Community-Based Outbreaks in Vulnerable Populations of Invasive Infections Caused by Streptococcus pneumoniae Serotypes 5 and 8 in Calgary, Canada

Background Outbreaks of invasive pneumococcal disease (IPD) typically occur within institutions. Beginning in 2005, we detected an increase in serotype (ST) 5 and ST8 IPD cases, predominantly in homeless persons living in an open community. Methodology/Principal Findings CASPER (Calgary Area S. pneumoniae Epidemiology Research) surveillance study of all IPD (sterile site isolates) in our region (pop ∼1,100,000). Interviews and chart reviews of all cases and all isolates phenotypically analyzed and selected isolated tested by multi-locus sequence typing (MLST). Conclusions/Significance During 2005–2007, 162 cases of ST5 IPD and 45 cases of ST8 IPD were identified. The isolates demonstrated phenotypic and genotypic clonality. The ST5 isolates were sequence type (ST) 289 and demonstrated intermediate susceptibility to TMP-SMX. The ST8 isolates were predominantly ST1268, with a susceptible antimicrobial susceptibility profile. Individuals with ST5 IPD were more likely to be middle aged (OR 2.6), homeless (OR 4.4), using illicit drugs(OR 4.8), and asthmatic(OR 2.6). Those with ST8 were more likely to be male (OR 4.4), homeless (OR 2.6), aboriginal (OR7.3), and a current smoker (OR 2.5). Overlapping outbreaks of ST5 and ST8 IPD occurred in an open community in Calgary, Canada and homelessness was a predominant risk factor. Homelessness represents a unique community in which pneumococcal outbreaks can occur.


Introduction
Streptococcus pneumoniae (pneumococcus) is a leading cause of morbidity and mortality worldwide. Bacteremia, meningitis and pneumonia are the most common manifestations of invasive pneumococcal disease (IPD). Otitis media, sinusitis and nonbacteremic pneumonia are the most common noninvasive forms.
We describe two large outbreaks of IPD in Calgary, Alberta, caused by serotypes 5 and 8 (ST5, ST8. These outbreaks took place in the community, but preferentially affected specific populations. The early features of these outbreaks were previously reported [9,10]. The ST5 outbreak spread across western Canada and some features have been reported elsewhere [11,12,13].

Outbreak investigation
Active, population-based surveillance of IPD has been conducted in the city of Calgary and surrounding area since January 1, 1998 by the Calgary Area Streptococcus pneumoniae Epidemiology Research (CASPER) team [14]. The area is an integrated, publicly funded health region (Calgary Health Region, CHR) that includes the hospitals and outpatient facilities within the city of Calgary, as well as medical centres serving rural areas surrounding the city (population 1,111,614 in 2007).
Patients were deemed to be CHR residents on the basis of home postal code listed in the admission and laboratory records. If this information was unavailable, patients with provincial health care numbers were considered to be CHR residents if the culture specimen was submitted to a collection site within the CHR boundaries. This study was approved by the Conjoint Health Research Ethics Board of the University of Calgary and written, informed consent was obtained from persons with IPD were enrolled in the study.
Cases of IPD (infections with positive cultures from normally sterile body fluids) are identified through active laboratory surveillance in Calgary Laboratory services (CLS), a centralized laboratory service that provides clinical microbiology services to all hospitalized and ambulatory patients in the CHR. Although IPD is a notifiable disease in Alberta requiring notification of local public health officials of all cases to permit appropriate public health follow-up and reporting (passive surveillance), the CASPER surveillance project is separate from public health and collects more detailed information on cases and organisms.
For all IPD cases since 2003, a patient interview was conducted as well as a detailed chart review for all cases for 30 days after the diagnosis. Demographic and pre-existing health status information was collected, as were details of the clinical course and outcome.
Increased cases of serotype specific IPD (serotypes 5 and 8) were identified, starting in 2005 for ST8 and 2006 for ST5. These cases occurred only in adults so the analysis in this manuscript includes only persons aged 16 years or more.
The cases of ST5 IPD occurred frequently in the homeless population and a targeted intervention was directed to try to control transmission. A one-week vaccination campaign, with the 23-valent pneumococcal polysaccharide vaccine (PPV23), was conducted by the CHR Public Health program in December 2006. The PPV23 was offered at several homeless shelters and agencies. Approximately 655 doses of vaccine were administered. In the autumn of 2006, before the outbreak, 339 doses of pneumococcal vaccine had been provided to the homeless population, as that group was newly considered eligible for vaccine in the province. The outbreak was determined to be over 5-6 months later when the number of cases returned to baseline. During, the vaccine campaign, a survey to determine nasopharyngeal colonization with pneumococcus, with a brief questionnaire, was performed at the main homeless shelter in Calgary. Pneumococcal isolates obtained from this survey were also tested for antibiotic susceptibility and serotype.

