Canada Communicable Disease Report

1 April 2008

Volume 34

Number 04

Outbreak of invasive group A streptococcal disease in two hospitals, Ontario, 2003

L Nolan, MD, MHSc, FRCPC (1), R Schertzberg, MLT, ART, CIC (2), S Wilson-Clark, BScN MHSc (3), A McGeer, MD, FRCPC (4), R Pellizzari, MD, MSc, CCFP, FRCPC (5), C Steingart, MD, FRCPC (2)

1. Region of Waterloo Public Health, Waterloo, Ontario
2. Grand River Hospital, Kitchener, Ontario
3. Canadian Field Epidemiology Program, Public Health Agency of Canada, Ottawa, Ontario
4. Mount Sinai Hospital, Toronto, Ontario
5. Toronto Public Health, Toronto, Ontario

Invasive group A streptococcal disease is caused by Streptococcus pyogenes. Typically group A streptococci cause mild human diseases, such as pharyngitis and cellulitis. However, this pathogen can cause invasive and sometimes life-threatening illnesses such as necrotizing fasciitis, meningitis or streptococcal toxic shock syndrome. There has been a marked increase in serious group A streptococcal disease since the 1980s(1). Currently in Ontario, about 2 to 3 per 100,000 population will acquire invasive group A streptococcal disease in a year, with about 13% of these cases having streptococcal toxic shock syndrome(2). The case-fatality rate for invasive group A streptococcal disease is 13%; however, streptococcal toxic shock syndrome has been reported to have case-fatality rates as high as 81%(2).

Group A streptococci are spread by direct contact with secretions from the nose or throat of an infected person, or by contact with infected wounds on the skin. People are infectious when they are ill, but transmission from asymptomatic carriers is possible. Serious illness is more common at the extremes of age and in people who are immunosuppressed. Close contacts of cases of invasive group A streptococcal disease are more likely to have invasive disease than the general population⁽²⁾. Therefore, chemoprophylaxis has been considered as a means to eradicate nasopharyngeal colonization and prevent potential disease transmission.

In February 2003, an increase in invasive group A streptococcal disease was noted in Waterloo Region. Ten cases had been reported to Public Health over a period of 5 weeks when generally only 15 to 25 cases are reported each year (Figure 1). The community-acquired cases were not epidemiologically linked.

Figure 1. Number of reported invasive group A streptococcal cases by month, Region of Waterloo, 1996-2004

On 5 February, 2003, infection control staff from two hospitals reported a cluster of five post-operative group A streptococcal cases to the Region of Waterloo Public Health. Three of these cases had been previously identified as part of the cluster of invasive group A streptococcal disease, and their symptoms had begun in the community.

Note that some initial cases did not remain on the final line listing as they were found to be non-invasive or acquired outside of Waterloo Region.

Methods

An outbreak team with staff from both hospitals, including an administrator and infectious disease specialist as well as public health staff and external experts, was assembled to discuss the identification of post-operative cases of nosocomial invasive group A streptococcal disease at the two hospitals over the previous 2-week period. An investigation was conducted between 5 and 28 February, 2003, to determine the scope and source of the outbreak and to develop interventions.

Definitions

For the field investigation, a laboratory-confirmed case was defined as a person who lived in Waterloo Region (for a community-acquired case) or who had had a surgical procedure in a hospital in the Region (for a nosocomial case) with onset of symptoms between 1 January and 28 February, 2003, and the following:

• Group A streptococcus isolation from a normally sterile site (e.g. blood, cerebrospinal fluid, synovial fluid);

OR

• Group A streptococcus isolation from a non-sterile site with evidence of clinical severity.

Clinical severity was defined as necrotizing fasciitis, myositis, meningitis or hypotension with two or more of the following: liver function abnormality, renal failure, coagulopathy, acute respiratory distress syndrome and generalized erythematous macular rash.

Community cases were those whose symptoms began before hospital admission; nosocomial cases were those with symptom onset > 72 hours after admission or post-operatively.

