 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ARTICLE |
Departments of a Pediatrics
b Epidemiology and Biostatistics, University of California, San Francisco, California
c Department of Pediatrics, Comer Children's Hospital, University of Chicago, Chicago, Illinois
d Department of Pediatrics, Texas Children's Hospital, Baylor College of Medicine, Houston, Texas
e Department of Pediatrics, MassGeneral Hospital for Children, Boston, Massachusetts
f Department of Pediatrics, Children's Hospital of Philadelphia, Philadelphia, Pennsylvania
g Department of Pediatrics, Emory University School of Medicine and Children's Center for Digestive Healthcare, Children's Healthcare of Atlanta, Atlanta, Georgia
h Kaiser Permanente of Northern California, Oakland, California
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
METHODS. We conducted a retrospective cohort study of 989 consecutive pediatric patients (566 boys, 423 girls) who had Crohn disease (aged 0 to 17 years at diagnosis) by using the Pediatric IBD Consortium Registry. Uniform data were analyzed to compare the presentation and course of disease according to gender.
RESULTS. Median follow-up time was 2.8 years. Mean ± SD age at diagnosis of inflammatory bowel disease (11.5 ± 3.8 years) did not differ by gender. Compared with boys, girls had a higher prevalence of mouth sores at symptom onset and a higher prevalence of hypoalbuminemia at the time of diagnosis. Location of disease did not differ by gender. A higher proportion of girls had abnormal anti–outer membrane porin of Escherichia coli levels compared with boys. Girls were at increased risk for erythema nodosum/pyoderma gangrenosum and decreased risk for growth failure compared with boys.
CONCLUSIONS. Girls appear to have an overall more severe course of disease; however, boys are at increased risk for developing growth failure. Disease course and the impact of disease severity on growth according to gender in pediatric Crohn disease require prospective study.
Key Words: inflammatory bowel disease • growth failure • epidemiology • database • medications • complications • surgery • erythema nodosum • pyoderma gangrenosum
Abbreviations: CD—Crohn disease • IBD—inflammatory bowel disease • PCDAI—Pediatric Crohn's Disease Activity Index • ESR—erythrocyte sedimentation rate • OmpC—outer membrane porin of Escherichia coli • WBC—white blood cell • 5-ASA—5-aminosalicylic acid • OR—odds ratio • HR—hazard ratio • CI—confidence interval
Compared with male individuals, female individuals are reported to have a higher incidence of Crohn disease (CD). Variable reports on the magnitude of this increase exist in the literature. For example, incidences in female individuals have been reported at 6%,1 23%,2 and 45%3 higher than male individuals. In contrast, studies that were restricted to pediatric patients reported a male preponderance in several unique study populations.4–7 Consistent with these findings, Gollop et al8 reported an incidence of 0.5 per 100000 person-years in girls who were younger than 15 years and a higher incidence rate of 1.0 per 100000 person-years in boys who were younger than 15 years. This discrepancy is unexplained.9 Furthermore, a recent report documented a trend for a decline in the male/female ratio of CD as age at onset of disease increased.10
Gender differences in the course of CD are unclear. Studies have reported gender differences in mortality,11,12 the effect of smoking on colitis and need for immunosuppressant medications,13 the frequency of familial occurrence of inflammatory bowel disease (IBD),14 body composition,15,16 and risk for osteoporosis.17 We reported that female pediatric patients with CD are at increased risk for first intestinal resection compared with male patients.18
Identifying differences in the expression of disease by gender can enhance our understanding of the etiology, underlying genetic and pathophysiologic mechanisms, and the natural history of CD. We conducted a retrospective cohort study to examine the presentation and course of disease in male and female pediatric patients with CD.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Inclusion and Exclusion Criteria
A total of 1736 patients were enrolled in the Pediatric IBD Consortium Registry as of November 1, 2003 (data-retrieval date). We identified all patients who received a diagnosis of IBD before 18 years of age and had a final diagnosis of CD.18 Patients with a diagnosis of ulcerative colitis or indeterminate colitis as of November 1, 2003, were excluded from the study. Clinical history, physical examination, endoscopy, histology, and radiology were used to diagnose the type of IBD as previously described.4 Patients who were 
18 years of age at the time of diagnosis of IBD were excluded from the study.