Definitions
Aboriginal status was determined by patient interview and chart review.
Sepsis was defined as a systemic response to infection in the context of confirmed evidence of IPD. The systemic response was manifested by two or more of the following: temperature .38uC or ,36uC, white blood cell count .12,000/mL, ,/mL 4000/mL or 10% immature forms (bands), heart rate .90 beats per minute or a respiratory rate .20 breaths per minute or arterial CO 2 tension ,32 mmHg.
Septic shock was defined as sepsis associated with organ dysfunction: hypotension despite adequate fluid resuscitation and requiring use of vasoactive drugs, defined by systolic blood pressure greater than or equal to 90 mmHg; metabolic acidosis; oliguria; hypoxia, not explained by primary respiratory distress; confusion; renal impairment, creatinine ./ = to 177 mmol/L outside the context of pre-existing renal disease; coagulopathy, evidenced by platelets #100,000/mm 3 610 6 /L, and/or evidence of disseminated intravascular coagulation (defined by prolonged clotting times,) low fibrinogen level, and/or the presence of five and degradation products); liver involvement, alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin greater than or equal to twice the upper limit of normal; or acute respiratory distress syndrome.

Laboratory methods
Isolates were confirmed to be pneumococcus by standard methodology including Gram stain, colonial morphology on blood agar, bile solubility, susceptibility to Optochin and pneumococcal antibody agglutination (Phadebact Pneumococcus, Boule Diagnostic AN, Sweden). Susceptibility testing was performed by broth microdilution (PML Microbiologicals, bioMérieux Canada, Inc.St. Laurent , QC, Canada) and results were interpreted according to Clinical and Laboratory Standards Institute (CLSI) guidelines [15,16,17,18].
For the NP colonization survey at the homeless drop-in centre, a single NP swab was obtained from each participant, using the WHO recommended procedure [19]. Questionnaires were administered by trained nursing personnel. Serotypes were determined by the Quellung reaction, with use of commercial antisera obtained from the Statens Seruminstitut (Copenhagen, Denmark) at the National Centre for Streptococcus (Edmonton, Canada). Five isolates of each ST5 and ST8 (chosen from the chronologic list of isolates analyzing the first, last and the beginning of each quartile) were analyzed by multilocus sequence typing (MLST) at CLS. The sequences of the internal fragments of the 7 housekeeping loci used in the pneumococcal MLST scheme were determined for each isolate, as described elsewhere [20].

Results
A total of 1382 laboratory confirmed cases of IPC occurred in the CHR from January 1998 to December 2008. There was an average of 118 cases per year ( Table 1). The annual incidence of IPD is shown in Figure 1. Figures 2 and 3 show the number of cases of serotypes 8 and 5 IPD monthly, respectively, for the defined outbreak period from 2005 to 2007. During this period, 162 cases of IPD due to ST5 and 45 cases of ST8 occurred.  Table 5 describes the multivariate logistic regression analysis of demographic features, risk conditions and outcomes associated with serotypes 5 and 8 cases for all cases where complete was available.
The ST5 isolates demonstrated phenotypic and genotypic clonality during the outbreak. All isolates were susceptible to penicillin, erythromycin and ceftriaxone and all had intermediate susceptibility to trimethoprim-sulfamethoxazole (TMP-SMX). The MLST analysis of 5 representative ST5 isolates identified that all were sequence type ST289. This is similar to the previously described strain from Columbia also described by the Pneumococcal Molecular Epidemiology Network (Columbia 5 -19 strain) [21,22,23].
The ST8 isolates also demonstrated phenotypic and genotypic clonality during the outbreak. All isolates were susceptible to penicillin, erythromycin, ceftriaxone and TMP-SMX. The MLST analysis of 5 representative isolates identified sequence type ST1480 for the first isolate and ST1268 for the remaining 4 isolates.

