J. Trace ELern. Med. BioL. Vo[. 17 (2) 107-115 (2003) http://www.urbanfischer.de/journaLs/jtraceeLm © 2003 by Urban & Fischer Ver[ag Influence of the glass packing on the contamination of pharmaceutical products by aluminium. Part II1: Interaction container-chemicals during the heating for sterilisation Denise Bohrer*, PauLo Cicero do Nascimento, Regina Binotto and EmiLene Becker Departamento de Quirnica, Universidade Federal de Santa Maria, Santa Maria, Brasi[ Received September 2002 • Accepted April 2003 Abstract The interaction of chemicals with the container materials during heating for steri[isation was investigated, storing the components of parentera[ nutrition solutions individuaLLy in sealed glass ampouLes and in contact with a rub- ber stopper, and heating the system at 12! °C for 30 min. SubsequentLy, the aLuminium content of the solutions was measured by atomic absorption spectrometry (AAS). The assay was also carried out with acids, aLkaLis and some compLexing agents for A[. The containers were decomposed and also assayed for a[uminium. 30 different commerciaL solutions for parentera[ nutrition, stored either in glass or in plastic containers, were assayed measur- ing the aLuminium present in the solutions and in the container materials. The results of aLL investigated contain- er materials revealed an aLuminium content of 1.57% A[ in glass, 0.05% in plastic and 4.54% in rubber. The ster- iLisation procedure showed that even pure water was able to extract AL from glass and rubber, 22.5 + 13.3 pg/L and 79.4 _+ 22.7 pg/L respectiveLy, whiLe from plastic the a[uminium Leached was insignificant. The A[ reLeased from gLass ampouLes [aid between 20 pg/L for [eucine, ornithine and [ysine soLutions and 1500 pg/L for solutions of basic phosphates and bicarbonate; from rubber stoppers it reached LeveLs over 500 pg/L for cysteine, aspartic acid, gLutamic acid and cystine solutions, lon-exchange properties and influence of pH can explain the interaction of glass with some chemicals (saLts, acids and aLkaLis), but only an affinity for a[uminium could explain the action of some amino acids and other chemicals, as albumin and heparin, on glass and rubber, considering the a[umini- um reLease. Experiments with comp[exing agents for AL aLLowed to conclude that the higher the stability constant of the complex, the higher the A[ reLease from the container material Key words: aLuminium, contamination, glass, rubber, steriLisation, parenteraL nutrition Introduction ParenteraL nutrition (PN) is the administration of nutri- ents intravenously to patients that cannot be fed via gas- Dedicated to Professor Dr. Georg Schwedt on the occasion of his 60 m birthday. *Correspondence to: Denise Bohrer, Departamento de Quimica, Universidade Federal de Santa Maria, 97110-900 Santa Maria, RS Brasi[, Phone/Fax: +55(055) 2208870, E-mai[: ndenise@quimica.ufsm.br trointestinal tract. Products for parentera[ nutrition are commerciaLised in form of sterile solutions, and include eLectroLytes (Na, K, Ca, Mg, chloride, bicarbonate, acetate and phosphate), amino acids, carbohydrates (gLucose and poLyoLs), oLigoeLements (Cu, Zn, Cr, Mn), albumin, vita- mins and Lipids. ALthough the presence of A[ in pharma- ceutical products used parenteraLLy is known at Least for 18 years (I, 2), this is stiLL an unsolved problem today. AL has a toxic effect to children and adults with chron- ic renal failure and to patients on Long-term parenterat 0946-672X/03/17/02-%07 $15.00/0
`
`Eton Ex. 1012
`1 of 9
`
`