Outbreak investigation

Case finding was initiated on 5 February, 2003, and included active wound surveillance in both hospitals, active community surveillance for group A streptococcal surgical site infections with assistance from community physicians, and active post-discharge surveillance of all surgical patients. Specifically, infection control staff contacted each patient within 48 hours of discharge by telephone and administered a standard questionnaire. A review of all surgical patients and those admitted in the 2 weeks before 5 February was undertaken. Hospital admissions and emergency departments were monitored for additional cases. The coroner was alerted to the situation, and all deaths were reviewed to rule out related cases. M serotyping and sic gene sequence typing was performed on available strains of S. pyogenes⁽³⁾.

All nosocomial cases were reviewed for epidemiologic links. Epidemiologic links are defined as commonalities or links based on person, place or time, including history of staff and patient movements. Patients and health care workers routinely move between the two hospitals.

Control measures

On 9 February, one fatal case of post-operative necrotizing fasciitis was reported. The surgery for this case had occurred on 5 February, before control measures had been implemented. Upon identification of the fatality, control measures were enhanced. The outbreak team wanted to ensure that nosocomial transmission was interrupted while the investigation was ongoing. Control measures, started on 5 February and enhanced on 9 February, included peri-operative antibiotics (with doses before and after surgery) and suspension of elective surgery. This was accompanied by health care worker education, which reinforces the importance of routine infection control measures. The investigation team undertook an immediate inspection of the operating room to identify possible environmental sources and verify appropriate ventilation in each room within the operating suite. Thorough environmental cleaning of operating rooms was undertaken, with subsequent emphasis on locations where environmental screening was positive. The frequency of laundry removal was increased.

Environmental and epidemiologic investigations assisted in better understanding the scope and source of the outbreak. The findings of these investigations continued to inform and shape the control measures throughout the duration of the outbreak.

Health care worker screening

Screening for group A streptococcal carriage was undertaken for hospital staff and physicians. Screening included taking throat, peri-anal and vaginal swabs. Hospital and medical staff who tested positive were treated with appropriate antibiotics (cephalexin) and had repeat screening on days 14 and 28 after treatment to confirm eradication. If the person was still positive on follow-up screening, rifampin and a second-generation cephalosporin were prescribed. All family members of positive hospital personnel were also screened and treated if positive. The initial screening included hospital employees and physicians who provided care to the positive cases in the recovery room, operating room, invasive procedure clinics or in-patient units and was conducted over a 1-week period. On 9 February, screening was expanded to include personnel with any link to the circle of care, including staff from critical care, pharmacy, housekeeping and medical imaging. It was expanded to expedite the identification of possible carriers.

The screening process was carried out using extended hours at the Occupational Health Centre. Both individual and group education sessions were held. The swabs were coded numerically, and the staff being screened were assured that results would be handled in a confidential manner. A communication plan was developed to dispel fear, answer questions and keep the investigation transparent.

Environmental screening

Environmental testing in the form of settle plates was initiated on 9 February. At least two settle plates per 24 hours (one per 12 hour shift) were placed in all operating rooms, and operating room change rooms and lounges to detect environmental shedding of group A streptococci. Upon detection of the first positive settle plates, daily sign-in sheets were placed on the doors of all operating room change rooms and lounges.

Results

A total of six cases of nosocomial invasive group A streptococcal disease were confirmed in the cluster. Two presented as necrotizing fasciitis, one of which was fatal. A total of six cases of community-acquired invasive group A streptococcal disease were identified (Figure 2). Ultimately, four of the six nosocomial cases were determined to be linked (Figure 3). The remaining two nosocomial cases were not epidemiologically linked. No nosocomial cases occurred in individuals undergoing procedures after the control measures had been implemented.

Hospital personnel screening

Hospital personnel screening was a success as a result of the commitment of all personnel and administration in the hospitals. A total of 737 hospital personnel were screened (100% of the targeted individuals) of whom seven (1%) were positive. All family members of positive individuals were negative upon screening. Two out of seven individuals worked at both hospitals. One person required multiple courses of treatment to eradicate carriage, which were ultimately successful. This particular individual (who was epidemiologically linked with the four nosocomial cases in the outbreak) was asymptomatic and had no skin lesions.

