Previous PageNext Page

Chapter 1 – Introduction

1.1 Lecithin - an Emulsifier for Parenteral Use

Only a limited number of emulsifiers is commonly regarded as safe to use for parenteral administration, of which the most important are Pluronic F68® (Poloxamer 188), Tween 80® (Polysorbate 80) and lecithin. Compared with its synthetic alternatives, lecithin can be totally biodegraded and metabolised, since it is an integral part of biological membranes, making it virtually non-toxic. Other emulsifiers can only be excreted via the kidneys, and Pluronic F68®, owing to impurities contained, even has been reported to be the cause of endogenous pyrogen release [Huth et al., 1967]. The natural origin of lecithin produces, however, a rather complex composition, although in pharmacy in general well-defined singular excipients are favoured. The Ph.Eur. 1997 does not contain a monograph on lecithin, however, a monograph on soya lecithin will be contained in Ph.Eur. 1998, and USP23/NF18 [1995] only describes it as ‘a complex mixture of acetone-insoluble phosphatides’, but also does not specify content limits concerning its intended use. Nevertheless, lecithin is regarded as a well tolerated and non-toxic compound (which is also expressed by its GRAS-status [‘Generally Recognised As Safe’] approved by the FDA), making it suitable for long-term and large-dose infusion. As an emulsifier of intravenously administered fat emulsions, its composition and behaviour determine the structure and stability of the emulsion in a decisive way. Although extensive research in this field has been done, there is still disagreement about emulsion structure and the influence of the emulsifier. It is essential to understand the behaviour of lecithin in order to understand the behaviour of emulsions stabilised with it.

‘Lecithin’ is usually used as synonym for phosphatidylcholine (PC), which is the major component of a phosphatide fraction which is frequently isolated from either egg yolk (Greek ‘é’), or soya beans and is commercially available in high purity. Isolation and purification of lecithins from different sources are described in Kuksis [1985] and Prosise [1985]. PC is a mixture of differently substituted sn-glycerol-3-phosphatidylcholine backbones. As can be seen from Tab. 1, the structure of PC is variable and dependent on fatty acid substitution. In the sn-1-position, saturated acyl-groups, and in sn-2-position, unsaturated species are more common [Kuksis, 1985]. By dietetic means the fatty acid substitution of egg phospholipids can be altered in the sn-2-position [Hanahan, 1960]. Fatty acids of mainly 16-20 C in chain length dominate in egg PC. The sn-1-chain typically shows an average of 16 C, whereas the sn-2-chain shows an average of 18 C. Naturally occurring unsaturated fatty acids are almost entirely of all-cis-conformation [Gennis, 1989].

Palmitoyl/Oleoyl [16:0/18:1]-PC (POPC) 38.2% in Egg PC

Palmitoyl/Linoleoyl [16:0/18:2]-PC (PLPC) 21.8% in Egg PC

Stearoyl/Linoleoyl [18:0/18:2]-PC (SLPC) 11.22% in Egg PC

Stearoyl/Oleoyl [18:0/18:1]-PC (SOPC) 9.3% in Egg PC

Stearoyl/Arachidoyl [18:0/20:4]-PC (SAPC) 3.36% in Egg PC

Di-Palmitoyl [16:0/16:0]-PC (DPPC) 0.72% in Egg PC

Tab. 1 Different species of PC and their natural occurrence in egg yolk lecithin (mol/mol)
(values after Kuksis [1985])

1.1.1 Properties of Amphiphilic Lipids

Amphiphilic lipids consist of a rather polar ‘head group’ and comparably apolar residues. In the solid state, the molecules tend to orientate in a distinct way, which is that lipophilic tails and hydrophilic head groups take a separate, packed arrangement. The state of order and the length of the lipophilic tail depends on the conformation of the carbon chains which preferably will be ‘all-trans’ in the case of saturated acyl-chains, since in this case a potential energy minimum occurs. Thus, the carbon chains become extended to a maximum, whereas in the case of a carbon-carbon bond taking on the gauche-conformation, a kink in the chain occurs. A similar displacement can be observed if a cis-double-bond is located in the hydrocarbon chain (Fig. 1).

Fig. 1 Newman projections of the rotation about a carbon-carbon bond in an alkyl chain (left) and several C15-alkyl chain configurations (right) (adapted from Gennis [1989])

The shrinking of chain elongation owing to kinks or insertion of double bonds notably also increases the cross sectional area covered by each chain (‘’ in Fig. 4). Amphiphilic lipids typically often do not exhibit abrupt transitions from the solid to the liquid state, but do undergo ‘intermediate’ states, where properties of solid crystals and of liquids can be observed, as well [Lee, 1977]. This is why these ‘intermediate’ states are also known as ‘mesophases’ or ‘liquid crystals’. This so-called mesomorphic behaviour can be attributed either to temperature changes when e.g. heating causes ‘chain melting’ which means transformation of the alkyl chains into a less ordered state owing to increased occurrence of thermodynamically unfavourable chain kinks and consequently increased chain space requirement (‘thermotropic phase transition’), or changes in hydration state owing to the fact that only the polar head groups bind to water and become hydrated, which increases their respective space requirements and eventually results in changes in molecule packing (‘lyotropic phase change’). The resulting mesophases are usually named according to IUPAC nomenclature using the following abbreviations which describe long-range and short-range orders in lipid-water phases:

The following phases are most commonly encountered in lipid-water systems:

a) Lamellar liquid crystalline phase (L)

b) Lamellar gel phase (L)

c) Hexagonal I phase (HI)

d) Hexagonal II phase (HII)

e) Cubic I phase (QI)

f) Cubic II phase (QII)

The interdependence of lyotropic and thermotropic phase transitions of lipid-water systems can be summarised in the following fashion:

Fig. 2 Dependence of phase transitions of lipid-water mixtures on hydration and temperature (adapted from Brown and Wolker [1979], Gennis [1989] and Marsh [1990])

    Lamellar phases (L/L), Hexagonal phases (HI/HII), Cubic phase (QI) and Micellar solution (M)

Fig. 2 shows that certain preferences in phase behaviour according to substitution and temperature exist, which in most cases are reversible. It must also be stressed, however, that the thermal ‘history’ of the lipid plays also an important role at which phase changes occur. Similarly, in the case of phospholipid-water mixtures, occurrence of thermotropic or lyotropic phase changes is also well known. To have an idea which phases occur in a lecithin-stabilised O/W emulsion system, one has to start by looking at the phase diagram of lecithin alone.

1.1.2 Properties of Phospholipids

At low water contents of egg lecithin, a variety of phases are observable (Fig. 3). The surface area per molecule (‘S’ in Fig. 4) at 21°C has been determined as 59.3 Å2. With increasing water content, molecular area increases to about 71.7 Å2, and lamellar structures (L) become predominant. At 21°C and 40-44 wt% water content, excess water forms a second phase, in which the L phase is dispersed. Temperature increase leads only to melting of the L phase to an isotropic liquid at about 220°C [Small, 1986].