`

`108 D. Bohrer el at. nutrition (3-7). Moreover, AL can also be associated with impaired neurotogic development in preterm infants, which receive prolonged intravenous feeding (8-11). The American Societies for CLinical Nutrition (ASCN), for Parenterat and Entera[ Nutrition (ASPEN) (12) and for Pediatric Gastroenterology and Nutrition (ASPGN) (13) proposed to restrict the AL contamination of Large volume parenterats to a maximum of 25 pg/L, and although for small volume parenteraLs no maximum Level was proposed, the Label should inform how many IJg AL/L the product contains. Studies have been showing, however, that At found in these formulations is almost always over the pro- posed Limit (14-16). Several investigations (17-20) dealt with the presence of At in these formulations but none of them offered explanations for the origin of the At contamination. The results found by all authors revealed, however, that AL in these products can reach Levels over 5000 pg/L, and that even being produced by different manufacturers and in different countries, the distribution of the contamination in these products is the same: calcium gtuconate, phos- phate salts and otigoeLements are the most contaminated, followed by vitamins and albumin, while dextrose and amino acids are among the [east contaminated. Two possible sources can be suspected of contributing to the presence of At in these products: raw material and containers, not considering any step of the industrial pro- cessing of these formulations. In spite of the high proba- bility of the contamination by these sources, we found only two papers which related the contamination of albu- min (21) and gLuconate (22) solutions to their storage in glass containers. In previous works we investigated the influence of the storage time of the solutions in contact with the glass and plastic surface of the containers considering the shelf-Live of the products (23, 24), and the contribution of the AL present as impurity in the chemicals to the con- tamination of the solutions for PN (25). In the present work we investigated glass, plastic and rubber as container materials considering their interaction with the chemicals and possible A[ releasing during the heating for the steriLisation procedure. Experimental Apparatus A Varian SpectrAA-200 atomic absorption spectrometer equipped with a GTA-IO0 graphite furnace and an autosam- pLer (MeLbourne, AustraLia), a Trox class 100 cLean bench (Curitiba, BrasiL), a Phoenix AV 50 N 6281 autocLave (S~o Pauto, Brasi[), a Berghof BSB 939-IR sub-boiLing distiLLa- tion apparatus (Eningen, Germany) and a Digimed pHmeter D-20 (S~o Pauto, Brasit) were used. Reagents The water used throughout was distilled, deionised and further purified by a MiLLi-Q high purity water device (MiL- tipore, Bedford, USA). An AL standard solution containing 1000 mg/L At (Merck, Germany) was used to prepare the working standard solutions. HNO 3 (65%, 1.40 g/mL) from Merck was further purified by sub-boiLing distiLLation. Contamination control To avoid contamination, only plastic materials were used. ALL Laboratory ware (pipette tips, volumetric flasks, etc.) were immersed for at [east 48 h in a 10% (v/v) HN03/ ethanol solution and washed with MiLLi-Q purified water shortly before use. To avoid contamination from the air, all steps in the sample and reagent preparation were carried out in a class 100 dean bench. Procedures The assayed chemicals included besides salts, amino acids, glucose, albumin, heparin and some potyots used in parentera[ nutrition: acetic acid, hydrochloric acid, sodi- um hydroxide, sodium nitrate, sulfate and gLuconate, cal- cium chloride, magnesium chloride, zinc chloride, man- ganese chloride, copper chloride, chromium chloride, and the compLexing agents for A[ EDTA, NTA, oxalic acid and citric acid. ALthough heparin is not used for parenteraL nutrition, it was included in the experiment. The salts above cited that are not used in parenteraL nutrition were included in the experiment in order to compare the influ- ence of different cations and anions on the interaction with the container material. The substances were of "for analysis" quality and from different manufactures (Merck, Sigma, ALdrich, Riedet-de Hahn or Ajinomoto). For the analysis, solutions containing 0.1, 1.0 or 10.0% (m/v) of the substance were prepared in polyethylene volumetric flasks, and the AL content measured by flame or elec- trothermal atomic absorption spectrometry (FAAS or ETAAS, respectively), following the conditions described in Table 1. The chloride containing solutions were anal- ysed directly by FAAS or by ETAAS, after a treatment to separate the saline matrix according to a method earlier described (26), because chloride ions cause interferences on graphite furnace measurements. Solutions of amino acids were also prepared 0.1, 1.0 or 10.0% (m/v) according to their solubility or AL content. For the amino acids which solubility did not allow com- plete dissolution, HNO 3 was added to promote the solubit- isation. Heparin was dissolved to give a solution with the same concentration (5000 UI/mL) as the commercial product. At[ samples were prepared in triplicates and the mea- surement was carried out just after the preparation. Analysis of the containers Polyethylene ampoutes (10 and 20 mL), glass ampouLes (10 and 20 mL), bottles for amino acid solutions (100, 250 and 500 mL), and grey and red etastomeric closures for amino acid bottles, art from commercial products, were assayed for their At content. The glass containers were crushed into fragments about 1 mm in size and welt mixed. One hundred milligrams of the glass fragments was placed in a PTFE vessel with 5 mL of 48% (m/m) hydrofluoric acid and 5 rnL of water, and heated in a domestic microwave oven at a minima[ power J. Trace Etern. Med. Biol. 17/2 (2003)
`
`Eton Ex. 1012
`2 of 9
`
`

