Childhood chronic kidney disease (CKD) represents a spectrum of conditions, which result in renal impairment varying from mild renal insufficiency to end stage renal failure (ESRF) (Table 1). CKD can be defined as either a glomerular filtration rate of less than 60 ml/min per 1·73 m2 for greater than 3 months or kidney damage demonstrating pathological abnormalities or markers of damage – blood, urine or imaging investigations. Most studies in these children refer to those with moderate to severe CKD (stages 3–4) and ESRF (stage 5), which corresponds to a glomerular filtration rate being less than 15 ml/min per 1·73 m2 and requires renal replacement therapy of either dialysis or renal transplantation. The incidence of children with ESRF has been reported as nine per million child population overall with the highest incidence of new patients occurring between 10 and 15 years of age (Ansell et al. Reference Ansell, Feest, Byrne and Ahmad2003). Multiple aetiologies are responsible for CKD (Table 1), of which over 50 % are due to congenital abnormalities with the remainder being mainly due to hereditary conditions, glomerulonephrites or multi-system disease.
Table 1 Stages, aetiology and modes of presentation of chronic kidney disease (CKD)
GFR, glomerular filtration rate.
The presentation of CKD can be varied, either due to the primary renal disease or as a consequence of impaired renal function, with onset sometimes being silent with an insidious progression and symptoms only developing late in its course (Table 1). Even with optimal care, many of these children go on to develop ESRF and require renal replacement therapy. Management of CKD prior to renal replacement therapy is thus conservative with the main aims being to slow down disease progression, optimise renal function and minimise complications secondary to CKD.
Growth failure is a significant problem in CKD with up to 50 % of all patients with ESRF in childhood attaining adult heights below the 3rd centile (Fivush et al. Reference Fivush, Jabs, Neu, Sullivan, Feld, Kohaut and Fine1998). The cause of growth failure in CKD is multifactorial with linear impairment being a final common pathway of various factors including the aetiology of CKD, hormonal dysregulation, nutritional deficiency, metabolic acidosis, uraemia, chronic anaemia and persistent micro-inflammation (Kuizon et al. Reference Kuizon, Nelson and Salusky1997; Chan et al. Reference Chan, Williams and Roth2002). Whilst optimal conservative management aimed at minimising these complications has been shown to improve growth, the impact of each factor on growth remains unclear.
In the present review we shall summarise the various patho-physiological mechanisms of growth failure in CKD. In particular, we shall examine the importance of endocrine dysfunction and nutrition in growth failure and, finally, look at the close association of these factors with abnormalities in body composition in CKD.
Pathophysiology of growth impairment
Age of onset
CKD can result in impairment in each phase of development from in utero to adolescence, which can subsequently result in growth retardation, with studies suggesting that the degree of height deficit worsens with duration of disease (Norman et al. Reference Norman, Coleman, Macdonald, Tomsett and Watson2000). Intrauterine growth retardation is critical to linear growth and may occur in severe renal hypoplasia, whilst placental insufficiency causing intrauterine malnutrition may also affect renal morphogenesis in the first trimester and therefore retard fetal growth (Bassan et al. Reference Bassan, Trejo, Kariv, Bassan, Berger, Fattal, Gozes and Harel2000; Rodriguez-Soriano et al. Reference Rodriguez-Soriano, Aguirre, Oliveros and Vallo2005).
In infancy, untreated CKD is associated with severe growth retardation with loss in relative height being as high as SD 0·6 per month during the first year of life (Rees et al. Reference Rees, Rigden and Ward1989; Karlberg et al. Reference Karlberg, Schaefer, Hennicke, Wingen, Rigden and Mehls1996). Feeding problems experienced in these infants with its concomitant anorexia and uraemia results in protein–energy malnutrition, which in turn leads to increased protein catabolism and ultimately growth retardation, which may also explain the failure of some of these infants to exhibit catch-up growth.
