Identification of high-risk populations for serious infection due to S. pneumoniae will permit appropriately targeted prevention programs.
We conducted prospective, population-based surveillance for invasive pneumococcal disease and laboratory confirmed pneumococcal pneumonia in homeless adults in Toronto, a Canadian city with a total population of 2.5 M, from January 1, 2002 to December 31, 2006.
We identified 69 cases of invasive pneumococcal disease and 27 cases of laboratory confirmed pneumococcal pneumonia in an estimated population of 5050 homeless adults. The incidence of invasive pneumococcal disease in homeless adults was 273 infections per 100,000 persons per year, compared to 9 per 100,000 persons per year in the general adult population. Homeless persons with invasive pneumococcal disease were younger than other adults (median age 46 years vs 67 years, P<.001), and more likely than other adults to be smokers (95% vs. 31%, P<.001), to abuse alcohol (62% vs 15%, P<.001), and to use intravenous drugs (42% vs 4%, P<.001). Relative to age matched controls, they were more likely to have underlying lung disease (12/69, 17% vs 17/272, 6%, P = .006), but not more likely to be HIV infected (17/69, 25% vs 58/282, 21%, P = .73). The proportion of patients with recurrent disease was five fold higher for homeless than other adults (7/58, 12% vs. 24/943, 2.5%, P<.001). In homeless adults, 28 (32%) of pneumococcal isolates were of serotypes included in the 7-valent conjugate vaccine, 42 (48%) of serotypes included in the 13-valent conjugate vaccine, and 72 (83%) of serotypes included in the 23-valent polysaccharide vaccine. Although no outbreaks of disease were identified in shelters, there was evidence of clustering of serotypes suggestive of transmission of pathogenic strains within the homeless population.
Homeless persons are at high risk of serious pneumococcal infection. Vaccination, physical structure changes or other program to reduce transmission in shelters, harm reduction programs to reduce rates of smoking, alcohol abuse and infection with bloodborne pathogens, and improved treatment programs for HIV infection may all be effective in reducing the risk.
Citation: Plevneshi A, Svoboda T, Armstrong I, Tyrrell GJ, Miranda A, Green K, et al. (2009) Population-Based Surveillance for Invasive Pneumococcal Disease in Homeless Adults in Toronto. PLoS ONE 4(9): e7255. doi:10.1371/journal.pone.0007255
Editor: Landon Myer, University of Cape Town, South Africa
Received: September 28, 2008; Accepted: January 13, 2009; Published: September 29, 2009
Copyright: © 2009 Plevneshi et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: TIBDN surveillance from 2002 to 2006 was funded via unrestricted grants from Bayer AG. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Streptococcus pneumoniae is the most common cause of bacterial pneumonia, bacteremia and meningitis in adults, and is a major cause of morbidity and mortality in the general population , . Homeless adults may be at greater risk than other adults both because of underlying medical conditions that increase their risk of infection such as chronic liver disease or HIV infection –, and because communal living in shelters may be associated with transmission of pathogenic strains –. Authors of at least two reports of clusters of pneumococcal disease in shelters have recommended systematic vaccination of shelter residents , . However, no national guidelines currently include such a recommendation, and there are few data addressing the burden of illness associated with pneumococcal infection in the homeless.
The objective of this study was to describe the epidemiology of serious pneumococcal disease in homeless adults in metropolitan Toronto over a five year period.
The Toronto Invasive Bacterial Diseases Network (TIBDN) has conducted prospective, population-based surveillance of invasive pneumococcal disease in metropolitan Toronto, Canada (population, 2.5 million), since 1 January 1995 –. The surveillance network includes all hospital-based laboratories that provide clinical care to area residents. This comprises 25 licensed microbiology laboratories that serve 27 hospitals, long-term care facilities and out-patient offices. Personnel from these laboratories telephone the central TIBDN study office at the Mount Sinai Hospital in Toronto whenever S. pneumoniae is isolated from a sterile site or respiratory specimen. No laboratories serving this population used urinary antigen detection for the diagnosis of pneumococcal disease during the surveillance period. For each case, initial demographic data and the pneumococcal isolate are forwarded to the central TIBDN office. Additional clinical data, including patient co-morbidities, clinical course and outcome, antimicrobial therapy in the 3 months before presentation, and outpatient therapy for the current episode before the blood sample was obtained for culture, are acquired by chart review, patient interview, and by contacting the patient's attending physicians. Annual audits are conducted in each laboratory to ensure complete reporting. Surveillance and associated studies are approved by the research ethics boards of all participating institutions. In addition, invasive pneumococcal disease has been reportable in Ontario since January 1, 2001, with all cases of disease reported in Toronto investigated by Toronto Public Health. This study included all adult (> = 15 years of age) cases of invasive pneumococcal disease and laboratory confirmed pneumococcal pneumonia presenting between January 1, 2002 to December 31, 2006, with data from TIBDN and public health pooled to ensure accurate identification of homeless persons.