`

`A[uminium in parenterals 109 (:174 W) for :10 min. The procedure was repeated twice. After the total dissolution of the sample the volume was completed to 200 mL with water and A[ measured by FAAS. The plastic ampoules were cut into small pieces and a portion of :1 g was mixed with approximately 2 g of KNO 3 and heated in a platinum crucible at 250 °C for 3 hours. The ash was dissolved in 4 mL of 96% (m/m) H2SO 4 and diluted to 50 mL with water. AL was measured by FAAS. The decomposition of the rubber stoppers was carried out by heating 0.2 g of the rubber in a platinum crucible at 600 °C for 2 hours. The ash was dissolved by addition of 6 mL hydrofluoric acid :1 + :1 with water, and gently heated. The final volume was made up to :100 mL with water. The At measurement was carried out by FAAS. A blank experiment was carried out using only the reagents and including all steps of the three different decomposition procedures. Preparation of the samples for moist heat sterilisation As the analysis revealed a very low level of A[ in plastic containers, the heating procedure was carried out only with glass containers; moreover, the usual procedure for steritising parentera[s in plastic containers is not through heat but by using ethylene dioxide or radiation (27). :10 mL new glass ampoules for injectabtes (Schott do BrasiL) were titled with a 0.5% (m/v) solution of each substance selected for the assay. The ampoutes were sealed and heated at :121 °C in an autoclave for 30 rain. After cooling down they were opened and the A[ content Table I. Atomic absorption spectrometer operating conditions. Instrument Wavelength (nm) 309.3 Lamp current (mA) 10 Spectra[ stir width (nm) 0.5 Background correction Deuterium tamp Flclme Acetylene/nitrous oxide Graphite furnace Pyrolytic coated furnace with L'vov ptatform and argon as purge gas Temperature programme Step Temperature Time Gas flow (°C) (s) (L/min) i 85 5 3.0 2 95 35 3.0 3 120 I0 3.0 4 1100 20 3.0 5 1100 10 3.0 6 1100 2 0.0 7* 2500 0.7 0.0 8* 2500 1 0.0 9 2600 1 3.0 *read of the solution measured by FAAS or ETAAS. The AL atready present in the substances was discounted from the A[ measured. The experiment was carried out in triplicates. TweLve ampouLes fiLLed only with pure water were also submitted to the steriLisation procedure. Three new ampoules were assayed individuaLLy for their AL content as described in the previous section. Closure experiments Among the investigated commercial products for parenter- aL nutrition only the amino acids and albumin solutions were stored in bottles with etastomeric closures. Therefore the assay was carried out with amino acid solutions, and also with the compLexing agents EDTA, NTA, citric acid and oxalic acid, and with hydrochloric acid, acetic acid and sodium hydroxide. A rubber stopper (grey) was stored individually in 200 mL of each solution in a poLypropyLene flask and submitted to the steriLisation procedure as described above. The assay was also carried out with the closure in contact with pure water. The experiments were carried out in triplicates. Three pieces of these closures were also assayed to determine their AL content. Analysis of commercial products Commercial solutions of salts, glucose, heparin and albu- min and formulations containing amino acids were anal- ysed. With the exception of 10% and 20% NaC[, 10% and 19.1% KCL and ampouLes of oLigoelements (Zn, Cu, Cr and Mn salts), the AL was measured directly by FAAS or ETAAS (some samples had to be diluted to have their At concen- tration in the range of the calibration curves). The AL pre- sent in these salt solutions was determined according to the method mentioned above, after matrix separation (26). For all products, at Least 3 samples of the same batch were analysed, and the reported results correspond to the mean value calculated from these replicates. Results Containers and closures The analysis of containers and closures used for storing of commercial parentera[ solutions showed that At was pre- sent in all three types of materials: plastic, glass and rub- ber (Table 2). However, whereas in plastic containers only a small amount was found, probably from catalysts used for plastic potymerisation, At is present in high levels in glass and rubber. Aluminium oxide is added to glass to improve its chemical resistance, so that glass for par- enterals contains normally from 2.6 to 6.6% AL203 (28). Rubber can contain AI as well. Hydrate or calcined alu- minium silicate (clay) is used as filler in e[astomeric materials to improve hardness, abrasion resistance and density (29). In spite of the fact that other materials can be used as filler, AI was found in all rubber analysed in this work. Sterilisation The ampoutes used for this experiment presented an AL [eve[ of 2.14 + 0.47%. J. Trace E[em. Med. Biol. 17/2 (2003)
`
`Eton Ex. 1012
`3 of 9
`
`

