Nutrition Assessment of Patients With Inflammatory Bowel Disease   ,
From the * Departments of Nutrition and Correspondence: Charles N. Bernstein, MD, Section of Gastroenterology, University of Manitoba, 804F-715 McDermot Avenue, Winnipeg, Manitoba, Canada R3E 3P4. Electronic mail may be sent to cbernst{at}cc.umanitoba.ca. Background: Malnutrition among inflammatory bowel disease (IBD) subjects is well documented in the literature and may arise from factors including inadequate dietary intake, malabsorption, and disease activity. The aims of this present study were to complete a comprehensive nutrition assessment of IBD subjects. Methods: One hundred twenty-six consecutive adults with IBD completed anthropometric measures, 4-day food-record assessments, and biochemical markers of nutrition. Results: A high prevalence of inadequate nutrient consumption was observed: vitamin E (63%), vitamin D (36%), vitamin A (26%), calcium (23%), folate (19%), iron (13%), and vitamin C (11%). Several biochemical deficiencies were also observed. The prevalence of subnormal serum levels was hemoglobin (40%), ferritin (39.2%), vitamin B6 (29%), carotene (23.4%), vitamin B12 (18.4%), vitamin D (17.6%), albumin (17.6%), and zinc (15.2%). Dietary intake was not correlated with serum levels in all instances; there was a highly significant correlation between diet and serum values of vitamin B12, folate, and vitamin B6 for all IBD subjects, independent of disease activity, and for vitamin D among all IBD subjects in remission. Conclusions: Subjects with IBD have a high rate of iron deficiency and anemia, which are most likely not secondary to diet. Supplementing with iron should be warranted only if a true iron deficiency exists. The routine evaluation of serum vitamin B6 and vitamin D levels is recommended. Routine multivitamin supplementation is warranted in IBD in view of numerous dietary and biochemical deficiencies observed among adult IBD subjects. Even if subjects with IBD seem to be well nourished, they may harbor vitamin/mineral deficiencies. Inflammatory bowel disease (IBD) is a chronic inflammatory disorder that involves the colon in ulcerative colitis (UC) and may involve any part of the gastrointestinal tract in Crohn's disease (CD). The disruption of the gastrointestinal tract by inflammation and the associated symptoms of pain, nausea, and diarrhea lead to reduced food intake, reduced nutrient use, and ultimately to impaired nutrition status in these patients. Decreased food intake, increased requirements, malabsorption and maldigestion, increased intestinal losses, disease activity, surgical resections, and medications are all potential factors in the etiology of malnutrition in IBD.1–5 The prevalence of nutrition deficiencies and malnutrition in IBD has been well documented, especially in CD.6,7 The presence of malnutrition has been linked to adverse outcomes. Knowledge regarding the extent to which decreased or inadequate food intake contributes to the nutrition deficiencies seen in IBD assists in determining if and how much nutrient supplementation is required. Quantifying nutrient intake in the context of the overall nutrition status assists in determining if patients with IBD are meeting recommended levels of nutrient intake, and to what extent subjects require supplementation. It is important to assess nutrition status comprehensively, and this includes measuring body composition, dietary intake, energy expenditure, and biochemical indices of nutrition. The aims of the present study were to examine the nutrient intake of individuals with CD and UC in relation to the recommended nutrient intake values, anthropometric measures, and biochemical indices of nutrition, and to determine the extent to which reduced food/nutrient intake contributes to the nutrition deficiencies seen in this population. Furthermore, we aimed to compare the nutrition intake of individuals with CD to UC and of those with active vs inactive disease to determine the primary drivers of nutrition deficiencies, either by quantified dietary and supplement intake or measured serum values.