After 2 years of age, children who develop CKD exhibit an initial loss of height followed by growth parallel to their percentile after the disease process is better controlled (Schaefer et al. Reference Schaefer, Wingen, Hennicke, Rigden and Mehls1996). The degree of growth retardation is closely associated to renal dysfunction with a decrease in height standard deviation scores occurring once glomerular filtration rate falls below 25 ml/min per 1·73 m2 (Englund et al. Reference Englund, Tyden, Wikstad and Berg2003) Finally, in adolescence, onset of puberty may be delayed by up to 2 years in up to 60 % of affected children and the delay may correlate to the duration of CKD (Schaefer et al. Reference Schaefer, Seidel, Binding, Gasser, Largo, Prader and Scharer1990; Simon, Reference Simon2002). Even when they do enter puberty, a sub-optimal peak height velocity and shorter pubertal phase may result in a sub-optimal final height (Kleinknecht et al. Reference Kleinknecht, Broyer, Gagnadoux, Martihenneberg, Dartois, Kermanach, Pouliquen, Degoulet, Usberti and Roy1980).
Nutritional deficiency
Various studies have shown that these children frequently have protein, energy and nutrient-deficient diets regardless of the degree of renal impairment and this, in turn, can lead to both retarded growth and abnormal body composition (Salusky et al. Reference Salusky, Fine, Nelson, Blumenkrantz and Kopple1983; Abitbol et al. Reference Abitbol, Warady, Massie, Baluarte, Fleischman, Geary, Kaiser, McEnery and Chan1990; Norman et al. Reference Norman, Coleman, Macdonald, Tomsett and Watson2000). An association between energy intake and growth has been shown in children receiving haemodialysis or peritoneal dialysis (Tom et al. Reference Tom, McCauley, Bell, Rodd, Espinosa, Yu and Yu1999). Although energy requirements are aimed at 100 % of the recommended daily allowance for age, 120–140 % has been suggested for catch-up growth, but achieving this is often limited by anorexia secondary to uraemia. Anorexia is further exacerbated by hyperleptinaemia, which is shown to have a hypothalamic effect to reduce appetite and food intake and to increase energy expenditure (Warady et al. Reference Warady, Alexander, Watkins, Kohaut and Harmon1999; Mak et al. Reference Mak, Cheung, Cone and Marks2006). Protein deficiency is common in untreated children, which contributes to acidosis, uraemia, catabolism and impaired growth (Greiber & Mitch, Reference Greiber and Mitch1992; Meireles et al. Reference Meireles, Price, Pereira, Carvalhes and Mitch1999). This can remain a problem in children receiving peritoneal dialysis, where there is a significant loss of body protein into the dialysate effluent (Quan & Baum, Reference Quan and Baum1996). Further nutrient deficiencies in phosphate, K, Na, vitamin C, vitamin B6 and folic acid resulting from dietary restrictions could also contribute to growth failure and may require individualised prescription depending upon stage of renal failure and haematological and biochemical changes. In the UK, nutritional recommendations are based on dietary reference values (Department of Health, 1991) for food energy and nutrients. In the presence of normal height (>2nd percentile), energy and micronutrient requirements are based on chronological age but if height falls below the 2nd percentile, requirements for height and age may be used as baseline values and adjusted accordingly, as recommended by Coleman (Reference Coleman, Shaw and Lawson2001) (Table 2).
Table 2 Nutritional guidelines for the child with chronic kidney disease (CKD). Guidelines providing recommended ranges for energy and protein requirement. Further adjustment required based on individual nutritional assessment
* Estimated average requirement (EAR) based on Dietary Reference Values (Coleman, Reference Coleman, Shaw and Lawson2001).
CCPD, continuous cycle peritoneal dialysis; CAPD, continuous ambulatory peritoneal dialysis.