`

`110 D. Bohreret aL TabLe 2. A[uminium present as contaminant in commercial parenteraL solutions and in container materials. ALL solutions were within their guaranteed period of sheLf-Life. Product Alin solution Container ALin container _+SD (pg/L) (%) NaC[ 20% 149 + 10 glass ampoule 1.43 13 + 4 poLyethyLene 0.04 KCL 10% 68 _+ 6 glass ampoule 1.25 23 + 5 poLyethyLene 0.06 Magnesium sulfate 50% 560 + 85 glass ampoule 1.25 380 + 288 glass ampoule 1.43 Sodium acetate 2 meq/mL 45 _+ 7 glass ampoule 2.14 17 + 8 poLyethyLene 0.05 Potassium phosphate 2 meq/mL 988 + 76 glass ampoule 1.98 1325 + 142 glass ampouLe 2.45 Sodium phosphate 0.5 rno[/L 933 + 88 glass ampoule 1.65 879 + 203 glass ampoule 2.05 CaLcium gLuconate 10% 5621 + 1165 glass ampoule 1.51 5960 +_ 62 glass ampoule 2.21 Sodium bicarbonate 8.4% 833 + 141 glass bottle 0.99 922 + 102 gLass bottle 1.03 O[igoe[ements a 1129 + 33 gLass ampoule 2.14 O[igoeLements b 1854 + 744 glass ampoute 2.21 Amino acids 10% 164 + 6 glass bottle 0.82 rubber closure 3.91 Amino acids 10% 116 + 30 glass bottLe 0.76 rubber closure 4.23 Amino acids 10% 93 + 23 glass bottle 0.84 rubber closure 5.34 Amino acids 10% 65 + 13 glass bottle 0.89 rubber closure 5.70 Amino acids 10% 23 _+ 8 plastic bag 0.01 GLucose 50% 13 +_ 1 poLyethyLene 0.04 293 _+ 14 glass ampouLe 1.15 GLucose 25% 9 _+ 3 poLyethyLene 0.08 370 _+ 23 glass ampoule 1.87 ALbumin 20% 644 + 58 glass flask 0.67 rubber cLosure 4.06 149 + 23 glass flask 0.66 rubber closure 3.99 Heparin 5000 U]/mL 732 _+ 23 gLass ampoule 2.88 738 _+ 54 glass ampoule 3.03 aComposition: 22.0 mg ZnSO 4, 6.3 mg CuSO 4, 2.46 mg MnSO 4, 102.5 pg CrCL 3 per ampoule bComposition: 8.8 mg ZnSO 4, 1.60 mg CuSO 4, 123.04 pg MnSO 4, 20.50 pg CrCL 3 per ampoule SuppLiers: Abbott, Afiston, Aster, Baxter, Behring, B. Braun, Darrow, Fresenius, Fuji- sawa, Gayer, HaLex Istar, Hypofarma, Santisa, Roche, Zena[b The first step of the investigation on AI reLease from glass during the steri[isation procedure was carried out with pure water. The resuLts of ampou[es filled only with water showed that even pure water is capa- ble to extract AL from the glass surface (Fig. 1). A mean value of 22.5 + 13.3 pg/L A[ was measured in the water samples after heating for stefi[isation. Fig. 2 shows the A[ extracted from glass ampoules filled with solutions of the different investigated substances during the moist heat procedure. Among the salt solutions, those hav- ing a basic character showed the high- est interaction. Whereas chlorides and sulfates extracted a mean of 200 pg/L A[, basic phosphates and bicarbonate extracted more than 900 pg/L AI (see also Fig. 3a). AL released by the Na0H solution (4500 pg/L) confirms the stronger interaction of the alkaline solutions with glass. Among the amino acids, there were some that did not extract A[ from the glass ampoule, and others that were abLe to extract a large amount of AL Solutions of cys- tine, cysteine, aspartic acid, g[utamic acid and tyrosine extracted more than 200 pg/L respectiveLy, whereas soLu- tions of lysine, ornithine and leudne extracted as much At as pure water. Fig. 2 also shows AI extracted by action of complexing agents and the polyols mannito[, sorbito/and xy[ito[. Fig. 3 shows AL extracted by action of salts, grouped according to their constituents: different anions having the same cation in Fig. 3a, and differ- ent cations bound to the same anion in Fig. 3b. For this comparison, saEs not used in parentera[s were included; the aim was to evaluate the action of cations and anions respectively. Con- AI (Fg/L) loo-/ 80 -j 60-/ / 40-" / 20-" 1 2 3 4 5 6 7 8 9 10 11 12 Fig. 1. ALuminium extracted from glass ampou[es during steri[isation at 121 oc for 30 min by action of pure water. 3efore sterilisation 3. Trace Etem. Med. Biol. 17/2 (2003)
`
`Eton Ex. 1012
`4 of 9
`
`

