Invasive pneumococcal disease (IPD) is an important cause of morbidity and mortality among Canadian children.1 IPD is most frequent in Canadian children 6–23 months old and greater than 85% of isolates belong to serotypes included in the 7-valent pneumococcal conjugate vaccine (PCV7). Consequently, licensure of PCV7 in Canada in 2001 raised expectations that its routine use would rapidly decrease IPD in infants, as demonstrated in the United States.2,3 British Columbia (BC) (population 4.2 million) was one of the first Canadian provinces to implement a routine PCV7 program for infants 2 months of age, in September 2003. There was no catch-up program for older children, but PCV7 was also provided for children <5 years of age at increased risk of IPD, including First Nations children.4 Doses of PCV7 were routinely offered at 2, 4, 6 and 18 months of age,5 and given concurrently with other infant vaccinations.
We wished to assess the effectiveness of PCV7 vaccination in Greater Vancouver BC, because the incidence rate of IPD among 6–23 months old children was higher here than in 8 other large cities across Canada during 1994–1998,6 with a cumulative risk to age 6 years estimated at 1 in 260. The aim of this study was to measure the effectiveness of PCV7 during the first 2 years of routine use, in a large urban population.
The Greater Vancouver surveillance area included 2 large Health Authorities, one of which contains the City of Vancouver, with a population of ∼473,000 children 0–16 years of age or 58% of the pediatric population of BC. We identified all IPD cases arising in this population during the 24 months before the PCV7 program started (September 2001–August 2003) and the 28 months afterward (September 2003–December 2005). IPD was defined as isolation of Streptococcus pneumoniae from a normally sterile body site. Hospitalized and nonhospitalized cases were included. Four overlapping methods of case finding were used, including reviews of: (1) cases notified by law to the Health Authorities; (2) laboratory reports at local private laboratories and all hospitals admitting children; (3) relevant discharge diagnosis codes at the same hospitals; and (4) case reports and isolates referred to a central public health laboratory (BC Centre for Disease Control), which routinely requested isolate serotype determinations at the National Centre for Streptococcus, Edmonton, AB. Duplicate reports were removed.
A standardized case report was completed for each case, based on information in the health record of hospitalized cases or obtained from the physician of nonhospitalized cases. Postal codes were used to confirm residency in the study area; nonresidents were excluded. Census data were obtained from the provincial government. Case data were entered in duplicate into an electronic database (Access, 2002, Microsoft, Redmond, WA) and analyzed using SAS/STAT (SAS Institute, Cary, NC). The outcome of principal interest was the change in the IPD incidence rate for 6- to 23-month-old children between pre- and postprogram periods, expressed as the percentage reduction in the average annual period rates. Rates were considered significantly different if the 95% confidence intervals did not overlap. The rate change for serotypes included in PCV7 was estimated using the observed ratio of vaccine to total serotypes in each of the 2 periods, based on available serotype data. Rate changes were also examined for other age groups and subareas.
The study was approved by the research ethics boards of the University of British Columbia and both Health Authorities.
In total 276 cases were identified, 186 in the 2 years before the infant program began and 90 in the 2.3 years afterward (Table 1). For children 6–23 months old the average annual incidence rates per period were 125.5 and 34.1 per 100,000, respectively, a reduction of 72.7%. Considering only serotypes in PCV7, the incidence rates for this age group declined from 109.2 to 21.7 per 100,000, an 80.1% decrease. A significant rate decrease was also seen for 2-year-old children but not for other age groups (Table 1).
BC Children's Hospital (BCCH) played a prominent role in IPD management, with involvement in 32.1% of cases from Health Authority A, 82% of cases from Health Authority B (which includes Vancouver) and 94.9% of cases from Vancouver itself, for which it is the only pediatric facility. The preprogram IPD incidence rates for children 6–23 months old were 99.4 (95% CI, 73.0–132.1), 169.2 (124.8–224.3) and 235.7 (167.6–322.0) per 100,000 in the 3 settings, respectively. Postprogram rates were similar in the 3 settings, ranging from 27.6 to 38.0 per 100,000. The rate change for Vancouver (from 235.7 to 36.4 per 100,000) represented an 84.6% decrease in IPD overall and a 92.5% decrease for vaccine serotypes (from 208.8 to 15.6 per 100,000). The rate decrease occurred in the first year of the program in Vancouver (Fig. 1) and elsewhere. The small rate increase in 2005 was not statistically significant (Fig. 1). Of note, the rate of occult (simple) bacteremia before vaccine introduction was substantially higher in Vancouver than elsewhere for 6- to 23-month-old infants (187.3 vs. 72.7 per 100,000, respectively, P < 0.05), a major factor in the geographic rate differences for IPD. We believe that this resulted from adherence to a protocol for obtaining blood cultures from febrile infants at the Children's hospital, not matched at community hospitals elsewhere in the region.
Of 30 cases of IPD that occurred in 6- to 23-month-old children after program introduction, only 10 involved program-eligible children (born after July 1, 2003). One had infection in the neonatal period. Five had received 3 doses of PCV7, none of whom had infection with serotypes in the vaccine. Immunization history was uncertain for 4 children, one of whom had a vaccine-type isolate (6B). Four deaths occurred after the program commenced, none in eligible children.