Infants and young children with CKD frequently suffer from recurrent vomiting, anorexia and feeding problems (Hellerstein et al. Reference Hellerstein, Holliday, Grupe, Fine, Fennell, Chesney and Chan1987). Gastro-oesophageal reflux, which can significantly impair nutritional and medicinal delivery in an already undernourished child, has been reported in over 70 % of these patients, with further studies observing gastro-oesophageal reflux to have significantly prolonged reflux periods and intra-oesophageal pH to be abnormally decreased (Ruley et al. Reference Ruley, Bock, Kerzner, Abbott, Majd and Chatoor1989). This suggests the presence of both gastric and oesophageal dysmotility, further complicated by the presence of gastric dysrhythmias and delayed gastric emptying (Ravelli et al. Reference Ravelli, Ledermann, Bisset, Trompeter, Barratt and Milla1992). The uraemic state may also interfere with extra-renal secretion and degradation of regulatory peptides within the gut, with increases in serum gastrin and other peptides resulting in altered regulation of gastrointestinal motility, modulation of hunger and satiety (Hallgren et al. Reference Hallgren, Karlsson and Lundqvist1988). Prolonged enteral feeding is therefore very common and needed in order to reach adequate nutritional requirements, although this can be associated with feeding difficulties and it is subsequently important to offer support on non-nutritive sucking in infants and encourage normal feeding behaviours in the older child (Dello-Strogolo et al. Reference Dello-Strogolo, Principata, Sinibaldi, Appiani, Terzi, Dartois and Rizzoni1997).
Metabolic acidosis
Metabolic acidosis is common in CKD and often seen before glomerular filtration rate < 50 %. It is essentially due to impaired NH3 excretion, aggravated by nutritional protein load and altered electrolyte balance. This can then accelerate protein degradation by activation of the ubiquitin–proteosome pathway, stimulation of branched chain keto-acid dehydrogenase, stimulation of endogenous steroid synthesis and promotion of end-organ resistance to anabolic effects of growth hormone (GH) (Mitch & Price, Reference Mitch and Price2003). Metabolic acidosis has also been shown to directly stimulate bone resorption, whilst correction of this abnormality has been correlated with an improvement in bone formation (Domrongkitchaiporn et al. Reference Domrongkitchaiporn, Pongskul, Sirikulchayanonta, Stitchantrakul, Leeprasert, Ongphiphadhanakul, Radinahamed and Rajatanavin2002; Lemann et al. Reference Lemann, Bushinsky and Hamm2003).
Micro-inflammation
CKD is characterised by a chronic inflammatory state, with inflammatory cytokines being closely associated with protein metabolism. Recent studies have shown the existence of a malnutrition–inflammation complex, in which chronic inflammation leads to protein–energy malnutrition (Kaizu et al. Reference Kaizu, Kimura, Yoneyama, Miyaji, Hibi and Kumagai1998; Kalantar-Zadeh et al. Reference Kalantar-Zadeh, Kopple, Block and Humphreys2001). Malnutrition in CKD has been proposed to be a combination of uraemia and inflammation (Stenvinkel et al. Reference Stenvinkel, Lindholm, Heimburger and Heimburger2000) with resultant severe catabolism and potential growth failure. The roles of IL-6 and C-reactive protein in adult studies have been shown to have a negative correlation with muscle mass in CKD and particularly haemodialysis (Kaizu et al. Reference Kaizu, Ohkawa, Odamaki, Ikegaya, Hibi, Miyaji and Kumagai2003) due to cytokine activation (Roccatello et al. Reference Roccatello, Menegatti and Alfieri1998), which can induce protein catabolism and promotion of apoptosis (Carracedo et al. Reference Carracedo, Ramirez, Martin-Malo, Rodriguez and Aljama1998) via ubiquitin and capase-3 pathways (Carracedo et al. Reference Carracedo, Ramirez, Soriano, Martin-Malo, Rodriguez and Aljama2002; Raj et al. Reference Raj, Shah, Shah, Ferrando, Bankhurst, Wolfe and Zager2003). Strategies to reduce chronic inflammation, such as the use of regular l-carnitine in ESRF, may improve cellular defences, modulate the inflammatory cascade and, thus, indirectly maintain lean mass and growth potential (Pertosa et al. Reference Pertosa, Simone, Soccio, Marrone, Gesualdo, Schena and Grandaliano2005).