`

`A[uminium in parentera[s 111 AI (Fg/L) / 800- / 600- / 400 - / 200- o AI (#g/L) after efore 800 600 [ m ~ T I i T ~ ] 1 1 l l ] i L & .o AI (r~g/L) ' after before 400 200 o Heparin: 5000 UI/mL after ._ efore Fig. 2. A[uminium extracted from glass ampoutes during steri[isation at 121 °C for 30 min by action of so[ut]ons of different substances at a concentration of 0.5% (m/v). Resu[ts are a mean of three replicates. After = after steri[isation, before = before steri[isation. J. Trace ELem. Med. Blot. 17/2 (2003)
`
`Eton Ex. 1012
`5 of 9
`
`

`

`112 D. Bohrer et at. 1800 1200 (a) • iii 600 chloride acetate nitrate iii ii~iiiiiiiii~( iiiiiii!!~!,li!,iill/~ ii iiiiiiii sulfate phosphate gluconate bicarbonate phosphate rnonosodium disodium 1800 1200 600 O~ Na (b) K Zn Ca Mg Mn Cu Cr Fig. 3. Comparison of aluminium extracted from glass ampou[es by action of different cations and anions. (a) sodium salts of differ- ent anions, (b) ch[oride salts of different cations. 2000 1500 1000 AI 500 = I I !1 ii I I U I I I I I I I ] I II o#" 5000 #g/L~ ............. i N N o4 o after sterilisation oefore sterilisation Fig. 4. Aluminium extracted from grey rubber closure during steri[isation at 121 °C for 30 min by action of water, amino acids and comp[exing agents. Results are a mean of three replicates. Concentration of the solutions: 0.5% (m/v). 3. Trace Etem. Ned. Biol. 17/2 (2003)
`
`Eton Ex. 1012
`6 of 9
`
`

