|
740 Am J Health-Syst Pharm—Vol 64 Apr 1, 2007 CLINICAL REPORTS Measured versus estimated aluminum content REX A. SPEERHAS, B.S.PHARM., CDE, BCNSP, is Nutrition Support Clinical Specialist, Department of Pharmacy; and DOUGLAS L. SEIDNER, M.D., FACG, CNSP, is Director, Nutrition Support Team, Department of Gastroenterology and Hematology, Cleveland Clinic, Cleveland, OH. Address correspondence to Mr. Speerhas at the Department of the calculated dose using the concentra tions reported by the manufacturer. Methods. Fifty 2.5-mL samples were prepared for analysis and were obtained from PN solutions prepared for adult and pediatric/neonatal patients. Intravenous large-volume parenterals were used as controls. Samples were sent to two dif ferent reference laboratories. Batch 1 in cluded 26 samples, 22 from PN solutions, 1 control, and 3 paired (duplicate) samples selected from the 22 PN bags. Because 8 of the samples were lost to analysis, a second batch of samples was included. Batch 2 consisted of 24 samples from 4 PN solutions and 2 controls. Batch 2 was com posed of four aliquots from each solution that were divided equally and sent to the two laboratories. Results. Twenty-three values were used in the statistical analysis. The results showed that only two of the adult PN solutions equaled or exceeded the threshold set by the Food and Drug Administration (FDA) for measured aluminum exposure. The measured concentration of aluminum for all six of the pediatric and neonatal solu tions met or exceeded the FDA threshold; however, this value was much lower than what had been estimated using the la beled aluminum concentration at expiry. Conclusion. In PN solutions expected to have a moderately high concentration of aluminum, the measured amount of alu minum was far less than the amount that would be estimated by calculation using the labeled concentrations of aluminum in each of the ingredients. Index terms: Additives; Aluminum; Con centration; Contamination; Control, qual ity; Food and Drug Administration (U.S.); Labeling; Metals; Nutrition; Pediatrics Am J Health-Syst Pharm. 2007; 64:740-6 January 2000, the Food and Drug Administration (FDA) published a regulation on aluminum in additives used in parenteral nu trient (PN) solutions.1 It states that the aluminum concentration of in jectables that are used to compound PN solutions must be placed on all product labels. The concentration values are the result of analysis of these solutions at expiration and, therefore, they note the maximal amount of aluminum possible. FDA also requires manufacturers to place a statement in the "Warnings" section of the package insert that says that aluminum in excess of 4–5 μg/kg/day may result in toxicity. This regulation became effective on July 26, 2004. PN solutions are known to have a relatively high concentration of aluminum compared with other i.v. solutions. The primary source of aluminum contamination in PN solutions before the 1980s was the casein hydrolysate solutions that were used as the source of nitrogen. The amount of aluminum in PN solutions has dramatically decreased since crystalline amino acid solutions have replaced casein hydrolysates.2 Currently, the chief sources of alu minum are the glass containers in which PN ingredients are contained. Currently, the major sources of aluminum in PN solutions include phosphate, acetate, and calcium glu conate salts. PN administration has been shown to induce aluminum tissue loading that results in bone disease and neu rologic impairment.3Aluminum may also contribute to the pathogenesis of PN-associated liver disease.4 Since aluminum is primarily eliminated in the urine, premature neonates whose kidneys are underdeveloped and pa tients with renal disease are the most susceptible to aluminum toxicity. Pharmacists compounding solu tions can now calculate the maximal amount of aluminum in each PN solution as a result of the FDA regu lation on aluminum labeling. Therefore, it becomes the responsibility of the pharmacist to notify the prescriber of any danger of aluminum loading as a result of PN administration. It is currently not known how the calculated aluminum content compares with the actual amount of aluminum contained in each parenteral nutrition formulation. This study was conducted to directly measure the aluminum concentration in a select number of PN solutions and to compare this value with the calculated dose using the concentrations reported by the manufacturer. In addition, the actual amount of aluminum being received by patients was determined. Methods The institutional review board of the Cleveland Clinic approved this study. Fifty 2.5-ml samples were prepared for analysis. Samples were obtained from PN solutions prepared for adult and pediatric/neonatal patients before being dispensed from the department of pharmacy. PN solutions are compounded from large-volume parenterals in bags and individual small-volume parenterals in glass vials. Particular PN solutions were selected because they were thought to be at greatest risk for aluminum contamination because of the high amounts of calcium, phosphate, and acetate salts that they contained. Intravenous large-volume parenterals (Baxter Healthcare Corporation, Deerfield, IL) were used as controls to determine the potential