Growth failure is a common complication of Crohn disease (CD) and often precedes its diagnosis. In 1 study, 88% of 50 prepubertal children with CD were found to have decreased height velocity before diagnosis, often (in 42%) before intestinal symptoms were noted (1). Other studies report linear growth impairment at diagnosis in 36% to 65% of patients (2–4). A recent study of 123 children in England revealed a mean height deficit of −0.5 standard deviations (SDs) at diagnosis (5). Permanent growth failure also has been a concern. Deficits in ultimate adult height compared with predicted height have been reported in 19% to 37% of patients with CD diagnosed in childhood (5–7).
Various etiologies have been proposed to be the cause of poor growth in CD. These include poor energy intake, exposure to corticosteroids, and the growth inhibitory effects of the inflammatory process itself (8–12). Current therapy for CD includes frequent use of immunomodulators and biologics, modalities designed to minimize inflammation while decreasing the use of corticosteroids. Because many of the original descriptions of growth impairment in children with CD were in the preimmunomodulatory and prebiological therapy era, it is not known whether the widespread use of these agents improves growth outcomes. The aims of our study were to characterize linear growth in a contemporary population of newly diagnosed children with CD and to determine whether current therapies for CD have improved growth following diagnosis.
PATIENTS AND METHODS
In January 2002 the Pediatric Inflammatory Bowel Disease Collaborative Research Group established the Pediatric IBD Registry, an observational research program designed to examine and characterize clinical, laboratory, and humanistic outcomes associated with current and emerging treatments in newly diagnosed pediatric patients. All patient data for the present study were accessed from the registry. The centers participating in the registry prospectively record demographic, clinical, genetic, and laboratory information on newly diagnosed children and adolescents with IBD who have not yet reached their 16th birthday. A standardized data set is collected at the time of diagnosis and 30 days postdiagnosis, then quarterly thereafter, and submitted to a centralized data management center.
Severity of disease was defined by physician global assessment as inactive, mild, moderate, or severe. A Pediatric Crohn Disease Activity Index (PCDAI) (13) also was completed on each patient at diagnosis when component data were available. Standardized laboratory studies obtained on all patients at diagnosis included a complete blood count, erythrocyte sedimentation rate, and serum albumin.
To be eligible for this study, patients needed to have a diagnosis of Crohn disease and a Tanner stage of I to III for at least 12 months from the time of diagnosis. A total of 24 months of data were available for a subset of eligible subjects who remained Tanner I to III until that time. A further subset of individuals who remained Tanner I to II for the duration of the study also was examined.
All of the patients were managed according to the dictates of their physicians, not by standardized protocols. Immunomodulators and infliximab were used at the discretion of the attending clinician. No standardized corticosteroid dosing regimens were used, although in most patients corticosteroid therapy was initiated as oral prednisone or intravenous methylprednisolone at a dose of 1 to 2 mg · kg−1 · day−1. Tapering schedules were at the discretion of the primary clinician. A patient receiving more than 5 mg prednisone equivalent per day was considered to be receiving corticosteroid therapy. For analysis purposes, corticosteroid therapy in the first year was categorized according to duration as 3 months or less, more than 3 to 6 months, and greater than 6 months. Immunomodulators used included azathioprine, 6-mercaptopurine, and methotrexate. Although standard dosing regimens generally were used, 6-thioguanine levels were not uniformly monitored in all of the patients receiving azathioprine or 6-mercaptopurine.
Participation in the registry was approved by the institutional review board of each participating institution. Written informed consent was obtained from all parents or caregivers, and written assent was obtained from children when appropriate.
Height and weight z scores, as well as body mass index (BMI) percentiles based on the 2000 Centers for Disease Control growth charts, were determined at baseline, 12 months, and 24 months. Height velocity z score was determined at 12 and 24 months postenrollment, based on reference data from Tanner and Davies (14). Growth was considered to be impaired if height velocity z score or height z score was less than −1.0. Disease severity, laboratory values, and therapies were analyzed in relation to growth measures. Accurate height values 12 months before diagnosis were not available for most subjects; therefore, height velocity at diagnosis could not be calculated.
