture. This equates to at least four per 100,000 adults in 2008 and seven per 100,000 adults in 2010. This is a minimum incidence estimate only as we did not have information on the number of patients admitted to the children’s or private hospitals with iGAS infections; these additional cases would likely increase the incidence further. These rates are significantly higher than the rates of infection observed in other parts of Australia12 and in some European countries and USA.4,19–22 A New Zealand study has shown an annual incidence rate of greater than 8.1 cases per 100,000, driven largely by the indigenous Maori and Pacific Islander population.23 Approximately 21% of subjects in our study belonged to Pacific Islander or Indigenous Australian/New Zealand ethnic groups; our high incidence rate may be reflective of this. This study also highlights the increase in severity of iGAS infections over time. In 2010, more patients died, had a high APACHE II score and had TSS. In addition, a larger proportion of patients had no identifiable risk factors for iGAS infection (26% versus 12%) and had bacteraemia without a focus (21% versus 12%) in 2010. A recent study showed that children with bacteraemia without a focus had a lower risk for severe disease.24 In our study, only one child had bacteraemia without a
focus, in year 2010, and survived. In addition, bacteraemia without a focus was associated with a greater odds of death (OR 12.22; 95% CI 1.95–76.79). Skin breaches have been considered portals of entry for iGAS infections.4 A potential skin source was identified in more patients in 2008 compared to 2010. These observations support the hypothesis that there was an increase in the frequency of more severe infections in previously healthy hosts without an identifiable trigger in more recent times. In some countries, the increasing frequency of iGAS infections was attributed to the emergence of a new strain of S. pyogenes.25 In our study, there was no evidence for the emergence of a single clone contributing to the observed increase in frequency and severity of infections. However, similar to what is seen in New Caledonia,26 there is significant variation in the emm type distribution over time in Western Sydney, with 49% of infections in 2010 being caused by emm types that were absent in 2008. Additionally, we did not observe an association between emm types/clusters and type of infection (Table 3) or severity of infections. Emm 1 has been shown to be associated with more severe infections in some studies.4 In our study, the odds ratio of death in those who had emm 1 type was 1.79 (95% CI 0.17–18.82). The small number of isolates of this type may have reduced our ability to detect a statistically significant association between emm type and mortality.
Empyema Infective endocarditis Pharyngitis and/or quinsy
Pelvic inflammatory disease Septic arthritis
Septic arthritis with skin and soft tissue infection Skin and soft tissue infection
Toxic shock syndrome
Necrotising fasciitis Death
emm1(1) emm12(1) emm3(1) emm232(1), emm114(1)
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
370 Table 4
Pathology (2015), 47(4), June
SIVAGNANAM et al.
Coverage provided by the 30-valent M protein vaccine for infections caused by group A Streptococcus in Western Sydney
30-valent M protein vaccine coverage
emm cluster (frequency)
emm types (frequency)
Vaccine serotypes
A-C3 (7) E6 (4) E4 (4) A-C4 (4) A-C5 (3) E1 (2) Clade Y (1) E3 (1) D4 (3) E6 (2) E2 (4) D2 (1) E4 (1) E3 (2) D4 (2) E6 (1) E4 (2) D4 (6) E3 (4) E2 (1)
emm1(7) emm11(2), emm75(1), emm81(1) emm114(1), emm28(1), emm89(2) emm12(4) emm3(3) emm4(2) emm6(1) emm87(1) emm33(2), emm53(1) emm63(1), emm85(1) emm66(4) emm71(1) emm8(1) emm9(2) emm70(2) emm177(1) emm232(2) emm101(6) emm103(2), emm113(2) emm90(1)
Non-vaccine serotypes, protected by cross-opsonisation
Non-vaccine serotypes, not protected by cross-opsonisation Non-vaccine serotypes with no available data on cross-opsonisation
*
Proportion of infections in this study (n ¼ 55*) 47%
24%
4% 25%
Number of isolates available for emm sequence typing.