Secondary hyperparathyroidism and renal osteodystrophy
Renal osteodystrophy develops as a result of the effect of impaired renal function on Ca, P, vitamin D metabolism and parathyroid hormone activity (Mehls et al. Reference Mehls, Ritz, Kreusser and Krempien1980). Impaired phosphate excretion results in an elevation of serum phosphate and a reciprocal drop in Ca, stimulating the development of secondary hyperparathyroidism and renal osteodystrophy. Histologically, renal osteodystrophy results in widening of the growth cartilage zone due to the development of fibrochondroblastic, fibro-osteoclastic and woven bone resulting in narrowing of the cartilage layer and breakdown of the spongiosa layer. Although growth impairment does not occur initially, progressive destruction of the growth plate due to secondary hyperparathyroidism can lead to slipping of epiphyses and cessation of growth (Santos et al. Reference Santos, Carbajo-Perez, Rodriguez, Fernandez-Fuente, Molinos, Amil and Garcia2005).
Growth hormone/insulin-like growth factor 1 axis
GH exerts its somatotrophic effects through both direct and indirect actions via insulin-like growth factor (IGF)-1 by inducing differentiation of epiphyseal growth plate precursor cells towards chondrocytes, which in turn become responsive to IGF-1 and concomitantly express IGF-1 mRNA resulting in proliferation of pre-chondrocytes, osteoblast hypertrophy, bone remodelling and net mineralisation (Ohlsson et al. Reference Ohlsson, Bengtsson, Isaksson, Andreassen and Slootweg1998). IGF-1 affects somatic growth in fetal and postnatal development (Gohlke et al. Reference Gohlke, Huber, Hecher, Fimmers, Bartmann and Roth2005; Yakar et al. Reference Yakar, Pennisi, Wu, Zhao and LeRoith2005) acting systemically as a classical endocrine hormone and locally as a paracrine/autocrine growth factor (Dupont & LeRoith, Reference Dupont and LeRoith2001), whilst being influenced to variable degrees by GH, nutrition and sex steroids (Straus & Takemoto, Reference Straus and Takemoto1990; Thissen et al. Reference Thissen, Ketelslegers and Underwood1994). Serum GH levels and secretory rates in CKD have been shown to be high–normal in pre-pubertal children, probably due to attenuated bioactive IGF-1 feedback and reduction in the metabolic clearance rate of GH (Haffner et al. Reference Haffner, Schaefer, Girard, Ritz and Mehls1994; Tonshoff et al. Reference Tonshoff, Blum, Wingen and Mehls1995). The paradox of normal or elevated GH levels in the presence of growth retardation has led to the concept of uraemic GH resistance (Kaskel, Reference Kaskel2003). Despite increased GH concentrations, free IGF-1 levels and IGF-1 bioactivity are decreased in CKD due to the presence of IGF binding proteins, which have been shown to have multiple level interactions with the GH/IGF axis, which modulates their action (Frystyk et al. Reference Frystyk, Ivarsen, Skjaerbaek, Flyvbjerg, Pedersen and Orskov1999; Kiepe et al. Reference Kiepe, Ciarmatori, Hoeflich, Wolf and Tonshoff2005).
Pubertal dysfunction
Delayed or abnormal pubertal progression in children with CKD further increases the risk of growth retardation. These children exhibit elevated gonadotrophins with decreased or low–normal gonadal hormones resulting in a state of compensated hypergonadotrophic hypogonadism (Marder et al. Reference Marder, Srivastava and Burstein1983; Schaefer et al. Reference Schaefer, Hamill, Stanhope, Preece and Scharer1991). However, they may also have evidence of hypothalamic-pituitary-gonadal dysregulation with blunted response of luteinising hormone to gonadotrophins and decreased luteinising hormone pulsatility and bioactivity (Blackman et al. Reference Blackman, Weintraub, Kourides, Solano, Santner and Rosen1981; Oertel et al. Reference Oertel, Lichtwald, Hafner, Rauh, Schonberg and Scharer1983; Giusti et al. Reference Giusti, Perfumo, Verrina, Cavallero, Piaggio, Gusmano and Giordano1991). Finally, oestradiol and testosterone are reduced due to uraemia, either through direct toxic effects or hyporesponsiveness of the gland, which can thus further effect pubertal growth (Karagiannis & Harsoulis, Reference Karagiannis and Harsoulis2005).
