INTRODUCTION

Management of intestinal failure requires parenteral nutrition. For patients dependent on parenteral nutrition for longer than three months, home parenteral nutrition (HPN) is an opportunity for shortened hospital stays. The main indications for HPN in children include digestive tract diseases such as short bowel syndrome (59%), congenital enteropathies (10%), chronic pseudo-obstruction syndrome (9%) and inflammatory bowel diseases (5%) ¹. In the vast majority of children, intestinal failure begins during the neonatal period and usually persists throughout the first years of life. The main aim of the complex treatment of children with intestinal failure is to achieve optimal growth and weight gain, and so the provision of macro- and micronutrients is crucial. The high energy and protein demand of dynamically growing and developing pediatric patients together with the possible loss of nutrients due to diarrhea, impaired motility or high stoma output, presents a significant challenge in preparing nutritional admixtures. In the early stages after surgical resection, anti-secretory drugs are useful to decrease both stoma output and stool volume ². Total parenteral nutrition (PN) admixtures may be composed of different compounds which are mixed together, and so compatibility and stability must be considered ³. This is especially true for parenteral admixtures administered to premature infants, given the low final dose volume ⁴. The most critical parameters are the physical stability of these admixtures and the droplet size of the emulsions. If the droplet size exceeds the size of erythrocytes (6-8 µm), embolism can occur, which may have fatal consequences. The droplet size is influenced by many factors such as: the presence of electrolytes, especially those with a positive charge, can change the negative charge of the droplet surface of lipid emulsion (zeta potential), resulting in the coalescence of the droplets and even phase separation ⁵. All components, including vitamins, should mix together in order to avoid the requirement for patient manipulation in total parenteral nutrition (TPN) admixtures ⁶, and secondly, the presence of vitamins in PN protects lipid emulsions from oxidation ⁷. The second concern is the possibility of precipitation of calcium phosphate in pediatric parenteral admixtures. The most daunting problem for pharmaceutical preparation practices of PN is related to the day-to-day change of volume and formulation nutrients, due to changing clinical conditions and maturation. Considering the low dose volumes used in neonatology (as low as 1-2 ml), it is vital to ensure both physicochemical compatibility and adequate calcium and phosphorus supply ⁸. Further, newborn infants and small children require large amounts of calcium and phosphate in these low dose volumes which may increase the risk of precipitation of calcium phosphate ^{9) (10}. Precipitation can cause respiratory distress and pulmonary embolism ¹¹. Inorganic phosphate in PN has the potential of forming precipitates with calcium, especially when using calcium chloride. As such, calcium gluconate is the predominant calcium salt form employed in PN compounding due to its solubility profile with phosphate. Unfortunately, calcium gluconate contains higher levels of aluminum contamination than calcium chloride, resulting in an increased potential for aluminum toxicity in patients receiving traditional PN. The risk of aluminum toxicity is especially present in the neonatal population, in whom higher per kilogram of bodyweight amounts of calcium and phosphates are administered. Organic sodium glycerophosphate (NaGP), with the divalent calcium ion, has a lower propensity towards precipitation than inorganic phosphate, thereby allowing for calcium chloride utilization. Data presented in this study demonstrating NaGP and calcium chloride compatibility provide a clinical option for limiting aluminum contamination while providing sufficient calcium and phosphate to meet the needs of neonatal patients. Although NaGP is approved for use in Europe, it lacks full American Food and Drug Administration approval except on a temporary basis during phosphate drug shortages.

The aim of the study was to determine the physicochemical stability of admixtures for PN, prepared in one-chamber bags, designed for pediatric patients who receive PN at home. The stability of 30 compositions of admixtures up to eight days of storage at +4 °C was evaluated. The admixtures characterized contained increasing amount of electrolytes, and the consequent value of CAN (critical aggregation number) parameter ranged from 300 to 1402. The admixtures were prepared in the one-chamber bags Exacta-Mix Eva (ethylene vinyl acetate) Bag Parenteral (Baxa Ltd., United Kingdom) at the Hospital Pharmacy. Vitamins were added just before analysis. Three types of lipid emulsions: SMOFlipid, Lipofundin MTC/LCT or Omegaven were used. All admixtures were composed only of inorganic calcium salt. The composition of the parenteral admixtures under test was designed by clinicians from the Children's Memorial Health Center Institute in Warsaw, Poland.

