Aluminum Contamination of Parenteral Nutrition and Aluminum Loading in Children on Long-Term Parenteral Nutrition

Abstract

Background 

Children who are receiving parenteral nutrition are at risk of aluminum overload, which may contribute to such side effects as osteopenic bone disease. The aim of the present study is to determine the aluminum contamination of parenteral nutrition solutions and their components, and to assess the aluminum status of children on long-term parenteral nutrition.

Methods 

Aluminum concentrations were determined by graphite furnace absorption spectroscopy in components and in final parenteral nutrition solutions. The urinary aluminum excretion and plasma aluminum concentration were determined in 10 children on long-term parenteral nutrition.

Results 

The mean aluminum concentration in the administered parenteral nutrition solutions was 1.6 ± 0.9 μmol × l⁻¹(mean ± standard deviation (SD)). The resulting mean aluminum daily intake of the 10 patients was 0.08 ± 0.03 μmol × kg⁻¹ × day⁻¹.

Conclusions 

Compared to two previous studies performed in 1990 and in 1995 in our hospital, the aluminum contamination of parenteral nutrition solutions and the daily aluminum intake of the children seemed to decrease. However, the plasma aluminum concentration and daily urinary aluminum excretion of the children still remain above normal standards. The children had no clinical symptoms of bone disease but aluminum accumulation in tissue can not be excluded. To prevent this iatrogenic toxicity, the aluminum contamination of parenteral nutrition should be assessed regularly.

Patients who are receiving parenteral nutrition (PN) are at risk of aluminum (Al) overload toxicity. This condition was first demonstrated in patients with renal failure in whom there was an excess of Al in dialysis fluids. Neurological symptoms (encephalopathy) and bone disease have been described (¹). Parenteral nutrition (PN) has also been shown to induce Al loading (^1–3). Bone disease associated with Al overload can occur in adults on long-term PN. It is characterized by bone pain, osteoporosis, or patchy osteomalacia and reduced bone apposition rate (^2–4). Neurological impairment has been demonstrated in pre-term infants receiving PN (¹). In children on long-term PN, Al may contribute to the pathogenesis of liver disease and osteopenic bone disease, but the exact implication of Al overload remains uncertain (⁵).

The US Food and Drug Administration (FDA) has set an upper limit of 0.90 μmol × l⁻¹ for Al in all large-volume parenterals used in total parenteral nutrition therapy (⁶) and an upper safe limit of 0.18 μmol × kg⁻¹ × day⁻¹ for daily Al intake. The American Society for Clinical Nutrition (ASCN/American Society for Parenteral and Enteral Nutrition (ASPEN) Working Group on standards for aluminum content of Al in parenteral nutrition solutions has established a series of thresholds (upper safe limit, unsafe limit, and toxic limit) for Al intake for patients on long-term PN. (⁷). Both groups have published recommendations for accurate labeling of PN products. One way for a nutrition team to monitor Al contamination in components and PN solutions, and to evaluate the risk of Al overload in patients, is to regularly assess Al in plasma and urine samples and in final PN solutions and their components. The aim of this study is to determine Al concentrations in PN solutions and their individual components, and to assess the Al status of children on long-term PN. Results were compared to two previous studies (⁵) performed by the same pediatric nutrition team, to the FDA regulations and to the recommendations of ASCN/ASPEN Working group on standards for aluminum content of parenteral nutrition solutions (^6–7).

MATERIALS AND METHODS

Patients

Ten children, aged 1.5 to 16 years (mean ± SD = 8 ± 5 years) were studied. Patients had been receiving cyclic nocturnal home PN, 4 to 7 days a week for an average of 6.5 years. The oral intake of aluminum was not taken into consideration because these patients were primarily on intravenous feeding, and received less than 10% of their total energy intake by mouth. Digestive diseases were: short-bowel syndrome (6 cases), untractable diarrhea (2 cases), and chronic intestinal pseudo-obstruction (2 cases). No patients had impaired renal function. No patients showed clinical symptoms of bone disease. Table 1 summarizes detailed characteristics.