 

Fig. 3 Influence of temperature and water content on phases of egg lecithin (modified after Small [1986])

    C = Crystalline / H = Hexagonal liquid crystalline / L = Lamellar liquid crystalline / L = Lamellar gel / Q = Cubic disordered / R = Rhombohedral disordered

Depending on its intended use, lecithin is purified to different degrees. Yeadon et al. [1958], Hansrani [1980] and Rydhag and Wilton [1981] noted that the emulsifying properties of pure PC were by far inferior to mixtures of PC with its naturally occurring byproducts, e.g. phosphatidylethanolamine (PE). This is one of many different other compounds (often referred to as ‘minor components’) typically associated with PC (a selection is given below in Tab. 3 in which fatty acid substituents are abbreviated [R1, R2]). By way of comparison two commercial egg lecithins are also given, showing that composition is largely dependent on lecithin-source and degree of purification. The commercially used egg yolk lecithins designed for parenteral use (Lipoid E80® or egg phosphatide for Intralipid®) are enriched in PC at the expense of PE, compared with the crude product. In contrast to PC, which carries a zwitterionic charge in its head group, a few components are negatively charged (PA, PS) or uncharged at neutral pH values [Szoka and Papahadjopoulos, 1980]. This structural difference causes different hydration behaviour of the head groups, which is in turn responsible for different molecular head group space requirement and packing behaviour. Rydhag [1979] reported increased swelling of lamellar arrangements on incorporation of negatively charged components. Additionally, Hansrani [1980] found that various minor lecithin components like PA, PI, PS and especially the lysophospholipids were able to contribute to enhanced emulsion stability.
The so-called lysophospholipids are derived from hydrolytic cleavage of a fatty acid (where the sn-2 position appears to be preferred). Herman and Groves [1992] have shown that this process occurs according to first-order Arrhenius kinetics with different activation energies for PC and PE, whereas Håkansson [1966] reported that formation of free fatty acids was accelerated at acidic conditions where catalysis by hydronium ions occurred. Herman [1992] also pointed out that hydrolysis appears at the same rates for sn-1 and sn-2 positions and that subsequent intramolecular acyl migration may let the hydrolysis appear to have taken place at the sn-2 position. Ongoing hydrolysis results in cleavage of the second fatty acid residue, eventually yielding a molecule of the respective glycerophosphorylic compound, GPC or GPE.
Soya lecithins contain more PA and PI than egg lecithin, and are also reported to show excellent emulsifying properties [Rydhag, 1979], which led to development of various parenteral emulsions containing fractionated lecithin from soybean (see Tab. 5). Some reports on adverse reactions after parenteral application of soya lecithin-stabilised emulsions [Schuberth and Wretlind, 1961] were assigned to its higher PI content [Wretlind, 1976]. These emulsions successively vanished from the market, although other authors claimed that these toxic effects might be circumvented by further purification of soya lecithin [Mueller and Iacono, 1967]. Nevertheless egg lecithin has become the preferred lecithin type for parenteral use today.

Component

Formula

(%) in crude Soya lecithin

(%) in crude Egg lecithin

(%)
in
Lipoid E80®

(%) in Intralipid® lecithin

Phosphatidylcholine (PC)

33.0

66 – 76

77.7

(77 - 82)

87.1

Phosphatidylethanol-amine (PE)

14.1

15 – 24

7.8

(7.0 - 9.5)

7.8

Phosphatidylserine (PS)

0.4

1

n.a.

n.a.

Phosphatidylinositol (PI)

16.8

n.a.

n.a.

n.a.

Tab. 3a Distribution of selected phosphatides and lipids in crude and commercial lecithins (values for Lipoid E80® batch 19890-8 and tolerated ranges in brackets are given after product analysis sheet by the manufacturer, values for Intralipid® lecithin and egg lecithin [Kuksis, 1985], for soybean lecithin [Weber, 1985])

Component

Formula

(%) in crude Soya lecithin

(%) in crude Egg lecithin

(%)
in
Lipoid E80®

(%) in Intralipid® lecithin

Phosphatidic Acid (PA)

6.4

n.a.

n.a.

n.a.

Lyso-Phosphatidylcholine (LPC)

0.9

3 - 6

2.5

(max. 3)

2.5

Lyso-Phosphatidylethanol-amine (LPE)

0.2

3 - 6

< 0.5

(< 0.5)

2.5

Sphingomyelin (SPM)

n.a.

3 - 6

3.0

(2 - 3)

2.5

Tab. 3b Distribution of selected phosphatides and lipids in crude and commercial lecithins (values for Lipoid E80® batch 19890-8 and tolerated ranges in brackets are given after product analysis sheet by the manufacturer, values for Intralipid® lecithin and egg lecithin [Kuksis, 1985], for soybean lecithin [Weber, 1985])

The mesomorphic behaviour of lecithins depends on lecithin composition, ionic strength and pH. This points to the influence of charge and hydration on molecular packing orientation and phase shapes. The orientation of a PC molecule towards an interface is shown in Fig. 4. The acyl chains show a distinct tilt from the interface normal, because the polar region extends as far as the second C-atom of the sn-2 acyl chain and forces this part into parallel orientation to the interface. This molecular geometry favours orientation into lamellar phases (compare with Tab. 4). For each unsaturated site, the partitioning of both acyl chains increases, as chain displacement and consequently increased space requirements occur. Thus a less tighter packing, better mobility and lower melting of the fatty acid chains and of packed molecules can be observed. Accordingly, it is desirable to determine the fatty acid composition of a phospholipid fraction. Soybean lecithin, for example, carries more unsaturated fatty acids than the egg species [Rydhag, 1979 / Fiedler, 1996]. Furthermore, the space covered by the hydrated ‘head groups’ has strong influence on molecular orientation in organised arrangements (see Fig. 6 and Tab. 4). For the lecithin mixtures examined in this work, phase behaviour proves to be more complex than that described, for example, in liposome research literature, where exactly defined lipids are used.

Fig. 4 Schematic of PC showing its various regions and molecular orientation at an interface (from Gennis [1989])

The influence of the head group on molecular arrangement is illustrated by comparing PC with PE, the latter having a smaller head group [Gennis, 1989]. Thermotropic phase changes from and into lamellar phases are markedly different for both species (see Fig. 6). Also, packing adaptations are different: PE allows closer packing (Fig. 5a), whereas the more bulky head group of PC brings about different packing adaptations (Fig. 5b+c). Accordingly, the relation between the cross sectional area of the head group and the chains leads to different packing adaptations (see Tab. 4).

Fig. 5 Different packing adaptations in lamellar mesophases (after Gennis [1989]):
(A) Space requirement of ‘head group’ and chains are the same
(B) Head group requires more space than chains
(C) Tilted arrangement of chains allowing tighter packing

In Fig. 6, for PC a ‘ripple phase’ is observed as an intermediate state of transition from L to L. With PE, the ordered arrangement appears to melt directly into the liquid crystalline state. Different molecular shapes can be assigned to phospholipid molecules, explaining these preferential packing geometries and phases.