METHODS

STORAGE OF INCOMPLETE AND COMPLETE ADMIXTURES TPN

Parenteral admixtures were stored for up to eight days at refrigerated temperature (4 °C). Prior to analysis, mixtures were removed from the refrigerator, and stored for two hours at room temperature (21 ± 1 °C), the vitamins were added (t = 8 days) (Table I). After sampling, the mixtures were stored up to 24 hours at room temperature (21 ± 1 °C) in the dark, then analyzed (t = 8 days + 24h) (Fig. 1).

Figure 1 Schedule of analysis of TPN admixtures. 

ADDITION OF VITAMINS

Parenteral admixtures were completed through the addition of vitamins (Table I). Vitamins were added under non-aseptic conditions, to better mimic the expected conditions of this stage, e.g. if a patient is at home. Vitamins were prepared by dissolving the freeze-dried water-soluble vitamins (Soluvit(r) N) in the fat emulsion comprising a vitamin soluble in the oil phase (Vitalipid(r) N) or Cernevit(r) in 0.9% sodium chloride injection (a mixture number: 25, 27, 30). For this purpose, each vial of Soluvit(r) N was added to 10 ml of Vitalipid(r) Infant and stirred to dissolve the Cernevit(r) vial in 10 ml of 0.9% sodium chloride solution for injection, and stirred until dissolved. Thus, dissolved vitamins were added using a syringe with a needle with filter in a prescribed amount to the TPN admixture.

Table I Concentration of electrolytes (mmol/l), CAN and CaxP parameters of TPN admixtures 

RESULTS

VISUAL AND MICROSCOPIC OBSERVATIONS

Over eight days of storage, no visual changes were observed in the test samples stored at room temperature or in 4 °C. There was no creaming or discoloration. Neither precipitates nor flocculation were visible. A visual inspection was done for the assessment of large particle formation in the critical size 1-5 μm.

Under microscopic observation, the mean of the largest lipid droplet in μm out of 15 visual fields of 5 μm as the upper limit value for the emulsion stability was never reached by any sample, except one admixture (TPN 4). The mean value of the larger oily globules was about 2-3 μm (Table II and Fig. 2). There was no trend for the droplet size to increase or decrease over time of storage. In the unstable TPN 4 admixture, oil droplets in the range of 4-10 µm were detected.

Table II Microscopic observations of admixtures 

Figure 2 Microscopic observation of admixtures (scale 10 µm). 

OILY DROPLET SIZE DISTRIBUTION

Laser diffractometry method

Value of median (d0.5) of oily droplets size in the complete admixtures was 310-390 nm and 90% of oily droplets (d0.9) were under 570-680 nm. Despite the various composition and time of storage no oily globules lager than 1 µm were detected in any of all admixtures by using laser diffractometry method (Fig. 3).

Figure 3 Distribution of oily droplets of TPN 25 admixture during storage. 

Photon correlation spectroscopy

Z-average of oily droplets size was in range 252-372 nm, while the polydispersion index was 0.062-0.223, which indicates monodispersity studied systems (Fig. 4). The smallest oily droplets (Z-average approx. 250 nm) was recorded in admixtures 28, 29 and 30. No significant changes (± 30 nm) of oily droplets were noticed during storage TPN mixtures (Fig. 4). The TPN admixtures 2 and 10 have only at t = 24h larger oily droplets (Z-average about 550 nm), but in the others point of analysis Z-average was approximately 310 nm, so these admixtures were concluded as stable. Admixtures 7, 16, 22, 25 and 26 in a single time point reported a second peak in the range of 0.5 µm, indicating the presence of larger droplets of oil. However, the analysis in the other time points did not confirm these changes.

