Anemia is 1 of the clinical manifestations of the newly diagnosed patient with inflammatory bowel disease (IBD),1 and a drop in hemoglobin concentration often occurs with disease relapse.2 Nonetheless, the topic of anemia in pediatric IBD often receives less attention from the clinician in contemporary practice, than other extraintestinal disease complications, such as suboptimal nutritional status.3 It is likely that management of active disease takes precedence over other clinical concerns and complications of IBD, such as anemia. This is despite early evidence from adult studies suggesting enhancement in quality of life with improvement in hemoglobin concentration, independent of changes in disease activity.4
To date, the majority of evidence on the epidemiology of anemia in IBD patients comes from adult reports,5 as studies in children are sparse (Table 1). The majority of these previous pediatric studies were undertaken in small and selective patient groups, which may have introduced selection bias and preclude the extrapolation and generalization of the findings to the whole pediatric IBD population (Table 1). Furthermore, several of these studies were cross-sectional in design, having recruited treated children with longstanding IBD,6 which may have underestimated the true prevalence of anemia at IBD diagnosis7 (Table 1). Although most of the authors used hemoglobin concentration cutoffs to assess anemia, these varied markedly between studies (Table 1), and in several cases, the authors failed to justify their selection criteria or to cite an appropriate reference source.8–10 Because physiological levels of hemoglobin vary considerably with age and gender, especially during childhood,11 it is imperative to use only appropriate cutoffs.
Potential predictors or factors associated with anemia were addressed in very few of these studies (Table 1). Mack et al12 and, recently, Goodhand et al6 showed that increased disease activity was associated with anemia at diagnosis in pediatric IBD (Table 1). Other plausible predictors or factors associated with anemia, such as nutritional and growth status and delay in disease diagnosis, have not yet been addressed in any study apart from the one presented in an abstract in children with Crohn’s disease (CD)13 (Table 1).
Advances in diagnostic biomarkers14 and establishment of clinical guidelines recently have reduced delay in diagnosis and new treatment pathways are now available to promote both mucosal healing and in parallel provide nutritional support.15,16 Indeed, recent evidence suggests that a smaller proportion of IBD children now present as malnourished at diagnosis17 and nutritional status significantly improves at follow-up.18 Thus, a similar trend could be anticipated with reference to anemia at diagnosis and over the natural course of the disease.
Thus far, only 1 small, descriptive study7 has measured the prevalence of anemia at diagnosis and at follow-up, but the authors did not explore factors associated with anemia (Table 1). Identification of factors associated with anemia at diagnosis or over the course of the disease will allow the multidisciplinary team to identify early children with IBD at risk of anemia development and offer them prompt, preventive nutritional or medical intervention.
The aim of this study was 3-fold:
1. To measure the prevalence of anemia at diagnosis in a large representative population of children with IBD in the West of Scotland and to assess annual secular changes over a 6-year consecutive period.
2. To explore factors associated with anemia at diagnosis and after 1-year follow-up in a large cohort of contemporary children with IBD.
3. To explore the impact of exclusive enteral nutrition (EEN), the primary induction treatment for pediatric CD in United Kingdom, on hematological parameters and anemia prevalence in a cohort of newly diagnosed patients.
MATERIAL AND METHODS
Three cohorts of patients were included in this study. The first (cohort 1) encompassed all pediatric patients who were diagnosed at the Royal Hospital for Sick Children in Glasgow, United Kingdom, in the period between January 1, 2002, and December 31, 2007. As the Royal Hospital for Sick Children in Glasgow, United Kingdom, is the only tertiary children’s hospital in the West of Scotland (∼50% of Scottish population) with the only dedicated service responsible for diagnosis and management of all children with IBD in the region, our sample corresponded to the majority of the IBD population of the geographical region of the West of Scotland referred to pediatric services. In this cohort of IBD children, we calculated the prevalence of anemia at diagnosis and secular annual changes between 2002 and 2007. Any patient diagnosed before this period was excluded but was included in the second group (cohort 2) as described below.
The second group (cohort 2) included a convenient sample of all children with IBD (majority of patients from the first group [cohort 1] and others diagnosed before 2002) who were receiving active care at the IBD outpatient clinic in the period of the study and had data available at diagnosis and for a subgroup, with available data, up to 12 (±3) months follow-up. The aim was to explore factors associated with anemia at diagnosis and after 1 year follow-up.