for contamination of the samples with the needles, syringes, and containers used in this study. Each bag was gently agitated before withdrawal of the sample under a Class 100 laminar flow hood. Samples were withdrawn using an unused 3-mL plastic syringe (Becton Dickinson & Co., Franklin Lakes, NJ) with an 18-gauge needle (Becton Dickinson & Co.) after cleaning the entrance site with 70% alcohol and were placed in a plastic container previously rinsed with 20% nitric acid and deionized water. Samples were sent to two separate reference laboratories. Laboratory A samples were sent in prerinsed 50-mL urine collection containers (Sage Products, Cary, IL), and laboratory B samples were sent in plastic tubes provided by the laboratory. Batch 1 included 26 samples, 22 from PN solutions, 1 control from a bag of i.v. 0.9% sodium chloride, and 3 paired (duplicate) samples selected at random from the 22 PN bags. (The paired samples were included to assess laboratory consistency.) Because 8 of the samples were lost to analysis (including 2 of the 3 paired samples) and the sole paired sample results were variable, it was decided that a second batch of samples was needed to answer the research question. Batch 2 consisted of 24 samples from 4 PN solutions (2 pediatric and 2 adult) and 2 controls (i.v. dextrose 5% in 0.45% sodium chloride). This second batch of samples was composed of four aliquots from each solution that were divided equally and sent to two laboratories (reference laboratories A and B). Two of the 24 samples in batch 2 were lost to analysis. Both laboratories analyzed aluminum by inductively coupled plasma mass spectrometry. The laboratories noted no interferences when testing PN solutions using this method. This method of analysis was recommended by FDA for use by manufacturers of PN ingredients for their analysis as a result of the FDA mandate.1 Data collected included the PN volume of each patient’s formula, the result of the aluminum analysis (microgram per liter), the estimated aluminum (microgram) in each PN solution according to the labeled concentrations of the ingredients, and the patient’s weight. The measured aluminum concentration multiplied by the volume of the PN bag from which the sample was obtained gave the measured aluminum total dose. The estimated aluminum total dose was the sum of the aluminum content of all components of a PN bag as calculated from the labeled aluminum concentration. Finally, a measured aluminum exposure and an estimated aluminum exposure were calculated using the patient’s weight and compared with the amount of aluminum that places a patient at risk for aluminum toxicity (>4–5 μg/kg/day). Results are expressed as mean ± S.D., median, and 25th and 75th percentile values. The Wilcoxon rank sum test was used to test the significance of the difference between medians of the measured and estimated aluminum values. A value of p < 0.05 was considered significant. SAS 9.1 software (SAS Institute, Cary, NC) was used to analyze the data. Results Eight of the 26 samples were destroyed in shipping and could not be analyzed, leaving 18 samples for analysis. The 8 destroyed samples included one pediatric/neonatal PN solution (sample 2), 5 adult PN solutions (samples 7, 20–22, and 26), and 2 paired adult solutions (samples 13 and 15). The only paired samples from batch 1 available as a quality control measure were the adult PN solutions designated as samples 10 and 11 in Table 1. Samples in Table 2 are listed consecutively as number 27 through 32, with letter designations corresponding to the laboratory where the sample was analyzed. Each sample was paired with another identical sample (listed as 1 and 2) at each laboratory and is identified with the sample number followed by the laboratory (A or B) and then 1 or 2. Two samples from batch 2 were destroyed in shipping and could not be analyzed. The destroyed samples were both paired samples from pediatric/neonatal PN solutions (samples 29 A-2 and 30 A-2). After eliminating lost samples from batches 1 and 2 and using the mean for all paired samples, there were 23 values used in the statisti cal analysis. Examination of the results showed that only two of the adult PN solutions (samples 11 and 14) equaled or exceeded the threshold set by FDA for measured aluminum exposure, even though 13 of the 14 adult solutions had an estimated concentration that exceeded this level. The measured concentration of aluminum for all 6 of the pediatric/neonatal PN solutions met or exceeded the FDA threshold; however, this value was much lower than what had been es timated using the labeled aluminum concentration at expiry. The variability in paired samples can be seen in Table 3. Laboratory A results were sensitive to 5 μg/L or greater, and laboratory B results were sensitive to 3 μg/L or greater. As suming that all PN solutions contain aluminum, aluminum concentra tions of zero were assigned a value of 2.5 μg/L for laboratory A and 1.5 μg/L for laboratory B. In 9 of the 11 pairs, the intralaboratory difference in the results was less than 20%. Only paired samples 10 versus 11 and 28 A-1 versus 28 A-2 showed a greater intralaboratory variation. Laboratory A reanalyzed samples 10 and 11 and the results were within 20% of the original assay and were therefore