Subjects Consecutive adult subjects with IBD were recruited from the outpatient gastroenterology clinic of a single investigator (C.N.B.) at the University of Manitoba, Health Sciences Centre (Winnipeg, Manitoba, Canada). The Research Ethics Board of the University of Manitoba approved the study protocol, and all subjects gave their informed written consent before the start of the study. Subjects also signed a medical chart access form allowing the research team to access personal medical charts for information related to the research study. Subjects were asked to complete a brief medical history questionnaire to extract information regarding (1) IBD type; (2) year of diagnosis; (3) site of intestinal tract affected by IBD; (4) history of fistula, stricturing, or surgery; and (5) use of medications. The subject-provided answers from the medical questionnaire were verified using chart reviews. The activity of the subject's IBD was determined by using the Harvey-Bradshaw index (for CD) and the Powell-Tuck Index (for UC).8,9
Biochemical Indices of Nutrition Status
Anthropometric Measurements
Dietary Intake Subjects returned for a follow-up appointment with the dietitian within 1 month of their blood work. This was arranged in order to best match dietary intake with biochemical indices of nutrition status. At the follow-up appointment, the 4-day food record was reviewed with each subject to ensure completeness. Portion sizes of foods consumed were confirmed at the follow-up visit. In addition, the food-frequency checklist was used as a cross-check for accuracy of the food records. Nutrient analysis was performed using the Food Smart Nutrition Management Solutions, version 5.0 (Sasquatch Software, Vancouver, British Columbia, Canada) nutrient analysis program. Additional nutrient composition information was obtained directly from food manufacturing companies for foods not available in the Food Smart program. Nutrition intake from supplements was added into the final analysis. Macronutrient intake was analyzed and compared with Canada's Food Guide to Healthy Eating. Inadequate intake of nutrients was defined as <66% of the recommended dietary reference intake (DRI).13,14 Total caloric intake was compared with each subject's resting energy expenditure (REE). Intake of micronutrients was compared with the DRI values for vitamin A, vitamin D, vitamin E, vitamin B6, vitamin B12, vitamin C, folate, niacin, thiamine, riboflavin, calcium, iron, and zinc.
REE
Statistical Analysis
Subject Characteristics A total of 126 subjects enrolled in the study between November 2002 and April 2005 and had blood work completed (84, CD; 42, UC). The higher number of CD vs UC subjects enrolled in our study is a reflection of the general subject population characteristics normally seen at our clinic. Subject characteristics are shown in Table I. Of the 126 subjects, 105 provided a 4-day food record and 103 completed the REE assessment.
Anthropometrics
CD VS UC. There was a trend toward a lower BMI in CD vs UC (p = .08). There was no significant difference in the percentage of IBW between CD and UC. TSF was significantly lower in CD vs UC (p = .01). However, there was no significant difference in midarm muscle measure between disease type (p = .21). DISEASE ACTIVITY. There was no significant difference in BMI, percentage of IBW, TSF, and midarm muscle measure by disease activity. INTESTINAL INVOLVEMENT. Among subjects with CD, those with small bowel involvement and small bowel surgery did not differ in BMI, percentage of IBW, TSF, and midarm muscle measure compared with those subjects without small bowel involvement or surgery. REE and respiratory quotient (RQ). For CD, the median REE was 1615 kcal (range, 1130–2450) and for UC, REE median was 1690 kcal (range, 1330–2790; p = .03 for difference between CD and UC). However, REE was not significantly different by disease activity (p = .29). Among subjects with CD, there was no difference in REE between those with or without small bowel involvement (p = .57) or between those with or without small bowel surgery (p = .71). The RQ was not significantly different between CD and UC (p = .84). For CD, the median for RQ was 0.81 (range, 0.65–1.1) and for UC, the median was 0.80 (range, 0.73–0.93; p = .84 for comparison between CD and UC). RQ was also not significantly different between active disease and IBD in remission (p = .43). Among subjects with CD, there was no difference in RQ between those with or without small bowel involvement (p = .37) or between those with or without small bowel surgery (p = .3).