The third group (cohort 3) included children with CD who were treated with EEN at the Royal Hospital for Sick Children between September 1998 and July 2009 and had information on hematological parameters recorded at diagnosis and during the course of treatment. In all cases, patients followed an 8-week course of EEN using a nutritionally complete feed as described previously.19,20 The feeds provided the U.K. recommended daily energy and nutrient requirements for age or above if the patients were undernourished. Compliance with EEN was checked by regular dietetic review. Clinical response to treatment (i.e., achieved clinical remission, improved, and failed) relied on the clinician’s medical reports. In this newly diagnosed cohort of CD children, we assessed the effect of EEN support on anemia prevalence and hematological parameters.
The data collected for each group are summarized in Table, Supplemental Digital Content 1, http://links.lww.com/IBD/A234. Data on hematological parameters were retrieved from the hospital electronic records. For the last 2 groups of patients (cohorts 2 and 3), information on anthropometry, prescribed medication, systemic inflammatory markers, and disease characteristics were jointly retrieved from the hospital electronic patient database and from review of their medical and dietetic notes (Table, Supplemental Digital Content 1, http://links.lww.com/IBD/A234).
For all cohorts, patients whose hemoglobin concentration was not available close to the date of diagnosis (±7 days) and those with other significant concomitant illness were excluded. Patients who received blood transfusion or were on oral iron supplementation during the follow-up were included, but their results are presented and discussed accordingly. All patients were diagnosed according to established clinical, endoscopic, histological, and radiological guidelines.21 Disease location and behavior was classified according to the Montreal classification.22 However, because absorption of iron takes place in the small intestine and to gain statistical power, we classified CD phenotype into those with disease affecting the upper gastrointestinal tract (based on Montreal classification) and compared it with other phenotypes. Similarly, ulcerative colitis (UC) was classed as extensive and limited colitis encompassing left side colitis and proctitis.
Data Handling and Statistical Analysis
For the definition of anemia, gender- and age-specific hemoglobin cutoffs were used.11 Prevalence of anemia was calculated as the ratio of the number of anemic patients over the total number of patients. In our analyses, we arbitrarily describe severe anemia as a hemoglobin value < 10 g/dL. Accordingly, patients with a hemoglobin concentration lower than their age and gender cutoff for the definition of anemia but above 10 g/dL were classified as mildly anemic. Although this is an arbitrary threshold to use, it allows the assessment of changes not only in anemia prevalence but also in severity during follow-up. Continuous variables are presented either with means and SD or with medians and interquartile range (IQR) depending on the distribution of the data. Differences between groups were assessed with 2-sample t test and analysis of variance for parametric variables and Mann–Whitney U test or Kruskal–Wallis for nonparametric data. Differences in categorical data were assessed with χ2or Fisher’s exact test for 2 × 2 tables with small counts (n ≤ 5) per cell. Correlations between variables were assessed with Spearman rank correlation.
Potential discriminants of anemia incidence and the extent of its severity (cohort 2) were defined a priori at disease diagnosis and follow-up. At diagnosis, these included disease type, upper gastrointestinal tract disease phenotype in CD, extensive colitis in UC, age at diagnosis, gender, systemic biomarkers of disease activity (erythrocyte sedimentation rate [ESR], C-reactive protein [CRP], and serum albumin), nutritional (body mass index [BMI] z scores and weight loss) and growth (height z score) status, patient’s self-report of blood present in stool and time to diagnosis (i.e., time from symptoms onset, based on parental/patient reports documented in the referral letter to the pediatric gastroenterologist, plus the time passed since then and until confirmed diagnosis). We were advised by our statistician not to perform statistical analysis using the individual Montreal classification subgroups as a predictor of anemia because of multiple groups, unbalanced numbers, and clustering of participants’ disease location toward 2 predominant groups. At 1-year follow-up (cohort 2), these included disease type and upper gastrointestinal tract disease phenotype in CD, extensive colitis in UC, gender, systemic biomarkers of disease activity, nutritional (BMI z scores) and growth (height z score) status indices and their change since diagnosis at follow-up, documentation of oral iron supplementation in the medical notes, blood transfusion, incidence of anemia, and hemoglobin concentration at disease diagnosis. CRP was measured using a turbidometric assay after binding to a specific antibody on an Architect analyzer.
Changes in hemoglobin concentration during EEN (cohort 3) were assessed with 1-sample Wilcoxon test for nonparametric data. Predictors of hemoglobin concentration change during EEN (e.g., weight z score increase and hemoglobin concentration at EEN initiation) were tested using univariate linear regression analysis (cohort 3). All variables with P value less than 0.1 were entered in a multivariate model, and their independent association with the response (hemoglobin change) was assessed using backward stepwise regression analysis. A model with P values less than 0.05 for all associates of anemia was considered the final model. MINITAB 16.2.2 (Coventry, United Kingdom) was used for statistical analysis.
