Aluminum in Parenteral Nutrition Solution—Sources and Possible
`Alternatives*
`
`WINSTON W.K. Koo M.B. B.S.,F.R.A.C.P.,t LAWRENCE A.KAPLAN, PH.D., JACK HORN, B.S.,
`REGINALD C. TSANG, M.B.B.S., AND JEAN J. STEICHEN, M.D.
`
`From Children’s Hospital Medical Center, Chikren ls Hospital Research Foundation and Clinical Research Center, University of Cincinnati, Ohio
`
`ABSTRACT. The extent of aluminum (Al) contamination in
`parenteral nutrition (PN) solutions for infants is not known.
`Aluminum was measured in 136 samples from various commer-
`cially available components that are used with PN. Results
`showed Al content varied widely among different components.
`The same chemical may have a different Al content depending
`on the manufacturer. However, Al contents were similar among
`lots from the same manufacturer for the same chemical. Alu-
`minum contamination was arbitrarily classified as high (> 500
`μg Al/liter), intermediate (51-500 μg Al/liter) or low (≤ 50 μg
`Al/liter). The high group included most calcium and phospho-
`rus containing salts, 1 multivitamin preparation, folic acid,
`ascorbic acid and concentrated (25%) albumin. The interme-
`diate group included sodium lactate, potassium phosphates,
`zinc and chromium chloride, multitrace metal preparation, and
`5% plasma protein. The low group included amino acids, sterile
`
`water and dextrose water, chloride salts of sodium, potassium,
`calcium, copper and chromium, sodium phosphates, magnesium
`sulphate, zinc sulphate, vitamin B12, vitamin K1, 1 multivitamin
`preparation, soybean oil emulsion and heparinized (2 U/ml)
`saline. PN solutions made from high Al components may
`contain up to 300 μg Al/liter. Calcium gluconate contributed>
`80% of the total Al load from PN. Lowering of Al content in
`calcium gluconate in addition to use of specific low Al compo-
`nents offers the opportunity to significantly lower the Al con-
`centration of the final PN solution and theoretically may
`achieve an Al content as low as 12 μg/l. We speculate that Al
`contamination may occur because Al is present naturally in the
`chemical substance or added during the manufacturing process.
`(Journal of Parenteral and Enteral Nutrition 10: 591-595,
`1986)
`
`Aluminum toxicity, particularly with respect to im-
`paired bone mineralization, is well documented.1-6 In
`humans, there are two major groups of patients at risk
`for aluminum toxicity, namely, renal failure patients on
`chronic dialysis,!-3 and, patients receiving long-term par-
`enteral nutrition.4 4
`Although it is known that certain components of par-
`enteral nutrition solution such as calcium and phosphate
`may be contaminated with aluminum,’ there are no
`systematic data to document the extent of aluminum
`contamination from various components of parenteral
`nutrition solution and whether ’low’ aluminum contain-
`ing alternatives are available. This study therefore aimed
`to determine the extent of aluminum contamination of
`frequently used components of parenteral nutrition so-
`lutions used for infants; and, for components with high
`aluminum content, to explore whether there are alter-
`natives with low aluminum content.
`
`METHODS
`
`Aluminum (Al) concentrations were determined from
`individual components of parenteral nutrition (PN) so-
`
`Reprint requests: Dr. Winston W. K. Koo. Department of Pediatrics,
`2C300 Walter Mackenzie Health Sciences Centre. Edmonton, Alberta
`T6G 2B7. Canada.
`* Presented at the 10th Clinical Congress of The American Society
`for Parenteral and Enteral Nutrition, Dallas. TX. February 1986.
`† Recipient of National Institutes of Health Clinical Associate Phy-
`sician Award 3M01 RR 00123-21S1.
`
`lution used for infants (Table 1). Each component was
`obtained from commercial sources. Certain components
`such as preservative-free heparinized saline not directly
`used for PN purposes but which may be used during the
`period of PN for catheter &dquo;flushing&dquo; were were also
`measured for Al concentration. In addition, needles with
`an Al hub were also tested for their potential to introduce
`Al contamination of the PN solution. This was tested by
`flushing 25 to 100 ml of Al-free water or dextrose water
`through the needle and measuring aliquots of the flushed
`solutions for Al.
`Aluminum was measured by electrothermal (flameless)
`atomic absorption spectrophotometry.8-1o The aluminum
`content was measured by the method of standard addi-
`tion.i° Both internal (Ortho Diagnostic Systems Inc,
`Raritan, NJ) and external (Robens Institute, University
`of Surrey, Guilford, UK) quality control samples were
`used to monitor this method. The intra- and inter-run
`coefficients of variation for the aluminum measurement
`are both 13%. Minimum detection limit of our assay is
`J-Lg aluminum/liter.
`<5
`Data storage and calculation for mean and range of
`aluminum concentrations were performed with the
`CLINFO program of the National Institutes of Health,
`General Clinical Research Center. University of Cincin-
`nati.
`
`RESL’LTS
`
`Aluminum (Al) concentration was measured in 123
`samples from 16 different components (Table I ~ that are
`used in parenteral nutrition P’_~ ~. In addition, aluminum
`
`591
`
`Eton Ex. 1015
`1 of 5
`
`