Other factors associated with death in iGAS infections shown in prior studies include age 50, TSS, bacteraemia without a focus, chemotherapy, diabetes, and liver disease.4 In our study, whilst age 50, TSS, APACHE II 30, bacteraemia without a focus, bacteraemia, admission to intensive care unit, acute renal impairment and acute respiratory distress syndrome were associated with mortality on univariate analysis, only APACHE II 30 was significantly associated with death in the multivariate analysis. Diabetes mellitus, immunosuppression, and chronic medical conditions did not appear to influence outcome (Table 2). There is growing evidence to suggest a mortality benefit from adjunctive therapy with clindamycin and IVIG in iGAS infections.27,28 In our study, clindamycin and IVIG use were very low. However, there was an increase in usage of clindamycin in 2010 compared to 2008, particularly in patients with severe infections. In addition, all three patients who received IVIG were treated in 2010. Two of the three patients had severe disease requiring intubation and inotropes as soon as they presented to the emergency department. There may be a number of reasons for these low numbers. Firstly, there are no randomised controlled trials to suggest a clear benefit from clindamycin and IVIG use. Secondly, opinions vary on which subpopulations with iGAS are most likely to benefit from these treatments. However, increased use of clindamycin and IVIG in 2010 may be related to higher numbers of severe infections observed in 2010 as well as increased clinician awareness. In addition, infectious diseases physicians and intensivists are more likely to be involved in the care of patients with severe iGAS infections; they may be more likely to recommend clindamycin and IVIG. The fatality rate of iGAS is not negligible, particularly when the condition occurs in hosts without identifiable risk factors. Phase I/II studies among healthy adults have shown that a 26valent M protein vaccine, which was based on the predominant strains in USA, may be safe and immunogenic.29 A 30-valent M protein vaccine has been shown to evoke cross-opsonic bactericidal antibodies against non-vaccine serotypes resulting in greater vaccine efficacy10 The 30-valent M protein based
vaccine consists of 12 serotypes found in our population. As shown in Table 4, if vaccine serotypes alone were used to predict immunity, coverage would be 47% (26/55). However, the emm clusters that contain at least one representative included in the vaccine may demonstrate cross protection, except for the D4 emm cluster.7 Given this extension of coverage beyond the emm types contained in the vaccine, at least an additional 8 emm types representing 24% (13/55) of infections would be covered by the 30-valent M protein vaccine in our population. Data are not currently available on the bactericidal activity of vaccine induced antibodies against strains belonging to emm types 90, 101, 103, 113, 177 and 232 which accounted for 25% (14/55) of all our invasive infections. In addition, there were two isolates that belonged to emm 70; the vaccine has been shown to induce only low level antibodies against this type.10 Of the six patients with a fatal outcome, five isolates were available for emm typing. Four of the five were emm types with coverage in the 30-valent vaccine; data are unavailable regarding emm 232. Overall, the 30-valent M protein vaccine which covers 90% of invasive isolates in the US and 78% in Europe10 would have covered at least 71% of the emm types isolated in Western Sydney during this study. Limitations of our study include the small sample size, retrospective nature of data collection, limiting the study to the adult public hospital patients only, using laboratory information system as the sole surveillance tool to determine cases of iGAS infections and not determining the toxin profile of the strains as virulence may be related to toxin production rather than the emm types. Additionally, not all iGAS isolates from the study period were available for emm typing. However, the aim of our study was to describe the clinical and molecular epidemiology of iGAS infections in Western Sydney and to draw hypotheses which would require validation in future large, prospective studies. The study highlights the need to prospectively monitor GAS infections systematically. A recent study suggested potential benefit from penicillin prophylaxis in patients with recurrent leg cellulitis.30 Our data do not provide evidence of clustering to support the use of antibiotic prophylaxis in contacts of patients with iGAS infections. However,
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
EPIDEMIOLOGY OF INVASIVE GROUP A STREPTOCOCCUS INFECTIONS
postcodes and ethnicity alone do not adequately capture epidemiological links between patients. A secondary case in a household contact was observed in this cohort, both belonging to emm type 101; larger studies would be required to determine the incidence of iGAS infection among household contacts and the benefit of antibiotic prophylaxis in our population. In conclusion, there is evidence to suggest an increase in the frequency and severity of iGAS infections in metropolitan Sydney. There was no clear association between the increased frequency and severity and various emm types/clusters, although there were new emm types/clusters seen in 2010 that were not present in 2008. Comprehensive monitoring of iGAS infections through public health notification, contact tracing and emm sequence typing of these isolates will help determine the role of vaccination and antibiotic prophylaxis in our population. Acknowledgements: Mitchell Brown and Marion Nguyen for characterising a Lancefield group A positive, groups B, C, F and G negative, bacitracin susceptible Streptococcus dysgalactiae, emm type stG485, which was originally misidentified as GAS; this was excluded from the 2010 analysis. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Address for correspondence: Shobini Sivagnanam, 906 Dexter Ave N., Apt L539, Seattle, WA 98109, USA. E-mail: [email protected]
References 1. Carapetis JR, Steer AC, Mulholland EK, et al. The global burden of group A streptococcal diseases. Lancet Infect Dis 2005; 5: 685–94. 2. Aziz RK, Kotb M. Rise and persistence of global M1T1 clone of Streptococcus pyogenes. Emerg Infect Dis 2008; 14: 1511–7. 3. Dooling KL, Crist MB, Nguyen DB, et al. Investigation of a prolonged group a streptococcal outbreak among residents of a skilled nursing facility, Georgia, 2009-2012. Clin Infect Dis 2013; 57: 1562–7. 4. Lepoutre A, Doloy A, Bidet P, et al. Epidemiology of invasive Streptococcus pyogenes infections in France in 2007. J Clin Microbiol 2011; 49: 4094–100. 5. Lancefield RC. Current knowledge of type-specific M antigens of group A streptococci. J Immunol 1962; 89: 307–13. 6. Centers for Disease Control and Prevention. Emm types as proportions of total diseases isolates in six global regions. 28 Sep 2009; cited 19 May 2014. http://www.cdc.gov/ncidod/biotech/strep/emmtype_proportions.htm. 7. Sanderson-Smith M, De Oliveira DM, Guglielmini J, et al. A systematic and functional classification of Streptococcus pyogenes that serves as a new tool for molecular typing and vaccine development. J Infect Dis 2014; 210: 1325–38. 8. Centers for Disease Control and Prevention. Protocol for emm typing. 8 Feb 2008; cited 19 May 2014. http://www.cdc.gov/ncidod/biotech/strep/protocol_emm-type.htm. 9. Rogers S, Commons R, Danchin MH, et al. Strain prevalence, rather than innate virulence potential, is the major factor responsible for an increase in serious group A streptococcus infections. J Infect Dis 2007; 195: 1625–33.