TABLE 1: Characteristics of the patients

Nutrition

The hospital pharmacy or an external manufacturer prepared all PN solutions. Components were mixed under aseptic conditions and filled in EVA bags after filtration. Physicians prescribed each PN solution composition according to each patient's condition. Components and PN solutions were tested for aluminum content. Individual components were: Vinténe(Clintec, Amilly, France—batch number: 97010780 and 9701159), Vaminolact(Kabi Pharmacia, Guyancourt, France—batch number: 94831A01 and 95259A01), Glucose phosphate disodique 0.33 M (Phocytan, Aguettant, Lyon, France-batch number: 4170151), Nonan (Aguettant, Lyon, France—batch number: 421992101), Ivelip (Clintec, Amilly, France —batch number: 7071A81 and 6393A81), Medialipide (Bruneau, Boulogne, France—batch number: 9700156 and 9701272), and from the same manufacturer: Glucose 50% (batch number: 421819A01), Potassium chloride 7.46% (batch number: 16487), Potassium lactate 12.8% (batch number: 60701), Calcium gluconate 5% (batch number: 97013 and 97031), Dipotassium phosphate 1.7 M (batch number: 16551 and 16470), Oligoéléments formule B (batch number: T97009), trace elements formula D (batch number: T97017) (PCH, Paris, France).

Oral aluminum intakes were not included in the calculations of daily aluminum intake, because the patients were receiving primarily parenteral nutrition. Furthermore, since intestinal absorption of aluminum is very poor, we estimated that relatively little orally ingested aluminum was actually absorbed (⁶).

Measurements

Al concentrations were determined both in the individual components and in the final PN solutions. Because of their established low contamination levels (⁵), the Al contents of sodium chloride, magnesium chloride, zinc acetate, and sterile water were not evaluated. Plasma samples were taken at the end of the PN cycle and urine was collected over 24 hours.

Al was measured by graphite furnace absorption spectroscopy using a Perkin-Elmer 5100 PC. An Al hollow-cathode lamp was used to make measurements at the 309.3 nm resonance line with a spectrometer bandwidth of 0.7 nm. A calibration curve was constructed using standards at 10, 20, and 40 μg l⁻¹, prepared from a stock solution at 1000 mg l⁻¹. Dilutions were made with deionized water and 1% nitric acid. The minimum detection limit was 2 μg l⁻¹ and the variation coefficient was 4.5%. Each sample was measured twice. Sample containers were also checked for Al contamination and were found to be under the detection limit.

The daily intake of Al from PN was calculated as the product of the Al concentration of PN solution and the volume of PN solution administered. The urinary excretion of Al was calculated as the product of urinary concentration and the 24-hour urine volume.

Expression of Results and Statistical Analysis

All data are expressed as mean ± standard deviation. Statistical analysis was performed using unpaired Student's t test and r-test to assess correlation. Significance was set at P < 0.05.

Ethical Considerations

As the aluminum content analysis of the blood and urine was done in the context of regular checks made on patients on long-term parenteral nutrition, and, as no additional blood or urinary samplings were done specifically for the study, the local ethical committee did not require specific consent of the patients.

RESULTS

PN solutions

The mean Al concentration in PN solutions was 1.6 ± 0.9 μmol × l⁻¹. This result is significantly lower than the mean concentration reported in the study performed in the same hospital unit 10 years ago (P < 0.01) but not different from the 1995 study (P = 0.2) (⁵). Al concentrations in the components are summarized in Table 2 and compared to our two previous studies if the component and its manufacturer were identical. In the present study, concentrations varied from 0.5 to 30 μmol × l⁻¹. The highest Al concentrations were found in trace elements, calcium salts, potassium lactate, dipotassium phosphate, and AA solutions. In the final PN solution, the calcium additive provided about 50% of the total Al intake, the trace element solution 20%, followed by amino acids and glucose.

