Dietary calcium intake does not meet the nutritional requirements of children with chronic kidney disease and on dialysis

Ensuring adequate calcium (Ca) intake, along with maintaining vitamin D status and limiting dietary phosphate (P) exposure, is central to the dietary management of chronic kidney disease mineral and bone disorder (CKD-MBD) in children [1]. A low Ca intake is especially of concern in children with chronic kidney disease (CKD) who require Ca for bone accrual and growth and to prevent fractures [2], 90% of peak bone mass being accrued by 18 years of age [3]. Conversely, excess Ca may lead to coronary artery calcification which increases cardiovascular risk, the leading cause of death in this population [4, 5].

A large prospective cohort reported that children with even early CKD have a 2–3-fold higher fracture rate compared with that of their healthy peers [2]. The use of Ca-based P-binders was protective against fractures. Vascular calcification has also been identified early in CKD [6], and clinicians are increasingly favoring non-Ca-based P-binders to reduce Ca load, particularly for the adult dialysis population [7]. A study in young adults on dialysis [8] reported that the Ca intake from P-binders was nearly twice as high in those with coronary artery calcification compared with those without; no estimate of dietary Ca intake was made. However, extrapolating adult studies to pediatrics are not appropriate as children need a positive Ca balance [9], particularly at periods of rapid growth. Kidney Diseases Improving Global Outcomes (KDIGO) and the National Institute for Health and Care Excellence (NICE) hyperphosphatemia guidelines [10, 11] recommend that a Ca-based binder be used as first-line treatment in controlling hyperphosphatemia in children in order to meet the Ca requirements of the growing skeleton. International recommendations for Ca intake in children with CKD [1, 12] suggest that the total Ca intake from diet and medications should be 100–200% of normal requirements, apart from exceptional circumstances.

The aim of this study was to assess the Ca intake from food, formula feeds (FF; including infant milks, oral nutritional supplements and enteral tube formulas) and medications in children with CKD4–5D and controls. We quantified the Ca intake from diet, P-binders, and Ca supplements, expressed as a percentage of the UK Reference Nutrient Intake (RNI) [13], and compared this with the recommended Ca intake suggested by KDOQI [1]. The foods primarily contributing to Ca intake within this CKD population were identified and contextualized by comparison with controls and data from the UK national food consumption surveys [14, 15].

This is a single-center, prospective cross-sectional study that is part of a larger project assessing Ca balance in children and young adults with CKD [16]. Children from birth to 18 years of age with CKD4–5D attending renal out-patient clinics at Great Ormond Street Hospital were invited to participate. Age-matched controls were selected from siblings and friends of patients as well as children attending dermatology, ear, nose, and throat or plastic surgery minor operating lists at our hospital. All participants and/or their caregivers provided informed written consent and assent appropriate for their age. The study was approved by the research ethics committee and the NHS Health Research Authority.

Routine serum biomarkers were measured in CKD subjects at their clinic visit or prior to a mid-week hemodialysis session, and at the planned study visit or routine pre-operation assessment clinic for controls. These included serum total Ca, ionized Ca, P, magnesium, bicarbonate, intact parathyroid hormone (PTH), 25-hydroxyvitamin D [25(OH)D], and alkaline phosphatase. PTH concentrations were measured by the Immulite 2000 Intact PTH immunoassay (Siemens Healthcare Diagnostics, Frimley, Surrey, UK). 25(OH)D concentrations were analyzed by isotope-dilution liquid chromatography-tandem mass spectrometry (Waters Xevo TQ-S, Waters UK, Elstree, Herts, UK). Ca-based P-binders and sevelamer were prescribed as per NICE hyperphosphatemia guidelines [11].

Diet diaries for 3 consecutive days (including one weekend day) were completed by each child or their parents/caregivers, following instruction by the research nurse (SS). Participants were encouraged to describe the type, amount, and cooking method of all foods and drinks consumed, including packages and labels from manufactured food items where possible. The volume and composition of FF (oral and enteral) were recorded. To reduce participant burden and simplify day to day documentation, the full description of items that were repeatedly consumed was recorded once at the start of each diary (e.g., cow’s milk, semi-skimmed; bread, wholemeal thick-sliced). To ensure consistency in interpretation, one dietitian (LM) analyzed the diaries and (where necessary) converted food portions into weights using the UK Food Standard Agency Food Portion Sizes handbook [17]. Ca content was assessed with a software program, CompEat Pro (Nutrition Systems; www.​compeat.​co.​uk), which uses the reference UK food nutrient analysis database [18]. Manufacturers’ nutritional compositional data was used for FF. Calcium in drinking water (tap or mineral) was not included in the analysis, and the volume of water consumed was not recorded.