Descriptive statistics were calculated for all patient characteristics and outcome measures. Confidence intervals were estimated for all mean values at 5% error. PCDAI scores were compared for each of the physician global assessment levels using a 1-way analysis of variance. The primary outcome measures, height velocity z scores established at 12-month and 24-month follow-ups, were categorized as normal (≥−1.0) versus impaired (<−1.0). Paired comparisons of mean height z scores across study follow-up intervals (baseline to 12 months, 12–24 months) were evaluated with paired t test and with repeated-measures multivariate analysis of variance. Associations with height velocity were examined with analysis of variance (age, laboratory values) and logistic regression (sex, Tanner stage, disease extent, and disease severity). Multivariate logistic regression models were tested to evaluate the influence of subject characteristics at enrollment (age, sex, height, weight, disease extent, laboratory values, disease severity) on subsequent height velocity as established at 12-month follow-up. Therapeutic interventions and 12-month follow-up laboratory values and disease severity were added to the models to test the influence of therapy, adjusting for baseline effects. All α levels are 2-tailed. Statistics were calculated using JMP 5.1 (SAS Institute, Cary, NC).
We identified 176 children with CD who were Tanner I to III at diagnosis and at 1-year follow-up. The range of subjects recruited per site was 1 to 19, with a mean and median of 8.4 and 8.0 subjects per site. Two-year follow-up was available in 110 children, of whom 92 remained Tanner I to III at 24 months of observation (18 of 110 children progressed to Tanner IV at 24 months and were eliminated from growth analysis). Characteristics of the 1-year cohort are shown in Table 1. The mean age at diagnosis was 10.1 years ± 2.8 SD; 114 (65%) were males. Clinical and therapeutic status at diagnosis and 1 year are shown in Table 2. Disease severity by physician global assessment was mild in 33% and moderate or severe in 67% at diagnosis; at 1 year it was inactive or mild in 89% and moderate or severe in 11%. At 2 years (n = 92), 90% were inactive or mild, and 10% moderate or severe. At diagnosis, 105 children had complete PCDAI data available. Of these, moderate to severe disease (PCDAI >30) was found in 57% at diagnosis, 5% (n = 78) at 1 year, and 6% (n = 33) at 2 years. Analysis of variance showed an excellent correlation between physician global assessment and PCDAI values at both diagnosis and 12 months (P < 0.0001). Figure 1 depicts the frequency of cumulative use of various therapeutic interventions in the study period following diagnosis. Immunomodulators were used in 55%, 60%, and 86% of subjects by 6 months, 1 year, and 2 years following diagnosis. Infliximab was used in 10%, 19%, and 36% of subjects by 6 months, 1 year, and 2 years following diagnosis. There was no difference in medication usage between male and female patients.
Height z Scores
Figure 2 depicts height z scores at diagnosis, 1 year, and 2 years. The distribution of height z scores remained similar during the 2-year observation period. In tests of paired comparisons across follow-up intervals, no differences between means were statistically or clinically significant. The percentage of subjects with severely abnormal height z scores (<−2) declined somewhat, from 10% at diagnosis to 8% and 6.5% at 1 and 2 years, respectively. There was linear correlation between height and weight z scores at 1 year (r 2 = 0.48, P < 0.001, data not shown). For subjects who remained Tanner I to II during the first-year study period (n = 134), the mean height z score at 1 year was −0.45 ± 1.15. For those remaining Tanner I to II for 2 years (n = 58), the mean height score at 2 years was −0.40 ± 1.1. These values were similar to those noted for the entire study population.
