The aetiology and impact of malnutrition in paediatric inflammatory bowel disease

Introduction

Crohn’s disease (CD) and ulcerative colitis (UC), collectively known as inflammatory bowel disease (IBD), are idiopathic, lifelong, inflammatory, destructive conditions of the gastrointestinal (GI) tract (Ghosh et al., 2000; Shanahan, 2002). Crohn’s disease and UC can be diagnosed at any age, with the juvenile onset of the disease occurring with an average incidence of 5.2 per 10⁵ cases per year in the UK and Republic of Ireland (Sawczenko et al., 2001).

Nutrition has been implicated in all aspects of IBD, from disease pathogenesis (D’Souza et al., 2008; Tjonneland et al., 2009), diagnosis (Sawczenko & Sandhu, 2003) and treatment (Buchanan et al., 2009) to prognosis (Lopes et al., 2008). Intervention through dietary counselling and enteral or parenteral nutrition may also become necessary at some point to reconstitute nutritional balance or to induce clinical remission in active CD.

There are several parameters that could perturb nutritional status in IBD patients (Kleinman et al., 2004). In particular, suboptimal intake, as a result of anorexia and food aversions, in combination with increased energy and nutrient losses, as a result of malabsorption and gut losses, can disturb the nutrient equilibrium and increase risk of undernutrition (Fig. 1). Moreover increased nutrient utilisation and altered metabolism, as a result of activation of the immune system, tissue repair and drug–nutrient interactions, may increase nutritional requirements well above those established for the general population (Fig. 1).

Indeed, undernutrition is prevalent in IBD and presents in its different forms (Fig. 1). Protein-energy undernutrition, altered body composition, micronutrient deficiencies and poor bone health have been frequently seen in IBD patients. Additionally, in paediatric IBD patients, growth failure and pubertal delay have been associated with poor nutritional status (Fig. 1). This review provides a critical summary of the current knowledge and evidence of the incidence and aetiology of undernutrition in paediatric patients with IBD. It presents areas for further research and translates current evidence into clinical practice.

Protein-energy undernutrition

Many patients with IBD present features of protein-energy undernutrition at the time of diagnosis, and their status fluctuates during the disease course (Razack & Seidner, 2007; Hengstermann et al., 2008). A history of weight loss is a common presenting feature of the newly-diagnosed patient and accompanies almost every relapse (Weinstein et al., 2003; Kugathasan et al., 2007). This was seen in approximately 60% of newly-diagnosed CD children compared to 35% in UC patients (Sawczenko & Sandhu, 2003; Weinstein et al., 2003).

Children with IBD have abnormal anthropometry and, particularly those with CD, present low body mass index (BMI) and are underweight and thin compared to the national reference range or their healthy peers (Weinstein et al., 2003; Paerregaard & Uldall, 2005; Kugathasan et al., 2007). Apart from the higher incidence in CD, there is no consistent evidence to link specific disease characteristics (e.g. disease location, duration of symptoms before confirmed diagnosis) with being underweight.

Interestingly, recent data from adult and paediatric studies suggested that fewer patients are now seen with undernutrition compared to previous studies and, in fact, a large proportion of patients are overweight or obese at diagnosis, particularly UC patients. In a North American study of 783 newly-diagnosed IBD children, low BMI was seen in 22–24% of children with CD, and 7–9% of children with UC. By contrast 10% of the CD children and 20–30% of the UC patients had high BMI consistent with the definition of being overweight per obesity (Kugathasan et al., 2007). The present obesity epidemic in the general population and earlier disease recognition of IBD may explain these secular changes in patterns.

There is limited evidence on the progression of undernutrition after diagnosis. Unpublished data from a large sample of IBD children (n = 175) in our centre showed that 29% of the CD patients, 4% in UC and 11% in indeterminate colitis presented with BMI z-scores <−2 SD at diagnosis, whereas, at 6 and 12 months post-diagnosis, 5% and 1% of the CD children and 0% and 5% of UC patients had respective low BMI z-score values. Although none of the patients was obese (BMI z-score >2 SD) at diagnosis, at 6 and 12 months follow-up, over 12% of the UC patients and 3% of CD children had a BMI z-score consistent with the definition of obesity (BMI z-score >2). These results are in agreement with those reported by a French group (Vasseur et al., 2010) who recently showed that weight and BMI z-scores improved post-diagnosis and the prevalence of CD children with low BMI (<−2 SD) was half of that at diagnosis. Studies assessing protein-energy malnutrition in paediatric IBD populations using more robust techniques (e.g. nitrogen balance techniques) do not appear to be available.