Standard deviation scores (z scores) were calculated using the British reference data.23 A z score of less than –2 SD for height and BMI were consistent with the definition of short stature and thinness (undernutrition), respectively. Obesity was defined as a z score for BMI greater than 2 SD.
The study was registered with the local Research and Development office, which waived the need for full ethics application.
Prevalence of Anemia in an Incident Representative Population of Children With IBD (Cohort 1)
One hundred ninety-three patients were diagnosed between 2002 and 2007 corresponding to the majority of patients in the area of West of Scotland. For 9 children (4.6% of all), hematological parameters were not available close (±7 days) to the date of IBD diagnosis (n = 6) or were suffering from other significant comorbidities (n = 3) and were excluded. In total, 184 patients had measurements of hematological parameters at diagnosis and fulfilled the study inclusion criteria. The majority of participants were diagnosed with CD (n = 122; 66%); 51 (28%) suffered with UC; and 11 (6%) with IBD unclassified (IBDU). A higher proportion of children with CD were boys (61%), but in UC, there was an equal distribution of boys and girls (Table 2). The median age at diagnosis was 11.6 years (IQR: 3.7 years; minimum: 1.2 years; maximum: 17.3 years) with no significant differences (P = 0.150) between the disease types (Table 2).
Basic Hematological Profile and Prevalence of Anemia at IBD Diagnosis
The mean hemoglobin concentration at disease diagnosis was 10.9 (SD: 1.8) g/dL. One hundred thirty-two children (72%) presented with anemia at IBD diagnosis, of which 55 (30%) suffered from severe anemia (Table 2). There was no apparent change in the annual prevalence or extent of anemia severity in those patients diagnosed within the 6-year period between 2002 and 2007 (data available on request). The prevalence or severity of anemia was similar among the various types of IBD (severe anemia: CD versus UC versus IBDU, 29% versus 33% versus 27%; mild anemia: CD versus UC versus IBDU, 43% versus 36% versus 39%). Proportionally, more girls than boys were severely anemic (boys versus girls: severe anemia, 22% versus 40%; mild anemia: boys versus girls, 45% versus 38%; nonanemic: 33% versus 22%; P = 0.028), which was the case for all types of IBD (Table 2).
No patient had raised mean corpuscular volume (MCV > 100 fl). Three patients had MCV values greater than 90 fl. None of these 3 patients was severely anemic, 1 was mildly anemic (hemoglobin: 11.9 g/dL) and 2 had normal levels of hemoglobin. Seventy-two children (40%) had low MCV,11 and 135 (74%) had suboptimal levels (<27 pg) of mean corpuscular hemoglobin (MCH) consistent with hypochromic anemia. Microcytic anemia (suboptimal MCV) was more common in CD compared with UC (% suboptimal; CD versus UC: 47% versus 22%; P = 0.003) patients.
Associates of Anemia at Diagnosis and Follow-up (Cohort 2)
The medical case notes of 179 children who were receiving care during the study period (cohort 2) were available for review. Because of the small number of patients, the IBDU group (n = 9) was excluded from further comparative analysis. Newly diagnosed children with CD were older and a significantly higher proportion was thin (BMI z score ≤ −2 SD), presented short stature (height z score ≤ −2SD), and had abnormal inflammatory markers and suboptimal albumin concentration compared with the UC patients (Table 3). Similarly, CD children were more likely to report weight loss before disease diagnosis (CD versus UC: 77% versus 60%; P = 0.032), whereas those with UC reported more often the presence of overt blood in feces (CD versus UC: 49% versus 88%; P = 0.032) (Table 3). Seventy-seven percent patients with CD and 79% of those with UC were mildly (CD: 44%; UC: 46%) or severely anemic (CD: 33%; UC: 33%) at disease diagnosis (Table 3).
