A Pilot Study to Assess Alterations in Trace Element Status in Pediatric Patients with Malignancies.

Abstract 3183

Nutritional status of children with cancer is a surprisingly understudied area. While overt malnutrition is not common among US children with cancer, many experience weight loss or mild nutritional deficiencies. The Children's Oncology Group (COG) noted in one study that only 46% of responding institutions conducted a nutrition assessment on newly diagnosed children with cancer and only 52% had criteria for intervention (Ladas et al. 2006). Accordingly, the COG has called for more clinical studies in order to provide evidence-based guidelines for nutritional assessment and interventions for these patients. Zinc, a trace mineral used in hundreds of proteins in the body, is rarely assessed in these children. We hypothesized that there is increased incidence of zn deficiency in US children with cancer compared with the general population, and that low zn levels may be associated with increased incidence of common toxicities of treatment, including diarrhea, mucositis or infection. We report here an interim analysis.

We designed a prospective trial of children with cancers diagnosed and treated at Children's Hospital and Research Center Oakland (CHRCO) from January 2011 through August 2012. Fifty children aged 1 mo to 18 years were enrolled and followed over a six month period. Diagnoses included 33 leukemias, 2 lymphomas, 2 CNS tumors and 13 other solid tumors. Children were enrolled within two weeks of diagnosis and prior to start of any supplemental nutrition. Exclusion criteria included prematurity and known malabsorptive or zn wasting syndromes. A control group of 50 healthy children were also recruited. Approval was obtained from the CHRCO IRB.

Anthropometric measurements recorded included height, weight, mid upper arm muscle circumference (MUM C) and skinfolds. Laboratory markers measured included plasma zn, copper (cu), ceruloplasmin, serum albumin, c-reactive protein (CRP) and urine zn, cu and creatinine. Dietary assessment was performed using a self-administered food frequency questionnaire.

No significant differences were seen in baseline assessments between the case and control groups. Average body mass index (BMI), height z-score and weight z-score were the same between the two groups. However, subanalysis of the solid tumor group showed a decreased BMI of 16.3 vs. 18.5 in the control group, p=0.01. Similarly, there were no differences in dietary intake of major nutrients, with both groups averaging well above the estimated average requirement (EAR). One notable exception was Vitamin D which was low in both groups but significantly lower in the case group, 37% of EAR vs. 55% of EAR, p<0.01.

As expected CRP was elevated in the case group at baseline, 0.85 vs. 0.09 in the control group, p=0.01. Similarly cu was elevated at 142.9 ± 51.2 vs. 111 ± 31.2, p<0.01. However, zn was not decreased at baseline, 85.1 ± 20.8 vs. 84.5 ± 10.4. A larger proportion of the case cohort had low zn levels (<70ug/dl) than did the control group (16% vs. 8%) but this did not reach statistical significance. However, zn decreased significantly in the case group over a six month period from 85.1 ± 20.8 to 72.7 ± 15.2, p=0.002. This difference was most marked in the adolescent leukemia/lymphoma group who demonstrated 30% decrease in zn levels and 60% of these subjects had low zn levels at 6 mo. Cu levels also dropped over the 6 month period, from 142.9 ± 51.2 to 123.5 ± 41.9, p=0.01. We performed ANOVA to evaluate zn and cu in the case cohort over time. Zn dropped significantly with a p value of 0.0015, as did cu with a p of 0.0001.

Our data showed no significant deficiencies in plasma zn levels in children at diagnosis of cancer. However, plasma cu was elevated. Both zn and cu levels dropped over the initial 6 month course of treatment, most markedly in the leukemia/lymphoma group in contrast to a prior study which showed decreases only in a subgroup of solid tumors and not in the leukemia/lymphoma population (Malvy, et al. 1997). In multivariate analysis, variables correlating with low zn included plasma cu, age, MUMC and BMI. CRP did not correlate with plasma zn levels, suggesting that low zn was a consequence of true nutritional deficit rather than an acute inflammatory response. Not all the data are mature at this time and we have yet to evaluate the relationship between plasma zn levels and toxicities of cancer treatment. Further analyses are needed to identify the conditions contributing to and the consequences of zn deficiency in this population.