Figure 2. Number of invasive group A streptococcal outbreak associated cases, Region of Waterloo, January - February 2003

Figure 3. F Four linked nosocomial group A steptococcal cases, by onset date, Region of Waterloo,
January - February 2003

Epidemiologic investigation

The epidemiologic investigation found that the nosocomial cases had had surgery after the debridements of the two earliest community-acquired necrotizing fasciitis cases had taken place. It was initially hypothesized that the early debridements may have been related to the outbreak, but the evidence does not substantiate this.

Alone, epidemiologic findings did not clarify the linkages between cases. Three nosocomial cases had undergone procedures at one hospital, and two at the other hospital; one had had procedures at both hospitals. More than one health care worker had a positive screening result. Laboratory results, including environmental findings, were critical to understanding the outbreak. These results, in combination with the epidemiologic investigation, identified links between four of the six nosocomial cases and one health care worker carrier.

Environmental findings

There were no potential sources of contamination identified during the environmental assessment of the operating suite. Daily laundry services were implemented to avoid the potential for soiled linens to accumulate over the weekend and to reduce the likelihood of transmission due to environmental shedding.

In total, 412 settle plates were placed before the epidemio-logically linked carrier was identified and treated. Three of the 412 settle plates from the operating room change room and lounge at both facilities were positive for group A streptococci. All isolates from the settle plates were type M1. None of the operating room settle plates was positive. The sign-in sheets from the change rooms connected two of the three settle plates with one positive hospital staff member. One of three positive settle plates that were placed before treatment of the carrier could not be clearly epidemiologically linked to the positive carrier, who, according to history and sign-in sheets, had not been in the same room for 24 hours or more before the settle plate was placed.

Additional settle plate testing, further described below, was carried out after the carrier had been treated. Three additional positive settle plates (type M1) were identified, for a total of six.

Laboratory findings and additional screening

Laboratory findings are summarized in Table 1. Three community-acquired cases with necrotizing fasciitis and all six nosocomial cases were type M1. Two of the community-acquired necrotizing fasciitis cases and one nosocomial case had a sic gene type of SIC1.02. One community-acquired necrotizing fasciitis case and one nosocomial case were of sic gene type SICA. Three other nosocomial cases and one of two hospital personnel colonized with M1 group A Streptococcus had strains of sic gene type SICB. In total there were six positive settle plates, and five were typed. All five typed isolates from settle plates were of the same sic type (SICB). Isolates from one nosocomial case and one settle plate that were epidemiologically linked to the positive carrier were not available for typing.

After treatment of the carrier, an additional three settle plates tested positive, all at times when the carrier was known to be in the change room. The first of these plates (overall the fourth settle plate that was positive) was placed 36 hours after treatment of the epidemiologically linked carrier had been initiated. The fifth and sixth positive settle plates were placed more than 4 days after treatment of the carrier had been initiated.

Upon detection of additional positive settle plates, previously positive persons were re-screened, and all results were negative. On 20 February (9 days after the initiation of treatment), the previously positive epidemiologically linked carrier was asked to take settle plates home for a 24-hour period. Of these settle plates, one yielded group A streptococcus, despite the fact that the individual's nasal, throat and rectal swabs were culture negative. This individual was therefore given a second course of antibiotic treatment. Follow-up swabs remained negative and all subsequent settle plates were negative.

Table 1. Summary of group A Streptococcus isolate testing

*Not all isolates were available for sic gene typing.

The cluster of three type SICB nosocomial cases, the untyped nosocomial case, the type SICB hospital staff carrier, four of the five type SICB settle plates, along with the untyped settle plate were all epidemiologically linked.

No specific epidemiologic link could be found between the type SICA community case and the type SICA nosocomial case, or between the type SIC1.02 community cases and the type SIC1.02 nosocomial case.

