Aluminum in Large and Small Volume Parenterals Used in Total Parenteral Nutrition: Response to the Food and Drug Administration Notice of Proposed Rule by the North American Society for Pediatric Gastroenterology and Nutrition

BACKGROUND Aluminum was first recognized as a cause of low-turnover osteomalacic bone disease and encephalopathy in patients with uremia approximately a quarter century ago (1). By the late 1970s, low-turnover osteomalacia and bone pain, but not encephalopathy, were identified in adult patients with normal renal function who were receiving total parenteral nutrition (TPN) (2). By 1982, aluminum contamination of TPN solutions was identified, notably in the casein hydrolysate, which was the protein source at the time; aluminum was found in urine, blood, and bones of patients receiving these solutions (3). In histochemical stains of bone from TPN patients, aluminum was found to be deposited at the mineralization front, where new bone forms (4), a site identical to that seen in patients with uremia (5). The dynamic histomorphometric analyses of bone biopsy specimens from adult patients with uremia and those receiving long-term TPN treatment produced the identical finding: The quantitative aluminum concentration at the mineralization front of bone was inversely correlated with the rate of bone formation (4). Substitution of deionized water for dialysis and the use of non-aluminum-containing phosphate-binding gels in patients with uremia and crystalline amino acids for casein hydrolysate in adult patients receiving TPN led to gradual resolution of the bone pain and improved bone formation (6). In 1985, Sedman et al. (7) published the first report of aluminum loading in the serum, urine, and bones of preterm infants supported for 3 weeks by TPN (7). Sources of aluminum contamination were reported to be the calcium and phosphate salts, albumin and heparin. Moreover, this group first described aluminum contamination of infant formulas (7), although the significance of this latter finding is still controversial. Koo et al. (8-10) found that aluminum accumulated at the mineralization front in the bones of premature infants (8). Additional sources were identified, such as multivitamin infusates and trace element solutions (9), both used regularly in TPN therapy. Because of the inability to apply standard histomorphometric techniques to the bones of preterm infants, the actual toxicity of aluminum to the bones could only be inferred, although a case report in 1989 identified the occurrence of hypocalcemia when aluminum was chelated out of bone by use of deferoxamine (11). The presumed but not documented sequence of events was that the aluminum at the mineralization front interfered with bone uptake of calcium, and that once the aluminum was chelated from bone, the osteopenic bone took up calcium at the expense of serum calcium concentration in a situation analogous to the "hungry bone syndrome" of primary hyperparathyroidism (12). However, aside from that case report, no further evidence of aluminum toxicity to preterm infants was identified until 1997 when Bishop et al. (13) showed that preterm infants randomized to receive standard TPN containing 45 µg aluminum/kg per day had a lower score on the Bayley Mental Development Index at age 18 months than did age-matched infants who received a TPN solution that contained 4 to 5 µg aluminum/kg per day for a comparable period, 92 ± 20 (standard deviation) compared with 102 ± 17, respectively. In all 157 preterm infants without neuromotor impairment, exposure to aluminum for longer than 10 days resulted in a higher fraction of those who received the standard (high-aluminum) solution having mental development index values below 85 points, 38% compared with 17% (p = 0.03). Thus, although the encephalopathy documented in uremic adults and older children did not develop in the preterm infants, neurologic impairment occurred that could cause difficulties later in life. Preterm infants are more likely to retain aluminum because aluminum administered intravenously bypasses the normal gastrointestinal barrier to absorption of this element and is 95% bound to circulating plasma proteins, chiefly transferrin (14,15). As a result, only 5% of circulating aluminum is ultrafilterable by the kidneys (14). Additionally, renal function is developmentally impaired in preterm infants and does not reach maturity until approximately 34 weeks of gestation (7). Therefore, preterm infants have an additional burden of the inability to excrete the aluminum load that can be filtered. Furthermore, aluminum loading may be difficult to detect in preterm infants because of several factors, including inadequate intake of calcium and phosphate that could produce osteopenia and hypoxic or hemorrhagic insults to the brain that could cause neurologic dysfunction and mask aluminum-associated problems. The U. S. Food and Drug Administration (FDA) has been investigating this problem since 1986. In 1990 it published an intent to propose