OBJECTIVE: Influenza immunization is recommended for children with IBD, however safety concerns may limit uptake. This study assessed whether immunization was associated with adverse events in IBD patients using a population-based database of children with IBD.
METHODS: All children <19 years diagnosed with IBD in Ontario, Canada between 1999–2009 were identified using health administrative data, and matched to non-IBD controls. Self-controlled case series (SCCS) analyses determined health services event rates (outpatient visits, hospitalizations and emergency visits) in any 2-week risk period to 180 days post-immunization compared to a no-risk control period. Relative incidence (RI) was calculated for overall and IBD-related events and rates were compared between IBD cases and controls using relative incidence ratios (RIR).
RESULTS: A total of 4916 IBD patients were matched to 21 686 controls. IBD patients were more likely to have received immunization than controls (25.3% vs 13.2%, P < .001). No increased event rates existed in IBD cases during risk periods (pooled RI 0.95, 95% CI 0.84–1.07), including hospitalizations and emergency visits. There was a slightly higher event rate in IBD cases versus controls for days 3–14 (RIR 1.60, 95% CI 1.05–2.44, P = .03). IBD-related visit rates were lower in risk periods compared to control period (pooled RI 0.81, 95% CI 0.68–0.96).
CONCLUSIONS: There was no increase in health services use in the post-vaccine risk period in IBD patients, and there was evidence for a protective effect of influenza immunization against IBD-related health services use. Influenza immunization is safe in children with IBD and should be encouraged to improve poor coverage rates.
- health administrative data
- inflammatory bowel diseases
- influenza vaccines
- seasonal influenza
- vaccine safety
- CD —
- Crohn’s disease
- CI —
- confidence intervals
- ED —
- emergency department
- IBD —
- inflammatory bowel disease
- OCCC —
- Ontario Crohn’s and Colitis Cohort
- OHIP —
- Ontario Health Insurance Plan
- SCCS —
- self-controlled case series
- RI —
- relative incidence
- RIR —
- relative incidence ratio
- UC —
- ulcerative colitis
What’s Known on This Subject:
Yearly influenza immunization is recommended in patients with inflammatory bowel disease (IBD). However, concern regarding vaccine-related adverse events may limit uptake, and case reports in the literature detail disease flares after immunization.
What This Study Adds:
Influenza immunization rates in children with IBD are low but immunization did not result in increased outpatient visits, hospitalizations or emergency visits. Immunization was associated with fewer IBD-related visits in the post-vaccine period, which may indicate protection against IBD symptoms.
Children with chronic inflammatory bowel disease (IBD), including common subtypes Crohn’s disease (CD) and ulcerative colitis (UC), are frequently immunosuppressed as part of their ongoing treatment. Immunomodulator use ranges between 30% and 100% in North America,1 with rising use of immunomodulators and biologics in recent years.2 As a result, international clinical guidelines recommend influenza immunization for patients with IBD,3,4 which is consistent with the Public Health Agency of Canada recommendation for yearly influenza immunization in immunocompromised children or those with chronic diseases.5 Nevertheless, surveys of gastroenterologists have demonstrated inadequate knowledge of immunization recommendations and poor rates of administration of vaccines to their patients.6,7 In addition, observational reports have raised concerns regarding adverse events and IBD flares associated with influenza immunization,8–10 and epidemiologic evidence evaluating the safety of influenza immunization in IBD patients is lacking.11,12
Ontario, Canada, has among the highest rates of childhood-onset IBD in the world,13 and all cases are followed longitudinally in the Ontario Crohn’s and Colitis Cohort (OCCC), a population-based registry of pediatric IBD patients derived from provincial health administrative data. In 2000, Ontario introduced universal influenza immunization to provide funded vaccines for the entire population aged >6 months. These factors make Ontario an optimal setting to evaluate rare outcomes in children with IBD, such as adverse events after immunization. The aims of the current study were to assess the rate of physician-administered influenza vaccine among children with IBD in Ontario and to evaluate vaccine safety by assessing the likelihood of medical care requirement in the postvaccine period.
