CLINICAL NUTRITION (1986) C: 11-27 

In Vitro and Clinical Studies on Intravenous Feeding Mixtures 

Comprising Fat Emulsion, Amino Acid and Electrolytes S. S. 
Davis*, M. Galloway*, W. R. Burnhamf and L. Stevens-f *Pharmacy Department, University of Nottingham, 
Nottingham NG7 2RD, UK and tOldchnrch Hospital, Romford, Essex, UK (Reprint requests to S.S.D.) 
ABSTRACT 
This paper describes the effect of the amino acid mixture Freamine II” on the physical stability of Intralipid” 
fat emulsion. The amino acid-fat emulsion mixture was also administered to patients requiring intravenous nutrition and its 
clinical and biochemical effects were assessed. The amino acid-fat emulsion mixture was stable on storage for 48 hours as judged 
by flocculation and particle growth measurements made in the presence of 20 mmol!l monovalent cations at 2 mmol/l diavalent 
cations. 
Long-term storage of mixtures containing Freamine II’” is not recommended because of an interaction between the emulsifier 
and the amino acids. In a small open trial patients maintained their serum albumin concentrations. There was no evidence of 
impaired pulmonary function even in those subjects with impaired fat clearance. 
INTRODUCTION 
Lipid emulsions based upon vegetable oils (soybean oil, safflower oil) emulsified with egg or soy phosphatides have an important 
role in parenteral nutrition. Until recently, because of potential instability problems, they have been administered separately from 
other nutrients such as carbohydrates, amino acids, electrolytes and vitamins. 
Reports in the literature now suggest that it is possible to produce mixed nutrition systems containing all required components 
which will be stable for periods up to 48 h, provided proper consideration is given to the type and total quantities of added 
electrolytes and their effect on the physical stability of lipid droplets, as well as the known stability characteristics of vitamin 
preparations [ 1, 21. 
While added electrolyte is expected to have a deleterious effect upon emulsion stability, it has been suggested that amino acids 
could have a protective and stabilising effect [3]. The present studies were undertaken to examine the effect of the amino acid 
mixture Freamine II’” on the physical stability of the fat emulsion Intralipid’“, in the presence and absence of added monovalent 
and divalent electrolytes. Selected formulations were also evaluated clinically in patients requiring parenteral nutrition. 
MATERIALS AND 
Intralipid’” [soybean 
stabilised by egg lecithin (1.23,,)] was obtained from Kabi. A balanced mixture of essential and non- essential amino acids 
(Freamine II “) was obtained from Boots. The added electrolytes were all of analytical grade. In clinical studies Addamel “,, 
Vitlipid” and Solivito I( were obtained from Kabi and dextrose 109, from Travenol. 
Preparution of mixtures 
Parenteral nutrition mixtures were prepared under aseptic conditions and stored at 4’C in the dark in Viaflex ‘( containers. The 
formulations studied in vitro are given in Table 1 and the clinical mixtures are given in Table 2. 
In vitro tests 
Particle size analysis 
The particle size characteristics of the emulsion systems were determined as previously described [2] using three different 
methods that will cover the complete size range of particles in intravenous fat emulsion; photon correlation spectroscopy (10 nm- 
2 pm), Coulter Counter (600nm-20 pm) and light microscopy (10 pm-100 pm). The emulsions were also examined visually for 
separation of free oil. 
METHODS 
Measurement of surface charge 
in water (20’/,) emulsion 
The charge on the droplets was measured at 25°C (in 
21 
 
22 
INTRAVENOUS FEEDING MIXTURES COMPRISING FAT EMULSION, AMINO ACID AND ELECTROLYTES 
Table 1 Formulations used for in vitro tests on emulsion stability 
Intralipid’” Distilled Calcium Magnesium Sodium Potassium 
20”;, water Freamine II IL chloride chloride chloride chloride Formulation (ml) (ml) (ml) (mmol) (mmol) 
(mmol) (mmol) 
IA 50 100 - - 
B 50 50 50 - - - 
2A 50 100 - 0.25 - 
B 50 50 50 0.25 - - 
3A 50 100 - 0.25 - 
B 50 50 50 - 0.25 - 
4A 50 100 - - - 2.5 B 50 50 50 - - 2.5 5A 50 100 - - - - 2.5 B 50 50 50 - - - 2.5 6A 50 100 - - 0.25 2.5 - 
B 50 50 50 - 0.25 2.5 
Table 2 Intravenous feeding mixtures for clinical use 
3 litre bag containing: 
1 1 20% Intralipid 1 I FreamineII 11 10% dextrose +Addameln, Vitlipid”“, SolivitooQ” and additional potassium and phosphate 
and sodium (as potassium chloride, potassium dihydrogen phosphate and 304, sodium chloride respectively) according to 
requirements. All additions of electrolyte were within the stability limits for total added electrolyte load described previously [2]. 