TABLE 2: Aluminum concentration of PN solutions' individual components

Patients

The calculated mean daily Al intake was 0.08 ± 0.03 μmol × kg⁻¹ × day⁻¹. The PN solution Al concentration and the daily Al intake for each child are shown in Table 3. The mean plasma Al concentration was significantly higher than the upper normal plasma concentration (^8–19) reported in Table 3 (0.9 ± 0.5 μmol × l ⁻¹ vs 0.4 μmol × l ⁻¹, P < 0.001). Only one child had a plasma concentration less than 0.4 μmol × l ⁻¹. The mean daily urinary Al excretion was significantly higher than the upper normal value (3.3 ± 2.6 μmol × day⁻¹ vs. 0.3 μmol × day⁻¹, P < 0.001) (⁸). Eight of 10 children were above this standard. Detailed plasma and urinary Al results for each child are shown in Figure 1,2. Figure 3 illustrates the daily urinary excretion of Al as a function of PN duration. No correlation was found between the two parameters. Comparisons to previous studies, FDA and ASCN/ASPEN recommendations are made in Table 4.

TABLE 3: Al concentration in PN solutions and Al daily intake for each child

TABLE 4: Comparisons to previous study ( ⁵ ), FDA and ASCN/ASPEN recommendations and usual standards. An exponent letter indicates values significantly different from the present study

FIG. 1.: Percent contribution of each component to daily aluminum parenteral intake. Black bars: Amino acids 128 g/L. Gray bars: Amino acids 65.3 g/L.

FIG. 2.: Aluminum urinary excretion and plasma concentration in 10 children on long term parenteral nutrition.

FIG. 3.: Parenteral nutrition duration and aluminum urinary excretion.

DISCUSSION

Aluminum contamination of PN solutions has been recognized since the 1980s (³). It was higher then than it is at present. Previous studies from the 1980s estimated daily intakes of Al from 3 to 4 μmol × kg⁻¹ × day⁻¹ (^3–10), which is almost 50-fold more than the present mean intake. This change is a function of reduced contamination of PN and its additives. For example, the Al concentration in protein hydrolysate was 50 to 100 times above the Al concentration of amino acid solutions currently used (¹¹). Due to the awareness of Al toxicity, a better choice of components, and industrial improvements, PN contamination has been decreasing but still remains a matter of concern (⁴). The concentration of Al in PN has been reduced 2.5-fold compared to our previous study in 1990 (⁵), but it is not significantly lower than levels we observed in 1995 (unpublished data). Despite the lack of change between 1995 and the present, we found that the daily intake of Al was significantly reduced compared to both our previous studies. A lesser contamination of the individual components is unlikely to be the main reason. All studies point out that the components are unequally contaminated and that the variability is wide among manufacturers and batches (^3–12). This study still highlights these facts. Some components can be compared because the same manufacturers provided components in at least two of the three studies and the methods used to detect Al have not changed. The Vinténe amino acid solution appears less contaminated, as do the phosphate salts, while the glucose and potassium lactate seem to contain more Al. This variability and the small number of batches assessed in our studies do not allow us to make firm conclusions on the changes in individual components of PN solutions. In any case the most contaminated components remain the trace elements, calcium and phosphate. Because of the high calcium need in children, this component is the greatest Al provider (about 50% of the total Al intake) (^3–13). Trace elements are prescribed in small amounts but are an important source of Al because of their high contamination. This could be partially due to the acidity of the trace element solutions, which might promote Al transfer from glass bottles to solutions (¹⁴). In our study, potassium is no longer a major source of Al since potassium lactate, highly contaminated, has been replaced for the most part by potassium chloride in which the Al concentration is very low (< 4 μmol × l⁻¹). The switch from Vamine amino acid solution (Kabi Pharmacia), containing 4.3 μmol × l⁻¹(⁵), to Vaminolact solution likely contributed to the decrease in Al concentration in PN solution between 1990 and 1995. For example, NP pediatric solution (PCH, Paris, France), contained 10 times more Al than the product now used.