Ontario’s administrative databases are maintained by the Institute for Clinical Evaluative Sciences (ICES) through a comprehensive data-sharing agreement with the Ontario Ministry of Health and Long-Term Care. Approval for the current study was obtained from the research ethics board of the Children’s Hospital of Eastern Ontario. The OCCC is a longitudinal cohort of all children diagnosed with IBD in Ontario between 1994 and 2011, identified by using a validated algorithm14 and linked to other databases by using a unique patient identifier. To be included in the cohort, patients who underwent colonoscopy required at least 4 physician contacts or 2 hospitalizations on separate days over a maximum of 3 years with an associated International Classification of Diseases, Ninth Revision, code for CD (555.x) or UC (556.x) or an International Classification of Diseases,Tenth Revision, code for CD (K50.x) or UC (K51.x). Patients who did not undergo colonoscopy required at least 7 physician visits or 3 hospitalizations over 3 years. This algorithm identified children with IBD within Ontario’s databases with the following diagnostic accuracies: sensitivity of 89.6% to 91.1%, specificity of 99.5% to 100%, and positive and negative predictive values ranging from 59.2% to 76.0% and 99.9% to 100%, respectively.
The Registered Persons Database provided sociodemographic characteristics of subjects. Hospitalizations were captured by using the Discharge Abstract Database, collected by the Canadian Institute for Health Information with diagnostic codes 80% to 89% accurate,15 positive predictive values of 89.1% to 99.9%, and negative predictive values of 87.2% to 100%.16 For emergency department (ED) visits, we used data from the Ontario Health Insurance Plan (OHIP) before 2002 and the Canadian Institute for Health Information’s National Ambulatory Care Reporting System after 2002. Outpatient physician visits and immunization codes were obtained from the OHIP database, including influenza immunization by either a physician (G590, G591) or nurse practitioner (Q690, Q691). Compared with self-reported vaccination in 12- to 19-year-olds, these codes are 55% sensitive and 96% specific, with positive and negative predictive values of 94% and 68%, respectively.17 Sensitivity and negative predictive values were higher in patients with at least 1 chronic disease. The fair sensitivity is due to the availability of influenza vaccination from public health clinics, which are not recorded in the OHIP database. A separate validation study noted a 77.9% sensitivity in children with a chronic disease and 74.0% in healthy children aged 2 to 11 years.18 The live, attenuated influenza vaccine was not available in Ontario during the study period. We assessed safety of influenza vaccine with the use of Vaccine and Immunization Surveillance in Ontario, an analysis infrastructure created with the use of data from the ICES to monitor vaccine safety and efficacy in Ontario.19
We compared influenza immunization rates in IBD cases and non-IBD controls. Each IBD case was matched to 5 (or 4 where a fifth could not be found) controls according to gender, provincial administrative health region in which the subject resided, and date of birth (±30 days). The date of immunization was obtained by using billing codes for influenza vaccination. To assess safety of influenza vaccination, we first assessed rates of adverse events in specified time periods after vaccination in both cases and controls separately. We then compared the risk estimates for cases versus controls. We did so to overcome the healthy vaccinee effect, a recognized decrease in event rate before vaccination that can mask signals in the postvaccination period.20 This effect describes the higher likelihood of healthier patients to receive immunization and may result in a perceived low rate of adverse events at the time of immunization with an increased rate of events after immunization.19
For both cases and controls, we conducted a self-controlled case series (SCCS) analysis21 to assess the risk of immunization on subsequent adverse events (defined as a hospitalization, ED visit, or physician office visit for any reason) or, for IBD patients, IBD-related events; IBD-associated diagnostic codes are presented in Supplemental Table 4. For IBD cases, we included all children (aged 2–18 years), with at least 1 year of follow-up, diagnosed with IBD in Ontario from fiscal years 1999 through 2010. The longitudinal follow-up period started with date of diagnosis with IBD, defined as the date of first contact with a physician or first hospitalization with an associated IBD-specific diagnostic code. The end of follow-up was the subjects’ nineteenth birthday; March 31, 2011; the date of death; or the date the patient became ineligible for Ontario health care (due to migration out of the province). For control subjects, the longitudinal follow-up period started on the date of IBD diagnosis of their matched case.
Descriptive statistics were determined as means (±SD), medians (with interquartile range), or proportions where appropriate. We compared immunization rates in cases compared with controls by using χ2 analysis for categorical variables and Student’s t tests or Wilcoxon rank sum statistics for continuous variables.