and request for information (24). The American Society for Clinical Nutrition and the American Society for Parenteral and Enteral Nutrition responded with a combined statement (16). The American Academy of Pediatrics Committee on Nutrition, independently of the FDA and the requests for public comment, issued its first policy statement on aluminum toxicity in pediatrics in 1986 (17) and updated this policy again in 1996 (18). In both instances rigorous control by the FDA of aluminum contamination of parenteral solutions and infant formulas was recommended. The following constitutes our official response to the FDA's current notice of proposed rule, published in the Federal Register of January 5, 1998 (19). CURRENT STATUS OF ALUMINUM CONTAMINATION OF PARENTERAL SOLUTIONS The most recent development has been the confirmation of aluminum contamination outside of the United States. A two-part series has identified aluminum loading as a problem in parenteral nutrition solutions in Germany (20,21). The concentration of aluminum in TPN solutions continues to be highly variable. In preliminary studies in 1991, aluminum in the TPN solutions delivered to the neonatal intensive care nursery at the University of Texas Medical Branch contained 327 ± 130 µg/l, which would deliver approximately 30 µg/kg per day to a preterm infant (22). However, in 1994, infants in the intensive care nursery at that same institution were receiving approximately 5 µg aluminum/kg per day (23). At present, there is no method for determining whether large- or small-volume parenterals are manufactured with lower aluminum content. Therefore, the need to ensure some consistency in the aluminum content of TPN solutions is clear. SUMMARY OF THE FDA PROPOSAL The FDA is proposing to amend its regulations to add certain labeling requirements concerning aluminum in large- and small-volume parenterals used in TPN (19). The FDA is also proposing to specify an upper limit of aluminum permitted in large-volume parenterals and to require applicants to develop and submit for approval to the FDA validated assay methods for determining aluminum content in parenteral drug products. The FDA is proposing these requirements because of evidence linking the use of parenteral drug products containing aluminum to morbidity and mortality among patients on TPN therapy, especially premature infants and patients with impaired kidney function. SPECIFIC COMMENTS ON THE FDA PROPOSAL We agree with the FDA's view of the problem posed by aluminum contamination of solutions used for parenteral nutrition therapy and concur with the need to impose controls over the process. The FDA proposes to restrict the aluminum contamination of large-volume parenteral solutions (such as crystalline amino acid solution or dextrose and water) to a maximum of 25 µg/l. We support this proposal and believe that this should be readily achievable for all manufacturers of large-volume parenteral solutions, because most currently meet the standard. The FDA has data that most large-volume parenterals tested are already in compliance with this rule. If this requirement is not enforced, large-volume parenterals can continue to be made that could impose intermittent large aluminum loads in patients, with concomitant toxicity. We support the proposal that the package insert state that the drug product contain no more than 25 µg/l. For small-volume parenterals (including calcium and phosphate salts, the chief sources of contamination with aluminum in TPN solutions, multivitamins, heparin, and trace element solutions), we emphatically support the FDA proposal to require that the product's maximum level of aluminum at expiry (the date of expiration) be stated on the immediate container label of the small-volume parenteral and pharmacy bulk packages. The statement on the label is proposed to read: "Contains no more than XX µg/liter." The maximum level of aluminum at expiry would be expressed as the highest level of batches produced during the past 3 years, the highest level for the most recent five batches, or the maximum historical level, but only until completion of production of the first five batches after the rule takes effect. The FDA would also require a warning statement in the package insert that states that aluminum toxicity may occur in premature neonates (who have developmentally impaired renal function but who also require large amounts of calcium and phosphate infusions that contain aluminum) and in older patients with impaired renal function. In our opinion it is highly desirable for the FDA to implement these regulations as a first step toward limiting aluminum content of small-volume parenterals. In its response to the FDA notice of intent in 1990 (24), the combined working group of the American Society for Clinical Nutrition and the American Society for Parenteral in small-volume parenterals be initially restricted to the lowest published levels. The FDA has chosen not to require this, because a safe lower limit had not been established. Although