Inflammatory bowel disease (IBD), which includes Crohn disease (CD) and ulcerative colitis (UC), is a chronic, relapsing inflammatory disorder that primarily affects the digestive tract. Environmental factors in genetically susceptible individuals may contribute to the development of the disease. The specific factors that induce disease activity remain poorly defined. Infections and other environmental exposures that can vary by season have been postulated to trigger relapses (1–4).
Conflicting data have been reported regarding seasonal variation of IBD exacerbations in adult populations. Studies have found that UC exacerbations can peak in any season, including the fall (5–9), winter (5–7), spring (8–11), or summer (11). Others have found no association between season of the year and hospital admissions (12–15). CD exacerbations have been reported to peak in the fall and winter (16), although the majority of studies have not found any seasonality of disease activity (5,10,12–15). The natural course of IBD in children may differ from that in adults, potentially because of differential environmental effects. No study has focused solely on pediatric patients with IBD.
In the United States, traditional temporal seasons are more distinct in the northern region. Seasonality of disease activity may be more prominent in northern states if IBD exacerbations are related to a factor that varies by seasons. Exploring this possible association may help to elucidate the natural course of IBD.
The present multicenter study examined whether season is associated with disease activity in patients with pediatric-onset IBD and whether results are consistent across regions.
ImproveCareNow (ICN) is a multicenter network of health care providers established in January 2007 to improve the quality of care of children and adolescents with IBD. The ICN patient registry contains disease and treatment data collected prospectively and longitudinally during outpatient encounters. Patients were diagnosed and managed according to the usual practice of the primary gastroenterologist, although quality improvement methodology was used to reduce any unintended variation. Model care guidelines, emphasizing correct dosing of medications and frequency of laboratory tests, are available to all ICN sites, and include the recommendation that patients be seen at least twice yearly.
The present study analyzed data from patients diagnosed as having CD and UC before age 21 seen from December 2008 to November 2010 (8 seasons) at 12 secondary and tertiary centers in the United States. These centers were selected because of their near-complete enrollment of their patients with IBD during this time period.
We first compared the proportion of patients in continuous remission based on physician global assessment (PGA) during each season after combining the seasons across the 2 years. Physicians classified patients’ disease activity as quiescent (in remission), mild, moderate, or severe based on the clinical and laboratory data available at the time of the assessment (17,18). Clinical and laboratory indicators that may have influenced PGA classification included reported symptoms of abdominal pain, number of daily stools, level of functioning and well-being, height and weight indicators, physical examination, hemoglobin, hematocrit, erythrocyte sedimentation rate, C-reactive protein, albumin, and others as available to the clinician at the time of classification. To be classified as continuous remission, a given patient must have received an assessment of remission for all visits recorded during a given season. The proportion was calculated with the numerator defined as the number of patients seen who were in remission (by PGA) at all visits during that season. The denominator was defined as the total number of patients with at least 1 visit occurring in the given season, regardless of disease activity. Each patient was counted only once per season for the 8 seasons. Seasons were defined as winter (December, January, February), spring (March, April, May), summer (June, July, August), and fall (September, October, November) for each of 2009 and 2010. We tested for consistency of seasonal effects across regions and across years.
We then investigated whether any significant differences found in the proportion of patients in continuous remission each season were related to the distribution of visits in remission in each season. Correspondingly, we determined the distribution of visits with active disease in each season to evaluate for any seasonal effect. In this secondary analysis, we first evaluated whether the proportion of all visits with a PGA of remission was constant across the 4 seasons (eg, 25% for each season) after combining the seasons across the 2 years. The proportion of visits in remission each season was calculated with the numerator defined as the number of visits with a PGA of remission for that season and the denominator defined as the total number of visits with a PGA of remission across all seasons. Similarly, we then compared the proportion of visits with active disease (calculated as noted above for remission) that occurred in each of the 4 seasons to an expected rate of 25%. For this “visit-based” analysis, all visits for patients were included.
To investigate potential sex differences in patient visits across the seasons, we determined the proportion of girl and boy patients seen during each season. Similarly, we investigated potential race differences across the seasons. Patients were categorized as white, black, or other. The proportion of each category of patients seen was determined for each season.
To evaluate whether the effect of season may have varied by region, we performed a subanalysis stratifying northern versus southern region. Seven centers (Vermont Children's Hospital at Fletcher Allen Health Care, VT, 44°N; Barbara Bush Children's Hospital at Maine Medical Center, ME, 44°N; Advocate Lutheran General Children's Hospital, IL, 42°N; Massachusetts General Hospital for Children, MA, 42°N; Nationwide Children's Hospital, OH, 40°N; Inova Fairfax Hospital for Children, VA, 39°N; Children's Hospital and Research Center Oakland, CA, 38°N) were located in the northern region, and 5 centers (Pediatric Gastroenterology & Nutrition Associates, NV, 36°N; North Carolina Children's Hospital at University of North Carolina Health Care, NC, 36°N; The Children's Hospital at Oklahoma University Medical Center, OK, 36°N; Children's Healthcare of Atlanta at Egleston, GA, 34°N; Children's Medical Center at University of Texas Southwestern Medical Center, TX, 33°N) were located in the southern region using 37° latitude to bisect the country. Consistency of seasonal effect was also tested across centers. For all analyses, CD and UC were analyzed separately.