“Omitted from feeds stored over the weekend 
some cases over a range of pH values) using the technique of particle microelectrophoresis [2]. Distilled water (pH 6.4-6.6) was 
selected as the dispersion medium in these studies to provide high values for surface charge. Measurements were made 
immediately after preparation and after 28 days storage. A further indication about the surface characteristics of the fat droplets 
was obtained by incubating mixed emulsion systems with an equal quantity of human blood plasma for 15 mins at 37”, followed 
by the redetermination of surface charge. 
Accelerated stability test 
The breakdown of fat emulsions can be accelerated by methods such as centrifugation, shaking and freeze-thaw cycles. The 
freeze-thaw experiment is often a good predictor of long-term stability and was utilised in the present work [4]. A representative 
sample of each mixture (5 ml) was frozen in a deep freeze and then allowed to thaw at room temperature before particle size 
analysis. 
Clinical studies 
Patient selection and assessment of safety 
Sixteen patients at risk of or with established malnutrition were entered in the study (7 men, 9 women). Three had inflammatory 
bowel disease requiring surgical treatment, one had a perforated duodenal ulcer, 10 had gastrointestinal malignancy and one had 
leukaemia. Two of the patients with inflammatory bowel disease and six of the patients with malignant disease commenced 
intravenous nutrition before surgical treatment. 
All patients had daily clinical assessment and a note was made of their disability on the Karnofsky scale (Table 3). As 
aggregated or coalesced fat particles in the emulsion might theoretically cause fat embolism in the lungs, special emphasis was 
placed on the detection of this. When it was possible to move the patient to the pulmonary function laboratory, carbon monoxide 
transfer factor was measured during IV feeding and either before or after feeding. This would be expected to identify any 
Table 3 The Karnofsky rating scale 
0 Fully active, able to carry on all usual activities without 
restriction and without the aid of analgesics 1 Restricted in strenuous activity but ambulatory and able to carry out light work 
or pursue a sedentary occupation. Also patients who are fully active as in grade 1, but only with the aid of analgesics 2 
Ambulatory and capable of all self care, but unable to 
work. Up and about for more than 50% of waking hours 3 Capable of only limited self care, confined to bed or 
chair more than 50% of waking hours 4 Completely disabled and unable to carry out any self 
care and totally confined to bed or chair 
 
23 
Table 
4 
CLINICAL NUTRITION 
adverse effect on pulmonary gas exchange during infusion. When it was not possible to perform this test, isotope 
ventilation/perfusion lung scans were performed. 
All patients had daily blood and urinary urea and electrolyte estimations. Blood counts, liver function tests, serum albumin, 
cholesterol and triglycerides were measured twice weekly or more frequently if clinically appropriate. It was planned to study 
each patient for at least 14 days. 
Parenteral nutrition mixtures were normally infused over the 24 h immediately after preparation, except at weekends, when the 
feeds were kept refrigerated for up to 40 h before infusion. Vitamins were not added to feeds that were so stored. 
These studies were approved by the Ethical Committee of Oldchurch Hospital, Romford. 