we acknowledge the lack of a documented safe lower limit of aluminum in parenteral solutions, the reduction of aluminum levels would be hastened by such a requirement, pending revision as newer information becomes available. The FDA is also concerned about the daily amount of aluminum received by patients with impaired kidney function. The FDA cites the study by Bishop et al. (13) and lists 4 to 5 µg/kg as a possible safe upper limit for aluminum intake from TPN solutions. In their response to the 1990 FDA notice of intent, the American Society for Clinical Nutrition/American Society for Parenteral and Enteral Nutrition Working Group (16) proposed three working definitions of aluminum intake: safe, unsafe, and toxic. Safe was defined as the amount of aluminum taken in per patient per day that did not result in aluminum accumulation in tissues or fluids and was associated with no toxic effects. Thus the study of Heyman et al. (25) in 1986 put the safe limit of intake at 1 to 2 µg/kg per day for adults and adolescents. Unsafe was defined as the amount of aluminum intake per patient per day that resulted in tissue loading but no documented toxicity. The studies of Sedman et al. (7) and Koo et al. (8-10) were the references for the 15 to 30 µg/kg per day taken in by premature infants in the mid 1980s. Finally, toxic was defined as the amount of aluminum taken in per patient per day that resulted in tissue loading and symptoms of toxicity. Data from studies of TPN patients at the University of California Los Angeles defined the toxic effects of aluminum on bone formation; in these patients, TPN-associated bone disease developed after they received 60 µg aluminum/kg per day (3). Although we support the FDA proposal to include a warning label in the small-volume parenterals package insert that states that aluminum may be toxic, we think that health care professionals should have our best estimate of a toxic aluminum load. Eliminating these definitions from the rule does not constitute a best estimate. However, the newer information should modify the previous definitions of safe, unsafe, and toxic. Thus, the recent work of Bishop et al. (13) which demonstrated developmental delay in preterm infants receiving TPN therapy providing 45 µg aluminum/kg per day suggests that the lower limit for toxic effects in infants is at least at the aforementioned amount, and the previously published loads of 15 to 30 µg/kg per day reported in standard TPN solutions given to premature infants begin to approximate toxic levels. Bishop et al. (13) and Klein et al. (23) found no evidence for impaired neurologic development or impaired bone formation at parenteral aluminum intakes below 6 µg/kg per day. However, urinary aluminum content in the preterm infants receiving 2 to 6 µg aluminum/kg per day in their TPN solutions is comparable to that found by Sedman et al. in 1985 (7) (unpublished data), suggesting that aluminum continues to accumulate in tissues. If the FDA chooses to discuss this topic in the final rule, we suggest wording the statement as follows: "No aluminum toxicity to the brain or bone of preterm infants has been documented with intakes below 5 µg/kg per day; however, tissue loading may still occur at that rate of administration to preterm infants." We also support the FDA proposal that manufacturers develop validated assay methods that are reviewed and approved by the FDA. We emphasize that these methods should be those that are in general use, such as flameless atomic absorption spectroscopy with a graphite furnace (26), and that the methods should be sufficiently sensitive to detect aluminum at the microgram (parts per billion) and not the milligram (parts per million) level, in that the biologic effects of aluminum appear at the microgram level. The actions of the FDA should be expanded to include biologicals that are used for parenteral administration and could be used in TPN therapy. Specifically, albumin, plasminate, and any other colloidal volume expanders should be so regulated. Albumin is probably the most commonly used of these substances and was reported 13 years ago to be contaminated with variable quantities of aluminum (7,27). In preterm infants, burn victims (28,29), and perhaps other patient groups who may receive large quantities, albumin may be a source of aluminum contamination. The industry literature has in fact identified several sources of aluminum contamination of albumin during processing (30), and removal of aluminum, or at least substantial reductions, has already been achieved by one company (29). In conclusion, we strongly endorse the current FDA proposal as one that is in the best interest of the public health and safety. We strongly encourage the FDA to expand this activity to biological products that are similarly used. Finally, we advocate that the FDA be more specific in the definition of aluminum toxicity by load. Thus, the FDA could more closely guide the health care professionals who compound and administer TPN solutions.