All calculations were performed using SAS 9.3 (SAS Institute Inc, Cary, NC). For the primary analysis, data were analyzed using a generalized linear mixed model with a logit link to evaluate variation of disease activity by season with terms for season, year, region, site nested within region, season-by-year, and season-by-region interactions. For the secondary analyses, multinomial χ2 tests of proportions were used to test the distribution of percent of all visits in remission to hypothesized values of 25% for each season. A similar test was used testing the percent of all visits with active disease across seasons. The χ2 tests were used to determine whether there was a difference in sex or race categories for patient visits across seasons. A linear model with fixed effects including site, season, year, and season-by-site interaction was used to determine whether the seasonal effect was consistent across centers. P values <0.05 were considered significant.
Proportion of Patients in Continuous Remission Each Season
The study included 1325 patients with CD (43.5% girls; mean age 15 ± 3.5 years) and 587 patients with UC (53.7% girls; mean age 14.1 ± 4.1 years). PGA data were available for 99.1% (6102/6156) of visits of patients with CD and 98.7% (2394/2425) of visits of patients with UC. The median number of individual patient visits per year for patients with UC was 2 (1–10) in 2009 and 2 (1–15) in 2010. For patients with CD the median was 2 (1–13) in 2009 and 2010. In the primary analysis, the proportion of patients with UC in continuous remission throughout a season was highest in the summer and lowest in the winter (summer 67.2%, fall 64.3%, winter 55.1%, spring 60.6%; P = 0.01). There was a numerically greater proportion of patients with CD in continuous remission throughout the summer, but this was not statistically significant (summer 63.9%, fall 59.6%, winter 60.1%, spring 61.2%; P = 0.09). The results were consistent across years (2009, 2010) and both regions (northern, southern) because the 2-way interaction effects were not significant for either disease group (data not shown). Additionally, results were consistent across sites for patients with both UC (P = 0.80) and CD (P = 0.17).
Proportion of Visits in Remission and With Active Disease Each Season
In the secondary analyses, we analyzed the distribution of visits in remission and the distribution of visits with active disease across seasons combined during the 2-year period. For patients with UC, the distribution of visits in remission (n = 1440) across seasons was higher than expected in the summer and lower than expected in the winter (summer 28.7%, fall 24.8%, winter 20.5%, spring 26.0%; P = 0.0002). A similar pattern was found for patients with CD, with the proportion of visits in remission (n = 3688) being higher than expected in the summer and lower than expected in the winter (summer 28.1%, fall 25.6%, winter 22.6%, spring 23.7%; P < 0.001). The proportion of visits with active disease for both patients with UC (n = 954) (summer 24%, fall 25.4%, winter 24.5%, spring 26.1%; P = 0.81) and patients with CD (n = 2414) (summer 24.5%, fall 27.3%, winter 24.2%, spring 24%; P = 0.07) was similar across seasons. No differences were seen in the patterns above when analyzing within the northern and southern regions (data not shown).
Proportion of Visits by Sex and Race Each Season
In additional analyses, there were no significant differences in the proportion of female and male patients seen each season for patients with UC (P = 0.96) and CD (P = 0.83). Similarly, there were no significant differences in the proportion of white, black, or other patients seen each season for UC (P = 0.96) and patients with CD (P = 0.998).
Among patients with UC making ≥1 visits in a given season, we found the proportion of patients in continuous remission was greater during the summer than in the winter. For CD we found a similar but nonsignificant higher number of patients in continuous remission during the summer months. The proportion of visits with active disease was similar across seasons for both CD and UC, suggesting little, if any, variation in exacerbations by season. The greater proportion of remission visits seen in the summer months may be because of scheduling preferences for patients with stable, well-controlled disease to minimize the amount of missed school. Patients traveling greater distances for their visits may also prefer to schedule routine visits during the summer for ease of travel given potential weather hazards in the winter. Such an effect may be more of a factor in northern latitudes. This would inflate the proportion of patients in continuous remission during the summer, relative to other seasons. Taken together, these findings suggest little or no effect of season on actual disease activity in pediatric patients with IBD, at least as measured using outpatient data. In addition, although northern states are thought to have more distinct temperate seasons, we did not find any differences between northern and southern regions in the seasonal variation of disease activity. Although no effect was seen by region, many of the participating sites were relatively centrally located along latitude 37°, which may have mitigated any geographic variation. Additional analyses did not reveal any sex or racial differences in the distribution of visits across seasons that may have affected our results.