Typical packing adaptations are explained by means of the ‘critical packing parameter’ (CPP), which is the ratio of effective volume v, head group area S0 and chain length lc (CPP = v/S0lc). The CPP determines the preferred association structures assumed for each molecular shape. How the properties of phospholipid-water mixtures and, accordingly, their phase behaviour vary, is summarised in Tab. 4, where examples of typical molecular shapes, phase arrangements and CPPs are shown for the most prominent phospholipids.

Critical packing shape

Critical packing parameter

Phase formed

Lipid examples

< 1/3 (spheres)

1/3 – 1/2 (rods)

Lysophospholipids
(e.g. LPC, LPE, etc.),

free fatty acids (e.g. oleate, stearate, etc.)

1/2 – 1 (lamellar, vesicles)

Double-chained lipids with large head group areas and fluid chains:
PC, PS, PG, PI, PA, SPM

~ 1 (lamellar, planar bilayers)

Double-chained lipids with small head group areas, anionic lipids and saturated chains:
PE, PS + Ca2+

> 1 (hexagonal HII)

Double-chained lipids with small head group areas, non-ionic lipids and polyunsat. chains:
PE (unsat.), PA + Ca2+, PS (pH<4)

Tab. 4 Molecular shapes and association structures of phospholipids
(modified after Cevc and Marsh [1987] and Gennis [1989])

Lamellar lecithin phases show increasing inter-bilayer spacing on uptake of water (‘swelling’) to form myelins, which on further dilution form closed bilayer vesicles. These so called ‘liposomes’ show diverse size and number of bilayers (lamellae) [New, 1990]. Curvature of the phospholipid bilayers, which is inter alia determined by packing geometries, limits the smallest possible size of the bilayer and is also the reason for different distribution of the phospholipid molecules in the inner and outer bilayers [Vance and Vance, 1991]. Typical vesicular structures formed by phospholipids that favour lamellar phases are depicted in Fig. 7.

PC and its minor components exhibit preferentially micellar, lamellar or hexagonal phases depending on e.g. head group type, fatty acid substitution, pH, temperature and hydration [Cullis et al., 1991]. For the various lipids contained in lecithin-stabilised emulsions, one would expect lamellar (vesicular) and micellar phases, as well as monolayers at the oil/water-interface. Chapman [1975] stressed that in mixed PC/PE systems not a single phase but rather separation of phospholipids into different phases can occur (lateral phase separation), which could be proven by thermal analysis and electron microscopy [Cevc and Marsh, 1987]. Owing to the complexity of commercial lecithin composition, analytical data (e.g. thermoanalytical data) only yield limited, imprecise information, and it still remains a challenge to determine phases exhibited in lecithin-stabilised emulsions. Possible resulting influences on emulsion structure and stability are discussed in Section 1.2.3.
More detailed information and data on phase equilibria and transitions of phospholipids are available in Marsh [1990], Gennis [1989], Small [1986] and Cevc and Marsh [1987].

1.2 Parenteral Fat Emulsions

Intravenously-administered oil-in-water emulsions containing triglycerides as the dispersed phase and egg lecithin as the preferred emulsifier, are common in modern intensive care medicine. Containing essential, unsaturated fatty acids, such emulsions provide calories for patients which cannot be nourished orally. It is assumed that owing to their similar particle sizes and composition, parenteral fat emulsion droplets behave like chylomicra, which are natural-borne fat-globules circulating in the blood stream after oral intake of fat [Schoefl, 1968 / Thompson, 1974]. Accordingly, clearance from the blood is believed to be comparable with chylomicra in vivo. Parenteral emulsions are not only used as nutrients, but also as potential carriers or controlled delivery systems for poorly water-soluble, oil-soluble drugs [Jeppsson, 1976 / Prankerd and Stella, 1990 / Yamaguchi et al., 1995b]. Furthermore, reduced adsorption of drugs on infusion sets, reduction of local toxicity on infusion [Davis et al., 1985] or reduced drug hydrolysis [Repta, 1981] by incorporation of drugs into parenteral fat emulsions was reported. Renewed interest has emerged recently for the well-tolerated application of Amphotericin-B, an anti-fungal drug, using a liposomal formulation, which is commercially available as Ambisome®. Some promising attempts to incorporate this drug into commercial parenteral fat emulsions and thus also lower relatively high nephro-toxicity compared with solubilised drug were reported [Lipp et al., 1993]. However other reports could not confirm these findings [Schoeffski et al., 1996]. Other applications of lecithin-stabilised parenteral emulsions comprise x-ray [Vermess et al., 1977] and ultrasonic [Lanza et al., 1996] contrast-emulsions, as well as perfluorocarbon-emulsions as possible blood-substitutes [Magdassi et al., 1991 / Pelura et al., 1992 / Ni et al., 1996].

Kleinberger and Pamperl [1983] identified four generations of fat emulsions for parenteral nutrition (Tab. 5), to which a fifth was added a few years ago. The first parenteral fat emulsions used various emulsifiers, the first of which was egg lecithin in the late 1920s contained in Yanol® [Nomura, 1929]. After severe side effects had been reported, the emulsion was withdrawn by the end of the 1930s [Thompson, 1974]. McKibbin et al. [1943] examined lecithins from egg and soya and stated that combination with synthetic emulsifiers yielded the most stable emulsions. Lipomul® / Infonutrol® represents the second generation of parenteral fat emulsions now containing purified soya lecithins in combination with a synthetic emulsifier and the oil phase used consisted of cottonseed oil (Tab. 5). Again, pharmacological issues necessitated development of a new formula, since long-term administration of cottonseed-emulsions led to toxic side effects owing to gossypol contamination [LeVeen et al., 1961]. Soya oil and safflower oil therefore became the new triglyceride sources of choice (contained in Intralipid® and Abbolipid®) [Wretlind, 1964]. These products were produced with fractionated egg lecithin as the sole emulsifier, since the toxic behaviour of long-term Pluronic-administration was, and still remains, unclear (Tab. 5). The introduction of Intralipid® in the early 1960s thus marked the third generation, and many emulsions with similar composition (soybean oil, egg lecithin) have since been introduced. Other formulations used soybean phosphatides and other polyalcohols as isotonicity agents (see Tab. 5). But as soybean lecithin was reported to be more toxic than egg lecithin (refer to Section 1.1.2), this is now not used in commercial products, and glycerol is the only isotonicity agent.

