Canadian Medical Association Journal 1996; 154: 1041-1047
[Contents]
Hib vaccines were initially licensed in Canada in 1986,[6] the first one consisting of purified capsular polysaccharide (polyribosylribitol phosphate [PRP]). It was reliably immunogenic only from 24 months of age,[6] by which time about two thirds of Hib cases already occur.[3] In 1988 it was superceded by the first polysaccharide-protein conjugate vaccine (PRP-D [diphtheria protein conjugate]; ProHIBIT, Connaught Laboratories Ltd., North York, Ont.),[7] the greater immunogenicity of which allowed use from 18 months of age. About 40% of cases were potentially preventable using this vaccine.[7] During 1988-89 most provinces implemented universal vaccination programs using PRP-D. Vaccine efficacy in the target age group later proved to be only 74% to 88%.[8,9] Reductions in the incidence of Hib infections were reported among children in the target age group who received the vaccine[10,11] and among those too young to receive it,[11] presumably because of decreased Hib transmission from vaccinated children.[12] The most recent advance occurred in 1992, when all provinces and territories started programs using more potent PRP-conjugate vaccines given to infants 2 months of age and older.[13] With conjugation of the bacterial polysaccharide to protein carrier molecules such as tetanus or diphtheria toxoid, the immune response to PRP was converted from a T-lymphocyte independent pathway to a T-lymphocyte enhanced pathway, permitting antibody responses to develop in early infancy and establishing immunologic memory, the basis for durable, reinforceable serum antibody responses.[13]
Reports from the United States, where improved conjugate vaccines were licensed a year earlier than in Canada, indicated that infant-based programs rapidly reduced Hib infection rates.[14-16] The Immunization Monitoring Program, Active (IMPACT),[17] operated by the Canadian Paediatric Society with funding from the Laboratory Centre for Disease Control, Ottawa, was well-suited to observe the influence of the new programs. For the 10 pediatric centres involved in this Canada-wide network, there were two questions of interest: How rapidly and extensively would the caseloads decline, and what factors would account for new cases? For the latter question, if missed opportunities to vaccinate were found to be frequent, health authorities would be guided toward developing strategies to improve vaccine coverage. Similarly, if vaccination failures were found to be frequent with a particular product, health authorities would be prompted to consider available alternatives.
[Contents]
To obtain a more complete picture of temporal changes in the caseload, the centres determined approximate annual rates of hospital admission because of Hib infections from 1985 to 1990. No clinical information was collected. Four centres relied on laboratory reports of Hib isolates identified in blood and cerebrospinal fluid samples, excluding any duplicates; one centre relied on coded discharge diagnoses from its records department; and five centres relied on both laboratory reports and coded discharge diagnoses to identify cases.
To determine the caseload from Jan. 1, 1991, to Dec. 31, 1994, a trained nurse monitor at each centre conducted intensive case-finding activities. Cases were identified from hospital microbiology laboratory reports, and the numbers were periodically verified against lists of discharge diagnosis codes (038.41, 041.5, 320.0 and 482.2 of the International Classification of Diseases, 9th revision, clinical modification[18]) supplied by the hospital records departments. The monitors also visited wards, reviewed admission lists and contacted infection-control nurses to find new cases. Prospective case-finding was initiated at five centres in 1991 and at the remaining centres in 1993. Equal effort was exerted to discover all cases, regardless of whether ascertainment was conducted retrospectively or prospectively.
A definite case was one in which Hib was isolated from a sample of normally sterile body fluid (e.g., blood, cerebrospinal fluid or fluid from the pleural or joint space) in a child with an illness indicative of Hib infection. Culture of Hib from the surface of a swollen epiglottis was also considered evidence of definite infection. A probable case was one in which Hib antigen was detected in a sample of cerebrospinal fluid, urine or other normally sterile body fluid in a child with an illness indicative of Hib infection who had not received Hib vaccine within the preceding 14 days (a source of detectable antigen in itself).[19] A probable case was also one in which Hib was recovered from endobronchial secretions of a child with bronchopneumonia, except in the case of long-term airway intubation. Only type b organisms were accepted. Organisms and antigen were identified by hospital microbiology laboratories under routine conditions.
The monitors described cases in detail using a standard report form to abstract information from each child's hospital record. The information collected included age, sex, reporting hospital, admission date, illness manifestations, bacterial culture results (sites, organism), antigen test results, level(s) of care required, duration of admission and patient's condition at discharge. In addition, special efforts were made to document any prior Hib vaccination, including dates and product names. If vaccination details were not found in the hospital record, the monitors contacted the parents, health department or attending physician. The monitors then categorized cases as preventable or not, taking into account the recommended use of the specific product(s) given or available before disease onset. Vaccination failure was defined as the onset of culture-confirmed Hib infection more than 28 days after completion of age- appropriate vaccination.