Presently, the FDA has set an upper limit of 0.90 μmol × l⁻¹ for Al concentration of large volume parenterals (⁶). The ASCN/ASPEN Working Group on standards for Al content of PN solutions has defined three thresholds for Al intakes for patients on long-term PN (⁷). These limits, indicated in Table 4, are based on three different studies: a first study in which the patients showed no signs of tissue loading or dysfunction was used to determine the upper safe limit (¹⁵); intakes associated with Al tissue loading but no dysfunction were used to derive the unsafe limit (^3–16); and, a third study describing Al tissue overload clearly associated with bone disease was used to derive the toxic limit (¹⁷). In the present study, the mean Al concentration in PN solutions was above the FDA regulation on Al content for large volume PN solutions. Although only three children received solutions with concentrations below this limit, the mean concentrations were not significantly different (1.5 ± 0.9 vs 0.9 μmol × l⁻¹, P = 0.10). The daily Al intake for each child was well below the upper safe limit defined especially for premature neonates and patients with impaired kidney function by the FDA (0.18 μmol × kg × day⁻¹) (⁶). Nevertheless, the following statement in the FDA regulation is that tissue loading may occur at even lower rates of administration. The FDA used the article of Naylor et al. (¹⁸), who showed that 0.1 to 0.2 μmol × kg⁻¹ × day⁻¹ of Al does not cause a decrease in bone formation in pre-term infants. The infants in this study only received PN during their first three weeks of life, however, while the children in our study averaged 78 months of PN. Obviously, the duration of PN nutrition is likely to influence the bone concentration. Sedman et al. (³) have shown that bone aluminum concentration in children receiving IV nutrition therapy was higher when the duration of PN was longer than three weeks. For this reason, we compared our results to the ASCN/ASPEN standards. Although the mean daily intake is not significantly different from the ASCN/ASPEN upper safe limit (0.08 vs. < 0.07 μmol × kg⁻¹ × day⁻¹), there was a large variability among intakes, and when examined individually, 7 of 10 children were receiving a daily amount above this upper safe limit, but below the ASCN/ASPEN unsafe limit. The effects of intakes between safe and unsafe limits are not well defined.

None of our patients showed clinical symptoms of tissue accumulation of Al. However, abnormal tissue accumulation cannot be absolutely excluded in our patients. Although plasma Al decreased 2.2 fold compared to the 1990 study (⁵), it still exceeds the normal value. Indeed, only 1 of 10 children had a plasma Al below the normal value. By the same token, although urinary excretion of Al does not seem to increase with PN duration, the daily excretion was still about 10 times higher than the usual value, with 8 of our patients exceeding the acceptable limit. The only absolute proof of Al tissue loading would be biopsies, which were not be performed during this study for obvious ethical reasons. Although plasma Al and urinary Al excretion are poor predictors of Al bone overload (^19–21), plasma Al is increased in patients with bone overload, so its high value should alert physicians to a risk of toxicity (^20,21). Sedman et al. (³) have found high plasma and urinary Al concentrations in premature infants receiving intravenous feedings. They also found that bone Al concentration was 10 times higher in infants who had received 3 weeks of intravenous feeding than in those who had received infusions for a shorter time. Finally, if the Al absorption in gastro-intestinal tract is estimated to 0.9 to 1.8 μmol × day⁻¹ (¹³), 4 of our children on PN were receiving more Al than healthy adults on a normal diet.

The results of this study reinforce the idea that physicians should remain aware of any signs of Al toxicity in patients on long-term PN. Pharmacists making the PN solutions should regularly assess the Al concentration of both the components and the final PN solutions, choosing the least-contaminated components. These data should be transmitted to the care team so that the Al intakes of patients on long-term PN can be estimated and toxicity can be prevented.