Day 0 was defined as the date of initial diagnosis of IBD. Study end date for surgical patients was defined as date of first intestinal resection (partial small bowel resection, partial colectomy, or total colectomy). Study end date for nonsurgical patients was defined as date of the most recent clinic visit if this occurred before November 1, 2003, or November 1, 2003, when date of the latest clinic visit was after this data-retrieval date. Surgical patients were excluded from the analysis when study end date (for surgical patients) was on or before day 0 (date of initial diagnosis of IBD). A total of 989 pediatric patients with CD met these criteria.18
Description of Variables
Initial IBD diagnosis was defined as the initial classification of IBD (ie, CD, ulcerative colitis, or indeterminate colitis) in this cohort of children with a final diagnosis of CD.
Location of disease (determined at ±3 months from diagnosis date) was designated as (1) small bowel involvement without colon involvement, (2) small bowel and colon involvement, or (3) colon involvement without small bowel involvement. Disease location was established by abnormal histology or IBD-compatible radiology. Gastroesophageal involvement was not considered in this analysis. Granuloma on histology refers to granuloma reported on initial histology (±3 months from the time of diagnosis of disease).
Symptoms at onset of disease were patient and/or parent reported and included presence or absence of abdominal pain, diarrhea, fatigue, fever, joint pain, mouth sores, nausea, perianal disease, poor growth, rectal bleeding, dermatologic manifestations, vomiting, or weight loss. Severity of disease (determined ±1 week from diagnosis date) refers to severity of abdominal pain, diarrhea/hematochezia, limitation of activity, abdominal tenderness, perirectal disease, and extraintestinal manifestations. Severity of each of these parameters was rated according to criteria established by the Pediatric Crohn's Disease Activity Index (PCDAI)19 and then further dichotomized as normal or abnormal.
Laboratory parameters (obtained ±1 week from the date of diagnosis) included serum albumin level, hematocrit, erythrocyte sedimentation rate (ESR), white blood cell (WBC) count, and platelet count. Serum albumin levels, hematocrit, and ESR were each dichotomized into normal or abnormal. The criteria in the PCDAI19 were used to define normality for albumin, hematocrit, and ESR. WBC count and platelet count were dichotomized into elevated or not elevated. Elevated WBC count was defined according to age-established criteria (Riley Kidometer 2.2; http://rileychildrenshospital.com/physicians/practice-management/rk/kidometer.jsp). A platelet count of >450000/mL was categorized as elevated.
Serologic markers (anti–outer membrane porin of Escherichia coli [OmpC], anti–Saccharomyces cerevisiae antibody, and perinuclear antineutrophil cytoplasmic antibody) obtained any time before November 1, 2003, were coded as normal or abnormal according to established parameters (Prometheus Laboratories, San Diego, CA).
Medications included 6-mercaptopurine/azathioprine, methotrexate, thalidomide, cyclosporine, infliximab, 5-aminosalicylic acid (5-ASA; oral [mesalamine, sulfasalazine, balsalazide, or olsalazine], enema [mesalamine], or suppository [mesalamine]), corticosteroids (oral [budesonide, prednisone, or prednisolone], intravenous [methylprednisolone], or hydrocortisone foam or enema), antibiotics (clarithromycin, ciprofloxacin, metronidazole, ampicillin, or gentamicin), and tacrolimus. Patients were classified as exposed to a medication when the medication was started during the study period. Date of initiation of medication was recorded in the registry. Length of medication treatment was not considered.