RESULTS 
Storage stability 
Particle size analyses were conducted at 0, 24 and 48 h using the PCS and Coulter Counter methods. The PCS data are expressed 
in Table 4 as the mean diameter (z average). Slight changes in size were indicated for systems 4-6 especially at 48 h. The Coulter 
Counter data are given in Table 5 as the mean diameter and the percentage of particles below an arbitrary size of 1.5 pm. (Note 
the Coulter Counter diameter is larger than that obtained by PCS since it is a mean volume diameter rather than a number 
diameter and the lower measurement limit of the Coulter Counter used in the present studies was 
The stability of mixed emulsion systems on storage (Photon Correlation Spectroscopy data) All figures are mean 
diameter in microns 
Storage time 
at 4 C 
Formulation Oh 24h 48h 
1A 0.369 0.346 0.367 B 0.360 0.361 0.366 
?A 0.309 0.354 0.356 B 0.352 0.376 
3A 0.366 0.382 0.429 B 0.364 0.381 0.407 
4A 0.375 0.356 0.390 B 0.360 0.351 0.371 
5A 0.373 0.371 0.409 B 0.355 0.378 0.408 
6A 0.366 0.377 0.410 B 0.361 0.426 
about 500nm.) No significant changes in mean size nor percentage less than 1.5 km were observed. Microscopic and visual 
examination of the mixed samples failed to demonstrate the presence of free oil droplets in any of the samples. 
Surface charge 
The results shown in Table 6 indicate that the emulsion carried a negative zeta potential and that this was increased somewhat by 
the storage of the system in the presence of the amino acid solution. There was no significant change in zeta potential over a 28 
day period. 
Table 5 
The stability of mixed emulsion systems on storage (Coulter Counter data) Storage at 4°C 
Oh 24h 48h 
Mean Percent size below Formulation (pm) 1Spm 
1A 0.80 99 B 0.79 97 
2A 0.80 93 B 0.82 95 
3A 1.15 86 B 0.82 97 
~4 A 0.85 94 B 0.83 100 
5A 0.94 92 B 0.80 99 
6A 0.95 94 B 0.97 90 
Mean Percent size below (pm) 1.5pm 
0.82 90 0.75 90 
0.80 97 0.82 100 
0.83 98 0.82 98 
0.79 97 0.79 98 
0.80 99 0.80 99 
0.78 97 0.80 99 
Mean Percent size below (pm) 1.5pm 
0.84 99 0.81 99 
0.97 90 0.77 95 
1.05 90 0.80 99 
0.82 100 0.76 94 
0.81 98 0.80 99 
0.80 97 0.98 85 
I‘N R 
 
24 INTRAVENOUS FEEDING MIXTURES COMPRISING FAT EMULSION, AMINO ACID AND EI.EC’fRC)I.Y’I‘ES 
Table 6 Surface charge characteristics of emulsion formulations (pH 6.4-6.6) 
Od 28 d 
Zeta Zeta Mobility” potential Mobility” potential Formulation p/s/V/cm (mV) NV/cm (mV) 
1A 1.67 -25 - - 
B 2.15 -32 2A 1.35 -20 1.48 -22 B 1.97 -30 2.01 -30 3A 1.28 -19 1.54 -23 B 1.87 -28 1.71 -26 4A 1.85 -28 - 
B 2.34 -35 5A 2.13 -32 - - B 2.09 -31 - - 
6A 1.73 -26 1.47 -25 B 2.05 -31 1.89 -28 
“Standard deviation not greater than 0.25 mobility units 
The variation of mobility with pH for Intralipidm and Intralipid@ stored with Freamine’!” is shown in Fig. 1. It will be seen 
that Freamine@ produces an increase in zeta potential at all pH values examined. Incubation of the mixed emulsion systems with 
human plasma followed by dilution in water leads to 
4 
-E 
Table 7 The effect of incubation of emulsion systems with human blood plasma on surface charge characteristics (pH 6.4-6.6) 
Electro- Change phoretic Zeta in zeta mobility potential potential Formulation 
P/S/V cm (mV) (mV) 
1A 0.69 -10 + 15 
B 0.57 -9 +23 2A 0.81 -12 +8 
B 0.59 -9 $21 3A 0.84 -13 +6 
B 0.63 -9 + 19 
“Reduction in zeta potential due to adsorbed plasma protein. Value obtained by comparing equivalent systems in Table 6. 
a reduction in electrophoretic mobility (and therefore zeta potential) (Table 7). The change in zeta potential due to adsorbed 
plasma components can be calculated by comparing these data with those in Table 6. The systems originally containing amino 
acid demonstrate a greater reduction than those without amino acid. 