Body composition

There are several reasons why body composition in IBD patients may differ from that of healthy individuals. The secretion of pro-inflammatory cytokines may alter energy metabolism, protein turnover and energy substrate utilisation, whereas the use of corticosteroids increased body fat with catabolic effects on lean mass (Ma et al., 2003). Physical activity, on the other hand, was reported as being low in adult IBD patients and correlated inversely with fat mass (FM) in one recent report (Sousa et al., 2007), although evidence in paediatric patients is scarce. Whether there is an association between body composition and physical activity patterns in paediatric IBD remains to be explored.

There are few studies that have assessed body composition in IBD children. Body lean mass has been consistently reported as being significantly lower than healthy control groups (Boot et al., 1998; Sentongo et al., 2000; Burnham et al., 2005; Thayu et al., 2007), whereas gender-specific associations with FM were observed in some studies (Sentongo et al., 2000; Burnham et al., 2005; Thayu et al., 2007). In a cross-sectional study, Sentongo et al. (2000) found that boys and girls with longstanding CD presented lower fat free mass (FFM) by an average of 3.5 and 2.9 kg, respectively, compared to a control sample, whereas FM, although not different from the control group in CD boys, was 3.7 kg higher in CD girls. Similarly, in a cross-sectional study, Burnham et al. (2005) found that children with longstanding CD had a 6% deficit in lean mass compared to healthy children after accounting for stature, age, pubertal stage and race; however, FM was normal, suggesting nutritional cachexia. Recently, Thayu et al. (2007) in a well-designed study of newly-diagnosed children with CD, also reported gender associated differences with body composition. Fat mass and lean mass for height (adjusted for age, race and pubertal stage) were lower in female than in male patients. Compared to a cohort of healthy controls, body composition in girls was more consistent with wasting (low lean and FM), whereas, in boys, there was mostly preservation of FM and deficits in lean mass consistent with cachexia. No associations were observed between body composition, clinical activity, disease location or diagnosis delay. Interestingly, normalisation of BMI at 2 years of follow-up was not associated with a significant increment in FFM in children with CD (Sylvester et al., 2009), which implies that changes in body weight or BMI for age are not good proxies for body composition changes in IBD and the introduction of simple bedside techniques of body composition assessment, such as bioelectrical impedance analysis, for routine clinical use is required.

Nevertheless, interpretation of body composition data in disease has to be approached with caution because the underlying assumptions about the composition of body compartments may be invalid (Wells & Fewtrell, 2006). Most in vivo body composition methods used in previous IBD studies [e.g. dual energy X-ray absorptiometry (DXA)] have been tested and validated in healthy individuals or animal cadavers and their applicability in chronic illness is questionable given the changes that may occur in the hydration level and distribution of fluids within the body compartments (Williams et al., 2006). These errors will affect lean mass composition estimation because of the assumptions made during the DXA calculations. Assessment of the validity of these techniques in an IBD population and replication of these results with the application of more sophisticated methods needs to be explored (Reilly et al., 2010). At the same time, the validity and reliability of simple bedside methods of body composition more suitable for routine clinical use should be evaluated in paediatric IBD population. Development of IBD-specific prediction equations could improve the validity and accuracy of bioelectrical impedance analysis techniques, (Dung et al., 2007). The use of functional tests (e.g. handgrip strength) has been proposed as a proxy estimate of FFM in adult IBD patients, although these techniques lack specificity. Moreover, Wiroth et al. (2005) found that adult patients with CD in clinical remission have overall lower muscle performance than healthy controls, although this was independent of FFM levels. Studies that have evaluated the use of functional tests in paediatric IBD patients are lacking.