In children with either CD and UC, hemoglobin concentration was positively correlated with serum albumin concentration (Spearman rank correlation, UC: r = 0.59, P < 0.001; CD: r = 0.58, P < 0.001) and inversely with that of CRP (Spearman rank correlation; UC: r = −0.39, P < 0.001; CD: r = −0.45, P < 0.001) and ESR (Spearman rank correlation; UC: r = −0.38, P < 0.001; CD: r = −0.41, P < 0.001). Thus, a higher proportion of CD children with severe anemia had raised CRP (>7 mg/L) and suboptimal serum albumin (<35 g/L) compared with children who were not anemic (% raised CRP; severe anemia: n = 31 (89%) versus no anemia: n = 11 (48%); P = 0.002; % suboptimal serum albumin; severe anemia: n = 33 (97%) versus no anemia: n = 7 (29%); P < 0.0001) (Table 4). For children with UC, suboptimal levels of serum albumin were more common in children with severe anemia (Table 4). In children with CD but not UC, hemoglobin concentration was positively correlated with BMI z score (r = 0.26, P = 0.007) and time elapsed from symptoms onset to diagnosis (r = 0.33, P = 0.001). Indeed, anemic children with CD were younger and had a shorter diagnosis delay, and there was a higher proportion classified as thin compared with the nonanemic children, particularly those with severe anemia (Table 4). Female gender differentiated those patients who were severely from mildly anemic in children with CD (Table 4). Disease involvement at the upper part of the gastrointestinal tract was not associated with anemia or the extent of its severity in CD patients, whereas, in UC, extensive colitis differentiated children with severe anemia from nonanemic patients (Table 4). Descriptive statistics for the prevalence of anemia according to disease behavior and location (Montreal classification) is presented in Table, Supplemental Digital Content 2, http://links.lww.com/IBD/A235.
At 1-year Follow-up
One hundred thirty-nine of the UC and CD children mentioned above (82% cohort 2) had hematology measurements available between 9 and 15 months after diagnosis (median: 12.2 months; IQR: 1.9 months). For the remaining patients (n = 40), the duration of their follow-up data was too short (<9 months) to be included in the follow-up arm of this study. The majority of CD and UC patients had been treated with the 2 mainstream first-line treatments in United Kingdom, EEN and steroids, respectively (Table 5). Thiopurines had been used in 30% and 57% of the UC and CD patients, respectively (Table 5). Use of oral iron supplements was documented in the notes of 34% of the CD and 58% of the UC patients. Blood transfusion was administered in 5 (4%) patients since diagnosis (Table 5). Children with CD remained, on average, shorter and thinner than UC patients (Table 5). Twelve children (13%) with CD presented with short stature and only 1 was classed as thin at follow-up (Table 5). Although no UC patient was obese (BMI z score > 2 SD) at diagnosis, 5 of them (12%) were obese at 12 months follow-up. A higher proportion of CD than UC children had abnormal inflammatory markers (Table 5).
The proportion of patients with severe anemia decreased from 34% (n = 47) at disease diagnosis to 9% (n = 13), whereas the respective proportion of nonanemic patients was doubled at 12 months follow-up (n [%]; diagnosis: 27  versus 1-year follow-up: 54 ). This change was similar for both the diseases (Fig. 1).
Over half of the patients (n = 76; 68%) who were anemic (severe and mild) at IBD diagnosis (n = 112) remained so 1 year after diagnosis, whereas 9 of the 27 children who were not anemic at diagnosis were so at follow-up. Similar to the analysis at the time of disease diagnosis, disease activity, as reflected by systemic inflammatory markers, was the strongest covariate, which differentiated anemic from nonanemic patients at follow-up, particularly in children with CD (Table 6). A low hemoglobin concentration at diagnosis differentiated anemic (mildly and severely) from nonanemic CD patients at follow-up (Table 6). Similarly, a larger proportion of CD children who were severely anemic at follow-up were more likely to be severely anemic at diagnosis (Table 6). Change of BMI or height z score since disease diagnosis and usage of oral iron supplementation, as documented in the medical notes, did not differentiate patients who were anemic from those who were not for either of the diseases (Table 6).
Impact of EEN on Hematological Parameters and Anemia Prevalence in Children With Newly Diagnosed CD (Cohort 3)
Eighty-four children with de novo CD (girls: n = 35 [42%]; age, mean [SD]: 11.3 [2.3] years) started on EEN and had hemoglobin concentrations at treatment initiation and during the course of treatment. Of these, 47 had follow-up measurements at 4 weeks, 70 at the end of treatment, and 35 had measurements at all 3 time points of the study follow-up. When compared with EEN initiation, there was a statistically significant increase in the median hemoglobin concentration by 0.75 (g/dL) at the end of the 8-week treatment period (Table 7). Hemoglobin concentration did not increase significantly after 4 weeks on EEN, but a statistically significant increase, by a median of 0.35 g/dL, was observed between 4 and 8 weeks of treatment. Similar to the main cohort (cohort 1), prevalence of anemia at treatment initiation was 73%, of which 32% of the patients were classified as severely anemic (Table 7). At the end of EEN, prevalence of severe anemia was substantially decreased (P < 0.001) to 9% and that of mild anemia significantly increased to 57%. Similar changes were observed for the 35 CD children with a complete set of baseline and follow-up measurements (data available on request). Compared with treatment initiation, MCV and MCH values significantly improved in the middle and at the end of treatment but no statistical significant difference was found between the latter 2 time points (Table 7). No significant changes in erythrocyte counts were found during the course of the treatment. Weight significantly improved by a median of 2.5 (IQR: 3) and 4.3 (IQR: 4.2) kg at 4 and at 8 weeks on EEN, respectively (Table 7).
