What Is Known
- Quicker diagnosis and treatment of inflammatory bowel diseases has been found to lead to improved outcomes in children.
- Research on seasonality in onset of pediatric inflammatory bowel disease has had conflicting results.
What Is New
- First study to use independent discovery and validation cohort to address seasonality of inflammatory bowel disease presentation in children.
- Consistently significant findings in this work indicate that academic stress may contribute to onset of pediatric inflammatory bowel disease. This observation may aid the timely diagnosis of pediatric inflammatory bowel disease.
- No consistent significant relationship between seasonality and onset of pediatric inflammatory bowel disease was found in our study.
Epidemiologic data on high monozygotic twin discordance rates, migration studies, and the increasing prevalence of inflammatory bowel diseases (IBDs) strongly implicates their environmental origins (1,2). Perinatal seasonal variation in regards to IBD development (3–7) (ie, seasonality dependent perinatal origins from birth) and seasonal variation of disease onset (8–14) (ie, seasonal triggers) have been examined more commonly in the adult onset form of the disorders than in pediatric cases. Onset of pediatric IBD has been linked to seasonal variations by Dharmaraj et al (15), who found that IBD presentation had the highest incidence in the fall and lowest in the summer. More recent work from Brandvayman et al (16) did not corroborate these findings. These conflicting observations encouraged us to examine prenatal (month of conception), perinatal (month of birth), and disease onset (month of first symptoms) seasonality in a single-center cohort of pediatric patients with IBD from the Southern United States.
PATIENTS AND METHODS
The Baylor College of Medicine institutional review board approved protocol: Using Patient Data to Transform Care and Improve Outcomes for Children with Inflammatory Bowel Disease (H-26423) was utilized in this work. The medical records of pediatric patients with IBD, ages 6 to 18 years, diagnosed at Texas Children's Hospital and enrolled into our Improve Care Now Consortium Registry between 2009 and 2015 were examined. A discovery cohort of 169 patients (random list of patients presenting between 2009 and 2015, where data were already available from prior studies) and a validation cohort of 122 patients (rest of patients during examination years) were surveyed by 2 independent observers (Supplementary Table S1, Supplemental Digital Content, https://links.lww.com/MPG/B386), respectively. Only those cases were included into the analyses where the onset of symptoms attributable to IBD (such as fever, diarrhea, blood in stool, weight loss, abdominal pain) were clearly recorded in the medical record of the initial encounter. The month of disease onset was calculated based on the visit/encounter date and the length of symptoms. Cases were omitted from the study where onset of symptoms were not clearly recorded (the discovery cohort initially had 177 cases examined and the validation initially had 136 cases before exclusions). Only month of birth and extrapolated month of conception, age of onset, month of disease onset, sex, and working clinical diagnoses were collected from the medical records. Any notes in the patient's chart about premature birth were also used in the extrapolation of conception months. We applied Fischer exact test and Student t test in group comparisons. Level of significance was set at P <0.05.
We expected that each month and season would have the same probability of initial symptoms in the case of random event distribution. A Fischer exact test was first used to compare observed to expected seasonal IBD occurrence. Seasonality was assessed by dividing the year into 4 seasons defined as Winter (December–February), Spring (March–May), Summer (June–August), and Autumn (September–November). Since no seasonal pattern of incidence was discovered, we looked into the months of highest incidence. The peak 5 months were January, May, August, September, and December. We recognized that these are the months of semester initiation and end in variable school settings of the Houston Metro area. The Fischer exact test was then used to compare the 5 peak months of disease presentation to the next 5 months with highest presentation. Because our initial findings were an unexpected discovery, we decided to perform an unbiased replicate study in an independent validation cohort, which was executed by an observer not participating in the discovery work.
In the discovery cohort, we examined 177 charts of which 169 (95.5%) had onset of symptoms recorded. In the validation cohort, we examined 136 charts of which 122 (89.7%) had onset of symptoms recorded. The proportion of Crohn disease (CD), ulcerative colitis (UC), and indeterminate colitis (IC) was similar between the 2 cohorts compared to our newly diagnosed overall cohort of 488 patients between 2011 and 2016 (CD: 54.5%, 58.6%, 52.5%, respectively; UC: 30.1%, 28.4%, 29.5%, respectively; IC: 15.4%, 13.0%, 18.0%, respectively).
We found no association between season of conception, or birth and IBD in the discovery cohort (P > 0.5).
When comparing fall to summer presentation, we found no significant difference in the discovery cohort (P = 1.0) in contrast to Dharmaraj et al's (15) findings, where pediatric IBD presentation was highest in fall and lowest in the summer. Winter and spring presentation compared to summer and fall was significant in the discovery cohort (P = 0.0463) similar to studies done in Japan by Koido et al (10), but this only showed a nonsignificant trend in the validation cohort (P = 0.0958). A subanalysis by diagnosis comparing winter and spring incidence to summer and fall found significance for CD (P = 0.0463), but not UC (P = 0.1527) in the discovery group. In the validation cohort, however, this CD-related finding was found to be nonsignificant (P = 0.2158).