Discussion
Outbreaks of Streptococcus pneumoniae have been described in closed populations such as hospitals, daycares, schools, military barracks, nursing homes and prisons [1,2,3,4,5,6,7]. The ST5 outbreak in Calgary and across Western Canada is the first one described in a disadvantaged but relatively open community [13]. The outbreak in Calgary area began before the outbreak in Vancouver as described by Romney and colleagues [13]. In addition to the ST5 outbreak, there was an outbreak of ST8 IPD in Calgary that peaked in 2005, the year that an increase in ST5 cases began. This is the first described outbreak of ST8 IPD. Overlapping outbreaks of this kind have not been previously described.
Disadvantaged populations including aboriginals, homeless and impoverished people suffer higher rates of pneumococcal disease [24,25]. Outbreaks, however, have rarely been described in these populations [13,26]. Mercat and colleagues described an outbreak in homeless shelters in Paris in the late 1980s due to serotype 1 [26].
Both of the outbreaks occurred predominantly in persons at increased risk for IPD, although the particular factors differed for each serotype. The ST5 cases were more common in homeless persons and those reporting illegal drug use, while the ST8 cases were more common in males, homeless persons and aboriginals. The Calgary homeless population was estimated at 4,060 in 2008 and this number had increased 18% from 2006 [27]. This population is predominantly male (78%), caucasian (62%) with 29% between the age of 45 and 64 [27]. Eighteen percent of individuals in a survey in a subset of the population had been Calgary less than 1 year [28]. Self reported substance abuse in the homeless population was 85% and 24% had mental health issues [28].
Vaccination rates have traditionally been quite low in disadvantaged populations especially those who do not seek regular medical care. In August 2006 the Alberta Advisory Committee on Communicable Diseases recommended offering the 23-valent polysaccharide pneumococcal vaccine to homeless individuals and also those who are hepatitis C positive, and this position was further strengthened by a statement from the National Advisory Committee on Immunization in 2008 [29]. Both of the outbreak strains are be covered by the 23 valent polysaccharide pneumococcal vaccine, however the efficacy of this vaccine is limited [30,31]. The 7 valent conjugate vaccine (PrevnarH) available at the time of the outbreak does not cover the two outbreak strains [32]. The new 10-valent and 13-valent conjugate pneumococcal vaccines, (the latter now in routine use across Canada), cover these serotypes but are routinely so far only used in pediatric patients [33].
Serotype 5 is a particularly invasive serotype with invasiveness index 60 fold higher than some of the least invasive serotypes (3, 6A, 15) [34]. The ST5 strain has been particularly effective in this outbreak evidenced by the significant presence across Western Canada documented by both our jurisdiction and British Columbia [13]. Serotype 5 was otherwise uncommon in our area among invasive cases and in colonization surveys [35]. Serotype 8 was also uncommon among invasive isolates and accounted for only 1 of 678 colonizing strains identified over multiple years. [35]. However, ST8 is commonly an invasive serotype [34]. The majority of the ST8 sequence types were ST1268, with only the very first isolate tested being slightly different (ST1480 which only differs from ST1268 by 1 allele (spi)).
The demographic features, risk factors, clinical course and outcome differed for both ST5 and ST8 cases of IPD, compared with the rest of the adult cases of IPD before, during and after the outbreaks. In addition data from the survey questionnaire during the NP colonization survey several other factors that may be associated with acquisition of pneumococcal disease (Table 6), including cocaine and crack cocaine use, marijuana use as well as the sharing of bottles cigarettes, crack pipes and ''joints''. As suggested by Romney, crack pipes and other implements may be a route of transmission for pneumococcus [13].
Although both of the described outbreaks had their onset in the colder months of the year, a clear correlation between seasonality and this outbreak were not well-established (Figures 2 & 3). Crowding in the homeless shelters is more common in the winter months, but shelter use remains significant throughout the year.
Outbreak control measures utilized in the Calgary area included a vaccination campaign in the largest of the homeless shelters, December 2006, that did not result in a rapid resolution of the outbreak. The limited uptake of vaccine, limited vaccine effectiveness, a persistent susceptible population and ongoing crowding may have been factors in the persistence of the outbreak, especially for ST5 disease.
In summary, we describe overlapping outbreaks of ST5 and ST8 in an open homeless community. Significant risk factors that contributed to either or both of these outbreaks included homelessness, illicit drug use, smoking, aboriginal ethnicity and smoking. Targeted vaccination programs, preferably with proteinpolysaccharide conjugate vaccines that reduce both pneumococcal disease and transmission, may better control future outbreaks. Epilepsy or other seizure disorder n/a n/a 2.66 ns 15 not significant (P.0.15). 16 Missing values assumed non-smoker. doi:10.1371/journal.pone.0028547.t005