Figure 4 Distribution of oily droplets of TPN 4 admixture during storage PCS method. 

Zeta potential analysis

Zeta potential of all admixtures was in range -19 to -38 mV and did not change during the storage (± 4 mV) when compared with samples at t = 0 (Table III). The lowest recorded zeta potential of -40 mV was noted with TPN 23 while TPN 30 was seen to have a zeta potential of -21 mV (Table III).

Table III Zeta potential (mV) of TPN admixtures 

pH measurement

The pH values in complete TPN admixtures were in range 5.29-6.29. Compared with samples at t = 24h, these values did not change (± 0.05 of units) during storage (Fig. 5). The smallest pH value (pH 5.3) of this parameter was observed in TPN 22, 25, 27, 30.

Figure 5 pH values of the TPN admixtures, the effect of storage. 

DISCUSSION

All TPN admixtures under investigation were prepared using standard procedures in the hospital Pharmacy. Due to clinical needs, admixtures contained a much higher than normal physiological concentration of electrolytes: calcium (3-10 mmol/l Ca²+), magnesium (1-10 mmol/l Mg²+) and potassium (7-60 mmol/l K+) ions. CAN parameter of investigated admixtures was much higher than current in range 300-1400 mmol/l and CaxP (the products of multiplication of calcium and phosphate ions concentration) was in range 1-43 mmol2/l2. The highest CAN parameter was noticed in admixtures 30 (1402 mmol/l), whereas the smallest was found in admixtures 14 (322 mmol/l). All investigated admixtures were prepared using inorganic calcium salt to check if they are safe in PN for small children and if they can be used in clinical practice. Many authors suggest that calcium gluconate is preferred over calcium chloride when compounding PN because of its superior compatibility with inorganic phosphates. PN solutions containing calcium gluconate carry a higher aluminum load than equivalent solutions compounded with calcium chloride, leading to increased potential for aluminum toxicity ¹². Premature infants are particularly at high risk of aluminum accumulation and toxicity as they often require days of PN support and have immature kidneys that are incapable of excreting aluminum efficiently. Calcium gluconate and phosphate salts are known to be especially high in aluminum content and are often administered to premature infants in substantial amounts to promote bone mineralization ¹³.

The analysis of physicochemical subjected to 30 TPN admixtures prepared in two bags in the Pharmacy of the Children's Memorial Institute in Warsaw. Incomplete (no vitamins) TPN admixtures were prepared in the single-chamber bags Exacta-Mix Eva bag Parenteral. Each composition of the TPN admixture was prepared in two bags, which allowed for analysis at four points: first bag, at t = 24, that is to say 24 hours after preparation (including transport), and after 24 hours of storage at room temperature (t = 24h + 24h); second bag, after eight days of storage at 4 °C (t = 8 days) and also after 24 hours storage at room temperature (t = 8 days + 24h).

The term physical stability of the TPN mixtures stored for a long time has practical application in parenteral nutrition at home: you can prepare the patient "inventory" of mixtures for a limited time. In the present study, the ability to store the mixtures tested to eight days was evaluated. The most important parameter to evaluate the physical stability of the mixtures is the presence of a drop of oil ≥ 5 µm, because they exceed the size of the diameter of capillaries and, consequently, if they are introduced into the bloodstream in large quantities, they can lead to embolism, necrosis of the surrounding tissues or have an effect on the functioning of the organ ¹⁴. It is believed that a small amount of oil with droplet sizes greater than 5 µm is not dangerous for the patients, since lipases are present in the endothelium of blood vessels. Lipases are responsible for the biodegradation of fatty acid esters. Among the test methods used, the most reliable, detecting drops of oil with increased size in mixtures TPN, turned out to be microscopic observation. The significance of this method to determine the durability of mixtures of the TPN also stress the authors of many publications. The light microscope method is highly sensitive and practicable, with a simple equipment and a conventional method validated by photon correlation spectroscopy (PCS) and the Coulter method ¹⁵. Using a microscope with a 40-fold magnification allows the detection of particles approximately 1 μm in size or enlarged emulsion particles up to 100 μm in size. Furthermore, other non-lipid globules (such as particulate matters or precipitations) can also be detected using this method. The method provides an easy, sensitive, cost-efficient, time-sparing, and convenient way to test the physical stability of a lipid emulsion in the critical droplet size to indicate destabilization (large fat droplet assessment ≥ 1-2 μm), and it is suitable for drug incompatibility testing in parenteral admixtures.