In adult patients, an association between season of the year and disease activity in IBD was first noted by Myszor and Calam in 1984 with a retrospective survey of 50 patients with UC and 50 patients with CD (5). They found that exacerbations in patients with UC were more frequent from September to February. Other early studies also found seasonality of disease activity in patients with UC. In a prospective analysis of patients with UC, Sellu found higher rates of relapse in the fall and winter (6) and Riley et al found higher rates of relapse in the fall (7). Similarly, Karamanolis et al showed that UC exacerbations were most common in the fall and spring in patients from Greece (9). One analysis of patients with CD managed by 1 practitioner in Canada found higher rates of relapse in the fall and winter (16).
More recent studies, however, have failed to identify any seasonal variation in IBD exacerbations. Lewis et al analyzed 2773 patients with UC and 1587 patients with CD (10). There was a weak increased incidence of UC exacerbations in the spring, but overall they concluded that there was no association between season of the year and exacerbations for patients with CD or UC. Several other large-scale studies have failed to confirm any seasonality of IBD exacerbations (12–15). Therefore, our findings in the pediatric IBD population are consistent with the findings of the latter studies of adult patients with IBD.
The strengths of the present study include the multicenter prospective nature of our data collection during a 2-year period, the large sample size, and the geographically diverse population. There are several potential limitations to our study. Most important, the data used for the present analysis were collected during outpatient clinic visits. We do not have information on patients who did not have outpatient visits during a quarter, and hence could not include these patients in our analysis of continuous remission. Similarly, data regarding exacerbations managed during hospitalizations, during emergency department visits, and via telephone encounters were not available for analysis. Most relapses are managed in the outpatient setting, and ED or inpatient encounters generally receive prompt outpatient follow-up. We are also unable to account for patient adherence to medications that could vary by season. Although our sample is not a population-based sample, the majority of pediatric patients are managed at referral centers such as the ones included in our study, and the sites were specifically included only if they had enrolled most of their IBD population in the ImproveCareNow registry. The determination of an exacerbation of disease activity has differed among prior studies. We used PGA to define disease activity. Although it is a subjective measure, it has been used frequently as a primary outcome measure in pediatric IBD research and is the standard for other pediatric disease activity measures (19,20). A final limitation is that we measured disease activity at recorded visits, although the onset of disease exacerbations may have preceded the season at which the visit was made. This would tend to underestimate or shift any seasonal variation.
In conclusion, our study does not support any strong associations between season of the year and disease activity in pediatric-onset IBD. The higher proportion of patients with UC in continuous remission in the summer seems to be related to a higher proportion of remission visits in the summer, because the proportion of visits with active disease was similar across seasons. The exact role of environmental factors contributing to IBD activity remains to be determined, although it appears that any significant factors are likely consistent across season and region.
1. Sonnenberg A. Seasonal variation of enteric infections and inflammatory bowel disease
. Inflamm Bowel Dis
2. Kangro HO, Chong SK, Hardiman A, et al. A prospective study of viral and Mycoplasma
infections in chronic inflammatory bowel disease
3. Mee AS, Jewell DP. Factors inducing relapse in inflammatory bowel disease
. Br Med J
4. Sonnenberg A. Similar geographic variations of mortality and hospitalization associated with IBD and Clostridium difficile
colitis. Inflamm Bowel Dis
5. Myszor M, Calam J. Seasonality
of ulcerative colitis
6. Sellu DP. Seasonal variation in onset of exacerbations of ulcerative proctocolitis. J R Coll Surg Edinb
7. Riley SA, Mani V, Goodman MJ, et al. Why do patients with ulcerative colitis
8. Tysk C, Jarnerot G. Seasonal variation in exacerbations of ulcerative colitis
. Scand J Gastroenterol
9. Karamanolis DG, Delis KC, Papatheodoridis GV, et al. Seasonal variation in exacerbations of ulcerative colitis
10. Lewis JD, Aberra FN, Lichtenstein GR, et al. Seasonal variation in flares of inflammatory bowel disease
11. Bai A, Guo Y, Shen Y, et al. Seasonality
in flares and months of births of patients with ulcerative colitis
in a Chinese population. Dig Dis Sci
12. Don BA, Goldacre MJ. Absence of seasonality
in emergency hospital admissions for inflammatory bowel disease
13. Sonnenberg A, Jacobsen SJ, Wasserman IH. Periodicity of hospital admissions for inflammatory bowel disease
. Am J Gastroenterol
14. Sonnenberg A, Wasserman IH. Epidemiology of inflammatory bowel disease
among U.S. military veterans. Gastroenterology
15. Soncini M, Triossi O, Leo P, et al. Seasonal patterns of hospital treatment for inflammatory bowel disease
in Italy. Digestion
16. Zeng L, Anderson FH. Seasonal change in the exacerbations of Crohn's disease. Scand J Gastroenterol
18. Hyams J, Markowitz J, Otley A, et al. Evaluation of the pediatric Crohn disease
activity index: a prospective multicenter experience. J Pediatr Gastroenterol Nutr
19. Kappelman MD, Crandall WV, Colletti RB, et al. Short pediatric Crohn's disease activity index for quality improvement and observational research. Inflamm Bowel Dis
20. Turner D, Hyams J, Markowitz J, et al. Appraisal of the pediatric ulcerative colitis
activity index (PUCAI). Inflamm Bowel Dis