The mean daily dietary Ca intake from each completed 3-day diary was assessed and compared with the relevant age and sex UK RNI (defined as the amount of a nutrient that is enough to meet the needs of 97.5% of a group of healthy people in the population) [13].

The median daily dietary Ca intake for 4 groups of children were compared: controls, those with CKD4–5, CKD5D (on dialysis), and all children with CKD4–5D. This was expressed as mg Ca/day, mg Ca/kg body weight/day and as a percentage of RNI. The relative contribution of different food groups to dietary Ca intake was calculated for controls and all groups of children with CKD. The distribution was then compared with data for children from a UK cross-sectional survey of a representative sample of the general population, the National Diet and Nutrition Survey (NDNS) [14, 15].

The Ca from Ca-based P-binders and Ca supplements was quantified based on prescribed amounts, and the total combined daily Ca intake from diet, FF, and medications was calculated for all children as described above. Dietary P consumption was also assessed from the diet diaries, but it is acknowledged that this does not reliably reflect total P intake, as the amount from P-containing food additives cannot be quantified [19, 20].

Baseline characteristics of the study population are described in Table 1. The groups were matched for age, sex, and ethnicity, but, as expected, the median height and weight SDS of CKD4–5D children was significantly lower than that of the controls. The median BMI SDS for CKD4–5D and controls were not significantly different. Ninety-two percent of participants completed a 3-day diet diary and 100% completed 2 days.

In this study, we have shown that Ca intake from diet alone was inadequate in 76% of children with CKD4–5D, median dietary Ca intake being 81% RNI for age, and significantly lower than that for age-matched controls. Phases of rapid growth in childhood and adolescence are critical periods for bone mass accrual [3], giving rise to a greatly increased Ca requirement [9]. Studies in healthy pre-pubertal children receiving calcium supplementation showed increase in bone mass and augmented bone response to physical activity compared with non-supplemented controls [21, 22]. High calcium intakes seem to be protective against fractures in healthy adolescents [23]. Our data support the KDIGO [10] and NICE [11] recommendations for the use of Ca-based P-binders as first-line treatment for hyperphosphatemia in children with CKD and on dialysis.

These findings are in agreement with the Growth Failure in Children with Renal Disease (GFRD) study [24], where children with CKD achieved a Ca intake of 80% of normal requirements for age. Other studies have described similar or even lower Ca intakes [25‐27], even from CKD stage 3 [28]. It is likely that a poor appetite also contributes to the reduced Ca intake in those with CKD.

The relative contribution of different food groups to total dietary Ca intake in pediatric CKD has not been previously reported. The lower proportion of dietary Ca in CKD4–5D from dairy foods (40% vs 56% in controls), and the lower P intake assessed from the diet diaries, suggests that the reduction in dietary Ca intake may be an inadvertent consequence of advice to reduce P intake. Limiting dairy foods (along with avoidance of P food additives) often constitutes the first-line intervention for serum P control. In this cohort, all the children with CKD had been advised on dietary phosphate restriction. Excluding those exclusively on FF, cereals contributed 33% and 27% of dietary Ca intake in CKD and controls, respectively. Of note, mandatory fortification of most flour in the UK [29] (and the voluntarily fortification of foods such as breakfast cereals) increases the contribution of cereals to daily Ca intake in our patients. However, in countries where fortification is not common practice, the Ca intake in this population may be even lower. The contribution of FF to dietary Ca intake has not been previously studied. Children with CKD who were exclusively fed FF achieved a higher dietary Ca intake which was comparable with that of controls (p = 0.08).

A previous study reported a reduction in dietary Ca intake as renal function declines through all CKD stages [25]; we found this in those with CKD4–5, but not in those on dialysis, possibly due to the higher use of FF in the latter group. This finding emphasizes the importance of offering enteral tube feeding to improve overall nutrition (including Ca intake) in those struggling with a poor oral intake, as recommended by the Pediatric Renal Nutrition Taskforce (PRNT) [30]. Our results suggest a greater reliance on Ca-based medications with advancing CKD, perhaps as dietary P restrictions increase or are more strictly enforced, inadvertently reducing dietary Ca intake.

No previous dietary studies in children with CKD have quantified the contribution from Ca-containing medications. We have shown 76% of CKD4–5D patients received insufficient dietary Ca, and that Ca-based P-binders and Ca supplements improved Ca intake to a median of 177% RNI. While both KDIGO and NICE recommend Ca-based P-binders should be the first-line treatment for hyperphosphatemia in children with CKD [10, 11], the absorption of Ca from these binders is uncertain (reported to be between 20 and 30%). Absorption is generally lower than that from food, varies with the timing of administration in relation to food, and differs between Ca-carbonate and Ca-acetate [7, 31]. In addition, Ca absorption varies between 15 and 40% depending on the individual’s vitamin D status [32]. This variable absorption (which also applies to food-derived Ca) was not adjusted for in our analysis and has been identified as a future research need by the PRNT in their clinical practice recommendations for Ca and P for CKD2–5D [12].