Among children 6–23 months old, decreases were seen in isolates belonging to 6 of the 7 vaccine types after program introduction. Too few cases involving type 9V occurred to examine trends. Largest reductions involved types 14, 18C, 23F. No increase was observed in the frequency of infections with serotypes not included in the vaccine during the first 2 years of PCV7 use.
The impact of the new PCV7 vaccination program for infants in Greater Vancouver was remarkably swift, with a substantial decrease in IPD evident in the target age group during the first year. The uptake rate of PCV7 (3 doses) by the end of 2005 was estimated by the Health Authorities at ∼75%, despite the fact that the primary infant series offered in BC involves 3–4 injections per visit. No PCV7 supply interruptions occurred. The fact that only one-third of cases involving 6- to 23-month-old infants in the postprogram years (Table 1) occurred in program-eligible infants signals the likelihood of a further decrease in cases as the program continues. Only one of 10 cases in program-eligible infants involved a vaccine serotype and we were unable to determine if that child had been vaccinated. Disease rates in 2- to 5-year-old children should also decrease as the immunized cohort ages. The effect seen in 2-year-old children might have been the result of indirect protection (herd effect)3 but some of these children might have received PCV7 for risk conditions or through private purchase.
Our data conform with observations on IPD incidence for the whole of British Columbia, based on disease notifications. In children 12–23 months of age, disease incidence decreased from 135 cases per 100,000 in 2002 to 15 per 100,000 in 2005, an 89% reduction.7 Rate reductions were noted in 2- and 4-year-old children, but not in other age groups. A similar, rapid effect of PCV7 vaccination on IPD incidence rates was reported in Calgary,8 Alberta, the first Canadian province to implement an infant program (in September 2002). After 2 years of routine vaccine use, the disease incidence rate for 6- to 23-month-old infants had decreased 86% overall and 93% for vaccine serotypes,8 similar to the figures for Vancouver City.
Our data also highlight the importance of obtaining blood cultures to fully define the epidemiology of IPD. A higher baseline rate of IPD was detected in children <2 years of age in the Vancouver City population served almost exclusively by our pediatric hospital than in the surrounding population served by community hospitals. The BC Children's Hospital Emergency Department pediatricians follow a protocol for obtaining blood cultures from febrile children <2 years of age, which likely explains why they detected higher incidence rates of occult bacteremia and of IPD overall than were found by community physicians. The baseline incidence rate of IPD in Vancouver (235.7 per 100,000 for 6- to 23-month-old children) was much higher than rates previously reported for urban populations across Canada,6 which featured in estimations of the cost-effectiveness of PCV7 programs.9,10 If disease rates were underestimated, the cost-effectiveness of vaccination would have been underestimated as well.
Although this study featured robust case-finding in a large pediatric population, it had some limitations. PCV7 was available for private purchase for 2 years before program implementation, and so the baseline rate of IPD might have been under-estimated. However, no decrease in incidence rates was evident during the 2 preprogram years. The extent of private vaccine use is unknown. Serotype data were lacking for some isolates, and so the rate reduction for serotypes included in the vaccine was estimated rather than directly observed. Immunization histories were missing for some cases. The 2-year observation period after program introduction was likely too short to measure the full impact of vaccination, including indirect protection of other age groups.
We conclude that PCV7 infant vaccination in BC has been highly effective.
1. Scheifele D, Halperin S, Pelletier L, et al. Invasive pneumococcal infections in Canadian children, 1991–1998: implications for new vaccination strategies. Clin Infect Dis
2. Poehling KA, LaFleur BJ, Szilagyi PG, et al. Population-based impact of pneumococcal conjugate vaccine in young children. Pediatrics
3. Centers for Disease Control. Direct and indirect effects of routine vaccination of children with 7-valent pneumococcal conjugate vaccine on incidence of invasive pneumococcal disease—United States, 1998–2003. MMWR Morb Mortal Wkly Rep
4. Scheifele DW, Law B, Vaudry W, et al. Invasive pneumococcal infections among Canadian aboriginal children. Can Commun Dis Rep
5. National Advisory Committee on Immunization
. Statement on recommended use of pneumococcal conjugate vaccine. Can Commun Dis Rep
6. Bjornson GL, Scheifele DW, Halperin SA, et al. Population-based epidemiology of invasive pneumococcal infection in children in nine urban centers in Canada, 1994 through 1998. Pediatr Infect Dis J
7. Paulus S, David ST, Tang W, et al. Incidence of invasive pneumococcal disease after introduction of the universal infant immunization
program, British Columbia (2002–2005). Can Commun Dis Rep
8. Kellner JD, Church DL, MacDonald J, Tyrrell GJ, Scheifele D. Progress in the prevention of pneumococcal infection. Can Med Assoc J
9. McClure CA, Ford MW, Wilson JB, Aramini JJ. Expected benefits of pneumococcal vaccination in Canadian infants and children <5 years old. Can Commun Dis Rep
10. Lebel MH, Kellner JD, Ford-Jones EL, et al. A pharmacoeconomic evaluation of 7-valent pneumococcal conjugate vaccine in Canada. Clin Infect Dis