Height Velocity z Scores
Figure 3 shows the percentage of subjects with height velocity z scores of less than −2, −1 to −2, −1 to less than 0, and 0 or higher at 1 and 2 years. The percentage of subjects with height velocity z scores less than −1 declined from 45% (±7%) at 1 year to 38% (±9%) among the group at 2 years, a decline that did not reach statistical significance. Similarly, fewer subjects at 2 years had severely abnormal height velocity z scores (<−2) than those at 1 year (23 ± 7.5% vs 29 ± 6.5%) (P = not significant). Mean height velocity z scores increased by almost 1 full standard deviation, on average 0.96, from −0.71 at 1 year to 0.26 at 2 years (P < 0.03, n = 92). For subjects who remained Tanner I to II, the mean height velocity z score was 0.50 ± 2.9 at 1 year and −0.06 ± 2.8 at 2 years.
Age at diagnosis of 9 years or older, moderate to severe disease (at 1 year), and disease extent (isolated small bowel disease) were associated with abnormal height velocity at 1 year following diagnosis (Fig. 4). Sex and the following parameters at diagnosis were not predictive of height velocity at 1 year: height z score, hemoglobin, erythrocyte sedimentation rate, and albumin. Age at diagnosis younger than 9 years was associated with a higher mean height velocity at 2 years compared with children 9 years or older at diagnosis (P < 0.02). All subjects with abnormal height velocity z score at 1 year who were younger than 9 years of age at diagnosis achieved normal height velocity z score by 2 years.
Effects of Treatment Modalities
Among therapeutic interventions in the initial 1-year follow-up period, only corticosteroid use was associated with height velocity outcome at 1 year. Subjects whose corticosteroid use extended for 6 months or longer were more likely to demonstrate abnormal height velocity z scores, compared with those with less than 4 months of use (including no use) or 4 to 6 months of use (76% vs 31% and 38%, respectively; P < 0.001). As shown in Figure 4, adjusting for age at enrollment, disease severity at 1 year and disease extent, subjects with 6 months or more of corticosteroid use were almost 5 times more likely to have abnormal growth velocity z scores than those with less than 6 months of exposure (adjusted odds ratio [OR] = 4.5; confidence interval [CI] = 2.2–9.6; P < 0.0002).
Examination of corticosteroid use during the 2-year study period among the 74 subjects receiving these drugs revealed 3 usage patterns that had differential effects on growth at 2 years. The patterns included early and sustained corticosteroid use lasting for 1 year or more (early sustained, n = 33); early use for 3 to 6 months, followed by less frequent use and reintroduction at about 1 year (early intermittent, n = 35); or corticosteroid therapy beginning at 1 year (late, n = 8). As Table 3 indicates, early corticosteroid use is associated with abnormal height velocity in the first year of follow-up. However, by 2 years, subjects whose corticosteroid use was sustained in the first year but then discontinued demonstrated improved mean height z scores and proportionately fewer abnormal height velocity z scores at 2 years, despite having similar height velocity z scores at enrollment.
We examined the relation of immunomodulator use and infliximab to growth outcomes in our population. Most subjects received immunomodulators during the 2-year study period (86%), with 49% beginning therapy in the first 3 months of enrollment and 60% by 1 year. Doses were 6-mercaptopurine, mean = 1.2 mg · kg−1 · day−1, median = 1.1 mg · kg−1 · day−1; and azathioprine, mean = 2.3 mg · kg−1 · day−1, median = 2.1 mg · kg−1 · day−1. Use of immunomodulators was not associated with improved growth velocity at 1 year or 2 years. Of all subjects, 36% (n = 64) received infliximab, including 34 who began infliximab therapy in the first year (15 in quarter 1, 3 in quarter 2, 16 in quarters 3 or 4). The height z scores at diagnosis, 1 year, and 2 years for all subjects receiving infliximab were −0.34, −0.45, and −0.37, respectively. Infliximab use had no effect on height velocity z score at 1 year.