Population statistics were obtained from Statistics Canada, and annual incidence rates calculated using the estimated population on July 1 of each year. The population of homeless adults was estimated to be 5052, based on a single census conducted in April, 2006 .
Population vaccination uptake
A publicly funded 23-valent polysaccharide vaccination program was introduced into Ontario in 1996 . The seven-valent pneumococcal conjugate vaccine was licensed in Canada in June 2001, with a publicly funded program for infants initiated in Ontario in January 2005 .
Persons with invasive pneumococcal disease were classified as homeless if they had no fixed address, or gave their address as an emergency or transitional shelter. .
Invasive pneumococcal disease was defined as isolation of Streptococcus pneumoniae from a sterile body fluid with a compatible clinical syndrome. Sterile sites included blood, CSF, peritoneal fluid, pleural fluid, or abscess aspirate, but not bronchoalveolar lavage. Non-bacteremic pneumococcal pneumonia was defined as per Musher et al. , and required: (i) a clinical presentation including symptoms (eg. cough, sputum, fever) and physical findings consistent with pneumonia; (ii) radiographic confirmation of a pulmonary infiltrate; (iii) microscopic examination of a Gram stained sputum with at least moderate numbers of white blood cells per high power field and a predominance of Gram positive cocci in pairs or chains; (iv) a sputum culture that yielded S. pneumoniae but no other respiratory pathogen and (v) blood cultures were obtained and did not yield a pathogen.
All isolates were serotyped at the central study laboratory at the Mount Sinai Hospital, or the National Center for Streptococcus, Edmonton, Canada using commercial antisera (Statens Seruminstitut, Copenhagen, Denmark). Broth microdilution antimicrobial susceptibility testing was performed and interpreted by Clinical and Laboratory Standards Institute standards .
All data was entered in duplicate and analyzed using SAS for PC version 9.1 (SAS Institute, Cary, NC). Proportions were compared using chi-square or Fisher's exact tests and odds ratios calculated with 95% confidence intervals. Two comparisons of risk factors for invasive pneumococcal disease were conducted: one in which risk factors in cases occurring homeless adults were compared to all cases in other adults, and one in which cases in homeless adults were compared a cohort of other adults constructed by identifying, for each case in a homeless person, the four cases of disease in housed persons closest in age.
Over the five year period, there were 69 episodes of invasive pneumococcal disease in homeless persons, and 970 episodes in other residents of Toronto. During the time of the study, an estimated 5050 adults were homeless on any given day, with an estimated 27,000 persons homeless over the course of a year. Thus, the estimated rate of invasive disease in homeless persons was 273 per 100,000 per year, 30 fold higher than the concurrent rate in housed adults (9.0 per 100,000 per year). Homeless persons comprised 6.6% of all cases of invasive pneumococcal disease, but only 0.2% of the population of Toronto.
The typical seasonal pattern of invasive pneumococcal disease in Toronto, with highest rates in the winter months, and a nadir in July and August, was not present in cases in homeless persons (Figure 1). The higher proportion of invasive disease in homeless persons in the summer and fall could not be explained by differences in the age, differences in the occurrence of underlying illness or differences in the serotypes of infecting isolates (data not shown). However, smokers who were not homeless were also more likely to present with invasive pneumococcal disease in the summer months: 48 of 297 (16%) cases of invasive pneumococcal disease in smokers occurred in July and August, compared to 64 of 582 (11%,) in non-smokers, P = .03.
Bars represent cases in homeless persons, line represents cases in housed persons. The proportion of infections in homeless persons was significantly greater in summer (14/137, 10.2%) than in fall (23/305, 7.5%), winter (17/345, 4.9%) or spring (13/246, 5.3%), P = .05.