Complications of disease were dichotomized into present or absent and included presence or absence of an abscess, aphthous stomatitis, arthritis, clubbing, compression fractures or osteopenia/osteoporosis, erythema nodosum or pyoderma gangrenosum, fistula, renal calculi, sclerosing cholangitis, stricture, perianal fissure, or growth failure. Growth failure was defined as height for age below the fifth percentile or a decrease in height velocity below the fifth percentile.
Statistical Methods
The Mann-Whitney rank-sum test was used to compare year of diagnosis and age at diagnosis by gender. Fisher's exact test was used to compare geographic region, race/ethnicity, initial classification of disease, disease location, and initial histology in boys and girls. Odds ratios (ORs) from univariate logistic regression models were used to assess differences by gender in prevalence of symptoms at disease onset, severity of symptoms and signs at diagnosis, abnormal laboratory values at diagnosis, and IBD serologic markers. Hazard ratios (HRs) from univariate Cox proportional hazards models were used to quantify risk for treatment with a particular medication or development of a specific complication by gender. A multivariate model incorporating gender, age at diagnosis, and the interaction between gender and age at diagnosis was created to quantify further the risk for the development of growth failure. Additional multivariate models of gender and (1) symptoms at diagnosis (P < .30 on univariate analysis of risk for development of growth failure) and (2) disease location at diagnosis were created to further quantify the risk for growth failure. Kaplan-Meier survival analysis and the log-rank test were used to determine the cumulative incidence of growth failure and compare time to growth failure distributions by gender. The log-rank P value cited is based on the entire length of follow-up.
Registry data were collected and stored in Access 2000 (Microsoft, Redmond, WA). Statistical analyses were performed in SAS 8.2 (SAS Institute, Cary, NC) and Stata 9 (Stata Corp, College Station, TX). Statistical significance was defined as a P value of 
.05.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
|
Initiation of Medication Treatment
The risk for initiation of treatment with cyclosporine in girls was 2.72 times the risk in boys (95% CI: 0.93–7.97; P = .07). The risk for initiation of treatment with 5-ASA in girls was 0.88 times the risk in boys (95% CI: 0.77–1.01; P = .08). No gender differences were found for risk for initiation of treatment with corticosteroids (P = .20), methotrexate (P = .43), 6-mercaptopurine/azathioprine (P = .69), infliximab (P = .73), tacrolimus (P = .74), antibiotics (P = .92), or thalidomide (P = .96).
Complications of Disease
The cumulative incidence of growth failure in our cohort was 9.8% at 10 years from the time of diagnosis of disease (95% CI: 6.1%–15.5%). The cumulative incidence of growth failure was less in girls than in boys (Fig 1). On univariate Cox modeling, the risk for development of growth failure in girls was 0.28 times the risk in boys (95% CI: 0.12–0.63; P = .002). In a multivariate model that included gender, age at diagnosis, and the interaction of age at diagnosis and gender, girls remained at decreased risk for the development of growth failure (HR: 0.28; 95% CI: 0.12–0.63; P = .002). An interaction between gender and age at diagnosis in relation to the development of growth failure was not present (P = .98). Age at diagnosis was not associated with risk for the development of growth failure (P = .85). In a multivariate model that included symptoms at presentation (P < .30 on univariate analysis of risk for development of growth failure [poor growth, vomiting, and weight loss]), girls remained at decreased risk for the development of growth failure (HR: 0.28; 95% CI: 0.12–0.62; P = .002). In a multivariate model that adjusted for disease location, girls remained at decreased risk for the development of growth failure (HR: 0.32; 95% CI: 0.12–0.84; P = .02).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
We also found that girls were at increased risk for treatment with cyclosporine (HR: 2.72; 95% CI: 0.93–7.97; P = .07) and decreased risk for treatment with 5-ASA (HR: 0.88; 95% CI: 0.77–1.01; P = .08). These results did not meet statistical significance because the 95% CIs crossed 1 in each case; however, interpretation of the point estimates, 95% CIs, and P values in the context of a sample size of 989 patients shows clinically important results. Higher risk for treatment with cyclosporine and lower risk for treatment with 5-ASA suggest a more severe disease course in girls. Although clinicians have different practices when using these agents, we have no reason to believe that they would select a medication on the basis of gender. Because cyclosporine is not a first-line agent, its use most likely reflects disease severity that is refractory to first-line medical management.