Among subjects enrolled for 2 years, 25% had received infliximab for 1 year or more. Although a smaller proportion of these subjects (28%) displayed abnormal height velocity z scores at 2 years, compared with infliximab therapy less than 1 year (50%) and no infliximab use (37%), these differences were not statistically significant (Table 4). We also examined corticosteroid use in relation to infliximab use. Early sustained corticosteroid users received more frequent doses of infliximab, on average, than subjects in the early intermittent corticosteroid group, a trend that was consistent in all follow-up visits for year 1 and year 2. This effect was most pronounced at the 15-month follow-up visit, in which 33% were receiving infliximab compared with 9% in the early intermittent group (P < 0.02).
A total of 11 subjects received primary enteral therapy via nasogastric tube or gastrostomy during the first year. This therapy was initiated within 3 months of diagnosis in 7 (4%) and by 1 year in 10 (6%). Height z score for the 11 subjects was −0.36 (weight = −1.16) at diagnosis. By 1 year, the height z score only minimally improved to −0.32, with 30% of these children remaining under −1 SD.
The literature on growth abnormalities in children with Crohn disease largely dates from the 1980s and early 1990s, a period in which immunomodulators were rarely used and biological therapy had not begun. Our observations demonstrate that growth abnormalities continue to be present at diagnosis in a large number of children with Crohn disease, and that despite use of immunomodulators in the majority of patients, the abnormalities persist in many 1 to 2 years after diagnosis. Our large multicenter inception cohort of prospectively studied patients demonstrated a mean height z score of −0.5 SD at diagnosis, and the height z score of this population remained virtually unchanged during the 2-year period of observation. Profound growth failure (height z score <−2 SD) was present in 10% of the population at enrollment, and improved slightly during follow-up to 8% at 1 year and 6.5% at 2 years. The mean height z score of our patient population was identical to that recently reported from a smaller group of newly diagnosed children in England (5).
Many factors have been shown to play a role in the pathogenesis of growth abnormalities in children with Crohn disease, including chronic undernutrition, corticosteroid therapy, and the effects of the inflammatory process itself through the antagonistic effects on the growth process of circulating cytokines such as tumor necrosis factor-α (15–17). It is difficult to sort out the relative contributions of each factor because treatments often will affect more than 1 potential aspect of the growth process. What is concerning from our study is that despite disease severity markedly improving during the course of the study, our overall inception cohort did not demonstrate a significant improvement in height z score during the 2-year observation period.
Our data confirm previous observations that prolonged corticosteroid therapy is associated with poorer growth outcomes (6). Patients whose disease did not require corticosteroid therapy demonstrated the best growth. Patients with moderate to severe disease activity who had the slowest growth at diagnosis demonstrated good growth if they were fast responders to corticosteroid therapy and the medication was rapidly weaned. Patients with moderate to severe disease who were steroid dependent or steroid refractory demonstrated persistently poor linear growth.
Given that corticosteroid dependence is common in newly diagnosed children with Crohn disease (18), earlier use of immunomodulators and biological therapy is now common. In our subjects, 86% were receiving immunomodulators and 35% infliximab by 2 years following diagnosis. Patients receiving immunomodulators and infliximab had less corticosteroid exposure by 2 years, yet still exhibited subnormal growth. It is possible that the greater disease severity that prompted more aggressive therapy itself adversely affected the growth process. Walters et al (19) have reported that height velocity improves in patients with established Crohn disease who are treated with infliximab before puberty or early in their pubertal stages. We did find a numerical but not statistically significant improvement in the number of subjects with height velocity z scores less than −1 in those treated with infliximab for 1 year or greater compared with those treated for less than 1 year.
Primary therapy with enteral nutrition has been shown to improve growth parameters and treat Crohn disease with possible mucosal healing in children. Although we did not observe improved growth in our patients receiving enteral nutrition, the number of subjects receiving this therapy was too small to draw any conclusions.