Assessment of body composition
The complexity of nutritional management and its relationship to growth requires an accurate assessment of body composition in CKD and understanding of the potential problems of each method (Table 3). BMI was developed as a reliable method of assessing body fat and nutritional status in public health studies (Cole, Reference Cole1997). However, its value in the setting of chronic disease, where children may also suffer from growth retardation, has been questioned (Schaefer, Reference Schaefer2000). In addition to growth retardation, reduced levels of physical activity, ESRF with the need for dialysis and immunosuppressive therapy may further confound the interpretation of BMI. Furthermore, it is increasingly recognised that truncal obesity is associated with a higher risk of cardiovascular morbidity and that BMI, by itself, has limited value in assessing regional body composition (Bolton et al. Reference Bolton, Ionescu, Evans, Pettit and Shale2003).
Table 3 Limitations of methods of assessing body composition in chronic kidney disease (CKD)
Assessment of skin-fold thickness as a marker of fat mass, whilst widely used, is prone to significant inter- and intra-observer variation. Similarly, whilst waist circumference measurements are closely correlated to both abdominal fat mass and insulin resistance syndrome (Hirschler et al. Reference Hirschler, Aranda, Calcagno, Maccalini and Jadzinsky2005), fluid overload in CKD can result in overestimation of fat mass whilst abnormal regional body composition or abnormal ratio of subcutaneous:visceral fat could be unrecognised (Odamaki et al. Reference Odamaki, Furuya, Ohkawa, Yoneyama, Nishikino, Hishida and Kumagai1999).
Isotope dilution technique is often considered to be the gold standard for assessing body composition but this technique is not widely available and, in CKD, fluid overload may lead to inadequate equilibration with resulting underestimation of fat and lean mass (Wuhl et al. Reference Wuhl, Fusch, Scharer, Mehls and Schaefer1996). Densitometry estimates body composition from body density using either underwater weighing or air-displacement plethysmography (Sardinha et al. Reference Sardinha, Lohman, Teixeira, Guedes and Going1998; Fields & Goran, Reference Fields and Goran2000) whilst neutron activation analysis provides a non-invasive analysis of the total body content of major elements by directing neutrons at the subject and then measuring the time of decay of the excited neutrons (Ellis, Reference Ellis2000). Fluid overload and raised K and urea in CKD are again limitations to these techniques. Total body K counting measures the naturally occurring radioisotope 40K to calculate total body cell mass. However, this method relies on estimating total body K in lean muscle, which may be abnormal in CKD resulting in an overestimation of body cell mass (Blumenkrantz et al. Reference Blumenkrantz, Kopple and Gutman1980). Bioelectrical impedance analysis has been widely used to examine nutritional status in children and body composition in CKD (Nagano et al. Reference Nagano, Suita and Yamanouchi2000; Schaefer, Reference Schaefer2000). This method estimates body fluid compartment volume by measuring conduction of electrical current through body fluids and impedence by body components. However, concerns remain over the reliability, reproducibility and sensitivity to changes in regional composition of this method in individual children in these cohorts (Kyle et al. Reference Kyle, Bosaeus and De Lorenzo2004).