At EEN completion, 47 children (57%) achieved complete clinical remission, 26 (31%) improved significantly and 10 (12%) still had active disease. In patients who failed treatment, hemoglobin concentration did not increase significantly at any time point of follow-up, whereas, for those patients who improved or achieved clinical remission, hemoglobin concentration increased and the percentage of patients with severe anemia decreased at treatment completion (Table 8).
In univariate linear regression analysis, the concentration of hemoglobin at treatment initiation was the strongest predictor of its absolute change (R2 = 32%; P < 0.0001) at the end of EEN, followed by decrease in ESR levels (R2 = 32%; P < 0.001), increase in serum albumin (R2 = 26%; P < 0.001), and increase in weight z score (R2 = 11%; P = 0.004). Decrease in ESR levels and low hemoglobin concentration at EEN initiation were independently associated with improvement of hemoglobin concentration at the end of treatment, explaining 46% of the variation in multivariate analysis.
The current study aimed to assess the prevalence of anemia and potential associates, at disease diagnosis and after 1 year, in a large population-based cohort of children with IBD using age- and gender-specific hemoglobin cutoffs. To reduce selection bias, which may have occurred in previous studies (Table 1), we looked at all referrals over a defined period of time in the sole pediatric referral center in the West of Scotland. This is also mirrored by the demographic and disease characteristics of the population of the current study, which are similar to those of previous British epidemiological surveys,1,24 allowing the extrapolation of our results to at least the general pediatric IBD population in United Kingdom, North America,12,25 and similar populations.
At disease diagnosis, the prevalence of anemia was high with approximately three-quarter of the patients presenting with anemia and one-third of them suffering from severe anemia for all 3 cohorts. After 1-year follow-up, overall prevalence of anemia remained high but its severity changed with considerably less children suffering from severe anemia and more being mildly anemic. These findings are similar to 2 other recent reports,6,7 but, as these 2 studies did not differentiate between the different severity levels of anemia, it is impossible to comment on this aspect. Summarizing recent published evidence, it seems that anemia prevalence remains at the same levels described more than 2 decades ago despite significant advances in the diagnosis and management of IBD over the same time span (Table 1). Why this is the case is unknown and should be explored in prospective population cohorts. However, it may simply mean that the topic of anemia, particular mild anemia, receives low priority in the management of IBD patients and other clinical issues take precedence. The clinical importance of correction of anemia, particularly of mild severity, in children with IBD remains unknown, but, in previous studies, in adult IBD patients, this has been associated with poor quality of life,4 and in other conditions, resolution of anemia improved cognitive function.26
Severe anemia was more common in girls than boys, mainly in CD. It is unlikely that blood losses from menstruation in adolescent girls explain this finding as not only did we use gender- and age-specific hemoglobin cutoffs to define anemia but also we actually found that CD children with severe anemia were younger than children without or with mild severity of anemia.
Active disease, as reflected by the systemic inflammatory markers, was the strongest determinant of anemia and its severity at disease diagnosis and after 1 year of follow-up. Noteworthy in the cohort of patients on EEN is the improvement in hemoglobin concentration coincided with improvement of disease activity markers stronger than parallel nutritional rehabilitation, which suggests that, in active disease, optimal iron intake though EEN cannot compensate for gastrointestinal blood losses or the impact of disease activity on iron absorption and metabolism.27 In UC children where systemic inflammatory markers have poorer negative predictive validity, extensive colitis, a clinical index of disease severity, was a strong associate of severe anemia. Prevalence of anemia remained high at follow-up despite the high rate of immunosuppressants’ use and other disease-specific treatments, at follow-up, which suggests that additional anemia-specific treatment is needed for many patients.
Against our expectations we did not find an association between disease located in the upper gastrointestinal tract, where iron is absorbed, and anemia. These findings are in contrast to previous reports13,28 despite a similar prevalence of anemia among these studies. Impaired iron absorption might be 1 of the several causes of anemia in CD patients with involvement of the duodenum; however, in this and in a previous study, the prevalence of anemia was too high for this to be simply explained by iron malabsorption and the absence of a difference in anemia prevalence between CD and UC patients advocates that iron malabsorption is not the most important determinant of anemia onset in IBD. Moreover, a recent study in adults with mild and quiescent IBD found that iron absorption kinetics was similar to that of healthy individuals.29 Other factors such as blood loss from gastrointestinal bleeding and a diet poor in iron30 could equally contribute to anemia etiology. Unfortunately, because of the retrospective design of this study, no dietary assessment was available to address this possibility. An association between overt self-reported gastrointestinal bleeding and anemia incidence was not found although fecal occult blood cannot be excluded.