The SCCS design was used to assess risk of adverse events after immunization. Date of vaccination served as the index date for the risk period. In IBD cases, all-cause and IBD-related physician visits, hospitalizations, and ED visits were used as health services proxies for adverse events. For analytical purposes, we divided each individual follow-up period into 2-week intervals up to 180 days after the vaccination date, with a washout period between the exposed and unexposed periods (Fig 1). Our at-risk period after exposure is based on the concept that an IBD flare may occur up to 6 months after immunization. The control period was from the end of the risk intervals up to 30 days before the following influenza vaccination date. The period from IBD diagnosis up to 30 days before the initial influenza vaccination was also included in the control period. Our choice of the control period was based on the fact that it would be highly unlikely that acute inflammation from the vaccine would result in health service increases this far removed from the date of vaccination. Furthermore, we did not want to choose a control period so far removed from the vaccination that it could be influenced by the subsequent vaccination event through the healthy vaccinee effect. We calculated relative incidence (RI) with 95% confidence intervals (CI) of health services visits in the risk periods compared with the control period, controlling for within-individual clustering, patient age, and season. To assess whether IBD cases were more likely to experience events in the risk period than non-IBD patients, and to control for possible differential healthy vaccinee effects in healthy children and IBD patients, we calculated relative incidence ratios (RIR) with 95% CIs, which provide an estimate of the ratio of the RI in IBD children to the RI in controls within the SCCS model. Due to possible bias toward a lower RI in the first year after diagnosis (due to increased health services utilization after diagnosis22), a sensitivity analysis excluded immunizations administered in the first year after diagnosis. Statistical tests were 2 sided and with significance assigned at P < .05. Analyses were performed by using SAS version 9.2 (SAS Institute, Inc, Cary, NC).
Characteristics of Cases and Controls
The OCCC included 4916 children with IBD diagnosed between 1999 and 2010 who met inclusion criteria (2809 [57.1%] with CD, 1720 [35.0%] with UC, and 387 [7.9%] whose diagnosis was unclassifiable). These patients were matched to 21 686 non-IBD controls. Descriptive statistics of cases and controls are presented in Table 1. IBD cases were more likely than controls to live in urban regions (P < .001) and higher income neighborhoods (P < .001). They also had more physician visits per year of follow-up compared with controls (4.91 ± 7.96 vs 2.91 ± 19.53; P < .001).
Influenza Immunization Rates
During the full follow-up period, immunization was administered at least once in 25.3% of IBD patients compared with 13.2% of controls (P < .001). Although uptake was generally low, more IBD patients were immunized in all years in which they were eligible to receive the vaccine compared with non-IBD controls (3.8% vs 1.4%; P < .001), in at least 75% of eligible years (5.7% vs 2.0%; P < .001), and in at least 50% of eligible years (14.2% vs 6.0%; P < .001). The distribution of vaccine administration after date of diagnosis is displayed in Fig 2, and the number of immunizations received by cases and controls is reported in Fig 3. Vaccination coverage ranged from 4.6% to 23.5% in eligible IBD patients, and 0.18% to 12.4% in controls (Fig 4), with increased coverage after introduction of universal influenza immunization in 2000.
Adverse Events: Hospitalization, ED Visit, or Physician Office Visit After Immunization
In IBD patients, there was no increase in health services use in any risk period after immunization compared with the control period (Table 2). The pooled RI of any health services utilization during the risk period (3–180 days after immunization) was 0.95 (95% CI: 0.84–1.07; P = .38). The findings remained robust after a sensitivity analysis excluded the first 14 days after immunization to minimize the healthy vaccinee effect (pooled RI: 0.92 [95% CI: 0.82–1.04]; P = .19) and after eliminating the first year after diagnosis (Supplemental Tables 5 and 6). There was no increased health services utilization for IBD-related reasons in any risk period after immunization in patients with IBD (Table 3). There was significantly lower IBD-related health services utilization 15 to 30 days after immunization compared with the control period (RI: 0.45 [95% CI: 0.26–0.80]; P = .006). Pooled analyses demonstrated significantly lower IBD-related health services utilization in the combined postvaccine periods compared with the control period (RI: 0.81 [95% CI: 0.68–0.96]; P = .014). Pooled analysis of IBD-related hospitalizations and ED visits during days 3 to 180 after immunization did not reveal any significant difference (RI: 0.75 [95% CI: 0.47–1.21]; P = .24). In control patients, there was lower health services use in days 3 to 14 after immunization (RI: 0.64 [95% CI: 0.47–0.89]; P = .007) and days 31 to 45 after immunization (RI: 0.68 [95% CI: 0.52–0.91]; P = .008). Overall, the SCCS analysis in controls demonstrated lower health services use in the pooled risk period compared with the control period (RI: 0.90 [95% CI: 0.82–0.98]; P = .02).