We found that younger children (<9 years at diagnosis) appeared to have a greater improvement in height z scores than did older children. This may reflect a failure of some of our older children to enter their pubertal growth spurt on time. This would result in their z scores lagging age-matched reference peers whose height velocity would be increasing more at that time. Our study was further limited to using the subjects' chronological age and age-matched growth charts because we did not have bone age data available. We also found that height velocity z scores improved in the overall study population between 1 and 2 years (mostly accounted for by the growth in girls), although the aggregate height z score did not improve.
As noted above, the inflammatory response in Crohn disease plays a major role in growth failure. Interleukin-1 (20) is associated with anorexia in animal models of inflammatory bowel disease. Interleukin-6 suppresses growth and inhibits insulin-like growth factor-1 expression in rats with trinitrobenzene sulfuric acid–induced colitis (11). Tumor necrosis factor-α induces growth hormone insensitivity in rat models of colitis by downregulating the growth hormone receptor in the liver and impairing postreceptor signal transduction (12). Decreased synthesis of insulin-like growth factor-1 can follow and affect growth.
It also has been postulated that genetic polymorphisms may contribute to growth patterns in children with inflammatory bowel disease. Children with the −174 GG genotype (interleukin-6 gene polymorphism) were more likely to be growth retarded at diagnosis than those with the GC or CC genotype mediated by more severe disease activity (11). Polymorphisms of the tumor necrosis factor-α promoter gene (238 G/A, 863 C/A) have been shown to be associated with better height z scores and less growth retardation in children with Crohn disease independent of disease activity (21). The relation of the NOD2/CARD15 polymorphism to growth in pediatric patients with Crohn disease is not clear (22,23). An analysis of the influence of the OCT1/2 variants of the IBD5 locus revealed that the TC haplotype (OCTN 1 gene, SLC22A4 C/T, missense mutation, and OCTN2, SLC22A5-207G/C, promoter mutation) was associated with lower weight, height, and BMI at diagnosis (24). The present study did not include genetic analysis of our patients.
In summary, our study population of Tanner I to III children with Crohn disease had marked clinical improvement during the first 1 to 2 years of therapy, but growth abnormalities persisted in the aggregate, manifested as height z scores remaining at −0.5. It is noteworthy that height velocity significantly improved at 2 years in some subjects, and longer follow-up will be necessary to determine whether this trend will eventually result in overall improvement in final height attained. We found that prolonged corticosteroid use was a significant risk factor for poor growth, despite the fact that more than 80% of subjects received immunomodulators and more than one-third infliximab. Future treatment strategies further minimizing or even eliminating early corticosteroid use may be needed to improve growth outcomes.
The following research coordinators provided invaluable aid in the completion of this study: Miriam Lincoln, Sandra Hale, Rosa Negron, Gail Waltz, Kathy Grancher, Shari Huffman, Janet Trotta, Robin McLernon, Annette Langseder, Ruth Singleton, Tracey Williams, Rebecca Ehlert, Tracey Roiff, Ramona Bezold, Kelly Koslasky, Vivian Abadom, Myrna Miller, Melissa Metheney, and Sandra McRandal. Vicki Haviland-Wilhite coordinated the submission of the manuscript.
1. Kanof ME, Lake AM, Bayless TM. Decreased height velocity in children and adolescents before the diagnosis of Crohn's disease. Gastroenterology 1988; 95:1523–1527.
2. Kirschner BS. Growth and development in chronic inflammatory bowel disease. Acta Paediatr Scand Suppl 1990; 366:98–104.
3. Griffiths AM, Nguyen P, Smith C, et al
. Growth and clinical course of children with Crohn's disease. Gut 1993; 34:939–943.
4. Hildebrand H, Karlberg J, Kristiansson B. Longitudinal growth in children and adolescents with inflammatory bowel disease. J Pediatr Gastroenterol Nutr 1994; 18:165–173.
5. Sawczenko A, Ballinger AB, Savage MO, et al
. Clinical features affecting final adult height in patients with pediatric-onset Crohn's disease. Pediatrics 2006; 118:124–129.