`

`ALuminium in parenterals 113 sidefing the anions, the dibasic phosphates, bicarbonate and g[uconate interacted strongest, whereas among cations, the higher extraction occurred in solutions of chromium and copper salts. Fig. 4 shows A[ extracted from the elastomeric closures by action of water, amino acid and comp[exing agent solu- tions. It can be seen that the same behaviour, observed for the steriLisation of amino acid and compLexing agent solutions in contact with glass containers, occurred with the closures. Whereas cystine, cysteine, aspartic acid and glutamic acid interacted with rubber and withdrew AL, almost all other amino acid solutions had the same effect as pure water, i. e. they did not play any role in extracting A[ from the material. Commercial so[uUons Table 2 shows the results of the analysis of some commer- cial products. GLuconate and phosphate salts were among the most contaminated as already found by other authors. Also the o[igoe[ement solutions were highly contaminated by AL UnfortunateLy, we did not find calcium gluconate, phosphates and oLigoelements commercialised in plastic containers, therefore no comparison between these prod- ucts stored in glass and pLastic containers was possible. The investigation of the other products showed cLearly, that the same formulation can be more or Less contami- nated by AL depending on the container materiaL. As can be seen in TabLe 2, all products stored in glass containers were more contaminated than the same prod- ucts stored in plastic containers. However, among the products in glass containers, different contamination lev- els were found, with the same distribution observed in solutions of these substances after the sterilisation proce- dure carried out in this work. Discussion The aim of this work was to show that the action of chem- icals on the container material could lead to an A[ release into the solution, accelerated by the heat during the ster- i[isation procedure. This investigation was carried out only with glass containers and rubber stoppers, because the AL content measured in plastic containers was too Low to contribute to the contamination of the solution. The results revealed that the heating of the solutions in con- tact with glass and rubber could Lead to a contamination of the products by AL However, the action of chemicals on glass and rubber depends on their nature; the three following properties of a solution can Lead to At release: pH of the solution, sim- ilarity between At and any cation in the solution, and affinity to AL The dissolving action of alkaline solutions on glass sur- faces is known (30). Hydroxyl ions react with silica, according to the reaction: 4OH- + SiO 2 --> Si04- + 2H20. As AI aLso exists as an oxide in the glass structure and reacts with hydroxyl ions in the same way as silica, 40H- + At203 --> AI02- + 2H20, AI ions can be released into aLkaLine soLu- tions by simple dissolution of the glass itself. TabLe 3 shows the amount of A[ extracted from glass during steril- isation of solutions of mono- and dibasic phosphates of potassium and sodium, sodium hydroxide, and bicarbonate and the pH of these solutions. It can be seen that the higher the pH, the higher the amount of A[ extracted. Phosphate ions, that have themselves a high affinity for A[ (see below), showed strong interaction only in solutions with alkaline character. Whereas both sodium and potassi- um monohydrogenphosphate solutions, with pH 10, extracted 1450 and 1550 pg/L A[ respectively, the dihy- drogenphosphate solutions, having pH 5, extracted only 200 pg/L AI. Another possible way of withdrawing A[ from glass is the also well known cation-exchange property of glass surfaces. When the action of different cations (chloride salts) is compared (Fig. 3b), it can be seen that it is not related to the pH, because the pH of all solutions was around or below 7. The effect could be due to a cation exchange process between glass and solution. ALl cations having an ionic radius similar to A[ could be exchanged by AL The strongest interaction occurred with chromium and copper ions, that have the smallest ionic radii among the investigated cations. The exchange of A[ by Cr can be attributed not only to the size but also to a similarity between the ions. Both are highly-charged trivaLent cations, that are easily hydro[ysed in aqueous solution but remain unhydrolysed at low pH, and present an octahedra[ array of water molecules. The third possible way of withdrawing AL from glass, and in this case as well from rubber, is through complex formation by action of ligands showing affinity for AL According to the Pearson's theory of hard-soft-acid-base (HSAB) [igands (31), AL is a hard acid and therefore forms the most stable complexes with hard bases, e.g. phos- phate ions (P043-, HP042-, H2P04- ) and other oxygen donor groups. As the HSAB theory indicates favourabLe combinations of these [igands and metal ions, it could explain the high affinity of A[ for phosphate and gLu- conate ions. Besides this affinity, the pH also promoted the AL extraction by solutions of some phosphates, as already mentioned. The effect of [igands could also be seen in solutions of comp[exing agents. The conditions (temperature and time of contact) having been similar, it is possible to relate the action of these substances to their affinity for AL With an increasing stability constant of the A[ complexes, the amount of A[ released, either from glass or rubber, into the solution of the corresponding compLexing agent increased too. The log K of the A[ complexes are: EDTA Table 3. A[uminium released into solutions of different anions during the steri[isation procedure. Anion Cation pH solution At released (pg/L) OH- Na 12 4300 HC032- Na 9 1300 HP042- K 10 1550 HP042 Na 10 1450 HzPO ~ K 5 210 HzPO ~ Na 5 200 J. Trace E[em. Ned. Biol. 17/2 (2003)
`
`Eton Ex. 1012
`7 of 9
`
`