After no consistent seasonal pattern of incidence was discovered, we looked into this question by months of highest incidence. IBD symptoms presented significantly more commonly in the 5 months of academic semester start (August, September, January) and end (December, May) in the Houston Metro Area, than in the following greatest 5 months of incidence in the year (Fig. 1A; P = 0.0218). To further examine and validate this discovery finding, we decided to perform a validation study on independent pediatric patients with IBD by an unbiased additional observer. The 5 months of academic semester start (August, September, January) and end (December, May) stood out even more significantly from the rest of the months (Fig. 1B; P < 0.0001) in the validation cohort, than in the discovery cohort. A subanalysis by IBD diagnosis showed significant differences in this respect only in the validation cohort for both CD (P = 0.005) and UC (P = 0.0089), but not IC (P = 0.1308) (Supplementary Fig. S1, Supplemental Digital Content, https://links.lww.com/MPG/B386).
We did not find consistent seasonality for symptomatic presentation of IBD between our discovery and validation cohorts. The conflicting results in this respect even within a single study underscore the value of us utilizing the discovery and validation sequence of analysis. With this valuable approach, we observed an annual rhythm of disease onset, which consistently correlated with academic semesters. It is important to clarify that within the Houston Metro Area's Elementary, Middle, and High School systems, the beginning of January is the start of the second semester of a given academic school year, and the end of May is the culmination of the same semester (with standardized testing 1–3 weeks before the end). Late August is the start of the academic year's first semester (the stress of which we believe stretches into September) that ends in mid-December (again with finals, especially in High Schools). Personal parental experience in respect to school-related pediatric stress (including child and family) lead us to the discovery finding of the study. The nonhypothesis-driven discovery observation (ie, unbiased by observer prediction), and the employment of autonomous observers in independent discovery and validation cohorts, significantly adds to the reliability of our findings. In addition, the restricted season variation within Houston provides an advantageous location to observe academic relationship with IBD as weather is a limited confounding factor. This is indeed the first study to find a possible correlation between academic seasonality and onset of pediatric IBD.
Previous studies have established that academic examinations may cause an altered immune response (17,18). Murphy et al (17) detected decreased immune functioning in undergraduate students by measuring salivary cortisol concentrations during examination time versus nonexamination. In addition, students with a high stress perception had significantly more production of tumor necrosis factor α, interleukin 6 (IL-6), IL-1 receptor antagonist (IL-1Ra), and interferon γ (18,19). Higher levels of stress are also related to increased amounts of glucocorticoids (cortisol) and catecholamines (17,19–21), which if chronically present leads to decreased immune responses. Tian et al (21) identified how products of stress affect production of proinflammatory cytokines (which include tumor necrosis factor α, IL-1Ra, interferon γ) and anti-inflammatory cytokines. Although anti-inflammatory cytokines will initially be produced in greater number, chronic stress alters the hypothalamus-pituitary-adrenal axis to be cortisol resistant, which will ultimately cause relatively more proinflammatory cytokines to be produced (21). Therefore, the balance between anti-inflammatory and proinflammatory cytokines is disrupted by chronic stress, and susceptibility toward various inflammatory diseases may increase, including pediatric IBD (19). Our subanalyses did not indicate a consistently significant disease-specific difference in respect to academic stress–related IBD presentation.
Limitations to this study include its retrospective nature and being performed within a restricted geographical area. Prospective epidemiologic studies, however, do not guarantee unbiased results either. In addition, it is the defined, single-center geographical area, which is likely to have lead us recognize the critical finding of our study, because it allowed for defining the months of academic stress by following the schedule of the Houston area school districts. In respect to the higher incidence in January relative to the other months identified, viral and bacterial infections may also serve as possible confounding factors. Such infections have a greater occurrence in winter, which may work synergistically with academic stress to worsen the inflammatory responses, explaining the highest incidence of pediatric IBD presentation in January.
Academic stress was not able to be measured directly. Therefore, we acknowledge the provocative nature of our work with the goal of stimulating further research. Future clinical studies could investigate academic stress by measuring various immune factors, including cortisol levels by month.
Recall bias may have influenced the accuracy of our disease onset determination, but that should have occurred in each case equally, and is highly unlikely to have significantly affected both the discovery and the validation findings in a similar fashion. This conclusion repeatedly underscores the unique value of our discovery-validation cohort approach within this work.