Generation

Proprietary name

Oil phase
(w/v)

Emulsifier
(w/v)

Other excipients
(w/v)

I

Yanol® Castor oil (3%) Egg Lecithin

-

II

Lipophysan® Cottonseed oil (10%) Soya Phospholipids (2.0%) Glycerol (2.5%),
Tocopherol

II

Lipomul®/
Infonutrol®
Cottonseed oil (20%) Soya Phospholipids (1.2%) +
Poloxamer 188 (0.3%)
Glucose (4.0%)

III

Intralipid® Soya oil (10%/20%) 3-sn-PC from egg yolk
(0.6%/1.2%)
Glycerol (2.2%)

IV

Lipofundin S 10® Soya oil (10%) Soya Phospholipids (0.75%) Xylitol (5.0%)

Tab. 5 Composition of parenteral fat emulsions
(modified after Kleinberger and Pamperl [1983] and Thompson [1974])

In the late 1980s a new approach was made using as the oil phase a combination of middle-chain triglycerides (MCT) derived from coconut oil and possessing C8-10 chains, together with conventional long-chain triglycerides (LCT). Eckart et al. [1980] and Guisard and Debry [1972] claim MCT to have the advantage over LCT triglycerides of better availability owing to faster metabolisation and faster clearing. The side effects related to an ‘overloading’ of the reticulo-endothelial system (RES) in the liver with fat droplets observed with earlier products [Thompson, 1974 / Darby and Wallin, 1978], therefore, could be diminished [Bach et al., 1989]. Pure MCT oils were found to be less well tolerated, however [Bach et al., 1989]. The first emulsions of this kind to be marketed were Lipofundin® MCT, where conventional LCT and MCT were used in equal proportions. Fig. 8 shows typical fatty acid compositions of a conventional soybean emulsion (Intralipid®, LCT) and Lipofundin® MCT (MCT/LCT). There is no reported data on how these different triglyceride fractions effect emulsion structure and stability. It seems desirable to expand our knowledge of how different oil sources could effect the interfacial behaviour of lecithin, especially since the quest to improve tolerability and therapeutic effects has already led to new formulations containing further refined triglyceride sources. Omegaven-Fresenius® (containing -3-fatty-acids [fish oil] in the triglyceride-phase) and Structolipid® (containing randomly re-esterified MCT/LCT mixtures) are currently being introduced to the market or under clinical trials, respectively.

Fig. 8 Fatty acid composition of parenteral fat emulsions: Intralipid® (LCT, left) and
Lipofundin® MCT (LCT/MCT mixture, right), MCTs being C8 and C10 fractions.
(after Pharmacia&Upjohn [1997] and Bell et al. [1991])

-3-fatty-acids given intravenously were shown inter alia to lead to improved blood circulation [Pscheidl et al., 1992] and reduced occurrence of undesirable prostaglandin-related side-effects. This improved tolerability of parenteral emulsions compared with those having typical fatty acid composition [Bell et al., 1991 / Carpentier et al., 1997]. Structolipid® contains LCT and MCT enzymatically cleaved and randomly re-esterified instead of using physical mixtures of LCT/MCT. By doing so, improved metabolism of triglycerides and their fatty acids and overall better tolerability was reported, making them particularly useful in critically ill patients [Bell et al., 1991 / Sandström et al., 1993 / Dahn, 1995 / Hyltander et al., 1995 / Pscheidl et al., 1995].

Tab. 6 gives the formulations of current parenteral emulsions. Frequently, only 3-sn-phosphatidylcholine (PC) is quoted as the emulsifier, with variable PC contents, and minor components (see Tab. 3) are not referred to. Emulsifier composition has been demonstrated to change constantly owing to degradation during production and storage [Herman,1992], which has to be taken into account when speaking of ‘the emulsifier’. In most cases soya oil is used as the oil phase, sometimes admixed with MCT or safflower oil. In all cases, glycerol is the isotonicity agent, and sodium hydroxide or sodium oleate are employed to adjust pH to values between 7-8 (according to declaration by some manufacturers) as is desired for large-dose parenterals [Ph.Eur.1997, 1997]. In the case of drug-loaded emulsions, lipophilic drugs, especially narcotics or tranquillisers dominate.

Proprietary name

Oil phase
(w/v)

Emulsifier
(w/v)

Active ingredients

Other excipients
(w/v)

Abbolipid®
10% / 20%

Soya oil +
Safflower oil
(5% / 10% each)
3-sn-PC from egg
(0.74% / 1.2%)

-

Glycerol (2.5%),
Sodium Hydroxide

Deltalipid®
10% / 20%

Soya oil
(10% / 20%)
Egg phospholipids
(70-83% 3-sn-PC)
(1.2% / 1.2%)

-

Glycerol (2.5%),
Sodium Hydroxide,
Oleic Acid (0.03%)

Intralipid®
10 / 20

Soya oil
(10% / 20%)
3-sn-PC from egg yolk
(0.6% / 1.2%)

-

Glycerol (2.2%),
Sodium Hydroxide

Lipofundin®N
10% / 20%

Soya oil (fract.)
(10% / 20%)
Egg phospholipids
(min. 68% 3-sn-PC)
(0.8% / 1.2%)

-

Glycerol (2.5%),
Sodium Oleate,
-Tocopherol

Lipofundin MCT®
10% / 20%

Soya oil + Middle
chain triglycerides
(5% / 10% each)
Egg phospholipids
(min. 68% 3-sn-PC)
(0.8% / 1.2%)

-

Glycerol (2.5%),
Sodium Hydrox.,
Sod. Oleate, -Toc.

Lipovenoes®
10% PLR / 20%

Soya oil
(10% / 20%)
Egg phospholipids
(min.75-81% 3-sn-PC)
(0.6% / 1.2%)

-

Glycerol (2.5%),
Sodium Hydrox.,
Sodium Oleate

Lipovenoes®
MCT 10% / 20%

Soya oil + Middle
chain triglycerides
(5% / 10% each)
Egg phospholipids
(min.75-81% 3-sn-PC)
(0.6% / 1.2%)

-

Glycerol (2.5%),
Sodium Hydrox.,
Sodium Oleate

Salvilipid®
10 / 20

Soya oil
(10% / 20%)
3-sn-PC from egg yolk
(60%) (1.2% / 1.2%)

-

Glycerol (2.5%),
Sodium Hydroxide,
Sod. Oleate (0.03%)

Intralipid® 30

Soya oil (30%) 3-sn-PC from egg yolk
(1.2%)

-

Glycerol (1.67%),
Sodium Hydroxide

Diazepam Lipuro®
1% / 2%

Soya oil + Middle
chain triglycerides
Egg lecithin Diazepam
(5%)
Glycerol,
Sodium Oleate
(for pH adjustment)

Disoprivan®
1% / 2%

Soya oil 3-sn-PC Propofol
(1% / 2%)
Glycerol,
Sodium Hydroxide

Etomidat Lipuro®

Soya oil + Middle
chain triglycerides
Egg lecithin Etomidate
(2%)
Glycerol,
Sodium Oleate
(for pH adjustment)

Lipotalon®

Soya oil (10%) Egg phospholipids
(1.2%)
Dexamethasone-21-
palmitate (0.4%)
Glycerol

Propofol Abbot 1%

Soya oil

Egg lecithin

Propofol
(1%)

Glycerol,
Sodium Hydroxide

Propofol 1% Parke-Davis®

Soya oil

3-sn-PC

Propofol
(1%)

Glycerol,
Sodium Hydroxide, Oleic Acid

Stesolid®

Soya oil

Phospholipids from egg yolk

Diazepam
(5%)

Glycerol, Acetylated Monoglycerides, Sodium Hydroxide

Tab. 6 A selection of presently marketed parenteral emulsions in Germany
[Rote Liste, 1998]