Dietary Intake
Macronutrient intake. COMPARISON TO CANADA'S FOOD GUIDE. Table IV shows the percentage of IBD subjects that consumed inadequate portions of food items when compared with the recommendations of the Canada's Food Guide. A significantly higher percentage of subjects with CD consumed inadequate amounts of milk products compared with UC (p = .03). There was also a significantly higher percentage of CD subjects consuming inadequate amounts of meats and alternatives compared with UC (p = .02).
Protein, Fat, and Carbohydrate Fat. TOTAL FAT INTAKE. The median for the percentage of calories from fat consumed by all IBD subjects was 31.2% (range, 16.1–48.1). Fat consumption is considered excessive when >30% of total calories are consumed from total fat. There was no significant difference by disease type or disease activity (Table V).
SATURATED FAT INTAKE. The median for the percentage of calories consumed by saturated fat was 10.3% (range, 3.5–19.2). Excess saturated fat, defined as >10% of total fat calories, was consumed by 54.6% of all IBD subjects. Patients with UC consumed significantly more saturated fat compared with CD (p = .04). MONOUNSATURATED FAT. The median for the percentage of calories consumed from monounsaturated fat was 11.5% (range, 5.5–18.8). There was no significant difference by disease type or disease activity. POLYUNSATURATED FAT. The median for the percentage of calories consumed from polyunsaturated fat was 5.6% (range, 2.4–10.9). There was no significant difference by disease type or disease activity. Carbohydrate. The median for the percentage of calories consumed from total carbohydrate was 52.2% (range, 29.2–70.6). There was no significant difference by disease type or disease activity.
Biochemical Indices of Nutrition Status
We used serum ferritin as a measure of iron deficiency. We also assessed RBC mean corpuscular volume (MCV) and found that 20 subjects had a low MCV. Of these, 12 had subnormal serum ferritin levels (CD, 9; UC, 3; active disease, 5; remission, 7), whereas 8 had normal serum ferritin levels. Among the 8 subjects with normal serum ferritin and low MCV, 4 had active disease. CD vs UC. A sizeable percentage of all subjects with IBD was below the normal range for hemoglobin (40%), ferritin (39.2%), vitamin B6 (29%), carotene (23.4%), vitamin B12 (18.4%), vitamin D (17.6%), albumin (17.6%), and zinc (15.2%). The median serum vitamin B12, carotene, zinc, and B6 were all significantly lower in CD than UC (p < .05). More subjects with CD than UC had subnormal vitamin B12 and carotene levels. A significantly higher distribution of men with CD had hemoglobin below the normal range compared with men with UC (31.3% vs 6.3%, respectively; p = .05). Disease activity. Median serum levels of albumin, vitamin A, and vitamin B6 were all significantly lower in subjects with active disease than in subjects in remission (p < .05). More subjects with active disease than subjects in remission had subnormal serum levels of albumin and carotene. A significantly higher distribution of women with active IBD were below the normal range for hemoglobin compared with women with IBD in remission (41.2% vs 16.3%, respectively; p = .01). Intestinal involvement. Among subjects with CD, 66 (78.5%) had small bowel involvement. There was a significantly lower serum carotene level in patients with small bowel involvement compared to those without (1.3 µmol/L [range, 0.2–3.6] vs 2.1 µmol/L [range 0.6–3.8]; p = .006). There was a significantly lower serum vitamin B6 in patients with small bowel involvement compared to those without (4 µg/L [range, 2–47] vs 10 µg/L [range, 1–114]; p = .02). Among subjects with CD, 33 subjects (51.6%) with small bowel involvement were vitamin B6 deficient compared with 4 subjects (23.5%) without small bowel involvement (p = .005). There was a trend (p = .08) toward a higher percentage of subjects with small bowel involvement and carotene deficiency. Duration of disease. There was a significant difference in serum carotene among groups according to the duration of disease. Thirteen subjects with CD had disease duration >20 years (compared with no subjects with UC), and these subjects were significantly more likely to have a lower serum carotene (p = .03) than subjects with <20 years of disease. Medians and ranges for serum carotene are as follows: 0–5 years, 1.55 µmol/L (0.40–6.50); 6–10 