Background
Streptococcus pneumoniae causes a range of invasive and noninvasive infections and remains the primary agent of community-acquired pneumo­nia.[1] Until recently the only licensed vaccine available was the 23-valent pneumococcal polysaccharide vaccine (PPV23), which is protective against 23 of the 90 most commonly known serotypes2 but lacks efficacy in young children.
 
In 2001 a 7-valent pneumococcal conjugate vaccine, PCV7, was licensed for use in Canadian children. PCV7 targets serotypes 4, 6B, 9V, 14, 18C, 19F, and 23F, which are responsible for more than 80% of invasive pneumococcal disease (IPD) in this age group. As with conjugate vaccines against other polysaccharide-bearing pathogens, PCV7 has greatly improved immunogenicity and efficacy in children under 5 years of age. It also prevents carriage, which may contribute to herd immunity. 

In early 2002 the National Advisory Committee on Immunization recommended that healthy children up to 59 months of age receive routine vaccination with a four-dose schedule of PCV7.[2] In BC a universal infant immunization program that was introduced in September 2003 provided children with a four-dose schedule of PCV administered at age 2, 4, 6, and 12 to 15 months.[3] 

In addition, BC authorities continued to offer PPV23 to all persons older than  65 years. The expectation at the inception of the universal infant PCV7 immunization program was that within 4 years the program would prevent 112 cases of pneumococcal meningitis or sepsis annually in children under 5 years of age.[3]

Methods
IPD is a reportable condition in BC and has been monitored by the British Columbia Centre for Disease Control (BCCDC) since 1999. Invasive pneumococcal disease cases are defined by the isolation of Streptococcus pneumoniae from a sterile site, usually blood or cerebrospinal fluid.

For this study data on IPD cases were extracted from two sources: the provincial passive surveillance system (the integrated Public Health Informa­tion System or iPHIS) and the provincial laboratory databases. Together these data sources provided information on the demographics (patient age, sex, geographical location), episode date, type of infection, and serotype. 

Serotype analysis of clinical isolates was provided by both the National Centre for Streptococcus and BCCDC laboratories using MYSIS Healthcare Systems technology.

Annual incidence rates were cal­cu­lated by obtaining population estimates for the denominators from P.E.O.P.L.E 30 (Population Extrapolation for Organization Planning with Less Error, run cycle 30, BC STATS). Data management and calculations were performed using Microsoft Excel software. 

Trends over time were ex­plored with the chi-square test for trend, using the StatCalc calculator in EpiInfo 6. Trends were considered statistically significant if the P value was less than .05.

Results
Between 2002 and 2005 the rate of IPD reported in the overall population of British Columbia was relatively stable at between 7.5 and 8.5 cases per 100000 population (see Figure 1 [13]). In 2006 the overall rate increased 40% over 2005 to 10.4 cases per 100000. 

This increase was seen primarily among those aged 20 to 64 years and was related to an outbreak of serotype 5 IPD in British Columbia. For this age group alone the incidence rate in 2006 was 11.0 per 100000 (300 cases) compared with 4.6 per 100000 (117 cases) during 2002—a significant increase (x2 = 87.1; P < .0001). 

From 2002 to 2006 there was an in­crease in the incidence of cases with­in the 65+ years group (x2 = 2.6; P = .10) and the 10 to 19 years group (x2 = 0.3; P = .57), and a decrease in the 5 to 9 years group (x2 = 3.6; P = .06). 

Between 2002 and 2006 there was a statistically significant decrease in the incidence of IPD among children under 5 years of age (see Figure 2 [14]) from 54.4 to 13.6 cases per 100000 (x2 = 80.8; P < .0001). When individual age categories were examined, significant decreases in the rate of IPD were observed: the rate for infants younger than 1 year fell from 50.6 to 12.4 cases per 100000 (x2 = 16.6; P < .0001), while the rate for 1-year-olds fell from 134.7 to 14.9 cases (x2 = 68.5 P < .0001), and the rate for 2-year-olds fell from 40.4 to 12.2 cases (x2=10.5; P < .001). 

There were also nonsignificant decreases among 3-year-olds (27.9 to 14.3 cases), 4-year-olds (22.7 to 16.9 cases), and 5-year-olds (10.9 to 7.1 cases), with all having P values greater than .26.