Early Intralipid® emulsions and its generics were produced using 1.2% w/v emulsifier regardless of the oil content of the emulsions. It was assumed that this amount was necessary to stabilise 10% w/v of soybean oil and according to Ishii et al. [1990], it appeared to be the optimal concentration. Groves et al. [1985] proposed from calculations and experimental data that excess emulsifier was present and remained as multi-lamellar liposomes (MLVs) in the finished product. It could be shown that emulsions containing 10% inner phase held a substantial excess of lecithin [Hajri et al., 1990 / Férézou et al., 1994] forming separate phases besides emulsion droplets. This excess was suggested to be the reason for unfavourable dislipidemias found on long-term infusion [Lutz et al., 1990]. Deviation from the chylomicron-like behaviour, and thus a different metabolic fate than that of lecithin-coated oil droplets, was presumed. This excess was again suspected to consist of liposomes which caused enhanced formation of abnormal ‘Lipoprotein-X’, increased cholesterol release and other dislipidemic effects [Bach et al., 1996]. Consequently, most commercially available 10% emulsions that formerly contained 1.2% emulsifier, were reformulated and the lecithin concentration reduced to 0.6% w/v. In Intralipid 30® even threefold quantities of oil (30%) are incorporated using 1.2% lecithin, as further reduction of the lecithin/triglyceride ratio was sought.

Thus the historical changes in parenteral fat emulsion formulation were necessary for pharmacological reasons, not because of stability issues. Parenteral emulsions are often administered admixed with different amino acid, carbohydrate or electrolyte solutions, which is termed as ‘Total Parenteral Nutrition (TPN)’ and frequently causes incompatibilities. In the case of Intralipid 10® and Lipofundin®10% and 20%, admixtures are explicitly forbidden by the manufacturer, whereas for Intralipid 20® or Lipovenoes 20%® these are regarded as being possible taking due care. Emulsion formulation and its ‘stability’ is evidently still a sensitive issue and a matter not yet fully understood. All commercial lecithin-stabilised emulsions also show high, unexpected stability against steam-sterilisation which also may be related to their structure [Herman and Groves, 1992]. Instability problems during production or storage of lecithin-stabilised O/W emulsions do, however, occur [Magdassi et al., 1991 / Krafft et al., 1991 / Chaturvedi et al., 1992 / Lucks, 1993]. Differences in raw materials, equipment and preparation might, however, have been underestimated here. In the following sections, therefore, a brief overview of the decisive steps in parenteral emulsion production are given.

1.2.1 Preparation of Parenteral Fat Emulsions

The properties of the final emulsion are largely dependent on the preparation techniques applied. Schurr [1969] gives an introduction to large batch and Pscherer [1981] to industrial manufacturing of parenteral fat emulsions.

1.2.1.1 Preparation of Coarse Emulsions

To reduce the bioburden and possible pyrogens, both oil and water phases are filtered before they are mixed together to form the pre-dispersion [Pscherer, 1981]. Water-soluble components like glycerol or sodium hydroxide can directly be dissolved into the Water for Injection. Lecithin is practically insoluble in either oil or in water. By applying elevated temperatures (70-80°C), lecithin dissolves into the oil, which allows filtration of all of the components prior to premixing. High shear mixers (e.g. colloid-mill) or rotor-stator mixers are subsequently used to form pre-emulsions of rather wide particle size distribution of preferably  20 µm [Bock, 1994].
In lab-scale setups, rotor-stator mixers like the ultra turrax have frequently been used [Washington and Davis, 1987 / Bock et al., 1994]. However, magnetic stirrers [Rabiner et al., 1986] or high-shear mixers [Ishii et al., 1990] have also been used. Various methods of incorporation of lecithin into pre-dispersions were reported. Hansrani [1980], Rabiner et al. [1986], Washington and Davis [1987], Herman [1992] and Chaturvedi et al. [1992] used the lecithin dispersed in the (heated) water phase, whereas e.g. Ishii et al. [1990] and Bock et al. [1994] used the heated oil phase as a vehicle for the emulsifier. Bock [1994] assumed that lowered pH values during pre-emulsification were caused by incipient hydrolysis of the emulsifier, which he related to decreasing stability of the system. He stated that the phase-inversion method led to a markedly shorter pre-emulsification step and, therefore, to reduced hydrolysis. However, neither of the authors examined particle size after the pre-emulsification step, presumably because the subsequent homogenisation step was considered to be the more crucial process.

1.2.1.2 High-Pressure Homogenisation

Homogenisation involves concurrent reduction of particle size and narrowing of particle size distribution [Walstra, 1983]. This is of twofold importance. Injectables must avoid capillary blockage (see Section 1.2.2) and additionally are more stable against creaming when droplets are uniformly small and only subject to Brownian movement (see Section 1.2.2.4). To produce such emulsions of submicron particle size, size reduction must be highly effective. During emulsification, droplets are disrupted when they are deformed beyond a critical value for longer than a critical time. High energy input provides the tension necessary to break up a droplet by overcoming the Laplace pressure and interfacial tension.
The energy density (EV, [J/m3]) applied by the homogenising equipment can be described as:

where is the mean power density [W/m3] and the mean residence time [s] in the dissipation zone of the homogeniser. This could also be rewritten as (power input) [W] divided by (volume flow rate) [m3/s] [Schubert, 1997]. The resulting mean droplet diameter can be assumed to be a function of the energy density applied [Karbstein, 1994]. Effective droplet disruption could be achieved by exposing the droplets to high laminar or turbulent flow or shear stress of high energy density. High-pressure homogenisers have successfully been used for this purpose. They provide high PV and produce effective particle size reduction, especially in low viscosity dispersions [Karbstein, 1994], which can directly be related to the homogenisation pressure. Conventional valve homogenisers disrupt droplets by both shear forces existing adjacent to energy-bearing eddies in turbulent flow, and cavitation (owing to high local pressure fluctuations) which causes solvent vapour bubbles to form and collapse which in turn shatters the droplets [Walstra, 1983]. In submerged jet homogenisers, such as the ‘Microfluidizer®’ or the ‘Nanojet®’, turbulent flow more than shear or cavitation cause droplet disruption [Karbstein, 1994 / Schubert, 1997].

Particle sizes of homogenised emulsions are dependent on four key factors [Karbstein, 1994 / Schubert, 1997]:

Excess pressure or prolonged homogenisation can, however, lead to overprocessed emulsions, where droplet diameter increases [Pscherer, 1981 / Karbstein, 1994]. This coalescence is not a result of enhanced particle collision rates, since collision times in homogenisers are too short to allow film drainage between approaching droplets [Chesters, 1991]. It is rather a matter of insufficient emulsifier concentration at the oil-water-interface (discussed in Section 1.2.2.1). A plateau in particle size reduction with time is reached, where neither increasing temperature nor homogenising pressure lead to further droplet size reduction [Washington and Davis, 1987 / Lidgate et al., 1989 / Ishii et al., 1990]. The same authors report that reprocessing of the emulsion for several homogenisation cycles narrows particle size distributions further, but that major particle size reduction is already achieved after 2-3 cycles. Since heating during homogenisation may not alter only disruption and ‘recoalescence’ mechanisms but also enhance chemical degradation, temperature should be controlled carefully [Washington and Davis, 1987 / Bock et al., 1994].