Case reports were checked for completeness at the IMPACT data centre, Vancouver, and the information was then entered into a database. Dual data entry was performed to minimize transcription errors. All key facts were manually verified before transcribed cases were added to the master database. Descriptive statistical analyses were performed using SAS/STAT software (SAS Institute Inc., Cary, NC).
The centres determined their annual total discharge numbers for 1985 and 1993, using consistent 12-month intervals (fiscal or calendar years), to see whether a decline in the incidence of Hib infections may have reflected fewer hospital admissions in general.
This project was approved by the institutional review board at each of the 10 participating centres.
[Contents]
The age distribution of cases during 1991-94 is shown in Fig. 2. In 1991 the pattern was similar to that during the pre-immunization era,[3,4] children 6 to 18 months accounting for most cases (47.9%). In 1994 fewer cases were seen in all age groups, the fewest occurring among children 6 to 18 months (16.7%). The number of affected children in this age group decreased sharply (by 89.8%) from 1992 to 1993.
Of the 24 cases identified in 1994, 6 were classified as preventable. The parents refused vaccination in two, one of which was fatal, the attending physician inappropriately dissuaded the parents from having their children vaccinated in two cases, and Hib vaccination was not scheduled, for no apparent reason, in two cases, although the children's other vaccinations were up to date.
Among the nonpreventable cases in 1994 were seven children too young to have completed the primary series of Hib vaccination, two 12-year-olds who were too old to have qualified for earlier programs and one 6-year- old child not vaccinated because of HIV-related immunodeficiency who became infected despite intravenous immunoglobulin therapy. The remaining eight children had Hib infection despite having received age- appropriate vaccination.
Vaccine failures were encountered throughout the 1991-94 period (Fig. 3). All were culture-confirmed cases and numbered 36 in total, accounting for 10.6% of the cases seen during 1991-94. Although the eight vaccine failures in 1994 accounted for 33.3% of the cases seen that year, the actual number of vaccine failures did not increase over time (Fig. 3). Most (88.9% [32/36]) of the failures were attributed to vaccines no longer used: PRP accounted for 3, PRP-D for 27 and unnamed early products for 2. Of the four failures attributed to currently used products two occurred in 1993 and two in 1994; PRP-T (tetanus protein conjugate) vaccine accounted for three and oligosaccharide conjugate Hib vaccine (HibTITER, Lederle-Cyanamid Canada Inc., Markham, Ont.) for the other. The latter case was reported previously[20] and involved a child with impaired responsiveness to PRP antigen.
Of the vaccine failures in 1994 five involved children over 5 years old, all of whom had received vaccines no longer used (PRP-D vaccine in three, unspecified product in two). Three involved children 7 months to 5 years old (PRP-D vaccine in one and PRP-T vaccine [Act-HIB, Connaught Laboratories] in two). Of the two PRP-T vaccine failures one involved a child who had been given a single dose at 16 months and the other a 15-month-old child who had received three doses appropriately in early infancy; both children were concurrently immunocompromised by extensive burns, from which the second child died.
Some characteristics of Hib infections did not change appreciably from 1991 to 1994. Overall boys outnumbered girls (male:female ratio 1.7:1 [215:125]); this pattern was evident each year except 1994, when the ratio was 1:1.2 (11:13). Meningitis was the most common disease each year, accounting for 172 (50.4%) of the 341 cases during the 4 years. Epiglottitis was the next most common disease, occurring in 65 (19.1%) of the cases during the same period. Immunocompromised children accounted for a small proportion of cases (2.9%). The mean length of hospital stay remained constant (about 10 days). The case-fatality rate in 1994 was 4.2%, similar to that for the previous 3 years (2.5%).
In 1994 the total number of hospital bed-days for the 24 Hib cases was 222 days, 88 days of which were for intensive care. In 1991 Hib cases required 1723 hospital bed-days, 492 of which were for intensive care. The mean length of stay in 1991 was 10.6 days (95% confidence interval 9.4 to 11.7 days). If one assumes that the mean length of stay in 1985 was similar to that in 1991, Hib cases in 1985 resulted in an estimated 5141 days of hospital care. Thus, from 1985 to 1994 the annual number of hospital bed-days decreased by about 4919 days (95.7%).