Accelerated ageing and long-term stability tests 
All emulsion systems demonstrated a reduction in stability after one freeze-thaw cycle (Table 8). A 
3 
. . . 
U > In 
32 
Z .- 0 + % 
0 
= 
1 
PH Fig. 1 pH-electrophoretic mobility profile for Intralipidm 
emulsion droplets after storage with an amino acid solution (Freamine II)@. Mean data (standard deviation not greater than 0.25 
mobility units). Legend: Intralipid‘aj 0 
Intralipid stored in 207, Freamine II”“’ 0 
 
Table 8 Particle size analysis of mixed systems following one freeze/thaw cycle (Coulter Counter) 
Experiment 1 Experiment 2a 
Mean Percent Mean Percent diameter less than diameter less than Formularion (pm) 1.5pm (pm) 1.5 pm 
1A B 0.96 95 5.21 ‘1 
- 2A 1.25 81 0.99 95 B 4.41 22 3.25 31 SA 0.95 91 0.92 93 B 2.67 39 3.33 30 4A 1.30 
85 B 3.57 36 SA 0.92 90 B 2.45 55 6A 0.87 96 0.85 95 B 1.86 64 3.02 40 
“Repeat study on selected formulations 
dramatic difference is seen between those systems with and without Freamine’“. The latter were much more stable. Samples of 
the mixed system were stored for a period of 4 weeks in order to assess long- term stability and it was noted that those systems 
containing FreaminetR showed a distinct separation of free oil, while those systems without Freamine were more stable. 
Clinical studies 
Four patients died while on intravenous feeding for reasons related to their condition (extensive gastrointestinal malignancy). The 
extent of this had not been apparent before attempted surgical treatment. There was no reason, on clinical grounds, to indicate a 
contributory role for the feeding mixture. Two of these patients had post mortems and no lipid accumulation was found in the 
lungs. In two other patients the lung function tests were not completed; in one case because the patient went on a ventilator 
post-operatively. 
Data on the 10 patients who completed the 14- day study period are shown in Table 9. It can be seen that there was no 
evidence for respiratory impairment to intravenous feeding as judged by measurement of transfer factor or isotope lung scans. 
Apart from patients L.S. and J.O., no patient showed a significant fall in serum albumin concentration during the study. In these 
patients, changes in fluid balance in the first case and the effects of septicaemia in the second would appear to account for this. In 
CLINICAL. NUTRITION 25 
three cases with extensive malignancy and/or severe infection, there was an overall negative nitrogen balance over the period of 
study. Six patients had abnormal liver function tests (bilirubin, alkaline, phosphatase and AST) during intravenous feeding but 
only in two cases did this develop during intravenous feeding and, in both cases, this could have been due to other causes 
(infection, hepatic metastases). In one patient with markedly abnormal liver function tests, lipid clearance was mildly impaired 
but no deterioration in lung function occurred as a result. 
DISCUSSION 
Ztl vitro studies 
The results of particle size analyses on systems stored for up to 48 h at 4°C indicate that there is little or no change in the particle 
size of the mixed emulsion systems, whether formulated without electrolytes or with monovalent and divalent electrolytes or a 
mixture of both. Previous studies have indicated that fat emulsions mixed with amino acids and electrolytes can tolerate levels of 
monovalent and divalent electrolyte up to 15Omm0lI-~ and 2.5 mmoll-’ respectively, without problems of instability attributed to 
flocculation and coalescence of individual droplets [2]. Added amino acids would be expected to provide a protective effect by 
(a) removing Ca+ + from the solution if present, (b) increasing the buffer capacity of the system as well as the pH, and (c) by 
adsorption at the oil-water interface of the fat droplet, thereby increasing surface charge [3]. The electrophoretic mobility data 
indicate that a component of the amino acid mixture may well be adsorbed strongly at the interface leading to an increase in 
surface charge, electrostatic repulsion between fat droplets and thus an enhanced stability [5]. Incubation with plasma may well 
lead to removal of such a component from the interface since the mobilities of the fat droplets after incubation in plasma and 
subsequent dilution in water were very similar whether the system had been treated with amino acids or not. 