In contrast, boys were found on both univariate and multivariate modeling to be at increased risk for the development of growth failure compared with girls (Fig 1). These findings are in accordance with previous studies.20,21 Griffiths et al21 found that boys had less catch-up growth than girls during the course of disease. The cause of these observed gender differences in disease course is unclear and requires additional prospective investigation.
Disease Severity, Extraintestinal Manifestations, and Growth
The literature suggests that girls with CD have an overall more severe course of disease than boys; in contrast, body composition and growth seem to be more greatly affected in boys than in girls. Although Jess et al11 reported a trend for male gender to be associated with mortality in a population-based cohort of patients from Minnesota, the same author previously observed an increase in mortality that was most prominent in women who were younger than 50 years at the time of diagnosis from a population-based cohort in Denmark.12 Increased mortality was observed late in disease course and was attributed to death associated with severe CD.12 The observation that smoking hastened the onset of Crohn colitis and the need for immunosuppressant medications in women but not in men13 is additional evidence to support our conclusion that girls have a more severe course of disease.
Female gender has been associated with extraintestinal manifestations in patients with CD, specifically arthritis, erythema nodosum, and ocular and oral lesions.14,22,23 These extraintestinal manifestations are considered to be immunologically mediated and typically reflect intestinal disease activity related to active inflammation, lending additional support to our conclusion that girls have a more severe course of disease.14,24
Although studies that assessed body composition in CD yielded variable results, most studies suggested that male individuals are more affected than female individuals. Jahnsen et al16 reported that bone mineral content and lean body mass were significantly reduced in patients with CD and that body composition of men with CD was more strongly affected than that of women. Robinson et al17 reported that men with CD are at greater risk for developing osteoporosis than women. Geerling et al15 reported an altered body water distribution and body hydration in female patients, especially those with recently diagnosed CD; however, they found that male patients with long-standing CD had a significantly lower body weight, BMI, body fat, and percentage of body fat than in control subjects.
Growth impairment is a widely recognized complication of CD. The frequency of growth failure depends on its definition: commonly used definitions include height below the fifth percentile, a decrease in height velocity below the fifth percentile, or a fall off the child's growth curve. Decreased height-for-age has been reported in 39%, height less than the third percentile has been reported in 19%, and decreased linear growth velocity has been reported in 20%25 of patients. Kanof et al26 reported that decreased height velocity preceded the diagnosis of CD in 88% of their patients. Furthermore, 42% of their population had a reduction in height velocity before intestinal symptoms were noted. The cumulative incidence of growth failure in our cohort of pediatric patients with CD was 9.8% at 10 years from the time of diagnosis of disease (95% CI: 6.1%–15.5%). Comparison of these reports is difficult because of varying study populations, different definitions for compromised growth, and different methods of statistical analyses (eg, results reporting prevalence versus cumulative incidence).
Reports of the relative impact of disease severity on growth in patients with CD are contradictory. Some reports suggest that disease severity is the major determinant of growth in CD. One report21 implicated severity of gastrointestinal symptoms as the major factor influencing linear growth velocity during the first 2 years after diagnosis of disease in children with Tanner stages 1 and 2. Other studies5,27 have also found negative associations between linear growth and disease severity. In contrast, Sentongo et al28 reported that neither duration of disease nor average PCDAI differed by gender; however, significant differences in z scores for growth parameters in boys with CD compared with control subjects were observed, and these differences were not present in girls. The authors concluded that impaired growth in girls was more consistently associated with disease severity indicators than in boys. Taken together, our findings and current literature suggest that the impact of disease severity on growth differs by gender. Prospective longitudinal studies are needed to understand the impact of pubertal status and disease severity on growth by gender in children with CD.