Growth

Faltering linear growth is commonly encountered in children with IBD, and frequently precedes disease diagnosis. Approximately 23–25% of patients have presented with deviation from their growth velocity and height for age centiles, or as significantly shorter than their healthy peers (Kanof et al., 1988; Markowitz et al., 1993; Markowitz & Daum, 1994; Boot et al., 1998; Sentongo et al., 2000). Sentongo et al. (2000) found significant growth deficits in children with CD compared to healthy children, as well as failure to attain their genetic potential for linear growth, when their heights deficits were compared with their estimated midparental target height. This effect was more profound in boys. Girls did not differ from the healthy controls. The exact mechanisms by which growth impairment occurs in IBD are unclear. Undernutrition, the effect of circulating pro-inflammatory cytokines (Ballinger et al., 2000; Wong et al., 2006), long-term use of steroids (Alemzadeh et al., 2002) and genetic make up (Russell et al., 2006) have all been implicated (Fig. 1). On the other hand, bone age and puberty can be delayed in CD particularly in boys, which may explain to some extent the degree of growth deficits seen in IBD (Sentongo et al., 2000). This coincides with the fact that most of the patients eventually achieved final adult heights within their normal or slightly lower than their genetic potential (Alemzadeh et al., 2002; Sawczenko et al., 2006).

Bone health

Bone mineralisation is an important aspect in the care of the growing child with IBD, particularly because peak bone mass attained during adolescence was found to be most important determinant of lifelong skeletal health (Mauro & Armstrong, 2007; Sylvester et al., 2007). Osteopaenia and osteoporosis are important extraintestinal manifestations in IBD that may be related to increased risk of fractures (Klaus et al., 2002; van Staa et al., 2003; Card et al., 2004; Mauro & Armstrong, 2007; Sylvester et al., 2007). In adult studies, a 60–70% higher risk for vertebral and hip fractures incidence was found for IBD patients compared to healthy controls (van Staa et al., 2003; Card et al., 2004). There is no rigorous evidence to suggest that bones are more brittle in IBD children or that they experience more fractures compared to their healthy peers. Semeao et al. (1997) reported five cases of vertebral compression fractures in paediatric patients with CD, although Persad et al. (2006) found no difference in the prevalence of long bone fractures between paediatric IBD patients and their siblings. Similalrly, a recent study by Kappelman et al. (2010) showed that IBD children are no more likely to have experienced a diagnosed fracture at any site than their sex- and age-matched healthy controls. It is difficult to interpret these discrepancies between adult and paediatric studies. However, paediatric IBD patients may have more brittle bones that are at higher risk of fracture in adulthood and these may occur earlier than in healthy adults (e.g. before menopause). Moreover, it is also possible that vertebral fractures occur in IBD children but these may be asymptomatic and hence are undiagnosed (Stockbrugger et al., 2002).

Nevertheless, several studies have reported poor bone health in adult and paediatric patients with IBD (Compston et al., 1987; Boot et al., 1998; Semeao et al., 1999b; Schoon et al., 2000a,b; Jahnsen et al., 2003; Burnham et al., 2004; Walther et al., 2006). Total body, femoral neck and mainly lumbar spine bone mineral content or density were significantly lower than the reference cut-offs or compared to healthy individuals (Boot et al., 1998; Sentongo et al., 2000; van Hogezand et al., 2006; Sylvester et al., 2007; Schmidt et al., 2009). An inherent limitation of DXA is that it is two-dimensional and therefore does not measure true density but rather a ratio of bone mineral content over the area measured. This may lead to an underestimation of bone mineral density in paediatric patients with short stature and growth failure. This limitation was partially overcome by the introduction of peripheral quantitative computed tomography, which offers a three-dimensional imaging of the bone and soft tissues, although its application is currently limited to the extremities of the limbs. Recently, Dubner et al. (2009) observed substantial deficits in the trabecular volumetric bone mineral density and cortical bone geometry of the tibia in a newly-diagnosed cohort of CD children using peripheral quantitative computed tomography. Interestingly, trabecular volumetric bone mineral density deficits improved within 6 months of follow-up but cortical deficits further deteriorated.