1.2.1.3 Sterilisation by Autoclaving

As other sterilisation techniques are not applicable to emulsions, steam-sterilisation is favoured. Emulsions are sterilised according to pharmacopeial requirements (e.g. 121°C, 2 bar for at least 15 min). Thermal stress might, however, change the physical and chemical stability of the emulsions unfavourably. To minimise physical changes, rotating autoclaves are used to avoid unequal heat-distribution in the emulsions. To prevent excess thermal stress and to avoid refluxing on the upper surface of the bottles, readjustment to atmospheric conditions can be accelerated by spraying iced water onto the bottles in the autoclave [Schurr, 1969].

1.2.2 Stability of Parenteral Fat Emulsions

Stability of a parenteral emulsion is mainly used in terms of maintaining of its main physical property, namely the dispersed phase particle size distribution. Safety of application is to a high degree dependent on the particle size distribution of the oil phase, since particles larger than 5 µm given intravenously can lead to emboli in vivo [Hadfield, 1966 / Davis, 1974]. Therefore, USP23/NF18 [1995] limits particulate matter in parenterals to  25 µm and Ph.Eur.1997 [1997] states ‘controlling particle size range within suitable limits according to its intended use’. For injectables, this alludes to the diameter of the smallest capillaries of the body (~5 µm). The shelf life of commercial parenteral emulsions lies in a range of 2 years, although physical stability can be maintained for longer periods [Mueller et al., 1992 / Schuhmann, 1995]. Despite this, chemical and microbiological properties must also be controlled for toxicological reasons; one has to consider that these emulsions are often given to critically-ill patients for long periods.
According to the IUPAC definition, emulsions are disperse systems of two immiscible liquids or liquid crystals [International Union of Pure and Applied Chemistry, 1972]. They are, therefore, thermodynamically unstable (G0) and droplets tend to decrease their interfacial area by coalescence, leading to coarsening of droplet size and finally irreversible separation of the phases (‘coalescence’, ‘cracking’). Prevention of coalescence is thus used as a synonym for stabilisation of an emulsion. This can be achieved by use of an emulsifier to reduce the interfacial tension () and interfacial free energy (G) which are equivalent to tendency to coalesce [Hiemenz, 1986 / Stricker, 1987]. Although G can be seen as the driving force for coalescence, droplet approach is governed by other factors like steric hindrance [Fisher and Parker, 1988], electrostatic attraction and repulsion according to DLVO-theory [Derjaguin, 1989], hydration forces [LeNeveu et al., 1976] and emulsifier film viscoelasticity [Boyd et al., 1974]. This is also true for the interfacial barrier of lecithin-stabilised O/W emulsions [Hansrani, 1980 / Davis and Hansrani, 1985 / Fisher and Parker, 1988], where various aspects of stability need to be considered.

1.2.2.1 Stability against Recoalescence

During homogenisation of an emulsion, the dispersity of the internal phase is increased by breaking up its structure by means of high shear forces, cavitation impact, turbulence, etc. . Depending on the conditions existing, the droplets formed in situ will instantaneously start to ‘recoalesce’, if maintenance of this state of dispersion cannot be achieved [Dickinson and Stainsby, 1988], which is illustrated in Fig. 9. Only dissolved, molecularly dispersed emulsifier molecules are capable of stabilising the newly-created interface. If a high energy input during homogenisation creates too large an oil-water-interface to be either stabilised by this limited amount of free emulsifier or by too slow adsorption of emulsifier at the interface, recoalescence will occur. Low molecular weight emulsifiers (e.g. < 500) are transported more rapidly to the interface by diffusion and conduction, and can thus prevent coalescence on a time scale of milliseconds to a few minutes [Das and Kinsella, 1991 / Schubert, 1997]. Emulsifiers of higher molecular weight, e.g. proteins or polymeric emulsifiers, exhibit slower transport rates and thus less effective short time range coverage of a newly-created interface [Schubert, 1997 / Stang and Schubert, 1997]. According to Gibbs’ adsorption isotherm (Eq. 2), the emulsifier concentrations at an interface and in bulk solution at equilibrium are related via the interfacial tension:

[mol/m2] gives the excess concentration of the emulsifier at the interface, c [mol/l] its concentration in the bulk solution, [N/m] is the interfacial tension, R the gas constant and T the absolute temperature [K]. is approx. 25 mNm-1 for a pure triglyceride-water-interface [Fisher and Parker, 1988]. PC, as the major component of lecithin, has a very low aqueous solubility (Small [1986] quotes 4.610-10 mol/l for DPPC). Additionally, a second lecithin phase exists in aqueous dispersions in the form of liposomes. Eq. 2 predicts that now equilibrium between lecithin concentration at the O/W interface and in bulk solution as monomers lies greatly in favour of the interface, i.e. is large. Convective transport of lecithin in the form of liposomes to freshly-created interface during homogenisation appears intuitively more probable than passive diffusion processes.
Accordingly, stability refers to a short-time stability during homogenisation, where creaming becomes negligible. An emulsifier suitable for long-term stabilisation of larger droplets, therefore, will not necessarily be sufficient for production of small particle size. Owing to this fact, co-emulsifiers are frequently used for this purpose, which either enhance the initial effect of the main emulsifier, or often also show complementary effects regarding stabilising mechanisms (see Section 1.2.2.4).

Karbstein [1994] reported that increasing the fraction of the dispersed phase, , does not increase rate of recoalescence during homogenisation, although droplet collision events are increased, provided the emulsifier stabilises the newly-created interfaces sufficiently fast. This is also true for high , since now higher viscosity and less collision events are expected. Fisher and Parker [1988] suggested that although droplet break-up may be improved by higher temperatures, a more rigid emulsifier film allows better stabilisation and can be obtained by cooling the sample during homogenisation.

1.2.2.2 Stability during Autoclaving

Emulsions are readily destabilised and coalesce when stress (viz heat, freeze-thawing or shaking) is applied upon them [Becher, 1965]. Parenteral fat emulsions exhibit, however, a remarkable resistance against heat-stress during terminal heat-sterilisation, where 121°C is maintained for  15 min. This has been related to their structure, suggesting either the existence of a liquid-crystalline mesophase at the oil-water-interface [Groves et al., 1985 / Groves and Herman, 1993], which was suggested to prevent droplet coalescence by increasing the emulsifier film viscoelasticity. Washington and Davis [1987] stressed the importance of free fatty acids, since these would increase electrostatic repulsion between droplets.
Since heat-sterilisation is unavoidable in parenteral emulsion production, instability in this state must be circumvented. This necessitates avoiding coalescence and its various stages over a short time range, and also enhanced chemical degradation, where physical and toxicological properties are concerned. Effects observed at elevated temperature are accelerated chemical and physical degradation reactions, which appear to be interdependent [Herman and Groves, 1992].