When comparing IBD patients with control patients, we observed increased health services utilization 3 to 14 days after immunization in IBD patients (RIR: 1.60 [95% CI: 1.05–2.44]; P = .03). This finding was primarily due to outpatient visits, with no significant differences in hospitalization or ED visits. In IBD patients, the most frequent diagnostic codes assigned to outpatient visits during the 3- to 14-day postimmunization period were for CD (102 visits) and UC (37 visits). Otherwise, the reasons for outpatient visits did not differ significantly between IBD and control patients. In IBD patients, the top non–IBD-specific codes assigned at outpatient visits during this period were common cold (n = 23), anorexia/nausea/vomiting (n = 19), and spastic irritable colon (n = 11). For control patients, the top reasons for outpatient visits were common cold (n = 28), asthma/allergies (n = 15), and neuroses (n = 14). From 15 to 180 days, IBD cases had a lower RI relative to controls (pooled RIR: 0.85 [95% CI: 0.74–0.97]; P = .02).
We demonstrated low influenza immunization coverage in children with IBD in the context of a universal immunization strategy and supportive international clinical guidelines. We also observed that patients who were immunized against influenza were not at increased risk for hospitalization, ED usage, or outpatient physician visits for IBD-related or any other reason during the pooled at-risk period. Notably, IBD-related health services utilization decreased in the postvaccination periods compared with the control period.
Considering higher rates of overall physician care provided to children with IBD, it was not surprising that influenza vaccination coverage was higher compared with controls. Very few patients satisfied the recommendation of yearly immunization. Even if one were to conservatively consider the sensitivity of the influenza vaccine billing code to be 55%,17 fewer than one-half of children with IBD received 1 dose of vaccine in their follow-up period. Our findings are consistent with previous survey research demonstrating similar immunization coverage in other populations.23–26 Many unimmunized patients were unaware of the recommendation for yearly immunization,23,25,26 and a significant proportion of patients, ranging from 8% to 33%, reported concern about adverse events, which may have contributed to lower rates. Lack of accessibility to the vaccine may also have decreased uptake. The influenza vaccine is typically unavailable in Ontario pediatric IBD clinics, necessitating travel to family physicians or public health clinics. Stronger recommendations for immunization by providers and administration during regular IBD follow-up visits would improve uptake.
We found no evidence of worsening disease or adverse events in any 2-week period up to 6 months after vaccine administration in either IBD patients or controls. However, when comparing IBD patients with controls, we observed slightly higher overall health services utilization in the first 14 days in patients with IBD. This increase was due to an increased rate of outpatient visits, not hospitalizations or ED visits. The increased likelihood of outpatient visits after immunization in IBD patients compared with controls was due to an increased likelihood of IBD patients visiting their physician for IBD-related reasons, and it did not seem to be due to increased adverse events related to immunization. The overall reduction in health services utilization in the pooled period of days 15 to 180 in IBD patients compared with controls potentially suggests a stronger protective effect of influenza immunization in children with IBD compared with controls.
In addition, after immunization, there was a reduction in visits to physicians with associated IBD-related diagnostic codes. This finding may have been secondary to ascertainment bias, with symptomatic IBD patients visiting physicians more likely to receive immunization in addition to treatment of their IBD. The IBD treatment may have subsequently resulted in decreased visit rates in the postimmunization period. However, this outcome would not explain lower postvaccine visit rates than the control period, during which time patients should be at average level for disease activity. Our observation of lower IBD-related events after immunization may also have been secondary to the protective effects of the vaccine; fewer influenza infections may have resulted in a nonspecific immune priming effect (when a pathogen or immune elicitor alters immune response to other antigens), via the adaptive immune system,27 or protection against viral interaction with the gut mucosal barrier.
Although case reports of IBD flares after influenza immunization exist,8–10 observational studies demonstrated no association with flare of disease.24,28 However, these were underpowered to demonstrate rare events, nor did they document less serious events that may have prompted outpatient physician visits. Randomized controlled trials and observational studies in multiple sclerosis29,30 and rheumatoid arthritis31 also did not demonstrate increased risk of disease flare after immunization. The primary advantage of the current study is its large numbers of included children with IBD and its ability to assess the full spectrum of health system utilization. This is the first population-based study to demonstrate the safety of influenza by immunization, using SCCS methods. This method has several advantages over other observational designs. Multiple risk periods and control periods could be contained within a single patient if immunized in multiple years, thereby increasing power while controlling for clustering. By using individuals as their own controls, the design avoids the bias of using unvaccinated controls who are likely to be systematically different from vaccinated children.32
The use of health administrative data has limitations. Misclassification error is always possible. We used a validated algorithm, which performed well in 2 independent cohorts of Ontario children.14 However, the influenza vaccination code has demonstrated only fair sensitivity compared with self-reported vaccination rates, particularly in older, healthy adolescents.17 This finding is likely due to administration of influenza immunization by public health clinics. Therefore, we may have underestimated vaccination rates in Ontario patients. In particular, differential misclassification of immunization status may have contributed to the finding that IBD cases are immunized more frequently than controls. However, our analyses of adverse events after vaccine administration would not have been affected by billing code sensitivity because SCCS analysis assessed only vaccinated patients, and the code has a high specificity and negative predictive value.