`

`592
`
`TABLE I
`Sources of aluminum contamination in parenteral nutrition solution
`
`Eton Ex. 1015
`2 of 5
`
`

`

`TABLE I-Contuue~.
`
`593
`
`* Prepared under sterile conditions in pharmacy using normal saline (Abbott Labs) and heparin (1000 U/ml, O’Neal, .Jones and Feldman
`Company or 100 U/ml, Elkin Sinn Company).
`t Did not alter aluminum concentration of flush solution (aluminum &dquo;free&dquo; water, 50% dextrose water).
`
`was also measured in nine samples of protein containing
`colloid solution and in four samples of diluted heparin-
`ized (2 U/ml) saline which may be used in conjunction
`with PN. The degree of aluminum contamination was
`arbitrarily classified into high (> 500 ~g Al/liter), inter-
`mediate (51-500 Ag Al/liter) and low (~ 50 Ag Al/liter)
`ranges (Table I). The extent of aluminum contamination
`varied widely among different chemicals. Thus Al con-
`tent varied from <5
`4g/liter in sterile dextrose water to
`pg/liter in calcium gluconate. Al content also
`>5000
`varied among different salts of the same chemical. Thus
`~[g/
`the Al content of calcium salts may range from <20
`pg/liter in calcium
`liter in calcium chloride to >5000
`gluconate. Even the same product may have different
`degrees of aluminum contamination among different
`manufacturers, eg, 5-, 23-, and 373-fold differences from
`the lowest to the highest mean Al content for zinc
`chloride, potassium phosphates, and sodium phosphates,
`respectively. However, the Al concentrations for the
`same product were similar among lots from the same
`manufacturer.
`The contribution from the various sources of alumi-
`num in a high calcium and phosphorus PN solution
`tested for use in infantsll is shown in Table II. The
`calculated Al content of this solution when prepared with
`high Al components may reach almost 300 ~g/liter. Cal-
`cium gluconate contributed up to 88% of potential alu-
`minum load in the PN solution. It is theoretically pos-
`sible, that, by using available low Al components and
`calcium chloride as the source of calcium, a similar PN
`solution would have an estimated Al content as low as
`12 J-Lg/liter (Table II).
`The aluminum needle hub did not increase the Al
`content of water or dextrose water flushed through these
`large bore needles (Table I).
`
`DISCUSSION
`
`Aluminum is the third most abundant naturallv occur-
`ring element, and the most common metallic element,
`comprising approximately 8.8 0 of the earth’s crust.1:! It
`is therefore not surprising that aluminum contamination
`is widespread. However it is apparent from this study
`that the extent of aluminum contamination of nutrient
`products varied widely. Certain products such as calcium
`gluconate are heavily contaminated with aluminum
`
`whereas other products, such as sterile water and dex-
`trose water, are essentially aluminum free.
`Since some products eg, calcium gluconate, from a
`number of manufacturers have similar degrees of alumi-
`num contamination, the source of contamination may be
`from the aluminum present naturally in the chemicals.
`Variability in the aluminum concentration of other prod-
`ucts such as sodium phosphates from different manufac-
`turers would be consistent with either a variation in
`aluminum contamination of different sources or varia-
`tion related to the manufacturing process. The consistent
`range of aluminum concentrations of the same product
`among different lots from the same manufacturer is
`consistent again with uniform contamination at the
`source or in the manufacturing processes. In the prepa-
`rations of albumin solution for intravenous use, it has
`been reported that aluminum contamination originated
`from the extraction of filters and filter aids used to
`separate and clarify fractions obtained by the Cohn
`fractionation process. Alteration in the fractionation
`protocol was able to lower the aluminum contamina-
`tion.13 Thus at least in some instances, manufacturing
`process is a likely reason for aluminum contamination.
`There is no known physiologic role for aluminum in
`animals and humans. In contrast, aluminum toxicity in
`human is well documented.’-’,’ Markedly elevated alu-
`minum accumulation in bone may occur after as short
`as 3 weeks of parenteral nutrition in infants,’ supporting
`the thesis that infants receiving PN are at risk of alu-
`minum toxicity. From our data, it would appear that
`even elimination of aluminum contamination from a few
`products, such as calcium and phosphorus containing
`salts, could have major impact in markedly lowering the
`aluminum content of the final PN solution, thus dimin-
`ishing the potential toxic effects of aluminum.
`All calcium gluconate and calcium gluceptate solutions
`tested had very high aluminum content. Calcium chloride
`is much less heavily contaminated with aluminum and
`theoretically could be used to replace calcium gluconate
`as a source of calcium. The major potential disadvantage
`with use of calcium chloride instead of calcium gluconate
`is the high chloride content (60 meqjliter? of the final
`PN solution. The resultant chloride intake would be at
`or above the maximum recommended daily allowance of
`5.8 meq/kg for an infant ingesting 200 ml of milk/kg/
`day.14 The theoretical risk of hyperchloremic acidosis
`
`Eton Ex. 1015
`3 of 5
`
`

