| Sign In to gain access to subscriptions and/or personal tools. |
Safety and Stability of Lipid EmulsionsBaxter Healthcare Corporation Deerfield, Illinois
Retired, Professor of Pharmaceutics University of Illinois at Chicago
To the Editor: We read with interest the recent article in Nutrition in Clinical Practice by David Driscoll entitled "Lipid Injectable Emulsions: 2006."1 The article addresses 3 very important issues related to lipid emulsions: clinical safety, characterization of lipid emulsions using globule size measurements, and stability of lipid emulsions. These issues require further clarification and are discussed in this letter. The article did not mention the vast clinical safety record of lipid emulsions. Over 25 million patients worldwide have received the soybean-based lipid emulsions. This fat emulsion has been used clinically, and only rare adverse events have been reported. None of the adverse events reported to Baxter or Fresenius-Kabi have been related to lipid droplet size or fat embolism. We also performed an extensive review of the clinical literature on lipid emulsions, concentrating on autopsy studies, lung function, and reticuloendothelial function in humans. When adverse effects occurred, they were usually transient and associated with excessively high infusion rates. However, when administered according to recommended rates, lipid emulsions were demonstrated to be safe. Dr Driscoll refers to the "clinical consequences of infusing unstable emulsions" and raises concerns regarding embolic fat from lipid emulsions. However, no clinical data are presented to support this statement. Instead, 3 animal studies are referenced.2–4 The first, by Driscoll et al,2 evaluated total nutrient admixtures (TNA) containing lipid emulsions in adult guinea pigs. The emulsion used in these studies was formulated to be intrinsically unstable and not reflective of any clinically relevant product. The TNA was composed of 70% (by calories) lipid, 15% amino acids, and 15% glycerol. This TNA demonstrated spontaneous coalescence of the lipid particles with short-term storage (>72 hours). During this time period, PFAT5 (percent of fat globules >5 µm) increased from 0.004% (Time 0) to 0.161% (24 hours) to 2.418% (48 hours) to 7.861% (72 hours). Five animals were infused with fresh TNA (called "stable") and 5 animals received the TNA that was allowed to sit for 48 hours (called "unstable"). Despite use of this atypical and unstable TNA, no evidence of liver injury was presented, and there was no evidence of vascular obstruction in any organ examined (ie, liver or lung). There was evidence of tissue lipid uptake in the lungs of the 5 animals receiving the "unstable" lipid. We agree that unstable TNAs such as the one used in this study should not be compounded. However, the relevance of this TNA to formulations used clinically in the United States is questionable.
Driscoll et
al3,4
also reported on 2 studies using similar TNAs in rats. In the first
study,3 rats were
infused with the atypical and unstable TNAs (same as above) for 24 hours.
PFAT5 increased from 0.014% to 0.383% in the "stable"
group and from 0.682% to 2.655% in the "unstable" group. In the
second study,4 rats
were infused with similar TNAs for 24 and 72 hours. Hepatic malondialdehyde
(MDA) levels were slightly higher in animals receiving the unstable emulsions
(Table 1). However, glutathione
levels (a monitor of antioxidant status) were higher in both groups receiving
the unstable emulsion at 24 hours. Hepatic tumor necrosis factor
USP chapter <729> entitled "Globule Size Distribution in Lipid Injectable Emulsions"5 is a proposed chapter that has raised significant controversies among experts in the field of lipid biochemistry and is still under consideration. Thus, USP <729> should not be considered an accepted standard at this time. At the present time, it is unclear from the scientific literature what the most desirable physicochemical characteristics for stable lipid emulsions should be. It should also be noted that USP chapter <729> refers to lipid emulsions only and not to admixtures of the emulsions with other components. The mean globule size and PFAT5 are important characteristics of fat emulsions. Baxter was intimately involved in development of the technology to quantify these parameters for characterization of fat emulsions. Although these parameters are desirable for characterization of the globule size distribution of lipid emulsions, the data supporting their use for predicting stability and safety are lacking. It is critically important to define the difference between characterization and stability. Clearly, lipid globules that are large enough to block vascular supply and cause tissue injury would be undesirable. However, the specific values for the mean globule size or PFAT5 that predict these events are unknown. Dr Driscoll refers to the "stability" of lipid emulsions without defining stability. At present, there is no consensus on the definition for lipid emulsion stability. Should it be defined by visual, microscopic, particle size, oil separation, or another parameter? Webster's dictionary defines stable as "not readily altering in chemical makeup or physical state." However, it is also important to define stability by a change over time. The implication from the article is that PFAT5 is "indicative of stability" of lipid emulsions. We suspect that 5 µm was chosen because the smallest capillaries in the body are approximately 5 µm in size. There was concern that lipid globules could occlude these small capillaries. Thus, it seems that 5 µm was chosen for safety reasons. However, the data do not support the use of 5 µm for defining stability. Importantly, capillaries vary in size from 5 µm to 10 µm and red blood cells (distensible particles) are 6–10 µm in size but pass through the capillaries without difficulties. Fat globules are also distensible particles. It remains unclear