Clinical characteristics of invasive pneumococcal disease
Characteristics of invasive pneumococcal disease in homeless and housed adults are shown in Table 1. The most significant difference is in age distribution: homeless adults accounted for 3% (1/36) of cases of invasive pneumococcal disease in persons aged <30 years of age, 19% (49/256) of cases in those aged 30–50 years, 6% (10/157) of cases in those aged 50–60 years, and 1.5% (9/590) cases in those aged 60 years and older (P<.001). Homeless persons were more likely than others to be smokers, to abuse alcohol and use intravenous drugs, and to have chronic liver disease (see Table 1). Relative to age matched controls, they were more likely to have underlying lung disease (12/69, 17% vs 17/272, 6%, P = .006), but not more likely to be HIV infected (17/69, 25% vs 58/282, 21%, P = .73).
Of the 69 homeless persons who presented with invasive disease, six (9%) had received pneumococcal vaccine prior to admission. Six additional patients were vaccinated during or shortly after their hospitalization. Fifteen patients (22%) were known not to have been vaccinated; data were not available (either patients did not know their history, or they could not be contacted) for the remaining 42 (61%) patients. Similarly, influenza vaccination status was available for only 20 (29%) homeless patients with invasive pneumococcal disease; 10 (50%) of these had been vaccinated during the fall prior to their infection. In contrast, a pneumococcal vaccination history was available for 74% (719 of 970) patients who were not homeless, and an influenza vaccination history for 777 (80%): 28% had received pneumococcal vaccine, and 434 (45%) had been vaccinated against influenza.
The 69 episodes of invasive disease in homeless persons occurred in 58 patients: 6 patients had two episodes of disease, and 1 patient had six episodes. The proportion of patients with recurrent disease (7 of 58, 12%) was 5 fold higher than that for housed patients during the same time period (24 of 943, 2.5%, P<.001). One of eleven recurrent episodes may have been a relapse: it occurred 54 days after the first episode, and the infecting isolates were of the same serotype. All other episodes occurred more than four months apart, and, in all cases when isolates were available for typing, were caused by isolates of different serotypes. Three additional patients with episodes of invasive disease while they were homeless had another episode of disease during a time in which they had housing. Underlying liver disease was more common in the ten patients who had recurrent episodes of disease than in other homeless patients (6 of 10, 60% versus 12 of 48 (25%), P = .055), but the two groups of patients did not differ in other characteristics. The patient with six episodes of illness was a smoker and intravenous drug user in her 50s with HIV and hepatitis C co-infection, and known hepatic cirrhosis. The case fatality rate was 7 of 58 (12%) for first episodes of disease in homeless persons, compared to 3 of 7 (43%) for second episodes (P = .06).
Of 69 episodes of invasive disease in homeless patients, 8 (12%) were not admitted to hospital, and 7 (10%) left hospital against medical advice. In comparison, 114 (12%) of housed patients were not admitted to hospital, and 2 (0.2%) left against medical advice.
Clinical characteristics of laboratory-confirmed non-bacteremic pneumococcal pneumonia
Table 2 compares the clinical characteristics of laboratory-confirmed non-bacteremic pneumococcal pneumonia in homeless and housed adults. As for invasive pneumococcal disease, homeless persons were younger, less likely to be diabetic, more likely to have chronic liver disease and HIV infection, and much more likely to be smokers and to abuse alcohol. There were no statistically significant differences in patient characteristics or outcomes between homeless persons with invasive pneumococcal disease and those with laboratory confirmed, non-bacteremic pneumococcal pneumonia.
Isolates were available for 1309 of 1383 (95%) of episodes. The most frequently identified serotypes are shown in Table 3. Overall, 28 of 87 (32%) isolates associated with disease in homeless persons were of serotypes included in the 7-valent conjugate vaccine, and 72 of 87 (83%) isolates were of serotypes included in the 23-valent polysaccharide vaccine. There was no change over time in the proportion of infections in homeless persons caused by vaccines in the 7-valent conjugate vaccine (5/21, 24% of episodes in 2002/3 versus 9/26, 35% in 2006/7 were cause by these serotypes).
There were no differences in rates of resistance between isolates from homeless and housed adults (data not shown). Of the 88 isolates available from episodes of illness in homeless persons: 2 (2.3%) were non-susceptible to penicillin (both with MIC = 2 µg;ml), 18 (20%) were resistant to erythromycin, 3 (3.5%) were resistant to trimethoprim-sulfamethoxazole, and 1 (1.2%) was resistant to levofloxacin.
