Effects of Carrier Solution on Insulin Bioavailability av , MD* 
From the * Department of Medicine I,
Correspondence: Zdenek Ru Background: The aim of our study is to assess the influence of the base solution on the availability of biologic insulin expressed in the value of total insulin using the radioimmunoassay method and to monitor the influence of the base solution on the total insulin concentration over time. Methods: Total insulin in a saline and in a total nutrient admixture was measured using the RIA method. In 15 experiments, the application of the saline by a perfusor (8 IU Actrapid HM, 100 IU/1 mL + 20 mL saline) was carried out at time intervals of baseline, 5, 10, 30, 60, 90, and 210 minutes. The application of the total nutrient admixture (8 IU Actrapid HM + 20 mL total nutrient admixture) was carried out in the same way. The MANOVA, ANOVA and paired t test with Bonferroni correction were then used for statistical evaluation. Results: The average values of insulin concentration in saline at given time intervals were 21.5 ± 11.3 mIU/L (5.4% of the theoretically calculated concentration). The level of insulin in the total nutrient admixture did not change over time, and it reached the values of 115.2 ± 22.3 (28.8% of the theoretically calculated concentration). Changes in insulin concentration in time were found only in the saline (ANOVA time effect p < .001 for saline; p = .26 for total nutrient admixture). Conclusions: The availability of insulin was significantly higher in the total nutrient admixture solution than in the saline in the 3.5-hour experiment. The dependence of insulin concentration on time is present only in saline, and the main changes in insulin bioavailability are within first 60 minutes. The difference could be caused by smaller insulin absorption to the syringe walls and to the set, owing to the amino acids in the mixture. The question whether the amino acid concentration affects insulin bioavailability stays opened to other studies. Many intensive care patients are insulin resistant and need continuous insulin infusion, together with parenteral nutrition. In the last decade, it was very common to use total nutrient admixtures (all-in-one system). Clinical practice has taught us that the insulin requirements for hyperglycemia correction during parenteral nutrition using total nutrient admixture (TNA) are reduced when insulin is added directly into the TNA bag rather than administering continuous insulin in a saline infusion by means of a perfusor. The aim of the study was to assess the effect of the carrier solution on the availability of biologic insulin during the 3-hour experiment.
In the experiment, a perfusor (BBraun, Melsungen, Germany) was used with the original 20-mL polyvinylchloride syringe and extension infusion set (length 1.20 m, inner diameter 1.8 mm, volume 5 mL) made of the same material. Twenty milliliters of saline and the TNA (Nutrimix, BBraun, Melsungen, Germany) aspirated from original 3-L bags (composition: Aminoplasmal Hepa 10% 500 mL + glucose 20% 1000 mL + Nutralipid 20% 200 mL + Elotrace 100 mL) and insulin Actrapid HM (Novo-Nordisk, Copenhagen, Denmark; concentration 1 mL = 100 IU) were used consecutively for the experiment. First, 8 IU of insulin Actrapid HM was placed in syringe, and the system was filled using the infusion set with the experimental solution. The baseline samples (0.5 mL) were taken just after filling. Then the perfusor was set to 2 mL hour–1 and samples (0.5 mL each) were taken directly from the infusion set at the following intervals: 5, 10, 30, 60, 90, and 210 minutes. Fifteen sets of measurements were taken for both the saline and the TNA solutions. All the samples were placed on ice immediately after having been obtained and were stored at 4°C in closed syringes before measurements were taken. Immunoreactive insulin was determined by radioimmunoassay analysis with the commercially made set (CIS Bio International, Paris, France). Results in mIU/L are presented as means ± SD. Differences from the carrier solution were tested by MANOVA and time effect by ANOVA for repeated measures. If significant changes were found, a paired t test with Bonferroni correction was used for evaluating differences between the consecutive time points of the experiment. The value p < .05 was considered to be significant.
Table I shows insulin concentrations (IRI) in both the saline and the TNA solutions during the experiment. Insulin in the 2 carrier solutions tested differed significantly, giving higher insulin yield in the TNA solution compared with the saline at each time point (MANOVA group by time effect p < .001). Changes in insulin concentration depending on time were found only in the saline (ANOVA time effect p < .001 for saline; p = .26 for TNA solution). When the separate time points for saline were analyzed, major changes appeared at the beginning of the experiment (a decrease from 26 ± 8.47 at the baseline to 20 ± 7.57 at 5 minutes; p < .001). Significant changes were also determined between 30 and 60 minutes (21 ± 9.60 vs 26 ± 11.00); p < .05) showing a stepwise increase in insulin concentration.
The mean yield of insulin expressed as a ratio of measured to calculated (ie, 400 IU L–1) concentration at the baseline was 0.52 for saline and 0.70 for TNA.