Demographics: Age at Diagnosis and Race/Ethnicity
We did not find a gender difference for age at diagnosis of IBD, consistent with previous investigations.7,15 Furthermore, we did not find differences in the proportion of boys and girls by race/ethnicity. Cross et al23 did not find differences in the proportion of male and female patients in black compared with white patients in their study of patients with CD.
Initial Diagnosis, Location of Disease, Granuloma on Initial Histology, and Antibody Markers of IBD
Gender differences were not found by initial classification of disease or location of disease at the time of diagnosis. Wagtmans et al14 also reported no gender difference by disease location. In addition, we found no gender differences in the prevalence of granulomas on initial histology. Likewise, no gender differences were found in a report by Shaoul et al29 of 230 patients who had CD, 34% of whom had granulomas on biopsies obtained by colonoscopy and/or by surgical resection.
A higher proportion of girls were anti–OmpC positive compared with boys. Additional testing in a larger number of patients, incorporating disease activity indices and correlating these to absolute values of antibody titers and the number of positive antibodies by gender, is required to validate our findings and to define the clinical significance.
Limitations
Although difficult to assess accurately, disease duration before the date of diagnosis is an important variable to consider. Date of diagnosis of disease by a clinician was chosen as the starting time to maintain consistency among patients. Primary limitations of this study include those inherent to a retrospective cohort, secondary database analysis.18 The statistical models (Cox proportional hazards models and Kaplan-Meier survival analysis) take into account the follow-up time of each patient. Although our study is not population based, the nature of pediatric practice is that almost all pediatric patients with chronic disease including CD are followed by pediatric specialists in major regional medical centers such as those represented by the consortium, making our findings generalizable to most practices that care for children with IBD.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
We thank Joel Cutler for vision and leadership in raising awareness and support for issues that are relevant to children with IBD. The registry was supported by the indispensable efforts of a group of local study coordinators: Jennifer Cooper, Elizabeth Garnett, Catherine Geraci, Rachel Kreh, and Amy York. Advice and guidance pertaining to database development and management were provided by Traci Clemens, PhD (Emmes Corporation, Rockville, MD). We are grateful for local support, including generous support from the Wallace Family (Dr Winter), Jim Brooks (Dr Winter), the Barnett Family (Dr Kirschner), the Nathan Cummings Foundation (Dr Kirschner), John Fullerton and family (Dr Heyman), and the Marcus Foundation (Dr Cohen). We acknowledge the generous collaboration of the following associates who referred patients for our registry: Jeffrey Blumental, Tim Buie, Robert Cannon, Conrad Cole, Michael Durant, Mark Gilger, Ranjana Gokhale, Stefano Guandalini, Colleen Hadigan, Stephen Hardy, Jay Hochman, Alison Hoppin, Sandy Hwang, Esther Israel, Crain Jensen, Seiji Kitagawa, Ronald Kleinman, William Klish, Jeffrey Lewis, Larry Glen Lewis, Carlos Lifshitz, Petar Mamula, Marjorie McCracken, William Meyers, Kathleen Motil, William Mow, Anthony Olive, Dinesh Patel, David Piccoli, Edith Pilzer, Rene Romero, Philip Rosenthal, Gary Russell, Larry Saripkin, Bess Schoen, Robert Shulman, John Snyder, Gayathri Tenjalra, Ritu Verma, Xavier Villa, and Qian Yuan. Special thanks also go to Dr Michael Thaler (University of California, San Francisco) for invaluable advice in developing this manuscript.
|
FOOTNOTES |
|---|
Address correspondence to Melvin B. Heyman, MD, MPH, University of California, Department of Pediatrics, 500 Parnassus Ave, MU-4 East, San Francisco, CA 94143-0136. E-mail: mheyman{at}peds.ucsf.edu
The authors have indicated they have no financial relationships relevant to this article to disclose.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2007 American Academy of Pediatrics. All rights reserved. |
||
|
||