No outbreaks of pneumonia or pneumococcal disease were identified by either shelters or the department of public health in Toronto during this five year period. However, there was some evidence in our surveillance data of clustering of episodes of illness due to particular serotypes, suggesting that transmission may have occurred within shelters or other shared accommodation For instance, all five episodes due to serotype 7F (6% of isolates from homeless were of serotype 7F, compared to 2.9% among housed cases, P = .18) occurred over an eight month period; all three episodes due to serotype 11B (3% of isolates from homeless vs 0.7% others, P = 0.04) occurred over an eight month period. The first isolate of serotype 12F (20% of isolates from homeless vs 4% others, P<.001), was identified in September 2003, and all subsequent isolates occurred in residents of the largest shelter or those living on the street. In addition, all ten isolates of serotype 22F identified in homeless persons were resistant to erythromycin (MIC> = 64 ug/ml); compared to 14 of 79 (18%) other serotype 22F strains (P<.0001).
Homelessness is an important and growing problem in the developed world –. Previous studies have documented a significant burden of illness among homeless persons due to underlying chronic medical conditions, tuberculosis, HIV infection, trauma, and mental illnesses and addictions , –. This burden of chronic illness, and crowded living conditions in shelters, would be expected to be associated with an increased incidence of invasive pneumococcal disease. Nonetheless, the high rates of pneumococcal disease identified in this and a previous smaller study in Edmonton , are strikingly high; only patients with AIDS, hematologic malignancies and stem cell transplants have been identified as at similar or higher risk , . It is not possible to distinguish from this study the extent to which the increased incidence of disease is due to increased host susceptibility versus increased risk of transmission of pathogenic strains in crowded living conditions. Although most common underlying conditions do not increase the risk of invasive pneumococcal disease 30 fold –, , , many homeless persons have more than one underlying risk factor, and there are not data on risk in persons with multiple risk factors (eg. smoking, HIV infection and liver disease).
In keeping with the hypothesis that the host susceptibility to invasive pneumococal disease of homeless persons is much higher than that of other populations, we identified a five fold increased risk of recurrent disease in this homeless population. Recurrent invasive disease is known to be associated with serious immunodeficiency, in particular multiple myeloma and other malignancies, HIV infections, and chronic liver disease –. Increased host susceptibility is also likely the reason why, despite the fact that homeless patients with invasive disease are much younger than housed patients, their in hospital case fatality rate is almost identical. The fact that pneumonia was more prevalent in homeless than housed persons while bacteremia without focus was less prevalent may be explained by the dramatically higher proportion of smokers in the homeless population predisposing to lung infection. , .
The very high proportion of smokers (95%) among homeless persons with invasive pneumococcal disease may also explain the fact that rates of invasive infection in homeless persons do not decrease during the summer months in parallel to decreases in invasive infection in other adults. If homeless persons have higher carriage rates of S. pneumoniae , and the high carriage rate persists in the summer, transmission might also explain the relative excess of summer disease. However, one would expect crowding to be greater in winter months when shelters are used more often.
At least four previous outbreaks of pneumococcal disease in homeless shelters have been reported –. Disease in these outbreaks was due to isolates of serotypes 1, 5 and/or 8, serotypes that were rare in our population. There is some evidence in our data that transmission of isolates of other serotypes is associated with disease in homeless persons. We were unable to access data regarding shelter use over time, so that it is not possible to determine whether the apparent clustering of serotypes is due to transmission within or outside of shelters. Further study to define the risks for transmission of pneumococcal disease associated with the living conditions of homeless persons is warranted.
In those patients from whom a vaccination history could be obtained, pneumococcal vaccination rates were low. The difficulty in obtaining vaccination histories, and the serotype distribution of disease make vaccination programs for homeless populations a significant challenge . The coverage of 7-valent conjugate vaccines is less than 35%, and coverage with the 13-valent conjugate vaccine less than 50%. Although more than 80% of disease is caused by serotypes included in the 23-valent polysaccharide vaccine, the issue of hyporesponsiveness with repeated doses of polysaccharide vaccine in a relatively young population for whom consistent documentation of medical history is clearly of concern.  Further, the rate of recurrent disease suggests that this population is highly susceptible, so that vaccine efficacy may also be compromised , . Although vaccination was temporally associated with the termination of the two pneumococcal outbreaks in homeless shelters in which it was used as a control measure , , important questions about the efficacy of vaccination in this population remain.