A high risk of undernutrition, as reflected by a history of recent weight loss and low BMI z score at disease diagnosis, differentiated anemic from nonanemic CD patients and correlated with hemoglobin concentration, but it is difficult to delineate whether this was because of poor nutrition, increased disease activity, or both. However, combining these results with those from the cohort of CD during EEN, it seems that resolution of active disease rather than nutritional rehabilitation is the strongest predictor of hemoglobin concentration improvement. Interestingly, hemoglobin concentration and anemia incidence improved only after 8 weeks on nutritional therapy and not at 4 weeks, which shows that, during the second half of the EEN course, improvement of hemoglobin concentration because of nutritional rehabilitation and iron repletion might be more probable.
Unexpectedly, an inverse association between anemia and diagnosis delay was found in this study. It is possible that early onset of anemia accelerates clinical investigations and hence prompts disease diagnosis, rather than a diagnosis delay predisposing to anemia as we had originally speculated.
Because of the retrospective design of this study, it was not possible to measure additional hematological parameters to differentiate between the different types of anemia at diagnosis and follow-up, and thus far, there is no consensus on the best way to do so.31 Based on the size of erythrocytes, no patient suffered from megalocytic anemia, and this suggests that the primary causes of anemia in IBD were iron-deficiency anemia or anemia of chronic disease as shown by previous studies.6,7,32
Although, in the current study, anemia prevalence remained high at follow-up, a significant decline in the extent of its severity was observed. Nevertheless, more than half of the patients were still anemic at 12 months postdiagnosis, showing that a considerable number of patients will remain mildly anemic after a year of diagnosis despite radical improvement in their nutritional status with none presenting as malnourished. These results might indicate that gross markers of nutritional status (e.g., BMI) are not good proxies of micronutrient (iron) status; that changes in the hematological parameters are more sensitive than anthropometry to indicate changes in disease activity; or that correction of mild anemia attracts less attention from the clinical team. Indeed, in this study, the proportion of patients who received iron supplementation was low compared with the total number of anemic children at follow-up. It might be that clinicians are hesitant to treat mild anemia in the light of evidence, mainly from animal studies, to suggest excessive production of free radicals and oxidative stress at the site of tissue lesion with possible implications for increased risk of gastrointestinal inflammation and exacerbation of symptoms, although the evidence from human studies is inconclusive.5 In addition, mild anemia, as a marker or subclinical ongoing inflammatory process, warrants further investigation, particularly now with the routine use of fecal calprotectin in clinical practice.
Although some patients who received blood transfusion were included, the latter was not associated with anemia improvement at follow-up. It is possible that any favorable effect of blood transfusion on hemoglobin levels is either short term or counteracted by the profound effect of active disease on anemia.
This study did not find that oral iron was associated with a better anemia status but the results should be interpreted with caution given that the study is retrospective and we did not have any indication of the time period the children had received oral iron supplementation. This may have been since diagnosis or later following treatment of active disease and persistence of anemia. In addition, none of our patients had been treated with intravenous iron, which may explain the high prevalence of anemia at follow-up. The subgroup follow-up analysis relied on availability of hemoglobin measurements at 12 months and, hence, excluded relatively newly diagnosed patients or those who were well and did not need blood sampling. This may have introduced a sample selection bias at the assessment of anemia prevalence at follow-up. However, the number of such patients was very small as the majority had follow-up measurements as part of the routine clinical monitoring of disease progression. Follow-up was limited to approximately 12 months, and further long-term follow-up may have revealed that hemoglobin concentrations continued to improve.
EEN is the mainstream treatment of active pediatric CD in United Kingdom and elsewhere in Europe, but its use in North America remains exceptionally low.33 Moreover, there is currently much controversy about the optimal duration of an EEN course. Improvement of hemoglobin concentration only after 8 weeks on EEN advocates the long-term use of nutritional therapy. At 4 weeks of treatment, there was no significant improvement in the baseline hemoglobin concentration or prevalence of anemia, which might be because of the compensation from gastrointestinal blood losses in ongoing active disease and despite adequate iron intake from the feeds. In contrast, during the second half of the EEN course, where there is usually resolution of clinical symptoms and less gastrointestinal losses, EEN may promote nutritional rehabilitation, replenish body iron stores, and improve hematological profile. That was reflected by the substantial reduction in the prevalence of anemia only after the second half of the treatment. In patients who failed treatment, hemoglobin concentration did not significantly improve at the end of EEN, despite adequate nutritional intake of iron though feeds, which indicates that gastrointestinal losses through bleeding, mucosal sloughing, and perhaps malabsorption are the most important determinants of anemia in active disease.