A disease-associated effect is well documented, with poor bone health seen more often in CD than UC (Boot et al., 1998; Gokhale et al., 1998; Ardizzone et al., 2000). Disease location, duration and history of disease activity were risk factors in some but not all adult and paediatric studies (Semeao et al., 1999a; Sentongo et al., 2000; Habtezion et al., 2002; van Staa et al., 2003; Siffledeen et al., 2004). Burnham et al. (2004) reported that the difference in bone mineral content between CD children and healthy controls was eliminated when they used a regression model to account for differences in lean mass. Sylvester et al. (2009) showed that changes in bone mineral content during a period of 2 years post-diagnosis were positively associated with concomitant increments in FFM. These findings suggest that decreased mechanical stress may be an important factor for reduced bone health in children with CD and this opens a treatment opportunity to improve bone mass by optimising lean tissue gain through nutritional support and weight bearing exercise in IBD patients (Robinson et al., 1998). However, Dubner et al. (2009) did not observe improvements but rather a progression of deficits in the cortical bone deficits of newly-diagnosed CD children despite improvement in muscle cross-sectional area.

Bone mineralisation in IBD can be negatively affected by undernutrition; low vitamin D intake (Sentongo et al., 2002; Leslie et al., 2008); the effect of pro-inflammatory cytokines on bone formation, resorption and osteoblast maturation (Harris et al., 2009); and the long-term use of high steroid doses (Semeao et al., 1999a; Lopes et al., 2008). However, because delayed skeletal maturation and sexual maturation are commonly seen in children with IBD, particularly CD, it is important to express the results not as z-scores for chronological age but accounting for pubertal staging and the bone age of the children. Hill et al. (2009) showed that the bone mineral density z-scores deficits observed in CD when the results were expressed for the chronological age were significantly improved when the latter were calculating using bone age.

Delayed puberty

Delayed puberty is a common feature of young patients with IBD, more often in CD than UC, and in boys than in girls (Boot et al., 1998; Sentongo et al., 2000; Burnham et al., 2005). In a study of young IBD patients, menarche occurred at 16 years or later in eight of 11 female CD patients in whom disease onset preceded puberty, with three of them delayed beyond 20 years of age. By contrast, menarche occurred in all patients with juvenile-onset UC before their 14th birthday (Ferguson & Sedgwick, 1994). A mean delay in puberty of 0.7 years was found in Dutch children with IBD (Boot et al., 1998) and 1.5 years in CD children in the USA (Burnham et al., 2005). Delayed pubertal onset may influence linear growth and final adult height and could affect quality of life and self-esteem (Schwab et al., 2001), although these have not been addressed in a paediatric IBD population.

Undernutrition has always been thought to be the main reason for delayed puberty in IBD patients. However, puberty may be delayed despite a normal nutritional status. Observations in animal models of experimental colitis (Azooz et al., 2001) suggested that inflammation may have a direct adverse influence, independent of undernutrition, on the onset and progression of puberty (Ballinger et al., 2003) (Fig. 1), although relevant studies in IBD patients are lacking. In vitro studies suggested that pro-inflammatory cytokines [e.g. tumour necrosis factor (TNF)-α, interleuklin (IL)-1β, IL-6] can affect sex steroid production at the level of testes and ovaries (Terranova & Rice, 1997).

Micronutrient deficiencies and antioxidant status

Although clinical presentation of frank micronutrient deficiencies in IBD is very rare and largely limited to case reports (Myung et al., 1998), suboptimal circulating concentrations for virtually every vitamin, mineral and trace element have been reported previously, not only primarily in adult patients but also evident in the paediatric studies (Table 1). A consensus report of the North American Society of Paediatric Gastroenterology, Hepatology and Nutrition recommended routine screening for micronutrient status in paediatric patients with IBD (Kleinman et al., 2004).

Antioxidant trace elements (e.g. Zn, Se, Cu) and vitamins (e.g. vitamins A, E, C, carotenoids) were the main nutrients consistently reported at suboptimal circulating concentrations in paediatric IBD patients (Thomas et al., 1994; Hoffenberg et al., 1997; Bousvaros et al., 1998; Levy et al., 2000; Ojuawo & Keith, 2002) compared to healthy controls or the normal reference range (Table 1). Serum vitamin D has been reported to be low in adult (Leslie et al., 2008) and paediatric studies (Sentongo et al., 2002) and is an independent risk factor for poor bone health. A recent American study found that whole blood and particularly red blood cells folate concentrations were significantly higher in IBD patients compared to healthy controls despite a lower dietary intake, and this needs further exploration.