Dual energy X-ray absorptiometry (DXA) involves two photon beams passing through a subject's body to create a two-dimensional projection of a three-dimensional structure. In CKD, in addition to assessing bone mineral content, whole body DXA can also provide reproducible estimates of fat mass and lean mass in children and adults by measuring the degree of dual photon attenuation produced by each tissue type (Goran et al. Reference Goran, Driscoll, Johnson, Nagy and Hunter1996; Azocar et al. Reference Azocar, Cano, Marin, Delucchi and Rodriguez2004). Unlike the other methods described, DXA has the advantage of being more widely available in clinical practice. Furthermore, reference paediatric population data are increasingly becoming available for various populations and ethnicities (Ellis et al. Reference Ellis, Abrams and Wong1997; Van der Sluis et al. Reference Van der Sluis, Boot and Nauta2000). The main concern over the accuracy of DXA is its failure to measure total body water. Instead, the software used to calculate lean mass assumes the total body water to be 73 % of lean mass (Pietrobelli et al. Reference Pietrobelli, Formica, Wang and Heymsfield1996). This is particularly pertinent to CKD, as in the presence of fluid overload there can be an overestimation of lean mass.
Body composition in children with chronic kidney disease
There is increasing evidence that children with CKD have an altered fat mass and lean mass with nutrition, uraemia, chronic inflammation, physical inactivity and GH resistance contributing to the aetiology of abnormal body composition in these children (Axelsson et al. 2004; Pattaragarn et al. 2004). In affected adults, a high BMI has been reported with fluid overload and a low BMI has been associated with increased morbidity and mortality (Pifer et al. Reference Pifer, McCullogh and Port2002). In children, similar studies have shown an inverse bell-shaped association between BMI and mortality risk with a change of more than 1 SD being associated with a 6 % increase in mortality (Wong, 2000). However, BMI when corrected for height age is only moderately raised in children with CKD (Schaefer, Reference Schaefer2000). Measurement of skin-fold thickness in affected children has shown deficits in tricep skin-fold thickness (Orejas et al. Reference Orejas, Santos, Malaga, Rey, Cobo and Simarro1995), but Zivicnjak et al. (Reference Zivicnjak, Franke, Ehrich and Filler2000) demonstrated that fat distribution in CKD was disproportionate with an increase in truncal fat in comparison with limbs. Such findings have also been reported in adults who had a significant increase in visceral fat and a decrease in subcutaneous fat when compared with healthy subjects (Odamki et al. Reference Odamaki, Furuya, Ohkawa, Yoneyama, Nishikino, Hishida and Kumagai1999). Total body K, a potential marker of lean mass, is reported to be low in childhood CKD (Weber et al. Reference Weber, Michalk, Rauh, Romahn and Scharer1980), whilst in vivo neutron activation measuring total body N has also shown deficits in childhood CKD, although this change was not apparent when data were adjusted for height (Baur et al. Reference Baur, Knight, Crawford, Reed, Roy, Allen and Gaskin1994).
DXA data, corrected for height in children with CKD, demonstrated that BMI did not accurately reflect body composition, with patients exhibiting low lean mass with relatively high fat mass (Rashid et al. Reference Rashid, Neill, Smith, King, Beattie, Murphy, Ramage, Maxwell and Ahmed2006). Johnson et al. (Reference Johnson, Wang, Kaskel and Pierson2000) also reported reduced lean mass and high fat mass in children with CKD and an increase in lean mass and a decrease in fat mass over a 6-month period of recombinant human GH treatment. Similar results have been reported in pre-pubertal children with CKD, confirming the lipolytic and anabolic effects of GH, although neither of these two studies adjusted the data for height or body size (Boot et al. Reference Boot, Nauta, de Jong, Groothoff, Lilien, van Wijk, Kist-van Holthe, Hokken-Koelega, Pols and de Muinck Keizer-Schrama1998; Van der Sluis et al. Reference Van der Sluis, Boot and Nauta2000).
Conclusions
The present review seeks to explain the different mechanisms that affect growth in childhood CKD. Endocrine disturbances and nutritional problems play major roles in growth failure. Children with CKD may have abnormalities of their body composition, which may not be apparent on simple measurement of BMI. Evaluation by methods such as DXA needs to be performed carefully with due consideration of the concurrent growth retardation and pubertal delay.
Acknowledgements
R.R. is supported by a research grant from the Yorkhill Children Foundation.
