Abstract
Background
In the local pediatric intensive care unit, precipitation was observed in the intravenous catheter upon co‐administration of four drugs together with the buffered electrolyte solution (Plasma‐Lyte 148, Baxter). Co‐infusion of incompatible combinations represents a safety concern.
Aims
To reproduce the clinical case of precipitation. To further explore and understand the risk of precipitation, different combinations of the components as well as the corresponding electrolyte solution with 5% glucose (Plasma‐Lyte 148 with 5% glucose) should be investigated.
Methods
Physical compatibility of fentanyl, ketamine, midazolam, and potassium chloride was tested in combination with the buffered electrolyte solutions. The concentrations and infusion rates representative of children 10–40 kg were used to estimate mixing ratios. Analyses detecting visual particles (Tyndall beam) and sub‐visual particles (light obscuration technology) were undertaken. Measured turbidity and pH in mixed samples were compared with unmixed controls.
Results
Both midazolam and ketamine showed formation of visual and sub‐visual particles upon mixing with Plasma‐Lyte 148, respectively. Particle formation was confirmed by increased turbidity and a distinct Tyndall effect. pH in mixed samples mirrored the pH of the buffered electrolyte, suggesting that the solubility limits of midazolam, and in some ratios also ketamine, were exceeded. Midazolam also precipitated in combination with the glucose‐containing product that held a lower pH, more favorable for keeping midazolam dissolved.
Conclusions
Replication of the case revealed that both midazolam and ketamine contributed to the precipitation. Midazolam and ketamine were both evaluated as incompatible with the buffered electrolyte solution and midazolam also with the buffered electrolyte‐glucose solution and should not be co‐administered in the same i.v.‐catheter line. Fentanyl and potassium chloride were interpreted as compatible with both buffered electrolytes.
Keywords: fentanyl, ketamine, midazolam, pH, plasmalyte, precipitate
What is already known about the topic
Plasma‐Lyte 148 and Plasma‐Lyte 148 with 5% glucose are commonly used in pediatric intensive care, often co‐administered in the same intravenous catheter as drugs and other fluids. Information on their compatibility is very limited, especially for use in a complex therapy regime.
What new information this study adds
Midazolam and ketamine were found to be incompatible with Plasma‐Lyte 148 and midazolam also with Plasma‐Lyte 148 with 5% glucose. This information contributes to preventing precipitation in the catheter and thereby prohibit possible occlusion and infusion of large particles.
1. INTRODUCTION
Intravenous crystalloid fluids are widely used in the treatment of the pediatric intensive care patients for the replenishment of intravascular volume and restoration of hemodynamic stability. 1 The buffered electrolyte solution (Plasma‐Lyte 148, Baxter) and the buffered electrolyte‐glucose solution (Plasma‐Lyte 148 with 5% Glucose, Baxter) are two relatively new crystalloids that have recently been introduced in our local pediatric intensive care unit (PICU). These balanced fluids have high buffer capacity and have similar composition regarding electrolytes (Table 1), but only the one without glucose is isotonic with similar osmolarity as in extracellular compartment fluids. 2 The two buffered electrolytes also differ in product pH (Table 1). Buffered electrolyte solutions are used to provide daily needs of water, glucose and essential electrolytes and have been reported to reduce the incidence of hyperchloremia and metabolic acidosis compared with normal saline. 3 In buffered electrolyte solutions, some of the chloride anions have been replaced with buffers to reduce the acid–base balance disturbances. This is achieved by a lower chloride content that more closely matches that of human plasma through the substitution of the chloride ion with an anion, such as lactate, acetate, or gluconate. 4 These buffers are rapidly metabolized and excreted.
TABLE 1.