The accelerated test involving a freeze-thaw cycle was able to discriminate between systems containing Freamine R and those 
without. Good correlation was found between this test and long term storage stability of mixed systems. The destabilising effect 
of amino acid is not yet clear but it should be noted that Pamper1 and Kleinberger [6] have reported pronounced changes in the 
structure of fat droplets stored for extended periods of time with amino acid solutions. 
 
 
From the results obtained it can be concluded that Intralipid’” emulsion mixed with FreamineRl’ and levels of electrolyte 
components to 20 mmollll (monovalent cation) and 2 mmollll (divalent cation) should present no problems with regard to 
physical stability (flocculation and particle growth) for storage periods of up to 48 h. Indeed higher levels of electrolyte could 
well be permissible because of the known protective effect of amino acids against particle flocculation and coalescence. 
Long-term storage of mixed systems containing Freamine:” is not recommended because of an interaction between the emulsifier 
used in the fat emulsion and the added amino acids that leads to destabilisation. 
Clinic,al studies 
The number of patients entering the trial and who completed the study was small. This was due to the fact that intravenous 
feeding is reserved for those who cannot be adequately nourished by other means. As a result, only the very ill were included and 
such patients have a high complication and mortality rate. 
Those patients who did complete the study are evidence of the effectiveness of the treatment they received and this must 
include the intravenous feeding. Although three patients had a negative nitrogen balance this was related to the severity of their 
malignant disease plus associated sepsis and surgical treatment. It may be relevant that all of these patients also had abnormal 
liver function tests. In each case the serum albumin did not fall over the 
REFERENCES 
[l] Black C D, Popovich N G 1981 A study of intravenous 
emulsion compatibility. Effects of dextrose, amino acids and selected electrolytes. Drug Intell. Clinical Pharmacology 15: 
184-193 (21 Bumham W R, Hansrani P K, Knott C E, Cook J A, 
Davis S S 1982 Stability of a fat emulsion based intravenous feeding mixture. International Journal of Pharmaceutics 13: 9-22 [ 
31 Hardy G, Cotter R, Dawe R 1983 The stability and 
comparative clearance of TPN mixtures with lipid. In: 
Submission date: 27 March 85. Accepted after Revision: 24 June 85. 
CLINICAl. NUl‘RI’1‘ION 27 
period of study and the nitrogen balance returned to a positive level at the end of the two week study period. All of these patients 
were alive 12-18 months after intravenous feeding. 
Clinical and radiological measurements, together with measurements of carbon monoxide transfer factor showed no evidence 
of impaired pulmonary gas exchange associated with the intravenous feeding, even when lipid clearance was impaired. In those 
patients who could not co-operate in measurement of transfer factor, isotope scans of the lungs showed no change during 
intravenous feeding. Although the latter is not a very accurate method of assessing abnormalities, extensive pulmonary vascular 
obstruction by lipid should have been detected by this test if it had occurred. Other complications of intravenous feeding, such as 
abnormal liver function, were mild and did not cause clinical problems. The intravenous feeding did appear to prevent a 
deterioration in the clinical condition of the patient during the severest stage of their illness; this was evidenced by the clinical 
results as well as, in most cases, the biochemical data. 
It may be concluded that intravenous feeding with Intralipid R -Freamine II ‘( mixture used was both safe and effective in the 
small number of severely ill patients studied. 
ACKNOWLEDGEMENTS 
This work was supported by the Boots Company plc. 
[41 
PI 
(61 
Johnson I D A (ed) MTP Press, Lancaster, pp 241-260 Davis S S 1983 Stability of intravenous fat emulsions. In: Johnson I D A 
(ed) Advances in Clinical Nutrition. MTP Press, Lancaster, pp 213-239 Kawilarang G, Georghiou K, Groves M J 1980 The effect 
of additives on the physical properties of a phospholipid-stabilized soya bean oil emulsion. Journal of Clinical and Hospital 
Pharmacy 5: 151-160 Pamerl G, Kleinberger H 1983 Allgemeine charakteristika und Fragen zur Galenik von Fettemulsion. 
Infusionther 10: 108-l 17