Solutions for parenteral nutrition developed dramatically in the last 30 years. The multibottle system of amino acids, dextrose, and lipids is being overcome by the mixture of dextrose/protein or dextrose/protein/lipid–TNA (all-in-one system). In recent years, the TNA solutions in plastic multilayered bags prepared by companies or in hospital pharmacy have been generally used. In our study, we decided to compare commonly used TNA to saline. The reason why the insulin requirements are smaller when insulin is applied in a bag with TNA solution in comparison with saline is not clear. There are numerous possible explanations of this phenomenon: the change in insulin bioavailability (monomer vs hexamer forms, deposition, and formation of crystals), an accompanying pH change of the solution, its dilution, or insulin interaction with the particular components of TNAs, especially trace elements.1–13 Others blame the inner surface of the nutrition bag and the perfusor set for uneven insulin absorption. The main point of the experiment was to compare TNA to saline. We decided to choose the perfusor system over nutrition bags to exclude possible interactions between the material of the bag and insulin. Our results have shown a stabilizing effect of the TNA solution upon measured insulin concentration using the radioimmunoassay (RIA) method over time, although, according to our results, insulin availability from the TNA solution is 5 times higher in comparison with insulin application in saline. According to empirical clinical experience, when the patient is converted from continuous IV insulin application by perfusor to the TNA bag, the insulin dose is reduced less. If insulin is added to the TNA solution, it is mostly advisable to reduce the established doses of insulin by one third within 24 hours. It is possible that the interaction of insulin and the material of the bag for TNA solution further decreased the insulin bioavailability in clinical setting. The loss of insulin caused by absorption in the infusion set is often described in literature, with many different conclusions. According to some authors, the size of absorption losses does not in practice depend on either the material of which the bottles and sets are made, on the composition of the solution, or on the time between the addition of insulin and the use of the mixture. They claim that the absorption loss depends exclusively on the insulin concentration.4,13 Bigger losses are described at lower levels of concentration than at higher levels.1,13 The dependence of absorption on initial insulin concentration and on the ratio of the inner surface of the application system to the volume of the solution has been proved.14 An important point of the study is whether to dilute insulin with the solution or add insulin to the solution. In clinical practice, we usually add the insulin into the bag with TNA. When we use perfusor for continuous insulin infusion, we dilute the insulin by the solution, similar to the way the experiment was done. A number of authors have described an interaction between the drugs (insulin) and the material of which the bottles, bags and sets are made.11,14 The results are controversial but without importance for clinical practice. Some authors recommend using a higher dose of insulin or rinsing the set out with a 50-mL solution containing insulin before commencement of the application.1,9 Adding 0.1% to 1% human albumin, gelatin, or whole blood into the infusion fluid can reduce the absorption,1,3,6–8,10,12,13 and increasing dose of albumin does not reduce the insulin absorption further.8,10,12 In our experiment, amino acid solutions seem to be the most probable cause of higher stability and availability of insulin in the TNA solution, but we cannot exclude an influence of fat emulsion; therefore, we cannot state with absolute certainty that the improved bioavailability is solely caused by the amino acids. Some authors have had similar results and described a higher availability of insulin from TNA, depending on the sort of amino acid solution.4,9 Although a majority of authors have rejected the possibility of a more significant effect on the extent of absorption by the composition of IV fluid,1 some authors describe a reduction in absorption from the saline after adding KCl solution.14 The all-in-one system of total parenteral nutrition is often infused for 24 hours, and it is therefore useful to know the insulin stability within 24-hour period. The change of insulin concentration in time is controversial. Some studies show some dependence of bioavailability of insulin on time during the first 2 hours after adding insulin into the carrier solution.10,13 Other studies, however, deny any dependence on time.1,4 Kerchner et al7 state that the highest absorption losses from 0.45% NaCl solution occur during the first 20 minutes if blood is added to the solution and during the first 2 hours if there is no blood present in the solution. In our study, stability tests of insulin in TNA were performed over the period of 3.5 hours; it cannot be assumed that the results of the stability tests at 24 hours would be the same. Long-lasting insulin stability has been studied, and the results show the insulin concentration to be stable for >24 hours.7 The changes in insulin concentration depending on time were found only in the saline in our study. Major decrease of insulin concentration appeared at the beginning of the experiment at 5 minutes. Significant changes were also determined between 30 and 60 minutes, showing a stepwise increase in insulin concentration. For assessment of insulin concentration in the samples taken, 2 standard methods are generally used. The RIA method, which is cheaper but less accurate, always arrives at higher losses of insulin than the method of detection of insulin marked by an isotope.1,4–6 The difference in the measured insulin gains from TNA using both methods is enormous.3 Using the RIA method, a gain of approximately 20% of the theoretical insulin concentration is measured,4–6 whereas the use of tracer insulin leads to a gain of 90% to 95%.5,12,13 In our study, the RIA method was used. We believe that the lower accuracy of this method will not affect the quality of the study, as it was used in both types of solution and thus any possible fault is identical in both experiments.
The availability of insulin from the TNA solution is approximately 5 times higher than from the saline. Although the change of insulin concentration in time is present with the saline, it has not been found with the TNA solution in the 3.5-hour experiment. This could be caused by the interaction of insulin with the solution of amino acids, which can influence the insulin absorption to the walls of the application systems. It is open to further studies to confirm the results in a 24-hour period and to confirm whether insulin bioavailability is increased solely by the amino acids in TNA. Supported by a grant of the Ministry of Education, Charles University Prague, Faculty of Medicine in Pilsen, Czech Republic, CEZ: J13/98:111400001. Received for publication November 26, 2004. Accepted for publication July 29, 2004.
Journal of Parenteral and Enteral Nutrition, Vol. 28, No. 6,
439-441 (2004)
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
, Plze
Top
MATERIALS AND METHODS
RESULTS