Use of health administrative data does not allow for inclusion of clinical disease characteristics such as phenotype, severity, and extent. In addition, medication utilization is not available for children in Ontario. Therefore, we were not able to discern differences in immunization uptake or adverse events in patients receiving immunosuppressive therapy. Although we did not have access to serologic immunity testing in our cohort, most studies have demonstrated adequate immunogenicity to most influenza strains,24,33,34 and the reduced health services utilization in IBD patients compared with controls in the 6 months after administration suggests protection.
Our overall analysis demonstrated no association between influenza immunization and an increase in IBD-related physician visits, hospitalization, ED visits, or overall health services utilization, supporting the safety of the vaccine in this population. In fact, there may be a protective benefit for IBD patients. Importantly, immunization coverage is far from optimal, with few children receiving the vaccine yearly as per international guidelines. These findings represent an opportunity for quality improvement initiatives aimed at increasing vaccine uptake, which would likely reduce the burden of this chronic disease on the health system.
The authors thank Dr David Mack for his feedback on portions of the manuscript.
- Accepted February 22, 2013.
- Address correspondence to Eric I. Benchimol, Division of Gastroenterology, Hepatology and Nutrition, Children’s Hospital of Eastern Ontario, 401 Smyth Rd, Ottawa, Ontario, Canada, K1H 8L1. E-mail:
Dr Benchimol conceptualized and designed the study, interpreted the results, and wrote the initial manuscript; Dr Hawken conceptualized and designed the study, conducted the analyses, and reviewed manuscript drafts; Dr Kwong designed the study, interpreted the results, and reviewed manuscript drafts; Dr Wilson conceptualized and designed the study, interpreted the results, and reviewed manuscript drafts; and all authors approved the final manuscript.
FINANCIAL DISCLOSURE: Dr Benchimol is supported by a Career Development Award from the Canadian Child Health Clinician Scientist Program. Dr Kwong is supported by a Clinician Scientist Award from the Department of Family and Community Medicine, University of Toronto. Dr Wilson holds the Canada Research Chair in Public Health Policy. The Institute for Clinical Evaluative Sciences receives funding from the Ontario Ministry of Health and Long-Term Care. The results and conclusions are those of the authors; no official endorsement by the Ontario Ministry of Health and Long-Term Care should be inferred.
FUNDING: Internal funding.
- Rahier JF,
- Ben-Horin S,
- Chowers Y,
- et al.,
- European Crohn’s and Colitis Organisation (ECCO)
- Canadian Immunization Guide
- ↵Fields SW, Baiocco PJ, Korelitz BI. Influenza vaccinations: should they really be encouraged for IBD patients being treated with immunosuppressives? Inflamm Bowel Dis. 2009;15(5):649–651; author reply 651–652
- Benchimol EI,
- Guttmann A,
- Griffiths AM,
- et al
- ↵CIHI Data Quality Study of the 2009-2010 Discharge Abstract Database. Ottawa, Ontario, Canada: Canadian Institute for Health Information, Standards and Data Submission; 2012
- ↵Richards J, Brown A, Homan C. The data quality study of the Canadian Discharge Abstract Database. Proceedings of Statistics Canada Symposium. Available at: www.statcan.gc.ca/pub/11-522-x/2001001/session16/6282-eng.pdf. Accessed May 30, 2012
- ↵Benchimol EI, To T, Griffiths AM, Rabeneck L, Guttmann A. Outcomes of pediatric inflammatory bowel disease: socioeconomic status disparity in a universal-access healthcare system. J Pediatr. 2011;158(6):960–967. e1–4
- ↵Debruyn J, Wrobel IT. Influenza vaccination awareness and history in children with inflammatory bowel disease. Gastroenterol. 2009;136(5 suppl 1):A677
- Aaby P,
- Whittle H,
- Benn CS
- Rahier JF,
- Papay P,
- Salleron J,
- et al.,
- European Crohn’s and Colitis Organisation (ECCO)
- Miller AE,
- Morgante LA,
- Buchwald LY,
- et al
- Farrington CP,
- Nash J,
- Miller E
- Copyright © 2013 by the American Academy of Pediatrics