considered accurate according to the laboratory. Interlaboratory variability represents the percent decrease in the largest average of laboratory values compared with the smallest average. Observed interlaboratory variability was high, ranging from 4.2% to 94.7%, and was greater than 20% in five of the six results (83%). Because only 6 samples were available for comparison of values between laboratories, there was not enough power to perform statistical analysis. The average of paired samples was used for the final analysis. The adult and pediatric values were also combined for analysis. The median of the difference between the estimated and measured aluminum in adults was 17.5 μg/kg (p < 0.0001) and in pediatrics was 55.1 μg/kg (p = 0.0078). The median of the difference between the estimated aluminum and the measured aluminum for all values was 25.9 μg/kg (p < 0.0001). It should be noted that these results include sample 11 but not sample 10. These paired samples varied by more than 13-fold, but the difference between the actual and estimated concentrations for adults (calculation not shown) was found to be of the same statistical significance when sample 10 was used. The mean and median of the difference between measured and estimated aluminum values for the pediatric samples were much larger than for the adult samples. Thus, estimating the aluminum content of pediatric PN solutions would result in a larger misconception of the actual aluminum content. The aluminum concentration of all control samples was negligible, with results reported as <25 μg/L showing that contamination was not a problem with sample analysis. Discussion It is estimated that 3–5 mg of aluminum is ingested on a daily basis by most adults.5 The human body has natural protective barriers preventing aluminum intoxication. The skin and lungs are highly effective barriers to aluminum absorption as they block nearly all the aluminum that comes in contact with these surfaces. The gastrointestinal tract absorbs less than 1% of all aluminum that is ingested. Under normal circumstances, the small amount of aluminum that reaches the bloodstream is protein bound and is eliminated by the kidney.5 However, when administered intravenously, aluminum bypasses all of these natural barriers. Intravenous aluminum can result in tissue accumulation and toxicity, especially in patients with renal impairment (e.g., neonates, patients with renal disease). Patients who receive long-term parenteral nutrition (greater than three weeks) may also be at risk for aluminum overload.5 The FDA requirement to label all PN ingredients with their maximum possible aluminum content and to include a warning on the package insert of these ingredients regarding the threshold for aluminum toxicity was in part devised to make those who prescribe PN solutions aware of the risk of aluminum toxicity with this therapy. Aluminum contamination of pharmaceutical products can occur from several sources. Bohrer et al.6 showed that the raw chemicals of the PN ingredients were contaminated with aluminum, but the amount of aluminum noted in the PN ingredi ents was much more than what was expected. The same authors later showed that aluminum is leached into solution during the steriliza tion procedure when an ingredient is packaged in a glass container with a rubber stopper.7 Baumann8 studied individual chemical solu tions in glass and showed that glass is a major source of aluminum and that complex-forming anions, such as calcium and phosphorus, more actively extract the aluminum from the glass. Bohrer et al.9 also showed that the amount of aluminum contaminating PN ingredients pack aged in glass increased over time. Aluminum exposure from PN solutions can therefore be affected by the type of ingredients used to com pound the formula and how these ingredients are packaged and stored. The major reason the measured dose of aluminum was less than the estimated value was that the micronutrients used to compound these solutions were far from their date of expiry. For example, after our analysis had been performed, we contacted the manufacturer of the calcium gluconate that we used and learned that the product has a shelf life of 24 months and that the product that we had in stock had an expiration date ranging from 18 to 22 months before that time. Since the aluminum leaches out of the glass containers over time and, as a general practice, pharmacies use these ingredients early in their shelf life, it can be expected that the actual aluminum content of the fi nished PN solution would be less than the estimated aluminum exposure calculated from the labeled concentration.10 Another factor that may have affected our results is that a large number of samples were lost by one of the reference laboratories. Having said this, it seems unlikely that the results would have changed much because most of the paired samples from laboratory B did not vary to a great degree . The reported concentrations of aluminum at expiry of the ingredients used to compound PN solutions can help the pharmacist reduce the aluminum load provided by the PN solutions. Calcium, acetate, phosphates and multivitamin solutions from different manufacturers can vary drastically in aluminum concentrations. For example, calcium gluconate