There are a number of limitations to this study. The estimate of incidence is based on a single census of homeless persons conducted in April 2006; there are no standardized methods for such a census, and no tested means of validating the number obtained. However, the number is consistent with previous expert estimates . Further, the estimated incidence of invasive pneumococcal disease would still be of concern even if the census identified only 33–50% of the total homeless population. Census data by age for homeless persons are not available, so that it is not possible to calculate age-adjusted rates for this population. The relatively small number of cases annually means that, although there appeared to be no change in the serotype distribution of disease in homeless persons in association with the introduction of a pediatric vaccination program, our power to detect such a difference was low. The diagnostic challenge of pneumococcal pneumonia means that many more cases of non-bacteremic pneumococcal pneumonia occurred than we identified; the clinical characteristics and serotype distribution of undetected cases may be different from those in cases identified by our surveillance. Some data were not available for many of the cases occurring in homeless persons – in particular, we were not able to obtain data on prior pneumococcal vaccination in the majority of homeless persons, and we do not have information regarding stage of disease or treatment for those with HIV infection In addition, as noted before, our inability to obtain data on shelter use limited the interpretation of information regarding the potential of transmission of pneumococci in shelters.
Nonetheless, it is clear that the very high rates of invasive pneumococcal disease, the limitations of current pneumococcal vaccines, and the challenges of pneumococcal vaccination program delivery in homeless populations mean that the coordination of many different programs will be necessary to effectively reduce the burden of pneumococcal disease in this population. The provision of permanent housing and improved living conditions in crowded shelters might be expected to reduce transmission of this pathogen. Prevention and treatment programs for alcohol, smoking and substance abuse, and programs to improve HIV diagnosis and care delivery might prevent a fraction of cases; similarly, increasing influenza vaccination rates might be effective in preventing those cases secondary to influenza . While all of these programs may be necessary – and all with have benefits beyond pneumococcal disease – they are also relatively expensive and difficult to implement. Thus, studies of the effect of systematic or targeted pneumococcal vaccination programs against S. pneumoniae in homeless populations, and the development of more effective pneumococcal vaccines for adults are both urgently needed.
We are grateful to the many patients, families, physicians and hospital and public health staff who make this surveillance possible. Collaborating investigators in the Toronto Invasive Bacterial Disease Network are as follows: D. Chen, York Central Hospital; I. Davis and I. Kitai,, Rouge Valley Health System (Toronto); E. Devila, Peel Region Health Department (Brampton, Canada); H. R. Devlin and M. Muller, St. Michael's Hospital (Toronto); H. Dick, Vita-Tech Laboratories (Toronto); J. Downey and J. Powis, Toronto East General Hospital (Toronto, Canada); P. Garrod and N. Rau, Halton Healthcare Services (Oakville, Canada); R. Grossman and A. Sarabia, Credit Valley Hospital (Mississauga, Canada); F. Jamieson, Ontario Public Health Laboratory (Toronto); J. Kapala, Gamma Dynacare Laboratories (Toronto); S. Krajden, St. Joseph's Health Centre (Toronto); K. S. Lee and M. Baqi, Humber River Regional Hospital (Toronto); M. Loeb and F. Smaill, Hamilton Health Sciences Center (Hamilton, Canada); R. Lovinsky, D. Rose and J. Braithwaite, The Scarborough Hospital (Toronto); M. Lovgren and G. Tyrrell, National Centre for Streptococcus (Edmonton, Canada); A. G. Matlow and S. Richardson, Hospital for Sick Children (Toronto); B. Mederski and K. Katz, North York General Hospital (North York, Canada); Z. Moloo, D. Richardson, and C. Quan, William Osler Health Care Centre (Brampton); M. Naus, British Columbia Centers for Disease Control (Vancouver, Canada); K. Ostrowska Trillium Health Centre (Mississauga); P. Shokry, Markham Stouffville Hospital (Markham, Canada); A. E. Simor and M. Vearncombe, Sunnybrook Health Sciences Centre (Toronto); D. Sturman, Bridgepoint Hospital (Toronto); P. Van Nostrand, The Rehabilitation Institute of Toronto (Toronto); S. Walmsley and W. Gold, University Health Network (Toronto); D. Low, B. Willey, S. Pong-Porter, Toronto Medical Labs/Mount Sinai Hospital (Toronto); B. Yaffe, City of Toronto Public Health (Toronto); D. Yamamura, MDS Laboratories (Toronto); M. Silverman, Lakeridge Health (Oshawa, Canada); R. Robertson, Royal Victoria Hospital (Barrie, Canada); and G. Volkening, Southlake Regional Health Center (Newmarket, Canada).