This study is not without its limitations. The retrospective nature of data collection did not allow the calculation of validated clinical disease activity indices, which might have been more appropriate to use. Until recently, these indices were not routinely calculated for patients in daily clinical practice, and so the data for these indices was available only for the minority of patients involved in this study. Calculation of these indices retrospectively is invalid. Moreover, hematocrit levels are an integral component of the pediatric CD activity index,34 so that its use to explore relationships with anemia would be troublesome and potentially misleading. Taking all these considerations into account, therefore, in this study, we have used individual components of these disease activity indices (e.g., recent weight loss, height faltering, albumin value, and the presence of blood in stool), which were individually reported in the analysis with the majority of these being strong predictors of anemia. In addition, response to EEN was evaluated based on the global clinical assessment rather than pediatric CD activity index, but the correlation between these 2 assessments is strong based on previous work.34
Despite advances in disease diagnosis and management of IBD, prevalence of anemia remains high at disease presentation and after 1 year follow-up and does not parallel improvement in other aspects of the nutritional status of the patient. Why this is happening is not clear from the results of this study. If this is entirely because of an inevitable association between disease activity and anemia, the primary focus of the clinical team should remain the best possible management of the active inflammatory response but without losing the treatment window to intervene and improve anemia status when this is possible in quiescent disease.
The authors would like to acknowledge Dr. David Young, Senior Lecturer in Biostatistics, for his advice on the statistical aspects of this article. K.G., R.K.R., C.A.E., and P.M. designed the study; K.G., A.B., A.P., D.M., E.B., T.C., and R.T. collected the data; K.G. carried out statistical analysis; K.G. drafted the article; C.A.E., R.K.R., and P.M. coordinated study; all authors reviewed and accepted the final submitted manuscript.
1. Sawczenko A, Sandhu BK. Presenting features of inflammatory bowel disease in Great Britain and Ireland. Arch Dis Child. 2003;88:995–1000.
2. Hyams JS, Mandel F, Ferry GD, et al.. Relationship of common laboratory parameters to the activity of Crohn’s disease in children. J Pediatr Gastroenterol Nutr. 1992;14:216–222.
3. Gerasimidis K, McGrogan P, Edwards CA. The aetiology and impact of malnutrition in paediatric inflammatory bowel disease. J Hum Nutr Diet. 2011;24:313–326.
4. Wells CW, Lewis S, Barton JR, et al.. Effects of changes in hemoglobin level on quality of life and cognitive function in inflammatory bowel disease patients. Inflamm Bowel Dis. 2006;12:123–130.
5. Kulnigg S, Gasche C. Systematic review: managing anaemia in Crohn’s disease. Aliment Pharmacol Ther. 2006;24:1507–1523.
6. Goodhand JR, Kamperidis N, Rao A, et al.. Prevalence and management of anemia in children, adolescents, and adults with inflammatory bowel disease. Inflamm Bowel Dis. 2012;18:513–519.
7. Wiskin AE, Fleming BJ, Wootton SA, et al.. Anaemia and iron deficiency in children with inflammatory bowel disease. J Crohns Colitis. 2012;6:687–691.
8. Howarth LJ, Wiskin AE, Griffiths DM, et al.. Outcome of childhood ulcerative colitis at 2 years. Acta Paediatr. 2007;96:1790–1793.
9. Khan K, Schwarzenberg SJ, Sharp H, et al.. Role of serology and routine laboratory tests in childhood inflammatory bowel disease. Inflamm Bowel Dis. 2002;8:325–329.
10. White JM, O’Connor S, Winter HS, et al.. Inflammatory bowel disease in African American children compared with other racial/ethnic groups in a multicenter registry. Clin Gastroenterol Hepatol. 2008;6:1361–1369.
11. Dallman PR, Siimes MA. Percentile curves for hemoglobin and red cell volume in infancy and childhood. J Pediatr. 1979;94:26–31.
12. Mack DR, Langton C, Markowitz J, et al.. Laboratory values for children with newly diagnosed inflammatory bowel disease. Pediatrics. 2007;119:1113–1119.
13. Thayu M, Leonard MB, Baldassano RN, et al.. Prevalence of anemia in incident pediatric Crohn disease. J Pediatr Gastroenterol Nutr. 2005;41:547–548.