[Contents]
The introduction of vaccination programs appears to be the principal explanation for the decrease in Hib cases. When interpreting annual case totals, one must consider natural variation in Hib prevalence,[3,21] the magnitude of which was not defined at participating centres for a period before the introduction of Hib vaccines. A steady decrease in overall prevalence was evident following the introduction of PRP-D vaccine in 1988, but data for individual centres showed some variability through 1991 (Table 1). The sharpest interannual decline in combined case totals was noted following the introduction in 1992 of the more potent PRP-conjugate vaccines, which can be given from 2 months of age. All provinces and territories implemented infant-based vaccination programs in 1992. All of the IMPACT centres experienced sustained decreases in case totals thereafter. The greatest decrease after 1992 was observed among children 6 to 18 months of age. A similar effect on Hib cases was observed in the United States following the implementation of infant-based programs there in 1991.[14,15]
Other changes may have contributed to reduced Hib caseloads at the participating centres. The total number of hospital admissions decreased by about 10% from 1985 to 1993, reflecting bed closures and possible changes in referral patterns for serious infections. However, the number of admissions because of invasive pneumococcal infections, another target of surveillance at the IMPACT centres, did not decrease from 1990 to 1994 (data not presented). Population-based incidence data are awaited to confirm our observations. Our data provide no evidence that Hib has become less virulent over time: the death rate and syndrome mix in 1994 were unaltered from previous years.
The impact of reduced Hib caseloads on hospital bed utilization has been substantial. An estimated 4919 fewer days of hospital care were required in 1994 than in 1985, freeing up 1.3 beds on average at each of the 10 centres. Savings in costs for long-term care of the 15% of meningitis survivors who have deafness, mental impairment or other sequelae[5] must also be substantial. The cost- effectiveness of Hib vaccination has previously been demonstrated.[22] This case-series study has both advantages and disadvantages. A fundamental limitation is the absence of denominator data. The participating IMPACT centres have defined catchment areas (regions, provinces) but no precise information about the proportion of Hib cases arising therein that are referred to them. The proportions may differ depending on illness syndrome, severity, local physician availability or parental preferences. Thus, an attempt to estimate the populations served for purposes of calculating disease incidence would be unwise. On the other hand, the data from this case series are likely more complete than the passively reported case data used for most calculations of disease incidence. The superiority of active laboratory-based case surveillance over passive reporting of Hib cases was recently documented in US centers.[23] Only one or two carefully instructed monitors gathered data at each IMPACT centre. The period of intensive surveillance involved varying proportions of retrospective and prospective work at each centre, but complete case ascertainment was attempted in both contexts. The desired case information was readily available from patients' medical records. For both retrospectively and prospectively described cases, nurses obtained information from immunization records directly from parents or health care providers when details were lacking in the hospital records. The ascertainment of cases for 1985 to 1991 was not as uniform or complete as it was for the later years, but the particular case-finding methods used by the centres were applied fully and consistently during this period. We assumed that the microbiology laboratories at the centres could reliably identify Hib organisms. All of the participating laboratories take part in accreditation exercises.
The results of our study indicate a decrease in the number of Hib infections in most areas of Canada. However, Saskatchewan, New Brunswick, Prince Edward Island and the Yukon and Northwest territories were not represented directly. Nevertheless, the IMPACT centres are well distributed across Canada, and all have large catchment areas, some spanning whole provinces or multi-province regions. Together they account for over 80% of the country's tertiary care pediatric beds.[17] We believe that most Hib cases are admitted to tertiary care pediatric centres, at least in the participating areas.
Our data provide reassurance about the efficacy of Hib vaccines. Although some vaccine failures were documented in children given older products, only four involved products used since 1992. Three of these four were associated with host immunocompromise. Similar observations have been reported by others regarding the rarity of failures with current vaccines and the frequency of immunologic defects in those who acquire a Hib infection despite vaccination.[24,25] More children will need to be followed for extended periods before currently used products can be said to be fully assessed. Health care providers who encounter vaccination failures should report case details to their local health department.
Of special note are the data for 1994. Although so few cases of Hib were identified, there were missed vaccination opportunities, some with tragic outcomes. Hib infections remain severe and potentially fatal. Widespread vaccination using conjugate vaccines has reduced Hib colonization rates among young children;[12] however, circulation of the organism continues. Consequently, herd immunity is a poor substitute for personal immunity. All health care professionals who counsel parents regarding childhood vaccination should emphasize the importance of protecting infants against Hib and ensure that the vaccine is administered in a timely fashion to all who need it.
[Contents]
[Contents]