years, 1.60 µmol/L (0.2–4.9); 11–20 years, 1.8 µmol/L (0.6–4.1); and >20 years, 0.70 µmol/L (0.4–3.1). The number and percentage of subjects below normal for each group were 0–5 years (n = 6; 14.3%), 6–10 years (n = 12; 28.6%), 11–20 years (n = 4; 13.8%), and >20 years (n = 7; 63.6%; p = .02). Use of medication. Eighty-five percent of the subjects were using medications (Table I) at the time of the study. Vitamin A levels were significantly higher among users of prednisone or azathioprine than those who used no medications or 5-aminosalicylic acid (5-ASA) only. Medians and ranges for vitamin A for all groups were as follows: no medications, 631.5 µg/L (414–1062); 5-ASA, 729 µg/L (325–1195); azathioprine, 869 µg/L (280–2096); prednisone, 860 µg/L (460–1538); and other medications, 895.5 µg/L (376–1527; p < .0002). There were no significant differences in the distribution of subjects below and within the normal ranges according to the type of medications used. History of bowel resections. Among subjects with CD, 38 had a small or large bowel resection or both. Subjects with large bowel surgery only had significantly lower levels of serum ferritin (8.5 µg/L; 95% confidence interval [CI], 2–41; p = .01) than those with either no surgery (37 µg/L; 95% CI, 6–173), small and large bowel surgery (30 µg/L; 95% CI, 9–212), and small bowel only (43 µg/L; 95% CI, 15–128). Subjects with small bowel surgery only had significantly lower levels of serum carotene (1.0 µmol/L; 95% CI, 0.2–2.6; p = .03) than those with no surgery (1.5 µmol/L; 95% CI, 0.4–3.8), small and large bowel (1.15 µmol/L; 95% CI, 0.5–1.8), and large bowel only (1.7 µmol/L; 95% CI, 0.40–3.10).
Vitamin D Deficiency The distribution of subjects between normal and below normal ranges for serum vitamin D was not different when using 35 and 50 nmol/L. However, when using 75 nmol/L, there were a significantly higher percentage of subjects with vitamin D deficiency among IBD subjects (52% patients in the fall/winter vs 20.8% patients in the spring/summer; p = .01).
Correlations
Multivariate analysis. A multivariate analysis was undertaken to control for gender, IBD type, IBD status, ileum involvement, medication use, surgical history, duration of disease, and use of vitamin/mineral supplementation. Further analysis was completed to control for serum vitamin B6, vitamin B12, carotene, ferritin, and zinc. Risk for vitamin D deficiency (<75 nmol/L) tended to be higher in UC (odds ratio [OR], 3.40; 95% CI, 0.86–13.5). The use of medication tended to lower the risk of vitamin D deficiency; use of 5-ASA (OR, 0.13; 95% CI, 0.014–1.23; p = .08), use of prednisone (OR, 0.08; 95% CI, 0.008–0.860), use of an immunomodulator (OR, 0.14; 95% CI, 0.015–1.2). If serum vitamin B6 was low, there was a higher risk for vitamin D deficiency (OR, 4.69; 95% CI, 1.22–17.94; p = .02).
We set out in this study to complete a comprehensive nutrition investigation of subjects with IBD. To our knowledge, our study is the first to use various anthropometric and biochemical parameters, as well as to assess energy expenditure and a complete diet analysis in IBD subjects stratified for disease type and disease activity. We reported that by using BMI and percentage of IBW (anthropometric measures alone), our IBD subjects were in fact within a healthy weight range, with no significant differences between active IBD and IBD in remission. According to these data, one could conclude that IBD subjects present to an outpatient specialty clinic, generally, as well nourished. Results on weight and BMI among IBD subjects vary in the literature. Some studies are in agreement to our observations in that BMI of both CD and UC subjects are within a healthy BMI range, whereas other studies report weights below the ideal range, significant reductions in muscle and fat stores, and the existence of a considerable degree of undernutrition even among CD who are not acutely ill.12–14 Our data may reflect a modern era of care with greater attention to nutrition, and hence most patients are in fact presenting as well nourished. A key finding from our work is that our IBD patients appeared well nourished according to anthropometric measures; however, several micronutrient deficiencies were identified, namely, iron, vitamin B6, and vitamin D deficiencies. These micronutrient deficiencies were not necessarily secondary to an inadequate diet.
