| Sign In to gain access to subscriptions and/or personal tools. |
Safety and Stability of Lipid Emulsions—ResponseHarvard Medical School Boston, Massachusetts
The author responds: My August 2006 invited review entitled "Lipid Injectable Emulsions: 2006"1 was an update of a previous invited review published in the August 2001 issue of NCP from 5 years earlier.2 Dr Zaloga and Baxter cite in their Letter to the Editor 3 important issues that "require further clarification": (1) clinical safety, (2) characterization of lipid emulsions using globule size measurements, and (3) stability of lipid emulsions. I will address the main issues of their concern regarding the experimentally derived data in order to clarify these matters, including the serious concern about Intralipid (Baxter Healthcare, Deerfield, IL) as currently provided in plastic containers.
Zaloga et al mention the "vast clinical safety record" of lipids worldwide. There is no dispute that this is true when their lipids were exclusively packaged in conventional glass bottles; however, the safety record with respect to newly introduced Intralipid in plastic containers3 should not be commingled into this database, because the pharmaceutical aspects of the current emulsion are very different indeed. Shortly after the plastic-container-based version of Intralipid was introduced in the United States (third quarter of 2004), we analyzed the globule size distribution (GSD), and for the first time, we observed a large-diameter fat globule population, expressed as the volume-weighted PFAT5 values, to be 10–100 times higher than all others manufactured in glass, including when Intralipid was available in glass containers.4 Comparatively speaking, it was clear the newly introduced emulsion was, at the very least, pharmaceutically inequivalent to all previous parenteral nutrition (PN) emulsions investigated using the technique proposed in Method II of USP proposed Chapter <729>5 and is uniquely a very coarse IV emulsion. Thus, the safety record before June 2004 may not be applicable to the current dosage form. Zaloga et al cite several "flaws" in various infusion studies during the time in which we were clearly devising the best animal model, outcome variables, and study conditions. Because these were early studies, they were, of course, "exploratory." Having had numerous discussions with seasoned investigators and lipid injectable emulsions' manufacturers about globule sizing methods and toxicity of unstable parenteral emulsions since 1985, it became clear the problem was multifaceted. First, globule sizing techniques were variable and often company-specific according to experience. Second, without a quantifiable and validated method of globule size analysis of parenteral emulsions, how could toxicological studies (eg, LD50, median lethal dose) be conducted? Third, a reproducible unstable lipid emulsion model did not exist. Fourth, beyond intravascular accumulation, there were no known outcome parameters at that time. Those with intimate knowledge in the field have known these "deficiencies" for some time but accepted them in the absence of a clear alternative. From a clinical perspective, I found this situation unacceptable and decided to explore potential solutions. The data are now clearly emerging from our animal infusion studies that, in the Sprague-Dawley rat model, the liver, as with neonates,6 is the principal organ adversely affected by the infusion of unstable vs stable lipid emulsions, now confirmed by the PFAT5 criterion.7,8 For example, we have most recently presented an in-depth study of 2 groups (stable, n = 16 vs unstable, n = 17) of rats and showed that not only was there evidence of oxidative stress in liver tissues (according to MDA) and increased blood aspartate aminotransferase levels indicative of liver injury present, which confirmed earlier findings,7 but the with the application of an additional and highly sensitive hepatic injury marker, glutathiones-transferase, this indicator enzyme is also significantly higher in the unstable group.8 Further, C-reactive protein, a marker of the systemic inflammatory response, was also significantly higher in the group receiving the unstable emulsion by our PFAT5 criterion.8 Unexpectedly, Baxter's criticism of the infusion admixture used in these studies and its applicability in the United States as "not reflective of any clinically relevant product" suggests that they do not consider the clinical use of lipid-containing PN admixtures as low osmolality formulations intended for peripheral vein administration "clinically relevant." Further, when investigating the pathophysiological effects of infusing unstable lipid emulsions, it is not reasonable to have used a test formulation that is not "intrinsically unstable." Finally, is Baxter's position a reasonable one, vis-à-vis potential safety concerns, that "the unstable TNAs appeared to produce effects... seen experimentally using preconditioning... [and therefore] support the safety of lipid emulsions approved for human use"? Finally, our greatest clinical concern is the implications of this uniquely coarse formulation when used in the critically ill, premature infant population. The FDA specifically requires a black box warning on all lipid injectable emulsions approved for use in the United States, outlining the potential for significant morbidity and mortality due to their "poor clearance of IV fat emulsion." We were very concerned about the effect of the abnormal GSD on clearance of lipid injectable emulsions in the new plastic containers in this patient population. In fact, we recently presented preliminary evidence9,10 of significantly impaired plasma clearance in premature infants who received Intralipid in the new plastic containers (n = 72), compared with those who received the same Intralipid formulation packaged in conventional glass (n = 50). When linked to the study published in 1995 showing that hypertriglyceridemia in neonates was associated with significant increases in liver dysfunction and growth retardation,6 the clinical significance of this coarse emulsion in plastic is very concerning.