However, it is necessary to verify microscopic observation by other methods, so in this paper TPN admixtures were also analyzed using laser diffraction (LD) and photon correlation spectroscopy (PCS). Particle size measurement by means of a LD MasterSizer E enables the detection of particles from 0.05 to 80 µm, indicating size distribution. In turn, the use of ZetaSizer Nano ZS apparatus to measure the particle size using the method of dynamic light scattering allows the determination of particle size in the range of 0.6 nm-6 µm, taking advantage of the differences in the speed of the Brownian movement of particles in the medium in which they are suspended. The studies used both methods in view of the fact that, if the particles are too large, this method cannot determine the PCS size. An increase in oil droplet size by LD was not observed in any of the admixtures tested, despite their presence under microscopic observation. Some studies using PCS (no trends over time) detected the presence of a second peak and increased polydispersity index.

While mixtures 2 and 10 in the midpoint of the test (t = 24h) indicated an increased average droplet size of the oil phase (approximately 200 nm - by PCS), and mixtures 7, 20, 22, 25 and 26 showed the presence of a single time point of the second peak of about 5 mm, this change was not confirmed in the remaining time points. Although the results obtained by methods LD and PCS do not allow to detect irregularities at all the time points (the presence of larger oil droplets), yet on the basis of microscopic observations the mixture was physically unstable with composition 4. When the results obtained by the PCS and LD, it can be said that the median value of the oil droplets (Dv50; LD method) and an average oil droplet sizes (Z-Average, a method PCS) is slightly different. Z-average values were approximately 50 nm smaller. Measurement of zeta potential of TPN admixtures showed no significant changes during storage for 24h at room temperature (Table III). Zeta potential values were in the range denoted generally TPN admixtures. Another important factor indicating the stability of parenteral admixtures is pH, which decreases over time in PN admixtures because of the hydrolysis of fat triglycerides. Additional chemical reactions yielding base or acidic products also affect the pH. For the lipid stability and lecithin emulsifier, a pH range of 5-8 is necessary. The negatively charged surface (phosphate moiety) prevents the coalescence of the lipid globules. A pH below 5.0 favors lipid instabilities ¹⁶. In this study, no significant changes were noted in pH during storage of TPN admixtures (Fig. 5). Considerably lower pH of the mixtures 22, 25, 27 and 30 as compared to other systems is caused by using a different amino acid preparation (18 Vamin EF).

CONCLUSIONS

Based on the results of this study, and despite the increased concentration of electrolytes, the excess of CAN (297 to 1402) and the presence of inorganic calcium salts, the physicochemical stability of the parenteral admixtures tested has been demonstrated. The exception was mixture 4, which was found to have coalescence of drops of oil visible during microscopic observations and at some point of time by the PCS, so the composition of TPN 4 should be modified. An important observation was that inorganic calcium salt in PN can achieve the same stability profile as the organic salt in use in the clinic today. To apply the above examined compositions in clinical practice is their preparation under the same conditions using the same ingredients and packaging. Changing any factor (packaging, manufacturer, or replacement of the inorganic salt to organic) can cause a lack of physical stability and requires reconsidering physical stability. The research of physical stability indicates that, except for admixture 4, the test compositions can be used in nutrition at home with prolonged stability of prepared TPN admixtures, provided that they are implemented in the same manner and using the same components.