`

`114 D. Bohrer et al. 16.5, NTA 11.1, citric acid 7.9, ma[ic acid 7.6 and oxalic acid 6.1 (32). The action of some amino acids could also be explained in this way, considering that their effect could not be caused by the pH of the solutions, which was not higher than 8. Amino acids with polar side chains are able to form chelate complexes with metallic ions and extracted more A[ from glass and rubber than pure water. In spite of the fact that stability constants of complexes between A[ and the majority of amino acids are not known, the ability of amino acids to complex metal ions could be one of the reasons for the high AL contamination rates. In the Litera- ture, only the stability constants of A[ complexes of aspar- tic and g[utamic acids are listed. Cystine, on the other hand can be considered an exception: no metallic com- plexes with cystine are reported (32, 33), cystine showed, however, a high interaction with glass and rubber. Among the other investigated substances albumin, heparin, glucose, mannito[, sorbitoL and xyLitol, there seems to exist an affinity for A[ too. Albumin is a protein responsible for the unspecific transport of metals in the human circulatory system, therefore it should be able to form stable complexes with metallic cations. The other substances in this group have either -OH or -COOH groups. These oxygen donors can therefore act as basic [igands binding meta[[ic cations such as A[. Conclusion From the results obtained it can be concluded that, with- out other sources of At in the industrial processing, the steri[isation of PN solutions by their heating in glass con- tainers or in containers with rubber stoppers, may be mainly responsibte for the presence of At in these soLu- tions. However, the A[ release occurs only under certain conditions, e.g. an elevated pH, the presence in the soLu- tion of cations with similar properties to A[ or Ligands able to bind AL SoLutions that meet these characteristics should be better stored in containers made of materials different from grass and rubber, or should not be submitted to a steriLisation process by heating, or otherwise shoutd be stored in glass containers/containers with rubber closures that would not contain AL in their composition. AcknowLedgements The authors are grateful to CNPq (ConseLho Naciona[ de Desenvo[vimento TecnoLOgico) PADCT N o: 62.0261/97-8, Brasi[ for financial support, and to the ParenteraL Nutri- tion Service of the Santa Maria University Hospital for the commercial formulations. References 1. KLein GL, Alfrey AC, MiLler NL, Scherrard D3, Hazlet TK, Ament ME, and Coburn 3W (1982) ALuminum Loading during total parentera[ nutrition. Am. 3. CLin. Nutr. 35:1425-1429 2. KLein GL (1995) ALuminum in parenteraL solutions revisited- again. Am. 3. Ctin. Nutr. 61 (3): 449-456 3. Greger JL, and SutherLand JE (1997) ALuminum exposure and metabolism. Crit. Rev. Ctin. Lab. Sci. 34 (5): 439-474 4. ALfrey AC (1993) Aluminum toxicity in patients with chronic renal failure. Ther. Drug Monit. 15 (6): 593-597 5. Ott SM, Maloney NA, Klein GL, Alfrey AC, Ament ME, Coburn 3W, and Sherrard DO (1983) Aluminum is associated with low bone formation in patients receiving chronic parentera[ nutrition. Ann. Intern. Med. 98:910-914 6. Vargas JH, Klein GL, Ament ME, Ott SM, Scherrard D3, Horst RL, Berquist WE, Alfrey AC, S[atopolsky E, and Coburn 3W (1988) Metabolic bone disease of total parenteral nutrition: course after changing from casein to amino acids in par- enteral solutions with reduced aluminum content. Am. J. CUn. Nutr. 48:1070-1078 7. Sedman AB, Klein GL, and Merrintt RJ (1_985) Evidence of aluminum loading in infants receiving intravenous therapy. N. EngL J. Med. 312:1337-1343 8. Bishop NJ, Morley R, Day JP, and Lucas A (1997) Aluminum neurotoxicity in preterm infants receiving intravenous-feed- ing solutions. N. EngL J. Med. 336 (22): 1557-1561 9. KLein GL (1998) Metabolic bone disease of total parentera[ nutrition. Nutrition 14:149-152 10. Koo WW (1996) Laboratory assessment of nutritional metabol- ic bone disease in infants. C[in. Biochem. 