`

`594
`
`TABLE II
`Distribution of aluminum rA.lJ in a &dquo;high calcium and phosphorus&dquo;parenteral nutrition solution
`
`Maximum dextrose content = 12.5 g/dl for peripheral vein infusion, 20 to 25 g/dl for central vein infusion.
`* Estimated from Table I. For those components in which the aluminum concentration was below the detection limit, their contribution to
`the aluminum contamination of the final solution were arbitrarily assigned as zero.
`t High chloride content with use of calcium chloride.
`I Multivitamins used: MVC 9 + 3 as high aluminum component and MVI 12 as low aluminum component.
`
`may be unacceptable. It is possible to lower aluminum
`contamination and lower chloride load with the use of
`equal mixtures of calcium gluconate and calcium chlo-
`ride ; or the use of nonchloride salts of other cations eg,
`sodium or potassium acetate. Whether these changes in
`the composition could affect the solubility of calcium
`and phosphorus in the nutrient infusate and/or the acid-
`base status of the patients remains to be tested.
`Heparin solutions have been reported to be quite heav-
`ily contaminated with aluminum.’ However when pre-
`pared in concentrations suitable for use in infant, the
`actual aluminum load appears minimal. Our data on the
`extent and variability of aluminum contamination of
`other products eg, protein containing colloid solutions,
`not generally used as parenteral nutrition, but neverthe-
`less used frequently in infants particularly for resusci-
`tation purposes, supports the findings of other published
`data. 15
`The presence of aluminum in the hub of large bore
`needles used in the preparation of PN solution does not
`appear to contaminate the final solution, presumably
`because of the small surface area of contact and the short
`period of contact during which various components
`(amino acids, sterile water, and dextrose water) are trans-
`ferred into the final bottle.
`We conclude that aluminum contamination of nutrient
`products is widespread. This might occur as result of
`natural contamination or during the manufacturing
`process. The presence of low aluminum contaminated
`products are freely available with the exception of a
`suitable intravenous calcium preparation. Thus, if a low
`aluminum containing calcium source is available, it is
`possible to markedly lower the aluminum content of PN
`solution. thus minimizing the potential risk of aluminum
`toxicity in infants requiring parenteral nutrition.
`
`ADDENDUM
`
`The aluminum content of MVI Pediatrics (Armour Com-
`pany) is 8 (6-9) ~g/liter, M (range) of four samples from two
`lots.
`
`1
`
`ACKNOWLEDGMENTS
`
`This work was supported by grants from National
`Institutes of Health 1R01 HD 18505-01Al, RR 00123,