6. Markowitz J, Grancher K, Rosa J, et al
. Growth failure in pediatric inflammatory bowel disease. J Pediatr Gastroenterol Nutr 1993; 16:373–380.
7. Sawczenko A, Ballinger AB, Croft NM, et al
. Adult height in patients with early onset of Crohn's disease. Gut 2003; 52:454–455.
8. Motil KJ, Grand RJ, Davis-Kraft L, et al
. Growth failure in children with inflammatory bowel disease: a prospective study. Gastroenterology 1993; 105:681–691.
9. Hyams JS, Moore RE, Leichtner AM, et al
. Relationship of type I procollagen to corticosteroid therapy in children with inflammatory bowel disease. J Pediatr 1988; 112:893–898.
10. Ballinger A. Fundamental mechanisms of growth failure in inflammatory bowel disease. Horm Res 2002; 58(Suppl 1):7–10.
11. Sawczenko A, Azooz O, Paraszczuk J, et al
. Intestinal inflammation–induced growth retardation acts through IL-6 in rats and depends on the -174 IL-6 G/C polymorphism in children. Proc Natl Acad Sci U S A 2005; 102:13260–13265.
12. DiFedele LM, He J, Bonkowski EL, et al
. Tumor necrosis factor α blockade restores growth hormone signaling in murine colitis. Gastroenterology 2005; 128:1278–1291.
13. Hyams JS, Ferry GD, Mandel FS, et al
. Development and validation of a pediatric Crohn's disease activity index. J Pediatr Gastroenterol Nutr 1991; 12:439–447.
14. Tanner JM, Davies PSW. Clinical longitudinal standards for height velocity for North American children. J Pediatr 1985; 107:317–329.
15. Zoli G, Katelaris PH, Garrow J, et al
. Increased energy expenditure in growing adolescents with Crohn disease. Dig Dis Sci 1996; 41:1754–1759.
16. Ballinger AB, Azooz A, El-Haj T, et al
. Growth failure occurs through a decrease in insulin-like growth factor 1 which is independent of undernutrition in a rat model of colitis. Gut 2000; 46:694–700.
17. Tracey KJ, Wei H, Manogue KR, et al
. Cachectin/tumor necrosis factor induces cachexia, anemia, and inflammation. J Exp Med 1988; 167:1211–1227.
18. Markowitz J, Grancher K, Kohn N, et al
. A multicenter trial of 6-mercaptopurine and prednisone in children with newly diagnosed Crohn's disease. Gastroenterology 2000; 119:895–902.
19. Walters T, Gilman A, Griffiths A. Linear growth improves during infliximab therapy in children with chronically active severe Crohn's disease. Inflamm Bowel Dis 2007; 13:424–430.
20. McHugh KJ, Collins SM, Weingarten HP. Central interleukin-1 receptors contribute to suppression of feeding after acute colitis in the rat. Am J Physiol 1994; 266(5 Pt 2):R1659–R1663.
21. Levine A, Shamir R, Wine E, et al
. TNF promoter polymorphisms and modulation of growth retardation and disease severity in pediatric Crohn's disease. Am J Gastroenterol 2005; 100:1598–1604.
22. Wine E, Reif SS, Leshinsky-Silver E, et al
. Pediatric Crohn's disease and growth retardation: the role of genotype, phenotype, and disease severity. Pediatrics 2004; 114:1281–1286.
23. Russell RK, Drummond HE, Nimmo EE, et al
. Genotype-phenotype analysis in childhood-onset Crohn's disease: NOD2/CARD15 variants consistently predict phenotypic characteristics of severe disease. Inflamm Bowel Dis 2005; 11:955–964.
24. Russell RK, Drummond HE, Nimmo ER, et al
. Analysis of the influence of OCTN1/2 variants within the IBD5 locus on disease susceptibility and growth indices in early onset inflammatory bowel disease. Gut 2006; 55:1114–1123.