29:429-438 11. Drisco[[ WR, Cummings JJ, and Zorn W (1997) Aluminum toxicity in preterm infants. N. Engl. 3. Med. 337:1090-1091 12. Klein GL, Alfrey AC, Shike M, and Scherrard D3 (1991) Par- entera[ drug products containing aluminum as an ingredient or a contaminant: response to FDA notice of intent. Am. 3. C[in. Nutr. 53:399-402 13. Klein GL, Leichtner AM, Heyman MB, and the Patient Care Committee of the North American Society for Pediatric Gas- troentero[ogy and Nutrition (1998) Aluminum in large and small volume parentera[s used in total parenteral nutrition: response to the food and drug administration notice of pro- posed ruLe by the North American Society for Pediatric Gas- troenteroLogy and Nutrition. 3. Pediatr. Gastr. Nutr. 27: 457-460 14. Koo WW, Kaplan LA, Horn J, Tsang RC, and Steichen 3J (1986) Aluminum in parentera[ nutrition solutions - sources and possible alternatives. 3. Parentr. Enter. Nutr. 10: 591-595 15. Berner YN, Shuler TR, Nielsen FH, FLambaum C, Farkouh SA, and Snike M (1989) Selected uEratrace elements in total parentera[ nutrition solutions. Am. J. Clin. Nutr. 50: 1079-1083 16. Chappuis P, Poupon J, and Arnaud 3 (1991) About the FDA recommendation for aluminum in parenteral solutions Am. J. CLin. Nutr. 54 (5): 551-953 17. Recknage[ S, Br~tter P, Chrissafidou A, Gramm H3, Kotwas J, and R~sick U (1994) ParenteraL aluminium loading in criti- cal care medicine. Part I: Aluminium content of infusion solutions and solutions for parentera[ nutrition. Infusions- ther. Transfusionsmed. 21:266-273 18. Hawkins NM, Coffey S, Lawson MS, and Delves HT (1994) Potential aluminum toxicity in infants fed special infant for- mu[a. J. Ped. Gastr. Nutr. 19 (4): 377-381 19. P[uhator-Murton MM, Fedorak RN, Audette R3, Marriage B3, Yatscoff RW, and Gram[ich LM (1999) Trace element contam- ination of total parenteral nutrition. 1. Contribution of com- ponent solution. J. Parenter. Enter. Nutr. 23:222-227 20. Popinska K, Kierkus 3, Lyszkowska M, Socha 3, Pietraszek E, Kmiotek W, and Ksiazyk 3 (1999) ALuminum contamination of parentera[ nutrition additives, amino acid solutions, and Lipid emulsions. Nutrition 15:683-686 3. Trace Bern. Med. BioL. 17/2 (2003)
`
`Eton Ex. 1012
`8 of 9
`
`

`

`A[uminium in parenterats 115 21. Victor R, Chan AK, and Mattoon M (1988) ALuminum con- tamination in albumin solutions from glass storage. Transfu- sion 28:290-291 22. Frey OR, and Maier L (2000) Polyethylene vials of calcium gLuconate reduce aluminum contamination of TPN. Annals Pharmacother. 34:811-812 23. Bohrer D, do Nascimento PC, Binotto R, and Pomb[um SC (2001) Influence of the glass packing on the contamination of pharmaceutical products by aluminium. Part I: Salts, glucose, heparin and albumin. J. Trace E[em. Med. Biol. 15:95-101 24. Bohrer D, do Nascimento PC, Binotto R, and Carlesso R (2001) Influence of the glass packing on the contamination of pharmaceuticaL products by a[uminium. Part II: Amino acids for parentera[ nutrition. J. Trace E[em. Med. BioL 15: 103-108 25. Bohrer D, do Nascimento PC, Binotto R, Becker E, and Pomb[um S (2002) Contribution of the raw material to the aluminum contamination in parentera[s. J. Parenter. Enter. Nutr. 26 (6): 382-388 26. Bohrer D, do Nasciment