has a reported aluminum concentration of 12,500 μg/L by one manufacturer and 5,212 μg/L by a second manufacturer. Both of these products are packaged in glass containers. A manufacturer of potassium acetate that packages its solution in glass reports an aluminum concentration of 50,000 μg/L, while another manufacturer that packages its potassium acetate in plastic reports an aluminum concentration of 180 μg/L. Another way to reduce aluminum contamination of PN solutions is to alert prescribers to use sodium phosphates (labeled aluminum concentration = 12,500 μg/L) instead of potassium phosphates (labeled aluminum concentration = 62,500 μg/L) whenever possible. The use of sodium phosphates instead of potassium phosphates in one of the PN solutions for a pediatric patient in this study (sample 4, Table 1) could have reduced the aluminum concentration below the FDA threshold. Ingredients with the lowest amount of aluminum should be used to prepare all PN solutions, especially in pediatric patients since they are at the greatest risk for toxicity associated with exposure to this element. FDA-mandated labeling of PN ingredients with aluminum concentration at expiry can help pharmacists reduce the potential amount of aluminum contamination in PN solutions and in most cases provide PN solutions with <4–5 μg/kg/day. Purchasing activities by pharmacies that compound PN solutions should should include a comparison of the labeled concentrations of aluminum for each ingredient and consider pur chasing the ingredients with the low est labeled aluminum concentration. Good medical practice requires the pharmacist to report to the pre scriber any potential harm that may result from the use of PN solutions. Prescribers should be reminded that all PN solutions are contaminated with aluminum. In addi tion, pharmacists should discuss with prescribers ways in which a parenteral nutrition formula can be adjusted to reduce the level of aluminum exposure. Finally, pre scribers should be reminded that aluminum toxicity, which includes neurologic and metabolic bone diseases, should be considered in patients on long-term parenteral nutrition, even when the exposure dose appears to be acceptable. This is the first study to show that the estimated value of aluminum exposure from PN solutions is significantly higher than the measured value in both adults and pediatric patients. We found the estimated aluminum exposure to be 7–10 times greater than the measured exposure. Only two of the adult PN solutions had a measured aluminum exposure that exceeded the FDA threshold. On the other hand, the actual aluminum exposure in all of the pediatric/ neonatal PN solutions tested was >4–5 μg/kg/day. Changes in pharmacy purchasing habits and in the prescribing patterns of pediatric PN solutions may reduce the amount of aluminum in these PN solutions; however, aluminum contamina tion of pediatric and neonate PN solutions should not be underesti mated. It should be kept in mind that the PN formulas selected for this study were not chosen randomly but rather selected because the amounts of calcium, phos phate, and acetate prescribed for these patients were relatively high. In all cases, the dose prescribed was needed to meet the patient’s metabolic requirements for cal cium gluconate and phosphate salts and was in the accepted range for physicochemical stability of these compounds in a PN admixture. Conclusion In PN solutions expected to have a moderately high concentration of aluminum, the measured amount of aluminum was far less than the amount that would be estimated by calculation using the labeled concentrations of aluminum in each of the ingredients. References 1. Food and Drug Administration. Alumi num in large and small volume parenter als used in total parenteral nutrition. Fed Regist. 2000; 65:4103-11. 2. Klein GL, Ott SM, Alfrey AC et al. Alu minum as a factor in the bone disease of long-term parenteral nutrition. Trans Assoc Am Physicians. 1982; 95:155-64. 3. Klein GL. Aluminum in parenteral solu tions revisited—again. Am J Clin Nutr. 1995; 61:449-56. 4. Klein GL, Berquist WE, Ament ME et al. Hepatic aluminum accumulation in chil dren on total parenteral nutrition. J Ped Gastr Nutr. 1984; 3:740-3. 5. Sedman AB, Klein GL, Merritt RJ et al. Evidence of aluminum loading in infants receiving intravenous therapy. N Engl J Med. 1985; 312:1337-43. 6. Bohrer D, Do Nascimento PC, Binotto R et al. Contribution of the raw material to the aluminum contamination in paren terals. JPEN. 2002; 26:382-8. 7. Bohrer D, Do Nascimento PC, Binotto R et al. Infl uence of the glass packing on the contamination of pharmaceutical prod ucts by aluminum. Part III: interaction container-chemicals during the heating of sterilisation. J Trace Elem Med Biol. 2003; 17:107-15. 8. Baumann L. Glass is the major source of aluminum contamination for parentera lia. Krankenhauspharmazie. 1998; 19:71- 4. 9. Bohrer D, Do Nascimento PC, Binotto R et al. Infl uence of glass packing on the contamination of pharmaceutical prod ucts by aluminum. Part I: salts, glucose, heparin and albumin. J Trace Elem Med Biol. 2001; 15:95-101. 10. Frey OR, Maier L. Polyethylene vials of calcium gluconate reduce aluminum contamination of TPN. Ann Pharmaco ther. 2000; 34:811-2.
|
Corporate Partnerships