Composition and physicochemical properties of the two buffered electrolyte products (P = Plasma‐Lyte 148 and PG = Plasma‐Lyte 148 with 5% glucose, Baxter)
| Ingredients | P | PG |
|---|---|---|
| Glucose monohydrate (g/L) | – | 55.00 |
| Sodium chloride (g/L) | 5.26 | 5.26 |
| Potassium chloride (g/L) | 0.37 | 0.37 |
| Magnesium chloride hexahydrate (g/L) | 0.30 | 0.30 |
| Sodium acetate trihydrate (g/L) | 3.68 | 3.68 |
| Sodium gluconate (g/L) | 5.02 | 5.02 |
| Amounts | ||
| Na+ (mmol/L) | 140 | 140 |
| K+ (mmol/L) | 5.0 | 5.0 |
| Mg++ (mmol/L) | 1.5 | 1.5 |
| Cl− (mmol/L) | 98 | 98 |
| Acetate (mmol/L) | 27 | 27 |
| Gluconate (mmol/L) | 23 | 23 |
| Osmolarity (mosmol/L) | approx. 295 | approx. 572 |
| pH | approx. 7.4 (6.5–8.0) | 4.0–6.0 |
Pediatric intensive care patients are often in need of numerous intravenous (i.v.) drugs and the number of i.v. drugs and fluids often outnumbers the number of i.v. access ports. Co‐infusion of incompatible drugs and fluids via the same catheter line and/or lumen may result in the formation of solid particles, that is, drug precipitates. 5 , 6 These precipitates may clot the i.v.‐catheter line and find their way into organs. 7 , 8 , 9 There have been reports of fatal outcome in children due to infusion of solid particles. 10 , 11 Documented compatibility information is scarce, especially for the pediatric population, and most studies only investigate the stability of two drugs in 1 + 1 ratio. 12 To the best of our knowledge, no compatibility studies involving buffered electrolytes in a complex therapy regime are available.
We present here a case of precipitation observed in the infusion line of a critically ill patient, with a body weight of 16 kg, who received a complex infusion regime containing the buffered electrolyte solution Plasma‐Lyte 148. The buffered electrolyte solution (10.83 ml/h) was co‐administered in the same i.v.‐catheter line (22 Gauge) with fentanyl (50 μg/ml, 0.64 ml/h), ketamine (10 mg/ml, 4.8 ml/h), potassium chloride (1 mmol/ml, 0.2 ml/h), and midazolam (5 mg/ml, 3.2 ml/h). The nurses noted that the syringe pump was alarming due to high pressure, and visual inspection of the i.v.‐catheter line clearly showed signs of precipitation. Different actions were undertaken; first, the ketamine‐infusion was transferred to one of the other lumen of the patient's i.v.‐catheter lines (22 Gauge), but still a white precipitate could be observed in the catheter. It was first when the buffered electrolyte solution was stopped, that no precipitation could be seen in the catheter‐line. In order to keep rehydrating the patient, the fluid was changed to the corresponding product with 5% glucose.
The aim of the current study was to investigate the precipitating factors in our clinical case and to explore and understand the risk of precipitation using buffered electrolytes in a complex therapy regime. Other clinically relevant mixing ratios and scenarios were included in the study to warrant safe use of both buffered electrolyte solutions with and without 5% glucose.
2. METHODS
The buffered electrolytes investigated were Plasma‐Lyte 148 (Baxter, Oslo, Norway) and the corresponding product Plasma‐Lyte 148 with 5% glucose (Table 1). An overview of the drug products used in this study, their composition and physico‐chemical properties is presented in Table 2.
TABLE 2.
Overview of drug product information (manufacturer information) and physico‐chemical information
| Drug product (manufacture) | Active ingredient a | Excipients a | pH product a | Active ingredient b | |
|---|---|---|---|---|---|
| pKa | Solubility | ||||
| Fentanyl 50 μg/ml (Hameln) | Fentanyl citrate | Sodium chloride, hydrochloric acid or sodium hydroxide, water for injection | 5.0–7.5 | 8.43 | 0.74 mg/ml |
| Potassium chloride 1 mmol/ml (B. Braun) | Potassium chloride | Water for injection | 4.5–7.5 | – | >100 mg/ml |
| Ketalar 10 mg/ml (Pfizer) | Ketamine hydrochloride | Benzethonium chloride, sodium chloride, water for injection | 3.5–5.5 | 7.5 | 0.046 mg/ml |
| Midazolam 5 mg/ml (B. Braun) | Midazolam hydrochloride | Sodium chloride, hydrochloric acid 10%, water for injection | 2.9–3.7 | 6.6 | 0.1 mg/ml |
| Sterile water (Fresenius Kabi) | Water for injection | – | 6–7 | – | – |
Summary of Product Characteristics.