14. Quail MA, Russell RK, Van Limbergen JE, et al.. Fecal calprotectin complements routine laboratory investigations in diagnosing childhood inflammatory bowel disease. Inflamm Bowel Dis. 2009;15:756–759.
15. Gerasimidis K, Nikolaou CK, Edwards CA, et al.. Serial fecal calprotectin changes in children with Crohn’s disease on treatment with exclusive enteral nutrition: associations with disease activity, treatment response, and prediction of a clinical relapse. J Clin Gastroenterol. 2011;45:234–239.
16. Otley AR, Russell RK, Day AS. Nutritional therapy for the treatment of pediatric Crohn’s disease. Expert Rev Clin Immunol. 2010;6:667–676.
17. Kugathasan S, Nebel J, Skelton JA, et al.. Body mass index in children with newly diagnosed inflammatory bowel disease: observations from two multicenter North American inception cohorts. J Pediatr. 2007;151:523–527.
18. Cameron FL, Gerasimidis K, Papangelou A, et al.. Clinical progress in the two years following a course of exclusive enteral nutrition in 109 paediatric patients with Crohn’s disease. Aliment Pharmacol Ther. 2013;37:622–629.
19. Buchanan E, Gaunt WW, Cardigan T, et al.. The use of exclusive enteral nutrition for induction of remission in children with Crohn’s disease demonstrates that disease phenotype does not influence clinical remission. Aliment Pharmacol Ther. 2009;30:501–507.
20. Gerasimidis K, Talwar D, Duncan A, et al.. Impact of exclusive enteral nutrition on body composition and circulating micronutrients in plasma and erythrocytes of children with active Crohn’s disease. Inflamm Bowel Dis. 2012;18:1672–1681.
21. Inflammatory bowel disease in children and adolescents: recommendations for diagnosis—the Porto criteria. J Pediatr Gastroenterol Nutr. 2005;41:1–7.
22. Silverberg MS, Satsangi J, Ahmad T, et al.. Toward an integrated clinical, molecular and serological classification of inflammatory bowel disease: Report of a Working Party of the 2005 Montreal World Congress of Gastroenterology. Can J Gastroenterol. 2005;19(suppl A):5–36.
23. Cole TJ, Freeman JV, Preece MA. Body mass index reference curves for the UK, 1990. Arch Dis Child. 1995;73:25–29.
24. Armitage E, Drummond HE, Wilson DC, et al.. Increasing incidence of both juvenile-onset Crohn’s disease and ulcerative colitis in Scotland. Eur J Gastroenterol Hepatol. 2001;13:1439–1447.
25. Heyman MB, Kirschner BS, Gold BD, et al.. Children with early-onset inflammatory bowel disease (IBD): analysis of a pediatric IBD consortium registry. J Pediatr. 2005;146:35–40.
26. Nissenson AR. Epoetin and cognitive function. Am J Kidney Dis. 1992;20:21–24.
27. Nemeth E, Tuttle MS, Powelson J, et al.. Hepcidin regulates cellular iron efflux by binding to ferroportin and inducing its internalization. Science. 2004;306:2090–2093.
28. Gryboski JD. Crohn’s disease in children 10 years old and younger: comparison with ulcerative colitis. J Pediatr Gastroenterol Nutr. 1994;18:174–182.
29. Lomer MC, Cook WB, Jan-Mohamed HJ, et al.. Iron requirements based upon iron absorption tests are poorly predicted by haematological indices in patients with inactive inflammatory bowel disease. Br J Nutr. 2012;107:1806–1811.
30. Lomer MC, Kodjabashia K, Hutchinson C, et al.. Intake of dietary iron is low in patients with Crohn’s disease: a case-control study. Br J Nutr. 2004;91:141–148.
31. Oustamanolakis P, Koutroubakis IE, Kouroumalis EA. Diagnosing anemia in inflammatory bowel disease: beyond the established markers. J Crohns Colitis. 2011;5:381–391.
32. Revel-Vilk S, Tamary H, Broide E, et al.. Serum transferrin receptor in children and adolescents with inflammatory bowel disease. Eur J Pediatr. 2000;159:585–589.
33. Levine A, Milo T, Buller H, et al.. Consensus and controversy in the management of pediatric Crohn disease: an international survey. J Pediatr Gastroenterol Nutr. 2003;36:464–469.
34. 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.
35. Thomas DW, Sinatra FR. Screening laboratory tests for Crohn's disease. West Med J. 1989;150:163–164.
© Crohn's & Colitis Foundation of America, Inc.