Iron Deficiency and Anemia Blood loss is associated with gastrointestinal diseases such as IBD and is considered one of the leading causes of iron deficiency.16 According to our results, we assume that the iron deficiency and anemia observed are most likely secondary to blood loss associated with IBD. Our observation of iron deficiency and anemia reinforces the importance of identifying the cause of anemia before initiating therapy, especially among IBD subjects where ferritin, an acute-phase protein, may be elevated in the presence of inflammation. We should acknowledge, however, that our data provide minimal estimates of iron deficiency because only serum ferritin and hemoglobin measures were used. In view of this, our results may be underestimating the prevalence of iron deficiency, and more extensive tests to measure anemia such as erythrocyte zinc protoporphyrin or transferrin receptor would be beneficial. Interestingly when the MCV data were analyzed, 20 patients (16%) showed microcytosis, of which 12 (10%) had both low MCV and low serum ferritin. Of the additional 8 subjects with low MCV, but having normal serum ferritin, 4 had active disease. Therefore, a possible underestimation of iron deficiency among our IBD subjects seems to be similar among those with active disease and those in remission.
Serum Vitamin B6 Deficiency In our study, serum measures of vitamin B6 were positively correlated to diet, initially leading us to believe that a low oral intake of dietary/supplemental vitamin B6 may be contributing to the low vitamin B6 levels observed. However, only 6% of our subjects consumed inadequate amounts of vitamin B6, and so the role of reduced dietary intake is unlikely. Furthermore, malabsorption is most likely not a pathogenic factor in low serum vitamin B6, because subnormal values were observed in both CD and UC and there was no significant difference between disease types. At this stage, it is difficult to conclude why vitamin B6 levels were subnormal among our IBD subjects, although it is not obvious that disease activity, dietary deficiency, and malabsorption were likely contributors. It has been suggested that an increased consumption in the synthetic pathway of inflammatory mediators may contribute to the subnormal serum vitamin B6 observed among IBD patients.20 Interestingly, in rheumatoid arthritis, a disease of chronic relapsing inflammation, vitamin B6 levels may be influenced by inflammation.21 Therefore, inflammation may in part explain the subnormal vitamin B6 levels observed IBD. The presence of vitamin B6 deficiency is of concern as it can contribute to hyperhomocysteinemia, an independent risk factor for both arterial and venous thrombosis.20 Among IBD patients, a low B6 level has been associated with hyperhomocysteinemia.20 Although the etiology of hyperhomocysteinemia is considered to be multifactorial (including genetic, nutrition, and life-style factors), a deficiency of 1 or more B vitamins (folate, vitamin B12, and vitamin B6) plays a role.22 Interestingly, only a small amount of homocysteine is cleared via the metabolic pathway dependent on B12 and folate, and most of the homocysteine is converted to cystathionine by a vitamin B6–dependent enzyme.23 Serum homocysteine levels (and methyl malonic acid levels) were not measured in our study but may be valuable parameters to measure in IBD subjects with subnormal vitamin B6 in future work.