1.2.2.3 Stability against Chemical Degradation

Possible chemical degradation within parenteral fat emulsions includes oxidation of unsaturated fatty acid residues in the triglyceride or lecithin molecules [Kemps and Crommelin, 1988] and hydrolysis of phospholipids [Grit et al., 1989 / Herman, 1992].

Oxidation of double bonds within the lecithin may influence the nature of the O/W interfacial film. As a possible mechanism, Magdassi et al. [1991] proposed that despite oxidation, polymerisation of molecules could also take place and subsequently desorption from the interfacial layer. The detection of initial formation of conjugated dienes is reported by New [1990] using UV absorption at 230 nm. Magdassi et al. quote that oxidation level could be determined by monitoring the ratio of UV absorption at 233 nm and 215 nm, whereas other authors use the ratio of the absorptions at 234 nm and 270 nm to follow oxidation of unsaturated fatty acids [Belitz and Grosch, 1982]. Despite trace amounts of the natural antioxidant tocopherol contained in soy bean oil, all production steps must be carried out under inert gas purging.

Hydrolysis of phospholipids leads to formation of free fatty acids, lyso-phospholipids and glycerophosphorylic compounds (Fig. 10). Grit et al. [1989] and Herman and Groves [1992] determined hydrolysis rates of PC and PE in aqueous media at various temperatures, which were lowest at pH 6.5. However, different hydrolysis rates were determined, according to the concentration of additional phospholipid components [Grit et al., 1991]. The degradation products are more water-soluble than the original phospholipids, which may influence emulsifier distribution at the O/W interface, and also promote further degradation as they cause lowering of pH [Boberg and Håkansson, 1964 / Kawilarang et al., 1980 / Herman and Groves, 1992]. Accordingly, adjustment of pH of an emulsion before autoclaving provides physiological pH in the finished product, but also allows less hydrolysis to occur in the final product during storage, where pH is about 6-8.

Hydrolysis products need to be minimised since lyso-phospholipids exhibit haemolytic effects in vitro and have, therefore, to be regarded as toxic [Saunders, 1957 / Weltzien, 1979]. It has been demonstrated, however, that LPC has a stabilising effect on PC vesicles by adhering to the bilayer [Kumar et al., 1989]. This could reduce the concentration of free lyso-compounds in the emulsion, thus explaining why toxicity in vivo does not occur despite up to 8% lyso-compound concentration [Muehlebach et al., 1987], or 8.2% LPC and 3% LPE [Herman, 1992] have been reported. Hansrani [1980] also reported a stabilising effect of LPC on lecithin-stabilised emulsions from admixing experiments using a droplet-surface coalescence model.

Hydrolysis of triglycerides to fatty acids and di- or mono-glycerides is also theoretically possible in these emulsions. The resulting hydrolysis products again would be surface active, with possible changes in emulsifier film properties and an influence on pH of the bulk water phase.

1.2.2.4 Stability against Physical Degradation

Emulsions are inherently unstable. Comprehensive discussions of the physical stability of emulsions are available elsewhere [Becher, 1988 / Dickinson and Stainsby, 1988 / Friberg and Larsson, 1997], and only the major processes involved are summarised in Fig. 11. The driving force of coalescence is always the reduction in Gibbs free energy (G), achieved by reducing the size of the O/W interface (A) of tension (), despite the associated decrease in entropy, (TS).

Disruption of a continuous soya oil phase into droplets of 300 nm within water increases the interfacial area by approx. 105-fold, whereas the presence of 0.5% w/v lecithin reduces only approx. 68-fold [Mueller and Heinemann, 1993]. Therefore, additional factors affecting the mechanisms of coalescence must be considered for lecithin-stabilised soya oil-in-water emulsions. Droplet coalescence irreversibly changes the quality of dispersion and is a first-order process [Cockbain and McRoberts, 1953 / Van den Tempel, 1953], where the rate-determining steps are either the drainage of the continuous liquid film between two approaching droplets to its rupture thickness, or the probability of film collapse occurring as a consequence of mechanical distortion when equilibrium thickness has been reached [Fisher and Parker, 1988]. At surfactant concentrations well above the CMC (which is certainly the case in lecithin-stabilised emulsions), the rate of film drainage, -dD/dt, follows approx. the Reynolds equation (Eq. 4) for two rigid discs spaced by the interfacial film [Fisher and Parker, 1988].

is the net interdroplet attraction force, is the viscosity of the continuous phase, and is the radius of the discs. The processes which counteract film drainage and thus inhibit coalescence are:

Coalescence is accompanied by flocculation and sometimes also creaming. The first describes the adhesion of dispersed droplets, which yet remain single bodies. The latter describes the movement of the less dense phase (in the case of O/W emulsions i.e. triglycerides) to the upper region of the container. In both cases the droplet number and their size distribution remain unchanged, and although distribution of the droplets becomes inhomogeneous, this can often readily be reversed by gentle agitation. Fisher and Parker [1988] stated that interfacial multilayers of surfactant may enhance flocculation, but allow easy redispersion since multilayers are sensitive to shear force. This could cause enhanced mechanical stabilisation against coalescence during droplet collision. However, formation of larger aggregates may be facilitated, which in turn promotes faster creaming and may thus facilitate coalescence. This is especially true for highly polydisperse systems where different creaming rates produce enhanced droplet encounter rates [Dickinson and Stainsby, 1988]. Small particles, which remain in Brownian dispersion, are therefore separated from the larger particles.

The contribution of negative surface charge (Zeta potential) of lecithin-stabilised emulsion droplets to their ability to withstand flocculation or coalescence has been frequently investigated. There is substantial agreement that increasing Zeta potential gives rise to enhanced stability against various stress factors as predicted by DLVO-theory [Kawilarang et al., 1980 / Washington and Davis, 1987 / Washington et al., 1989 / Ishii et al., 1990 / Rubino, 1990 / Chaturvedi et al., 1992]. However, enhanced stability was not reported in all cases; sometimes coalescence was enhanced or flocculation increased after addition of electrolytes or charged lipids [Rubino, 1990 and Muchtar et al., 1991]. Davis et al. [1985] and Mueller and Heinemann [1993] stressed that also the mechanical barrier functions of the lecithin film have to be taken into account in order to estimate stability of an emulsion.

Phase inversion and Ostwald ripening are considered to be of no importance for parenteral fat emulsions. In the latter case a pronounced solubilisation of the oil phase in surfactant micelles in the aqueous phase would be required to allow migration of oil through the aqueous phase [Dickinson and Stainsby, 1988]. This is clearly not possible owing to PC’s extremely low solubility in water.

1.2.2.5 Stability against Microbiological Degradation

Microbiological contamination always has to be considered when water is the continuous phase of an emulsion and is, of course, completely unacceptable for intravenous application. Since preservatives are not allowed as additives in parenteral emulsions [Ph.Eur.1997, 1997], even refrigerated storage of opened containers will not suffice to prevent microbiological degradation. Only manufacture under aseptic conditions with terminal steam-sterilisation can ensure the product’s sterility.