`and RR 00068 (CLINFO).
`The authors would also like to thank Dr. A. Patterson
`for his encouragement in pursuing this study, and Mrs.
`Nancy Gau for her technical assistance.
`
`REFERENCES
`
`1. Parkinson IS, Ward MK, Feest TG, et al: Fracturing dialysis
`osteodystrophy and dialysis encephalopathy: an epidemiological
`survey. Lancet 1:406-409, 1979
`2. Ellis HA, McCarthy JH, Herrington J: Bone aluminum in haemo-
`dialyzed patients and in rats injected with aluminum chloride:
`relationship to impaired bone mineralization. J Clin Pathol 32:832-
`844, 1979
`3. Ott SM, Maloney NA, Coburn JW, et al: The prevalence of bone
`aluminum deposition therapy. N Engl J Med 307:709, 1982
`4. Ott SM, Maloney NA, Klein GL, et al: Aluminum is associated
`with low bone formation in patients receiving chronic parenteral
`nutrition. Ann Intern Med 98:910-914, 1983
`5. Goodman WG, Gilligan J, Horst R: Short-term aluminum admin-
`istration in the rat: effects on bone formation and relationship to
`renal osteomalacia. J Clin Invest 73:171-181, 1984
`6. Alfrey AC: Aluminum. Adv Clin Chem 23:69-91, 1983
`7. Sedman AB, Klein GL, Merritt RJ, et al: Evidence of aluminum
`loading in infants receiving intravenous therapy. N Engl J Med
`312:1337-1343, 1985
`8. Rostyniak PJ: An electrothermal atomic absorption method for
`aluminum analysis in plasma: Identification of sources of contam-
`ination in blood sampling procedures. J Anal Toxicol 7:20-23, 1983
`
`Eton Ex. 1015
`4 of 5
`
`

`

`9. King SW, Savory J, Wills MR: The clinical biochemistry of alu-
`minum. Crit RevClin Lab Sci 14:1-20, 1981
`10. Bertholf RL, Brown S. Renoe BW, et al: Improved determination
`of aluminum in serum by electrothermal atomic absorption spec-
`trophotometry. Clin Chem 29:1087-1089, 1983
`11. Koo WWK, Tsang TC, Steichen JJ, et al: Parenteral nutrition for
`infants: Effect of high versus low calcium and phosphorus content.
`J Pediatr Gastroenterol Nutr (in press)
`12. Anderson WA, Haupin WE: Aluminum and aluminum alloys. IN:
`Kirk-Othmer Encyclopedia of Chemical Technology, Grayson
`M , Eckroth D (eds). John Wiley &
`Sons, New York, 1978, pp
`
`129-188
`13. Milliner DS, Feldman F, Shinaberger JH. et al: Aluminum contam-
`ination of albumin-replacement solutions. N Engl J Med 312:1389-
`1390, 1985
`14. Committe on Nutrition, American Academy of Pediatrics: Com-
`mentary on breast feeding and infant formulas, including proposed
`standards for formulas. Pediatrics 57:278-285. 1976
`15. Milliner DS, Shinaberger JH. Sherman P. et al: Inadvertant alu-
`minum administration during plasma exchange due to aluminum
`contamination of albumin-replacement solutions. N Engl J Med
`312:165-167, 1985
`
`595
`
`Eton Ex. 1015
`5 of 5
`
`