Suboptimal dietary intake, increased utilisation, malabsorption and increased enteric losses have all been postulated as causes of these nutritional deficiencies (Fig. 1). Some studies have linked nutritional deficiencies with clinical disease activity and inflammatory markers (Table 1) but whether micronutrient depletion plays an important role in the pathogenesis and perpetuation of the mucosal lesions or is the result of these needs to be studied further. However, it should be noted that changes in the plasma concentrations of many antioxidants and their association with systemic and clinical activity indices (Kuroki et al., 1993; Bousvaros et al., 1998) can be an epiphenomenon of the acute phase response in inflammatory conditions such as IBD (McMillan et al., 2000). Although reduced serum concentrations of micronutrients are often used to define deficiency states, these concentrations may better reflect disease activity and inflammation rather than being biomarkers of body tissue deficits (Galloway et al., 2000). A prime example is the transient decrease in plasma retinol binding protein and accordingly transported vitamin A plasma concentrations in the presence of the acute phase response in inflammatory conditions.

It has been proposed that the assessment of micronutrient body stores using serum concentrations in inflammatory condition is erroneous and that the use of other indices of body micronutrient stores which are independent of the effects of the acute phase response, such as red blood cells, is required (Vasilaki et al., 2009).

Several studies have drawn an association between the body’s antioxidant defence system and the pathogenesis and gut injury seen in IBD (Grisham, 1994; Rezaie et al., 2007). IBD is characterised by aggregation of inflammatory cells (granulocytes, monocytes and neutrophils) at the site of the intestinal lesion and the production of reactive oxygen species is part of the normal immune properties of these cells. The damaging action of these free radicals is normally counteracted by the body’s defence mechanisms. Uncontrolled production coupled with reduced removal from an impaired endogenous antioxidant defence system may induce tissue damage (Lih-Brody et al., 1996; Sturniolo et al., 1998). There is good evidence that IBD patients have increased oxidative stress (Levy et al., 2000; Wendland et al., 2001), which may cause damage to biological macromolecules (D’Odorico et al., 2001; Wendland et al., 2001) and possibly intestinal lesions. Wendland et al. (2001) found that lipid peroxidation, a marker of oxidative damage, was higher in adult CD patients, compared to healthy controls. Plasma antioxidant vitamins were low despite no profound difference in the dietary intake of antioxidants between CD patients and healthy controls, although this can be explained simply by the effect of the acute phase response on micronutrient plasma concentrations in some patients with active disease as explained above. Following these observations, clinical trials found a reduction in markers of oxidative stress and improved antioxidant status after supplementation with antioxidant micronutrients (Geerling et al., 2000b; Aghdassi et al., 2003). Nevertheless, a positive association between improvement of disease activity and restoration of oxidative status has not been established.

Anaemia in inflammatory bowel disease

Overt or occult intestinal bleeding is a major symptom and a drop in haemoglobin occurs with almost every flare up in CD and UC. However, the topic of anaemia in IBD has received little attention from clinicians. The prevalence of anaemia in paediatric IBD is in the range 41–88%, depending on the characteristics of the population studied and the definition used (Gryboski, 1994; Khan et al., 2002). Clinical disease severity was a strong predictor of anaemia in some but not all studies, as was the type of disease, gender, upper GI involvement, nutritional status and growth (Thayu et al., 2005; Mack et al., 2007; Gerasimidis et al., 2008b).

Two predominant types of anaemia have been identified in the context of IBD. Iron-deficient anaemia and the anaemia of chronic disease accounted for the majority of the cases. Iron-deficient anaemia was the major cause of anaemia in IBD, and this was ascribed to a negative iron balance from excessive iron loss through GI bleeding, increased epithelial sloughing, reduced dietary intake found in adult patients (Lomer et al., 2004), and impairment of iron absorption (89), particularly when the disease was located at the major sites of iron absorption (Gryboski, 1994; Thayu et al., 2005). Similarly, anaemia of chronic disease, associated with the production of inflammatory cytokines in chronic inflammation, had significant systemic effects on iron absorption (Semrin et al., 2006), the proliferation of erythroid progenitor cells, the production of erythropoietin, and the life span of red blood cells (Weiss, 2005; Weiss & Goodnough, 2005). Anaemias associated with vitamin B₁₂ and folate deficiency, or drug associated anaemia as a result of the long-term use of medication to manage IBD, have occasionally been reported, although these are uncommon (Dyer et al., 1972; Burbige et al., 1975). Patients treated with methrotrexate, an antagonist of folate metabolism, need background prophylactic supplementation with oral folate. Of utmost importance are those patients whose terminal ileum has been resected because B₁₂ absorption takes place at this site. Regular monitoring of the B₁₂ blood concentration and adequate dietary intake in these patients has been recommended (Gasche et al., 2007).