Conceived and designed the experiments: IA GJT KAG DEL AM. Performed the experiments: AP AM KAG AM. Analyzed the data: AP AM AM. Contributed reagents/materials/analysis tools: TS IA GJT. Wrote the paper: AP TS IA AM DEL.
- 1. Ortqvist A, Hedlund J, Kalin M.Streptococcus pneumoniae: epidemiology, risk factors, and clinical features. 2005 Semin Respir Crit Care Med 6: 563–574.
- 2. Greene CM, Kyaw MH, Ray SM, Schaffner W, Lynfield R, et al. (2006) Preventability of invasive pneumococcal disease and assessment of current polysaccharide vaccine recommendations for adults: United States, 2001–2003. Clin Infect Dis 43: 141–150.
- 3. Burman LA, Norrby R, Trollfors B (1985) Invasive pneumococcal infections: incidence, predisposing factors and prognosis. Rev Infect Dis 7: 133–142.
- 4. McKenzie H, Reid N, Dijkhuizen RS (2000) Clinical and microbiological epidemiology of Streptococcus pneumoniae bacteremia. J Med Microbiol 49: 361–366.
- 5. Nuorti JP, Bulter JC, Farley MM, Harrison LH, McGeer A, et al. (2000) Cigarette smoking and invasive pneumococcal disease. NEJM 342: 681–689.
- 6. Pastor P, Medley F, Murphy TV (1998) Invasive pneumococcal disease in Dallas County, Texas: results from population based surveillance in 1995. Clin Infect Dis 26: 590–595.
- 7. Shariatzadeh MR, Huang JQ, Tyrrell GJ, Johnson MM, Marrie TJ (2005) Bacteremic pneumococcal pneumonia: a prospective study in Edmonton and neighbouring municipalities. Medicine (Baltimore) 84: 147–161.
- 8. Mufson MA, Kruss DM, Wasil RE, Metzger WI (1974) Capsular types and outcome of bacteremic pneumococcal disease in the antibiotic era. Arch Int Med 134: 505–510.
- 9. Mercat A, Nguyen J, Dautzenberg (1991) An outbreak of pneumococcal pneumonia in two men's shelters. Chest 99: 147–151.
- 10. De Maria A, Browne K, Berk SL, Sherwood EJ, McCabe WR (1980) An outbreak of type 1 pneumococcal pneumonia in a men's shelter. JAMA 244: 1446–1149.
- 11. Birtles A, McCarthy N, Sheppard CL, Rutter H, Guiver M, et al. (2005) Multilocus sequence typing directly on DNA from clinical samples and a cultured isolate to investigate linked fatal pneumococcal disease in residents of a shelter for homeless men. J Clin Microbiol 43: 2004–2008.
- 12. Daneman N, McGeer A, Green K, Low DE, Toronto Invasive Bacterial Diseases Network (2006) Macrolide resistance in bacteremic pneumococcal disease: implications for patient management. Clin Infect Dis 43: 432–438.
- 13. Kumar D, Humar A, Plevneshi A, Green K, Prasad GV, et al. (2007) Invasive pneumococcal disease in solid organ transplant recipients – 10-year prospective population surveillance. Am J Transplant 7: 1209–1214.
- 14. Kumar D, Humar A, Plevneshi A, Siegal D, Franke N, et al. (2008) Invasive pneumococcal disease in adult hematopoietic stem cell transplant recipients: a decade of prospective population-based surveillance. Bone Marrow Transplant 41: 743–747.
- 15. Homelessness Report card, City of Toronto Public Health. Street Needs Assessment staff report(http://www.toronto.ca/housing/streetneeds.htm) and Housing and Homelessness Report card, 2003. (http://www.toronto.ca/homelessness/index.htm; accessed June 19, 2008.
- 16. Al-Sukhni W, Avarino P, McArthur MA, McGeer A (2008) Impact of public vaccination programs on adult vaccination rates: two examples from Ontario, Canada. Vaccine 26: 1432–1437.
- 17. National Advisory Committee on Immunization.Active Immunizing Agents, Part 4 of Canadian Immunization Guide, 7th edition, 2006. available at http://www.phac-aspc.gc.ca/publicat/cig-gci/p04-pneu-eng.php.