Vitamin D Deficiency There is substantial evidence that suboptimal vitamin D status is common in IBD patients, particularly in CD.25–27 Our study confirms these previous reports. Our results also support previous results that compare CD and UC in that we also did not observe a significant difference in serum 25-hydroxyvitamin D between CD and UC, regardless of the cutoff value used.28,29 As with other serum nutrition deficiencies observed in our study, there may be several contributing factors to the etiology of vitamin D deficiency. Vitamin D deficiency may arise from malabsorption, reduced intake, or inadequate exposure to sunlight.27 Among patients with CD, vitamin D deficiency may also arise secondary to small intestinal resection, which may be a result of fat-soluble vitamin malabsorption due to steatorrhea.30 Although malabsorption is thought to be a contributor to vitamin D deficiency,31 our results showed similar prevalence rates between CD and UC. We would have expected a higher prevalence of vitamin D deficiency among CD if malabsorption were the most likely contribution factor. Furthermore, when a cutoff of 75 nmol/L was applied, even a larger percentage of UC subjects (79%) were vitamin D deficient vs CD subjects (69%; not significant). Vitamin D status is highly dependent on exposure to UVB sunlight exposure27 and has been cited in the literature as being strongly influenced by geographical and seasonal differences.24,25 Interestingly, seasonal variations in the onset and exacerbation of IBD have been detected; December seems to be a peak time for the symptomatic onset of UC in the northern hemisphere, and higher relapse rates of CD have been observed in the fall and winter.24 In view of this, we further analyzed our data to see if season affected our results. We observed a higher percentage of patients with vitamin D deficiency among patients recruited during the fall and winter months only when 75 nmol/L was used as a cutoff. It would be interesting to pursue this question and study the onset or relapses of IBD in conjunction with season and vitamin D levels. We also observed a high prevalence of inadequate dietary vitamin D intake among our total IBD group (36%), supporting previous work reported at our center.11 This may be due to an inadequate intake of milk products, which may be secondary to a common complaint of lactose intolerance among this population group. It is unclear from our study whether this was secondary to a change in diet due to the disease itself or an already existing dietary habit before diagnosis. Interestingly, a significant positive correlation between dietary vitamin D intake and serum values was observed only among CD subjects and IBD subjects in remission. Therefore, the role of dietary vitamin D cannot entirely explain the vitamin D deficiency observed among our IBD subjects. Nevertheless, our results may warrant the routine recommendation of vitamin D supplementation in IBD.
Patients with IBD presenting to a clinic may seem well nourished but micronutrient deficiencies may still exist. It has been reported that patients with IBD have a high rate of iron deficiency and anemia. According to our results, iron deficiency is most likely not secondary to diet, and therefore, supplementing with iron should be warranted only if a true iron deficiency exists; the true cause of anemia should be determined before initiating therapy. Our results suggest the need for routine measurement of serum vitamin B6 (and possibly homocysteine) levels among IBD. Furthermore, our study supports that vitamin D supplementation is warranted in IBD, whereas a prospective study in high-risk individuals assessing the role of vitamin D supplementation and vitamin D status on IBD diagnosis or relapse rates would be of interest to pursue. Multivitamin supplementation is warranted in IBD because several nutrition deficiencies are observed via biochemical and dietary markers, regardless of disease type or disease activity. The study was funded by a grant from the Innovations and Opportunities Fund, Health Sciences Centre, Winnipeg, Manitoba. The authors thank Dr Donald Duerksen and Dr Khursheed Jeejeebhoy for their review of this manuscript and their helpful comments. Received for publication June 6, 2006. Accepted for publication February 23, 2007.
Journal of Parenteral and Enteral Nutrition, Vol. 31, No. 4,
311-319 (2007) This article has been cited by other articles:
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Top
MATERIALS AND METHODS
RESULTS
2 test was used to analyze
proportions of subjects above or below normal limits for serum or diet
parameters. A Spearman correlation test was performed to analyze the
association between the nutrient intake (nutrient intake per 1000 kcal) and
the corresponding serum value. Significance level was set at 
= .05.
Multiple logistic regressions and multiple linear regressions were used to
model the effects of all the groups adjusted for one another, use of vitamins,
and other possible confounders. 