We emphasize that the most common/shared form of stress that manifests in children is academic, which peaks during the beginning and end of each semester in the geographical location studied. In addition, academic stress may be more impactful in older populations, which can be seen in the notable difference when comparing the significance of the 5 months identified between our discovery and validation cohorts (Table 1). Our discovery cohort had a younger median (13) and mean (12.39) age compared to our validation cohort's median (15) and mean age (13.79). Therefore, this may account for the greater significance found in our validation cohort (P < 0.0001) compared to the significance of the discovery cohort (P = 0.0218) if we speculate that older pediatric patients (in high school) may have more academic stress than younger children. In addition, this age difference may account for the significance when comparing winter and spring to summer and fall for onset of symptoms of CD in the discovery cohort, but lack of this finding in the validation cohort. Younger children may be more susceptible to CD inducing environmental factors present in winter and spring than older pediatric patients in the Houston Metro area. This speculation may warrant further studies in the future.
In conclusion, our findings implicate that onset of IBD may be in part linked to academic stress when added cytokine imbalances may tip the genetically, epigenetically and “dysbiotically” susceptible host (22) toward developing the diseases.
The recognition of the highest incidences of IBD presentation during the peaks of academic stress may alert primary care physicians to recognize the alarm signs for the diseases, and prompt timely referral to the gastroenterologist. Such practice may optimize patient care, since early initiation of appropriate therapies has been associated with favorable outcomes in pediatric patients with IBD (23).
The authors would like to acknowledge the Gutsy Kids Fund, including philanthropic donations from the Karen and Brock Wagner family. The authors would also like to acknowledge Raymond Tjhia for his aid in formatting figures for submission.
1. Kellermayer R. Epigenetics and the developmental origins of inflammatory bowel diseases. Can J Gastroenterol
2. Fofanova TY, Petrosino JF, Kellermayer R. Microbiome-epigenome interactions and the environmental origins of inflammatory bowel diseases. J Pediatr Gastroenterol Nutr
3. Jung YS, Song CS, Kim ER, et al. Seasonal variation in months of birth and symptom flares in Korean patients with inflammatory bowel disease. Gut Liver
4. Shaw SY, Nugent Z, Targownik LE, et al. Association between spring season of birth and Crohn's disease. Clin Gastroenterol Hepatol
5. Sonnenberg A. Date of birth in the occurrence of inflammatory bowel disease. Inflamm Bowel Dis
6. Sørensen HT, Pedersen L, Nørgård B, et al. Does month of birth affect risk of Crohn's disease in childhood and adolescence? BMJ
7. 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
8. Araki M, Shinzaki S, Yamada T, et al. Age at onset is associated with the seasonal pattern of onset and exacerbation in inflammatory bowel disease. J Gastroenterol
9. Basaranoglu M, Sayilir A, Demirbag AE, et al. Seasonal clustering in inflammatory bowel disease: a single centre experience. Expert Rev Gastroenterol Hepatol
10. Koido S, Ohkusa T, Saito H, et al. Seasonal variations in the onset of ulcerative colitis in Japan. World J Gastroenterol
11. Aamodt G, Bengtson MB, Vatn MH. Can temperature explain the latitudinal gradient of ulcerative colitis? Cohort of Norway. BMC Public Health
12. Romberg-Camps MJ, Hesselink-van de Kruijs MA, Schouten LJ, et al. Inflammatory bowel disease in South Limburg (the Netherlands) 1991–2002: incidence, diagnostic delay, and seasonal variations in onset of symptoms. J Crohns Colitis
13. Aratari A, Papi C, Galletti B, et al. Seasonal variations in onset of symptoms in Crohn's disease. Dig Liver Dis
14. Moum B, Aadland E, Ekbom A, et al. Seasonal variations in the onset of ulcerative colitis. Gut
15. Dharmaraj R, Jaber A, Arora R, et al. Seasonal variations in onset and exacerbation of inflammatory bowel diseases in children. BMC Res Notes
16. Brandvayman Y, Rinawi F, Shamir R, et al. Associations of seasonal patterns and vitamin D levels with onset and flares of pediatric inflammatory bowel disease. Minerva Pediatr
2017; [Epub ahead of print].
17. Murphy L, Denis R, Ward CP, et al. Academic stress differentially influences perceived stress, salivary cortisol, and immunoglobulin-A in undergraduate students. Stress
18. Maes M, Song C, Lin A, et al. The effects of psychological stress on humans: increased production of pro-inflammatory cytokines and a Th1-like response in stress-induced anxiety. Cytokine
19. Elenkov IJ, Chrousos GP. Stress hormones, proinflammatory and antiinflammatory cytokines, and autoimmunity. Ann N Y Acad Sci
20. Webster Marketon JI, Glaser R. Stress hormones and immune function. Cell Immunol
21. Tian R, Hou G, Li D, et al. A possible change process of inflammatory cytokines in the prolonged chronic stress and its ultimate implications for health. ScientificWorldJournal
22. Kellermayer R. Challenges for epigenetic research in inflammatory bowel diseases. Epigenomics
23. Walters TD, Kim MO, Denson LA, et al. Increased effectiveness of early therapy with anti-tumor necrosis factor- (vs an immunomodulator in children with Crohn's disease. Gastroenterology