371
10. Dale JB, Penfound TA, Chiang EY, et al. New 30-valent M protein-based vaccine evokes cross-opsonic antibodies against non-vaccine serotypes of group A streptococci. Vaccine 2011; 29: 8175–8. 11. The Working Group on Severe Streptococcal Infection. Defining the group A streptococcal toxic shock syndrome. Rationale and consensus definition. JAMA 1993; 269: 390–1. 12. O’Grady KA, Kelpie L, Andrews RM, et al. The epidemiology of invasive group A streptococcal disease in Victoria, Australia. Med J Aust 2007; 186: 565–9. 13. Knaus WA, Draper EA, Wagner DP, et al. APACHE II: a severity of disease classification system. Crit Care Med 1985; 13: 818–29. 14. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40: 373–83. 15. Carapetis J, Robins-Browne R, Martin D, et al. Increasing severity of invasive group A streptococcal disease in Australia: clinical and molecular epidemiological features and identification of a new virulent M-nontypeable clone. Clin Infect Dis 1995; 21: 1220–7. 16. Lithgow A, Duke T, Steer A, Smeesters PR. Severe group A streptococcal infections in a paediatric intensive care unit. J Paediatr Child Health 2014; 50: 687–92. 17. Norton R, Smith HV, Wood N, et al. Invasive group A streptococcal disease in North Queensland (1996-2001). Indian J Med Res 2004; 119 (Suppl): 148–51. 18. Carapetis JR, Walker AM, Hibble M, et al. Clinical and epidemiological features of group A streptococcal bacteremia in a region with hyperendemic superficial streptococcal infection. Epidemiol Infect 1999; 122: 59–65. 19. Montes M, Ardanuy C, Tamayo E, et al. Epidemiological and molecular analysis of Streptococcus pyogenes isolates causing invasive disease in Spain (1998-2009): comparison with non-invasive isolates. Eur J Clin Microbiol Infect Dis 2011; 30: 1295–302. 20. Darenberg J, Luca-Harari B, Jasir A, et al. Molecular and clinical characteristics of invasive group A streptococcal infection in Sweden. Clin Infect Dis 2007; 45: 450–8. 21. Lamagni TL, Darenberg J, Luca-Harari B, et al. Epidemiology of severe Streptococcus pyogenes disease in Europe. J Clin Microbiol 2008; 46: 2359–67. 22. O’Loughlin RE, Roberson A, Cieslak PR, et al. The epidemiology of invasive group A streptococcal infection and potential vaccine implications: United States, 2000-2004. Clin Infect Dis 2007; 45: 853–62. 23. Safar A, Lennon D, Stewart J, et al. Invasive group A streptococcal infection and vaccine implications, Auckland, New Zealand. Emerg Infect Dis 2011; 17: 983–9. 24. Gauguet S, Ahmed AA, Zhou J, et al. Group A streptococcal bacteremia without a source is associated with less severe disease in children. Pediatr Infect Dis J 2015; 34: 447–9. 25. Meehan M, Murchan S, Bergin S, et al. Increased incidence of invasive group A streptococcal disease in Ireland, 2012 to 2013. Euro Surveill 2013; 18: 20556. 26. Baroux N, D’Ortenzio E, Amedeo N, et al. The emm-cluster typing system for Group A Streptococcus identifies epidemiologic similarities across the pacific region. Clin Infect Dis 2014; 59: e84–92. 27. Carapetis JR, Jacoby P, Carville K, Ang SJ, Curtis N, Andrews R. Effectiveness of clindamycin and intravenous immunoglobulin, and risk of disease in contacts, in invasive group a streptococcal infections. Clin Infect Dis 2014; 59: 358–65. 28. Linner A, Darenberg J, Sjolin J, Henriques-Normark B, Norrby-Teglund A. Clinical efficacy of polyspecific intravenous immunoglobulin therapy in patients with streptococcal toxic shock syndrome: a comparative observational study. Clin Infect Dis 2014; 59: 851–7. 29. McNeil SA, Halperin SA, Langley JM, et al. Safety and immunogenicity of 26-valent group a streptococcus vaccine in healthy adult volunteers. Clin Infect Dis 2005; 41: 1114–22. 30. Thomas KS, Crook AM, Nunn AJ, et al. Penicillin to prevent recurrent leg cellulitis. N Engl J Med 2013; 368: 1695–703.
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.