Taking into account body size, dietary Ca intake, expressed as mg per kg body weight, decreased with increasing age in controls and children with CKD4–5D (as shown in Supplementary Fig.1). It is difficult to attribute any importance to this as Ca requirements are influenced by a number of factors that are dependent on age rather than weight. The reported difference in dietary Ca intake (as a % RNI) between controls and those with CKD4–5D was still significant when adjusted for height age.

RNI for Ca was not met by 44% of controls, suggesting that a low Ca intake in the general pediatric population may be a concern, particularly in those over 10 years of age. This finding was also reflected in the latest report from the NDNS, where an increasing proportion of UK children in the age group 11 to 18 years had an insufficient dietary Ca intake [33]. The wide variation in published international requirements for Ca [12] and the increased availability of data on Ca balance in children may suggest that the UK RNI figures which were published in 1991 are no longer an appropriate reference against which to benchmark Ca intake. This has been addressed by PRNT [12] who propose a new term, Suggested Dietary Intake (SDI), which reflects the range of recently published international requirements derived using the most reliable methods for assessing mineral status (and does not include the UK RNI). Only 11% of our CKD cohort had a dietary Ca intake within the SDI, compared with 24% achieving UK RNI as the latter is set lower for most age groups. Although international guidelines [10‐12] recommend that clinical decisions are based on serum Ca (or ionized Ca) levels, none of the 7 children with a total Ca intake below RNI was hypocalcemic, and only one patient having > 2 × RNI was hypercalcemic, suggesting that serum Ca may not be a useful guide to the adequacy of Ca intake. The true Ca requirements for any individual depend on their age, growth, and rate of bone turnover [12]. To fully interpret our results, and understand the implications of high and low Ca intakes in children with CKD, careful Ca balance studies are needed to determine Ca absorption, state of bone mineralization, and fate of the ingested Ca. Meanwhile, as recommended by KDIGO, trends in Ca, P, PTH, and ALP, taken together, are the best available measure for the clinical management of MBD [10]. Indeed, as shown in a recent study in children and young adults with CKD 4–5D [34], tibial cortical bone mineral density (BMD) was negatively associated with parathyroid hormone (PTH; r = − 0.44, p < 0.001) and alkaline phosphatase (ALP; r = − 0.22, p = 0.03) and positively with serum calcium (r = 0.33, p = 0.001). On multivariable linear regression PTH, ALP and calcium together predicted 57% of variability in tibial cortical BMD measured by peripheral quantitative CT scan, whereas dual energy X-ray absorptiometry did not correlate with biochemical measures.

Forty-one percent of our CKD4–5D cohort received > 200% RNI for Ca as a result of the contribution from Ca-based P-binders. However, in this study, we did not record adherence to medications and this may have resulted in an overestimation of total Ca intake, as adherence to P-binder medication is notoriously poor [35, 36]. It is also not certain how much calcium is absorbed from these binders [7]. In clinical practice, in the presence of high serum calcium, we would carefully check the patient’s dosage and timing of administration of Ca-based P-binders in relation to their dietary phosphate intake. In our experience, patient education often results in a lower overall dosage, when binders are being used unnecessarily with low phosphate meals. In addition, patients would be instructed not to give calcium carbonate at the same time as ranitidine or other H2 blocker as gastric acidity is required for the dissociation of calcium carbonate and its phosphate-binding effect [7].

Limitations of our study include the incompleteness of the diet diaries (3 CKD4–5D and 3 controls completed only 2 rather than 3 days). The use of a semi-quantitative prospective 3-day diet diary is considered to be an appropriate dietary assessment tool for Ca or P intake [12, 37], but imposes a high participant burden with possibly incomplete or selective recording [37, 38]. In an attempt to limit potential errors due to coding and interpretation of the diaries, a single dietitian performed all the analyses. Nutrient databases and dietary assessment software packages to calculate Ca intake are considered reasonably accurate when compared with direct chemical analyses [38], but a food database does not fully account for Ca intake from food additives, food fortification, or from drinking water. Ca intake from dialysate was also not included in the Ca analysis.

In conclusion, 76% of the CK4–5D study population from a large pediatric tertiary care hospital had a dietary Ca intake below 100% RNI for age. A reduction in Ca intake from dairy foods is likely to account for this. In our cohort, 80% of patients had Ca-based P-binders and 9% also had Ca supplements, increasing Ca intake to 177% RNI and thus meeting the KDOQI guidelines for CKD of 100–200% recommended intakes for the normal population. Just as with healthy children, an adequate Ca intake is required to meet the needs of the growing skeleton in children with CKD.