Discussion

While community-acquired invasive group A streptococcal disease is common, 12% of invasive group A streptococcal disease infections in Ontario are hospital acquired⁽⁴⁾. Expert recommendations suggest that one nosocomial case of invasive group A streptococcal disease in a post-operative patient should trigger an investigation, and if one or more possibly linked additional cases are found within 1 month the situation should be treated as an outbreak until typing results are available. Health care worker screening should be instituted if personnel appear to be epidemiologically linked to the group A streptococcal transmission⁽²⁾. An asymptomatic staff carrier will be identified in most, but not all, outbreaks associated with post-surgical disease⁽⁵⁾.

In this outbreak, the rapidity of the onset and its seriousness required intervention before epidemiologic links could be fully investigated. The concurrent increase in community cases, particularly community M1 serotype cases, meant that typing results were critical in understanding the epidemiology of the outbreak. The definitive link between staff and cases was not clear until all subtyping results were available. Although it has been recommended that only staff with epidemiologic links to cases should be screened, treated, and excluded from work⁽²⁾, the speed and complexity of this outbreak meant that all staff who might have been in contact with any of the cases were required to participate in staff screening before being able to return to work and while the outbreak investigation was going on. All carriers' family members were evaluated for asymptomatic infection before there was evidence regarding who may have been epidemiologically linked to the cases.

Full participation in staff screening can be challenging in a group A streptococcal outbreak⁽⁶⁾. This process was a large undertaking, involving screening of 737 hospital personnel from two hospitals. Our high compliance may have been related to the prompt initiation of broadly applied screening, the support by senior staff, and clear protocols to guarantee confidentiality. Providing education and answers to all questions was important in gaining trust. The number of staff members (seven) who were identified as colonized with group A streptococcus is consistent with the expected rate of 1% carriage in healthy adults in the community^(7,8).

Numerous hospital personnel work in both hospitals, including operating room, recovery room and critical care staff. On the basis of person, place and time, there was no one unifying explanation that linked all cases in this cluster, although some cases were linked through common hospital personnel or operating room location. Post-operative outbreaks are almost invariably linked to carriers who were in the operating room at the time of surgery⁽⁴⁾. In this cluster, it proved to be true for the four nosocomial cases whose isolates were related by sic typing: the staff carrier who was identified had been in the operating room during the procedures on all four patients. However, several of the other hospital personnel who were colonized with group A streptococcus had also been present in the operating room during the procedure of at least one of the cases. No epidemiologic links were identified between any positive carriers and the two community-acquired necrotizing fasciitis cases that preceded the nosocomial cases.

Use of settle plates was extremely helpful in identifying ongoing risk of transmission and thus the need for ongoing control measures. They were also invaluable in the epidemiologic investigation. Settle plates have been used previously to determine whether carrier identification and treatment were sufficient^(2,9). Subtyping of environmental samples was a critical part of our investigation, permitting the linkage between a staff member and some of the cases and ensuring that asymptomatic shedding had ceased before control measures were discontinued. It is unknown why follow-up cultures were unsuccessful at detecting persistent colonization in the health care worker. Asymptomatic scalp shedding is hypothesized.

Elective surgery resumed after all staff screening was complete and all settle plate results were negative for a period of 1 week. There are a number of possible explanations for the single settle plate that appeared unlinked to the carrier. Two concurrent shedders had not been previously reported; however, a second staff member may have been transiently colonized. Another possible explanation is lack of compliance with sign-in sheets; self-reporting of locations visited by personnel is known to be prone to inaccuracy.

In this outbreak, subtyping of M1 strains was useful in confirming significant epidemiologic links. Since M1 constitutes 20% to 30% of all cases of invasive group A streptococcal disease^(2,10), the fact that two invasive cases are both serotype M1 does not mean that they are linked closely⁽¹¹⁾. Subtyping supported the identification of the epidemiologic links and the need for further treatment to successfully eradicate asymptomatic carriage, and it allowed us to separate cases of group A streptococcal disease that were not linked. The presence of more than one subtype in a cluster of cases has been previously reported⁽¹²⁾.