- 18. Hwang SW (2001) Homelessness and health. CMAJ 164: 229–233.
- 19. Musher DM, Alexandraki I, Graviss EA, Yanbely E, Eid A, et al. (2000) Bacteremic and nonbacteremic pneumococcal pneumonia: a prospective study. Medicine (Baltimore) 79: 210–221.
- 20. Clinical and Laboratory Standards Institute (2006) Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically; approved standard, 7th ed. CLSI document M7-A7. Clinical and Laboratory Standards Institute, Wayne, PA.
- 21. Turnbull J, Muckle W, Masters C (2007) Homelessness and health. CMAJ 177: 1065–1066.
- 22. Pritchard JW, Puzey JW (2004) Homelessness–on the health agenda in Wales? Rev Environ Health 19: 363–379.
- 23. Frankish CJ, Hwang SW, Quantz D (2005) Homelessness and health in Canada: research lessons and priorities. Can J Public Health 96: Suppl 2S23–S29.
- 24. Garibaldi B, Conde-Martel A, O'Toole TP (2005) Self-reported comorbidities, perceived needs, and sources for usual care for older and younger homeless adults. J Gen Intern Med 20: 726–730.
- 25. O'Toole TP, Gibbon JL, Seltzer D, Hanusa BH, Fine MJ (2002) Urban homelessness and poverty during economic prosperity and welfare reform: changes in self-reported comorbidities, insurance, and sources for usual care, 1995–1997. J Urban Health 79: 200–210.
- 26. Flannery B, Heffernan RT, Harrison LH, Ray SM, Reingold AL, et al. (2006) Changes in invasive pneumococcal disease among HIV-infected adults living in the era of childhood pneumococcal immunization. Ann Int Med 144: 1–9.
- 27. de Roux A, Cavalcanti M, Marcos MA, Garcia E, Ewig S, et al. (2006) Impact of alcohol abuse in the etiology and severity of community-acquired pneumonia. Chest 129: 1219–1225.
- 28. Kyaw MH, Rose CE, Fry AM, Singleton JA, Moore Z, et al. (2005) The influence of chronic illnesses on the incidence of invasive pneumococcal disease in adults. J Infect Dis 192: 377–386.
- 29. Turett GS, Blum S, Telzak EE (2001) Recurrent pneumococcal pneumonia. Arch Intern Med 161: 2141–2144.
- 30. Rodriguez-Creixems M, Munoz P, Miranda E, Pelaez T, Alonso R, et al. (1996) Recurrent pneumococcal bacteremia. A warning of immunodeficiency. Arch Intern Med 156: 1429–1434.
- 31. Einarsdottir HM, Erlendsdottir H, Kristinsson KG, Gottfredsson M (2005) Nationwide study of invasive pneumococcal infections in a population with a low prevalence of human immunodeficiency virus infection. Clin Microbiol Infect 11: 744–749.
- 32. King MD, Whitney CG, Parekh F, Farley MM, Active Bacterial Core Surveillance Team/Emerging Infections Program Network (2003) Recurrent invasive pneumococcal disease: a population-based assessment. Clin Infect Dis 37: 1029–1036.
- 33. McEllistrem MC, Mendelsohn AB, Pass MA, Elliott JA, Whitney CG, et al. (2002) Recurrent invasive pneumococcal disease in individuals with human immunodeficiency virus infection. J Infect Dis 185: 1364–1368.
- 34. Marra CA, Patrick DM, Marra F (2000) A cost-effectiveness analysis of pneumococcal vaccination in street-involved, HIV-infected patients. Can J Public Health 91: 334–339.
- 35. O'Brien KL, Hochman M, Goldblatt D (2007) Combined schedules of pneumococcal conjugate and polysaccharide vaccines: is hyporesponsiveness an issue? Lancet Infect Dis 7: 597–606.
- 36. Targonski PV, Poland GA (2007) Pneumococcal vaccination in adults: Recommendations, trends, and prospects. Cleve Clin J Med 74: 401–414.
- 37. Sisk JE, Whang W, Butler JC, Sneller VP, Whitney CG (2004) Cost-effectiveness of vaccination against invasive pneumococcal disease among people 50 through 64 years of age: role of comorbid conditions and race. Ann Intern Med 138: 960–968.
- 38. Christenson B, Hedlund J, Lundbergh P, Ortqvist A (2004) Additive preventive effect of influenza and pneumococcal vaccines in elderly persons. Eur Respir J 23: 363–368.