1.2.2.6 Stability against Admixing

Parenteral emulsions are often admixed with other injectable solutions to yield smaller total volumes for total parenteral nutrition (TPN) regimes. However, undesirable physical destabilisation is unpredictable, and prevents commercial pre-production of such TPN-regimes. Interactions of admixed components with parenteral emulsions have been discussed by Dawes and Groves [1978], Whateley et al. [1984], Davis et al. [1985], Washington et al. [1990], Washington et al. [1991], Washington [1992] and Mueller and Heinemann [1994]. It is important to consider compatibility with various amounts and kinds of admixed parenterals like carbohydrates, amino acids and ionic solutes, which can be regarded as an additional stress towards the emulsion. Since it still remains unclear, whether surplus emulsifier acts as a ‘buffer’ that enhances stability of emulsion droplets against degrading factors like ionic admixtures, the newly marketed 10% formulations (like e.g. Intralipid 10®) which contain reduced emulsifier excess, are believed not to be suitable for use in such mixed formulations as the respective 20% emulsions.

1.2.3 Model Theories of Parenteral Emulsion Structure

According to IUPAC, emulsions are dispersions of immiscible liquids and/or liquid crystals in other liquids [International Union of Pure and Applied Chemistry, 1972]. Boyd et al. [1974] reported the marked improvement in stability of an emulsion owing to the pronounced viscoelastic properties of liquid-crystalline mesophases as the interfacial film. Friberg et al. [1976], Rydhag [1979] and Rydhag and Wilton [1981] have all reported that phospholipids that form liquid crystalline phases, show lamellar layers at the oil-water interface which enhance emulsion stability. Rydhag [1979] showed that negatively-charged lipids increased the swelling of liquid-crystalline phases and related this finding to the enhancement of emulsion stability by using lecithins containing more negatively-charged components (PI, PS, PA). Groves and Herman [1993] postulated that reversible formation of viscous cubic phases could account for the emulsion’s stability against autoclaving. The systems reported by Rydhag [1979] and Rydhag and Wilton [1981] had, however, larger droplet size than typical for parenteral fat emulsions, which would result in a smaller O/W interfacial area. This causes a large emulsifier excess, which now forms the extensive liquid-crystalline structures reported [Westesen and Wehler, 1992].

Several authors have reported excess lecithin in parenteral emulsions, which could form liposomes beside the emulsion droplets [Groves et al., 1985 and Rotenberg et al., 1991]. Westesen and Wehler [1992] found that the dispersed phase consisted of ideal emulsion droplets covered by monolayers of the emulsifier. Groves and co-workers [1985] showed micrographs of multilamellar structures contained in the emulsions, whereas the latter authors reported the presence of unilamellar vesicles. However, Westesen and Wehler [1992] drew their conclusions from model systems containing higher lecithin/oil-ratios than in commercial systems. Groves and co-workers [1985] and Rotenberg et al. [1991] investigated centrifuged Intralipid® 10 samples, which possibly were changed in some way by centrifugation. Indeed, the mesophase which may form in the aqueous phase of an O/W emulsion and which theoretically may be able to adsorb on the interface as a multilayer, is not necessarily of the same composition as an interfacial mesophase, since the surface material may change over time [Fisher and Parker, 1988].

However, it still remains uncertain whether surplus lecithin (assuming an interfacial monolayer) contributes to the stability of the emulsions. If this were the case, it would thereby counteract the pharmacological demand for reduction of excess lecithin [Hajri et al., 1990]. This has been suggested by Groves et al. [1985], but contradicted by the reports of Chiba and Tada [1990] and Krafft et al. [1991], who claimed that excess lecithin may even destabilise emulsions. It has not been possible to clarify whether such excess emulsifier was inherent to lecithin-stabilised emulsions, or if it could be avoided. Although theoretically possible, the presence of micelles formed by lyso-compounds or free fatty acids in emulsions, has not been confirmed (e.g. [Westesen and Wehler, 1992]).

It is astounding that six decades of research in this field have resulted in substantial disagreement about the behaviour of the lecithin in emulsified structures and those not related with the interface. It is also not yet completely understood to which degree possible stabilisation mechanisms contribute to the stability of lecithin-stabilised O/W-emulsions during autoclaving.

1.3 Objectives of this Work

This work is intended to contribute a better understanding and a more precise definition of the ill-defined term ‘stability’ of parenteral fat emulsions. A combination of experimental methods was used here to give a more detailed insight into the structure of these ‘simple’ emulsions with their paradoxically ‘complex’ behaviour, which, undoubtedly is mostly due to the nature of the emulsifier.
Today’s knowledge of parenteral fat emulsions is derived from two sources: studies of commercial emulsions and experimental data from model emulsions. Even in the commercial systems different compositions and manufacturing procedures are used, making comparisons difficult. This is certainly the reason why previous work does not give a complete image of the structures present and how these influence stability and therapeutic efficacy. All studies on commercial samples reported here were undertaken with the respective up-to-date formulations (e.g. lecithin-reduced 10% formulations). Model systems made from ‘outdated’ composition were also investigated to confirm older reports, which dealt with former commercial formulations. As Westesen and Wehler [1992] stressed, lecithin-stabilised parenteral fat emulsions cannot be seen as mere monomodal oil-in-water-dispersions, since pharmacological (e.g. lipoprotein-x-like particles, hypercholesterolemia) and physical phenomena (e.g. ability to tolerate admixtures of drugs or i.v. solutions) are related to the existence of possible additional structures. Innovations such as admixture of amphiphilic or lipophilic drugs or incorporation of new fat components are new challenges to the developer, since emulsion structure and stability in these cases cannot necessarily be compared with conventional emulsions [Hamilton et al., 1996].
In this thesis, an elucidation of the structure of egg lecithin-stabilised fat emulsions is presented. The following questions define the objectives of this work:


To these ends
31P-Nuclear Magnetic Resonance Spectroscopy, X-ray analysis, Freeze-Fracture- and Cryo-Transmission Electron Microscopy were used together with Fourier-Transform-Infrared-Spectroscopy and Differential Scanning Calorimetry to characterise various emulsifier compositions as well as emulsions and separated phases. Asymmetrical Flow Field-Flow Fractionation helped to separate the systems and elucidate structure of liposome dispersions and model emulsions. To examine their physical stability, particle sizing using Photon Correlation Spectroscopy, Laser Diffractometry with and without PIDS and other techniques were used and judged regarding their advantages and disadvantages. Chromatographic analyses of the emulsifier composition together with Zeta Potential Measurements were used to follow chemical stability and its impact on the physical properties of the emulsions. The addition of minor components and their effect on interfacial properties using droplet/interface coalescence kinetics was investigated. Furthermore, emulsions of known properties were creamed and resuspended in different aqueous media to examine the influence of the latter on the stability of the dispersions.

Previous PageNext Page


Konvertiert vom Dissertationen Online Team im CCC der Universität Erlangen

dissertationen@ccc.chemie.uni-erlangen.de