Because many of the serological markers of the iron body stores are influenced by the acute phase response, their diagnostic value in IBD is poor and can be misleading. A decrease in serum iron and downregulation of transferrin is part of the acute phase response, causing functional iron deficiency, a state that can be misinterpreted as iron deficiency in a patient with active inflammation (Maherzi et al., 1996). On the other hand, inflammation can cause a false elevation of ferritin levels, bringing ferritin concentrations of patients with true iron deficiency into the normal range (Thomson et al., 1978). As a result, higher cut-offs and evaluation in conjunction with other haematological parameters to detect iron-depleted anaemic patients have been recommended for ferritin in active IBD (Guagnozzi et al., 2006; Gasche et al., 2007). Determination of serum transferrin receptor has been proposed as a potential diagnostic marker to distinguish anaemia of chronic disease from other types of anaemia in IBD (Kohgo et al., 1986; Revel-Vilk et al., 2000).

For the treatment of anaemia, it is important not only to identify the type and severity of the anaemia, but also its origin so that therapy can be targeted at the underlying mechanism and be tailored to the patient’s needs. Although guidelines on the diagnosis and management of iron deficiency and anaemia of chronic illness have recently been published (Gasche et al., 2007), these do not distinguish between adult and paediatric patients. According to these guidelines, oral iron supplementation is the mainstream treatment for patients with haemoglobin below the normal reference. The optimal dose has not been defined yet, although this should be approximately 100 mg of elemental iron per day. For adult IBD patients with severe anaemia (i.e. haemoglobin <10 g dL⁻¹), intolerance or inappropriate response to oral iron, intravenous iron is recommended with or without the use of erythropoietic agent. Erythropoietic agent with or without intravenous iron is also indicated for anaemia of chronic disease (Gasche et al., 2007). For intravenously infused iron, the exact recommended dose in adult IBD patients varies in the range 62.5–1000 mg depending on the iron derivative compound.

Aetiology of undernutrition in inflammatory bowel disease

The aetiology of undernutrition in IBD is multifactorial as are its manifestations. The inflammatory response (with the activation of the pro-inflammatory cascade and clinical manifestations of the disease) and medical and surgical therapeutic interventions can all affect determinants of nutritional balance (Fig. 1). These include poor nutritional intake, increased energy/nutrient requirements and altered metabolism, malabsorption, excessive GI losses and nutrient–drug interactions (Fig. 1). Beyond these major nutrition-associated determinants of undernutrition, there are other non-nutrition associated factors, such as the direct effect of pro-inflammatory cytokines on growth, bone and pubertal development, that can also interact independently (Fig. 1).

Reduced dietary intake

Compared to the inconsistent results from adult studies that assessed dietary intake in IBD populations (Geerling et al., 1998, 2000a; Filippi et al., 2006; Aghdassi et al., 2007), in paediatric CD patients, energy intake was reported lower than healthy controls and the national recommendations (Thomas et al., 1993), particularly during the active phase of the disease course (Pons et al., 2009). Children with active CD consumed on average 1757 kJ (420 kcal) less than their siblings (matched for height, sex and weight), whereas, for 21% of patients, the energy intake was lower than estimated energy requirements compared to 10% of the healthy control group (Thomas et al., 1993).

Dietary intake of micronutrients may also be compromised. Thomas et al. (1993), using 5-day weighed dietary records, found that significantly less CD children with active disease achieved their requirements for copper, zinc, folate and vitamin C compared to a matched group of their healthy siblings. Similarly, Pons et al. (2009), using a semi-quantitative food frequency questionnaire, estimated that children with active CD, but not those in clinical remission, did not achieve their requirements for iron and calcium.