Serotype distribution
Serotype-specific data acceptable for analysis were available for 70% to 83% of case reports in the iPHIS system for each year studied. The only age group with a statistically significant increase in infections with serotypes not covered by PCV7 was the 20 to 64 years group (see Figure 3 [15]) where there was a 6.7-fold increase from 25 to 176 cases (x2 = 145.2; P < .0001). Most of this excess was attributable to 117 serotype 5 cases reported in 2006.

For all other age groups there was no significant increase in cases of IDP associated with serotypes not covered by PCV7. Most importantly, in children younger than 5 years of age there was no evidence of strain replacement and this has continued to date  (see Figure 4 [16]). 

Strain replacement was not found in those younger than 1 year (7.47 to 4.97 cases per 100000 population; x2  = 0.1; P = .68), the 1-year age group (7.53 to 2.44 cases, x2 = 0.6; P = .42), the 2-year age group (2.43 to 0.00 cases, x2 = 1.6; P = .20), the 3-year age group (2.36 to 0.00 cases, x2 = 0.0; P =.99), the 4-year age group (0.00 to 2.41 cases, x2 = 1.8; P = .17), or the 5-year age group (0.00 to 0.00 cases, x2 = 0.0; P = 1.00).

Conclusions
Our findings support conclusions about two important and largely unrelated trends. First, a decrease in the incidence of IPD for children under 5 years of age is almost certainly related to the increase in vaccination coverage provided by the universal infant PCV7 immunization program in Bri­tish Columbia. 

Second, an increase in IPD in those 20 to 64 years is most likely not an example of serotype re­placement connected to the program because the increase occurred in a nonoverlapping age group.

The 2003 immunization coverage rate for the first group of infants offered the pneumococcal conjugate vaccine was 64%, a rate that increased to 77.6% in 2006.3 This increase in coverage may be improving herd im­munity.[4] 

It is of note that the impact of PCV7 on the prevention of IPD elsewhere has been greater than expected from original clinical trials,[5] which were used for approval of the vaccine and to guide the initial recommendations for the infant immunization schedule.[6] However, in BC we are not yet seeing herd immunity in age groups older than 5 years as described in the USA.[7] In British Columbia the IPD rate among those aged 65 years and older has not changed significantly in the past 4 years. This group continues to receive PPV23.

Regarding the increase in the rate of IPD in those 20 to 64 years, this trend can be adequately explained by the well-documented introduction of a clonal outbreak of serotype 5 IPD among indigent urban dwellers during 2006. The hospital and health authority involved in identifying the outbreak will soon be publishing an article describing the outbreak (personal communication with Dr Mark Romney, medical microbiologist, Providence Health Care, 16 August 2007).

This study had at least two limitations. First, while passive reporting of IPD by laboratories is generally fairly complete, there may be some underreporting. Second, not all isolates associated with a diagnosis of IPD were referred to reference laboratories for serotyping, which means that the serotype-specific trends are derived from the large subset of cases where such data were available. 

This study highlights the success of the universal infant PCV7 immunization program in British Columbia, where there has been a significant reduction in the incidence rate of IPD in children younger than 5 years. Since January 2007 BC has moved to a three-dose schedule of PCV7 administered at 2, 4, and 12 to 15 months. In order to evaluate this change, authorities will continue to monitor invasive pneumococcal disease in children under 5 years of age.

Acknowledgments
We would like to extend thanks to Ingrid Pocock, senior technologist, Bacteriology and Mycology Laboratory, BCCDC, for her work on streptococcal isolates during the study period.

Competing interests
None declared.

Mr Shahidi is an undergraduate science student at Simon Fraser University. He is also a scholar in the CIHR-UBC program for Translational Research in Infectious Disease (TRID). Ms Dhaliwal is a field epidemiologist with the BC Centre for Disease Control (BCCDC). Dr Tyrrell is an associate professor in the Department of Laboratory Medicine and Pathology and interim director of Microbiology and Public Health at the University of Alberta. Dr Hoang is a medical microbiologist and head of Bacteriology, Enterics, and Mycology at BCCDC. She is also a clinical assistant professor in the Department of Pathology and Laboratory Medicine at the University of British Columbia. Dr Patrick is director of Epidemiology Services at BCCDC and an associate professor in the Department of Medicine at UBC.