Peri-operative antibiotics have been found to be effective in protecting against group A streptococcal wound infection in previous outbreaks⁽⁹⁾. In this case, peri-operative antibiotics were given to all surgical patients in both hospitals during the outbreak period. It has been previously observed that early infection control measures, including active surveillance, may interrupt transmission and prevent morbidity and mortality⁽¹³⁾. There were no additional cases once the carrier had been excluded from the operating room and antibiotics initiated to eradicate carriage. It would be interesting to determine whether peri-operative antibiotics or rapid staff screening and removal of potential carriers was the more important control measure in stopping transmission. Since it is impossible for hospital operations to cease, and specimens from hospital personnel require 24 hours for processing and reporting, peri-operative antibiotic prophylaxis could be an essential control measure.

Conclusion

This nosocomial cluster in post-operative patients occurred during a period when increased community-acquired invasive group A streptococcal disease was occurring. Concurrently, one operating room staff member became colonized and subsequently caused an outbreak associated with four nosocomial cases. No source was determined for two unlinked nosocomial post-operative cases. Rapid identification of the outbreak, screening of hospital personnel, the use of environmental settle plates, antibiotic prophylaxis and treatment, and excluding the carrier from work were essential tools in controlling the outbreak and effectively stopping transmission.

Acknowledgement

We would like to thank Naideen Bailey and Nadine Parsons from the Region of Waterloo Public Health for participation on the outbreak team, Greg Tyrrell at the National Centre for Streptococcus for the specimen typing, and numerous staff of two local hospitals.

No external funding was received for this research.

References

1. Health Canada. Prevention and control of occupational infections in health care. CCDR 2002;28(S1):112-17.

2. Public Health Agency of Canada. Guidelines for the prevention and control of invasive group A streptococcal disease. CCDR 2006;32(S2):1-26.

3. Tyrrell G, Lovgren M, Forwick B et al. M types of group A streptococcal isolates submitted to the National Centre for Streptococcus (Canada) from 1993 to 1999. J Clin Microbiol 2002;40(12):4466-71.

4. Daneman N, McGeer A, Low D et al. Hospital acquired invasive group A streptococcal infections in Ontario, Canada, 1992-2000. Clin Infect Dis 2005;41:334-42.

5. Prevention of Invasive Group A Streptococcal Infections Workshop Participants. Prevention of invasive group A streptococcal disease among household contacts of case patients and among postpartum and postsurgical patients: Recommendations from the Centers for Disease Control and Prevention. Clin Infect Dis 2002;35:950-59.

6. Balram C. Report on the group A streptococcal disease investigation in Saint John, April-June 2004. Fredericton: Health and Wellness, New Brunswick, October 2004.

7. Ejlertsen T, Prag J, Pettersson E et al. A 7 month outbreak of relapsing postpartum Group A Streptococcus infections linked to a nurse with atopic dermatitis. Scand J Infect Dis 2001;33:734-37.

8. Ontario Nursing Home Association GAS Task Force. Guidelines for the management of residents with group A streptococcal infection in long term care facilities. October 1997.

9. Mastro TD, Farley TA, Elliot JA et al. An outbreak of surgical-wound infections due to group A Streptococcus carried on the scalp. N Engl J Med 1990;323(14):968-72.

10. Centers for Disease Control. Nosocomial group A streptococcal infections associated with asymptomatic health-care workers – Maryland and California, 1997. MMWR 1999;48(08):163-66.

11. Kakis A, Gibbs L, Eguia J et al. An outbreak of group A streptococcal infection among health care workers. Clin Infect Dis 2002; 35:1353-59.

12. Arnold KE, Schweitzer JL, Wallace B et al. Tightly clustered outbreak of group A streptococcal disease at a long-term care facility. Infect Control Hosp Epidemiol 2006;27(12):1377-84.

13. Hoe N, Nakashima K, Grigsby D et al. Rapid molecular genetic subtyping of serotype M1 group A Streptococcus strains. Emerg Infect Dis 1999;5(2):254-62.