Food aversions and special therapeutic diets (Gerasimidis et al., 2008a) to resolve or prevent exacerbation of GI symptoms, and anorexia mediated by the interaction of pro-inflammatory cytokines with appetite hormones (Barbier et al., 1998; Karmiris et al., 2006; Ates et al., 2008) may compromise intake in IBD patients, particularly during the active phase of the disease (Rigaud et al., 1994, 1994; Bannerman et al., 2001, 2001) (Fig. 1).

Altered energy/nutrient metabolism

Dietary intake, resting metabolic rate, physical activity expenditure and diet-induced thermogenesis are major components of the energy balance equilibrium and imbalance can cause undernutrition. Higher basic metabolic rate per FFM ratio has regularly been reported in CD compared to UC patients and healthy controls, although there is still controversy between studies (Azcue et al., 1997; Capristo et al., 1998; Filippi et al., 2006). This may indicate that IBD patients, mainly those with CD, have increased energy expenditure per unit of FFM but low lean body mass.

Moreover, Azcue et al. (1997) showed that children with CD fail to adapt their resting energy expenditure (REE) to their unit of lean mass, in contrast to anorexic adolescents who had significantly lower values than healthy controls, and perhaps this contributed to their undernutrition. The reason why this may happen remains unknown and could be attributed to inflammation and the action of pro-inflammatory cytokines. Interestingly, in contrast to previous expectations for a higher energy expenditure in CD with active disease, a weak negative association was recently found between REE per kg FFM^0.52 and the Paediatric Crohn’s Disease Activity Index, an association that may be explained by the confounding effect of response to undernutrition, body size and composition (Wiskin et al., 2009). In children, differences in body composition and size make calculation of energy requirements more complicated and expressing resting metabolic rate per kg may not correct for these disparities (Hill et al., 2007).

Few studies on the energy metabolism of patients with IBD found that the nonprotein respiratory quotient was significantly lower in CD compared to UC patients or healthy controls, suggesting an increased lipid oxidation rate in the former group of patients that may explain the lower FM found in other studies (Mingrone et al., 1996, 1998, 1999; Capristo et al., 1998). Al-Jaouni et al. (2000) also found increased fat oxidation in CD patients that correlated positively with disease activity. Diet-induced thermogenesis, a small component in the energy balance equation, was higher in one study in CD patients (Mingrone et al., 1999) which could explain the lower weight and higher risk of undernutrition in IBD. Although there was no difference in the resting metabolic rate between healthy controls and CD patients in remission, diet-induced thermogenesis was higher (6% versus 10%, respectively). However, diametrically opposite results were presented by Al-Jaouni et al. (2000) who furthermore found that diet-induced thermogenesis was lower in patients with active compared to inactive disease.

Increased GI nutrient losses

Nutrient and energy loss as a result of the maldigestion of food or malabsorption during the active course of the disease could potentially impact on the maintenance of energy balance, and explain undernutrition in IBD patients (Fig. 1). However, apart from some studies on specific micronutrients (Filipsson et al., 1978; Papazian et al., 1981; Leichtmann et al., 1991; Fujisawa et al., 1998) in patients with ileal resection or with bile acid malabsorption, (Davie et al., 1994), rigorous evidence is lacking to support the loss of dietary energy or other micronutrients as a result of malabsorption. Recently, a small study in Israel (Vaisman et al., 2006) found that malabsorption is a major contributor to being underweight in adult CD patients in remission. It was found that GI energy excretion was higher in an underweight group with CD than in a normal weight group, despite no differences between the two groups for dietary energy intake or resting metabolic rate. Similarly, in another study, malabsorption and increased faecal fat were observed in severely undernourished patients, which was attributed to the impaired gastric acid and pancreatic enzyme secretion in patients with CD (Winter et al., 2004). Interestingly, gastric acid and pancreatic enzyme secretion were severely impaired in 80% of these patients (Winter et al., 2004). Indeed, following nutritional rehabilitation, stool fat output and malabsorption were reduced with concomitant improvements in pancreatic enzyme synthesis, stores and secretion.