Zaloga et al claim that Chapter <729> "has raised significant controversies among experts in the field of lipid biochemistry," but I am aware of only select industry experts who consider this document controversial. Further, these issues were not brought forward as controversial after the first review was published in 2001,2 when the PFAT5 limit of <0.05% was first suggested. These guidelines have been repeated in subsequent publications.11 Baxter's claim that it "was intimately involved in development of the technology to quantify these parameters [mean globule size and PFAT5] for characterization of fat emulsions" is puzzling. In fact, I was invited in November of 1999 by the recently retired Dr Tom Barber to come to his lab and compare the technique of proposed Method II of USP <729> against their primary method at the time, laser diffraction. Dr Barber led the work at Baxter in the Chemical and Physical Sciences Division, including lipid emulsion stability, and is widely recognized as an expert in particle size analysis.12 In part, this work led to our collaborative (Baxter, Boehringer-Ingelheim, and B. Braun) publication in the International Journal of Pharmaceutics in 200113 where we made the following recommendations: "For commercial IVLE [IV lipid emulsions] from the manufacturer, we would suggest a mean droplet size (MDS) that does not exceed 450 nm, and an upper limit for PFAT >5 µm that does not exceed 0.05%. Based on our experience and the data shown in Table 1, these limits are easily met. Therefore, such a proposed range for both MDS and PFAT (> 5 µm) is pharmaceutically reasonable, in that the ranges are not only sufficiently broad, but are also likely safe, given the current use conditions and available data."
Finally, with respect to the question of "what droplet size would be
best to predict stability," the selection of 5 µm has been explained
on technical, physical, and physiologic grounds in great
detail.4 Further,
the question of "why use percent rather than quantity of globules at a
specific size to predict stability and safety" ignores the widely
acknowledged value of the "weighting" of emulsion droplet size
data. Number-weighting, as suggested by Baxter, assigns each droplet the same
weight irrespective of size, whereas volume-weighting takes into consideration
the volume of the droplet sphere (
Historically, most experts have viewed lipid injectable emulsions as either stable (fine) or unstable (coarse), and our group has applied the PFAT5 criterion as the principal indicator of stability, showing the differences between the 2 classifications to be several orders of magnitude. If Intralipid in plastic containers is, as claimed by Baxter, a stable but coarse native formulation, it doesn't meet the current generally accepted definition of stable or unstable. Our initial work in 1995 applied the PFAT5 criterion in assessing a wide array of total nutrient admixtures.14 The preliminary findings associating PFAT5 with visible evidence of phase separation was the first linkage between a quantitative analytical tool (light obscuration) and a crude but definitive outcome of emulsion instability (visually obvious phase separation). At the time when our study was being planned and conducted, iron dextran had been deemed compatible by Tu et al15 in 1992. Thus, on Baxter's renewed inspection and review of our 1995 study, as outlined in their letter, to suggest that the inclusion of iron dextran was a "critical flaw of the study" as it was "known to cause lipid emulsion instability" is puzzling and does not account for the historical evolution of PN and changes in guidelines over the years. That is, to cite A.S.P.E.N.'s 2004 version16 of the "Safe Practices for Parenteral Nutrition" guideline as being obvious to all of the devastating effects of iron dextran, and hence the dismissal of our primary research as flawed, is without merit. I contributed to the original section in the first A.S.P.E.N. guideline published in 199817 from our 1995 experience,14 which has not changed in the present 2004 version.16 Hence, the early work (1995–2000) by our group primarily focused on lipid injectable emulsions in total nutrient admixtures (TNAs), as it was and continues to be the most widely applied means of providing this dosage in the United States and probably so worldwide. Subsequently (2001–present), we have focused on animal models of infusion safety, as well as the ideal characteristics of the undiluted or native lipid injectable emulsion. In 2005, after assisting in the writing and adoption of Chapter <797> (on January 1, 2004) related to pharmaceutical compounding,18 I wrote an invited review because of the need to bridge parenteral compounding practices in accordance with pharmacopeial standards, with particular emphasis on the safety of PN admixtures.19 Subsequently, we have published 3 papers20–22 showing that, indeed, it is possible to compound TNAs that conform to the globule size requirements proposed in Method II of USP <729> (ie, PFAT5 <0.05%) with pharmacist-assigned beyond-use dates for up to 48 hours at room temperature. The only failures to date (ie, instability over time with PFAT5 >0.05%, coupled with physical evidence of instability) have occurred with TNAs made from Intralipid in plastic containers.22 Additional evidence has also shown that the undiluted emulsion in plastic, when given as a simulated neonatal infusion, is also less stable than US lipids packaged in glass.23 In conclusion, Baxter has provided their arguments and references and their personal experience regarding Intralipid injectable emulsion packaged in plastic. However, we have looked at many of the world's nutrition lipid emulsions13 and have not found any that even approach the PFAT5 levels seen with the new Intralipid plastic product. Although the coarse Intralipid product in plastic may be stable (as such) on the shelf, once opened and used under typical clinical conditions (ie, extemporaneously prepared as either a TNA for adults or a syringe dosage form for neonates), the coarse dispersion is less stable than the conventional fine emulsion we have grown accustomed to using in our patients. We have provided evidence for safety concern in a series of animal models with the use of these formulations and preliminary evidence that merit serious clinical concern in retrospective human data. The fact that hypertriglyceridemia is seen in neonatal usage, a finding specifically identified in the black box warning of the package insert, should elicit considerable unease among clinicians about this formulation in its new container.
* Dr Driscoll is a consultant or researcher in the area of lipids for Astra Zeneca, B. Braun, Biolink, GlaxoSmithKline, Hospira, and the Medicine Company companies. 
1 Driscoll DF. Lipid injectable emulsions: 2006. Nutr Clin
Pract. 2006;21:381
–386.
Nutrition in Clinical Practice, Vol. 22, No. 3,
370-372 (2007)
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
Top
Clinical Safety
Characterization of Lipids Using...
D3/6), therefore
appropriately weighting the distribution to the larger size fat globules that
are of greatest clinical concern with respect to stability and safety. In
toxicological studies, volume-weighting is necessary to establish the dose and
adverse effect(s) from coalesced fat globules as percentage of the total
concentration of lipid droplets.