what fat globule size would be best to predict stability. Why not use a 10 µm, 7 µm, or 2 µm size to evaluate stability? Why use percent rather than quantity of globules at a specific size to predict stability and safety? The optimal size of globules and the percent of globules at a specified size that can predict stability of lipid emulsions have not been systematically evaluated. Although the quantity of large fat globules may relate to stability, scientific data to support specific cutoff values are not known at this time. Dr Driscoll states that "if the normal population of these large fat globules is known in stable lipid injectable emulsions, higher amounts are associated with instability." This statement is not correct. It is possible to produce stable emulsions that have higher quantities of "large" lipid globules than current emulsions, and it is also possible to produce unstable emulsions (evidenced by spontaneous coalescence) that have lower quantities of "large" lipid globules. PFAT5 values can vary between lipid emulsions for a variety of reasons, including manufacturing processes. In fact, the globule size distributions of all commercial lipid emulsions are not identical. No limit for commercial preparations has been found that predicts instability. Commercial lipid emulsions must demonstrate months of stability under different conditions, with a wide variety of analytical evaluations before being approved for human use. Thus, by definition, commercial lipid emulsions are stable (ie, the characteristics of the emulsion change little over months of storage). The globule size distributions of all commercial lipid emulsions are not identical, with some commercial lipid emulsions having PFAT5 >0.05%. However, recent studies performed by our group using commercially available lipid emulsions (with baseline PFAT5 values ranging from approximately 0.003% to 0.3%) to prepare admixtures demonstrated that acceptable admixtures could be prepared and stored for 24 hours at 5°C plus 24 hours at 25°C (unpublished data) regardless of the PFAT5 level. Although PFAT5 increased slightly over the storage interval, there was no evidence of emulsion instability, as defined as the presence of irreversible phase separation according to visual inspection of the test article. Thus there are no data based on study of commercially available lipid emulsions indicating that a PFAT5 <0.05% is required for stability of the emulsion or for the stability of the admixtures prepared from them. Review of the literature supports the stability of current commercial lipid emulsions.6–11 The article states, "we have shown that when the PFAT5 level exceeds 0.4% (or approximately 10-fold higher than the proposed USP <729> limit), visually obvious phase separation or "cracking" occurs reflecting the terminal stages of emulsion stability." The referenced article by Dr Driscoll12 is the only published primary research article to make this claim. However, an inspection of the referenced study12 indicates that the study was performed using TNAs and that emulsion stability was limited by the excessive addition of polycations (primarily iron), which are known to cause instability of lipids and admixtures, and the large replicate-to-replicate variability reported in the study. Importantly, iron dextran is not recommended for use in TNAs by the American Society of Parenteral and Enteral Nutrition's (A.S.P.E.N.'s) "Safe Practices for Parenteral Nutrition" guidelines,13 because this trivalent cation is known to cause lipid emulsion instability. The investigators state, "Multiple stepwise regression analysis revealed that trivalent cation concentration was the only variable that affected the stability of nutrition emulsions." A PFAT5 of 0.4% lacked both sensitivity and specificity for predicting cracking of formulas containing iron dextran. The critical flaws in this study prevent any meaningful conclusion to be drawn between the PFAT5 value of a lipid emulsion and its stability. Dr Driscoll raises the issue of lipid globule size and the type of packaging of lipid emulsions (ie, glass vs plastic).14 Dr Driscoll's statement that, "it seems the stability problems associated with lipid injectable emulsions packaged in plastic are not necessarily a problem related to the container but rather are related to the manufacturer" is only partially correct. Although the PFAT5 of some lipids in plastic containers are >0.05%, there are no "stability problems" associated with lipid injectable emulsions packaged in plastic. However, we emphasize that it is not the manufacturer per se but rather the manufacturing process used at the specific plant where the lipid emulsion is manufactured that can affect the PFAT5 level. In addition, all commercially manufactured lipid emulsions are demonstrated to be stable over time (a requirement for their approval). We have tested lipid emulsions from different containers (ie, glass and plastic) and found that the container does not affect globule size. In conclusion, lipid emulsions represent an advance in the nutrition support of patients dependent on parenteral nutrition (PN). Lipid emulsions supply energy and essential fatty acids while minimizing the detrimental effects of excess carbohydrate administration. Extensive clinical experience with commercially available lipid emulsions indicates that they are safe and stable for patient administration.
* Dr Groves is a consultant for Baxter Healthcare 
1 Driscoll DF. Lipid injectable emulsions: 2006. Nutr Clin
Pract. 2006;21:381
–386.
Nutrition in Clinical Practice, Vol. 22, No. 3,
367-369 (2007) This article has been cited by other articles:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
(TNF)
and interleukin-1 β (IL-1) were lower in animals receiving the unstable
emulsion at 24 hours. At 72 hours, hepatic levels of glutathione, TNF, and
IL-1 were similar in both infusion groups. Plasma levels of acute phase
proteins (
Top