In theory, the absorption of specific nutrients should be impaired when the disease is located at the site of specific nutrient absorption or if this area has been resected. CD patients with ileal disease or resection are traditionally susceptible to vitamin B₁₂ deficiency as a result of inadequate absorption (Duerksen et al., 2006). Iron absorption may also be diminished in active paediatric CD because of the excessive production of hepcidin, a hepatic peptide mediating the absorption of iron at the level of the enterocyte (Semrin et al., 2006).

Apart from malabsorption in IBD, loss of nutrients can occur as a result of excessive intestinal mucosal sloughing, and through protein enteropathy from a ruptured, permeable gut. Studies using whole gut lavage have shown that disease activity was closely paralleled with gastrointestinal protein loss (Ferguson et al., 1998).

Drug–nutrient interactions

Several drugs used in IBD management can influence directly nutritional intake or can interfere with the absorption, metabolism and excretion of nutrients (Fig. 1). A prime example is the antagonistic interaction of methotrexate with folate metabolism and its inherent side effect of nausea. Prophylactic supplementation should be indicated as part of the mainstream management of these patients. The British Society of Paediatric Gastroenterology and Nutrition recommended the use of 5 mg, 24 h after methotrexate dose (Sandhu et al., 2008). Similalrly, the long-term effects of steroids on calcium excretion, bone resorption, growth body composition and nutritional intake are well recognised.

Conclusions

Undernutrition in paediatric IBD patients is prevalent and the recognition of its multifactorial aetiology is fundamental before any therapeutic approach is established. There are several messages from this review for those clinicians and health professionals who are responsible for the care of IBD children. Body weight and BMI are often compromised at CD diagnosis but quickly revert to normal levels after treatment. Longitudinal changes in body weight and BMI may not coincide with parallel changes in fat and lean mass. The use of simple bedside methods of body composition, such as bioimpedance analysis, may prove useful for monitoring changes of body composition in routine clinical practice. In the presence of active disease, plasma concentrations for several micronutrients are not representative of actual body stores and should be interpreted with caution and ultimately complemented by an extensive dietetic assessment. Because both nutritional and nonnutritional factors are implicated in the origin of anaemia, delayed pubertal development, growth failure and poor bone health in paediatric IBD patients, a multidisciplinary clinical team is the ideal management approach.

Future research should explore the different facets of undernutrition using appropriate techniques eliminating the limitations and flaws raised in previous studies. Research should be directed toward effects on indices of nutritional status of new treatment options that tackle the activation of the inflammatory cascade in its initial steps, as well as offer mucosal healing and nutritional rehabilitation. The impact of exclusive enteral nutrition on body micronutrient status should also be studied. It will be important to assess micronutrient body stores using better serological markers (e.g. measurements in red blood cells) that are not influenced by the acute phase response and explore their association with disease activity. In addition, the assessment of body composition compartments in paediatric CD should be studied using bedside methods after validation with more sophisticated techniques (e.g. isotope dilution and the use of a four-compartment model). Evaluation of simple functional tests (e.g. grip strength) as surrogate measurements of FFM in paediatric IBD patients may also be useful.

The origins of anaemia in paediatric IBD using more specific haematological parameters (e.g. serum transferring receptor) should be explored and the impact of biological therapy (e.g. anti-TNFα) on the management of anaemia of chronic disease investigated.

Other areas for study include the measurement of physical activity patterns in children with IBD, their association with body composition and the impact of exercise interventions on bone health and lean mass accretion.

Nevertheless, the inherent difficulties associated with clinical management and research in paediatric IBD populations should always be borne in mind. The heterogeneity in patients’ disease (e.g. disease location, disease manifestations) and demographic characteristics (gender, age at diagnosis) complicate management options and hinder the direct translation of research outcomes in clinical practice. Treatment in IBD paediatric patients should be individualised to the patient’s needs. Future multicentre studies within the remit of the existing managed clinical networks and with interactions with academia would enable the translation of research to bedside clinical practice.

Conflict of interests, source of funding and authorship

The authors declare that they have no conflicts of interest.

No funding is declared.

KG wrote the review, CAE and PMG revised the manuscript. All authors critically reviewed the manuscript and approved the final version submitted for publication.