Parent compound (weak acid or base) obtained from PubChem.
Prior to simulating the case, each of the drugs was tested separately with the buffered electrolyte to identify if any of the drugs were incompatible with the fluid in a simple system (Table 3). Since drugs and fluids in the clinical setting are given at different infusion rates based on the required dose (per kg body weight) and the concentration of the drugs, different ratios of drugs and fluids could potentially meet in the i.v.‐catheter line. In order to test clinically relevant pediatric mixing ratios (children with body weights from 10 to 40 kg), doses and infusion rates were calculated for each drug and fluid using information from UpToDate 13 and Koble. 14 Mixing ratios were calculated as earlier described by Nezvalova‐Henriksen et al. 5 Table 3 summarizes the mixing ratios tested. In two‐component drug+fluid combinations, equal parts (1 + 1) were tested in addition to one ratio where the fluid was in excess. None of the calculations resulted in mixing ratios containing more drug than fluid.
TABLE 3.
Overview of controls and tested mixing ratios for drug(s) + fluid(s)
| Drugs and fluids | Concentration | Control |
|---|---|---|
| Fentanyl (F) | 50 μg/ml | x |
| Ketamine (K) | 10 mg/ml | x |
| Midazolam (M) | 5 mg/ml | x |
| Plasma‐Lyte 148 (P) | – | x |
| Plasma‐Lyte 148 with 5% Glucose (PG) | – | x |
| Potassium chloride (KCl) | 1 mmol/ml | x |
| Sterile water for injection (SW) | – | x |
| Drug + Fluid(s) | Components mixed | Mixing ratios |
|---|---|---|
| F + P | 2 | 1 + 1 and 1 + 20 |
| K + P | 2 | 1 + 1 and 1 + 3 |
| M + P | 2 | 1 + 1 and 1 + 19 |
| M + PG | 2 | 1 + 1 and 1 + 19 |
| KCl + P | 2 | 1 + 1 and 1 + 12 |
| M + SW + P | 3 a | 1 + 2 + 3 |
| M + SW + PG | 3 a | 1 + 2 + 3 |
| Drug(s) + Fluid(s) | Components mixed | Mixing ratios |
|---|---|---|
| F + KCl + K + M+P | 5 b | 3 + 1 + 24 + 16 + 55 |
| F + KCl + K + M+SW | 5 b | 3 + 1 + 24 + 16 + 55 |
| F + KCl + K + M+PG | 5 b | 3 + 1 + 24 + 16 + 55 |
Simplified simulation of clinical case.
Replication of exact ratios from the clinical case.
As midazolam was identified early as incompatible, two‐component mixtures (same mixing ratios) were also tested with the buffered electrolyte‐glucose solution. Midazolam was further investigated in a three‐component mix where sterile water was used to mimic the volume of the other drugs in a simplified simulation of the case. These simulations were done for both types of buffered electrolytes. To replicate the case, all involved drugs were mixed in concentrations and volumes that were an exact representation of the ratio administered to the patient (Table 3). The mix of five components was studied for Plasma‐Lyte 148, the corresponding product with glucose, and finally, also one mix where the midazolam was replaced with sterile water. The intention with the two latter was to check if the formation of a precipitate could be pH‐mediated and to clarify if midazolam alone was responsible for the precipitation in the complex regime.
2.1. Sample preparation
All products were used as they were received from the pharmacy. Mixed samples were prepared by extracting the desired volume of each product and filtering the solution through a 0.2 μm sterile syringe filter (VWR) into sample tubes (15 ml; Coning). Aliquots of unmixed products were used as controls. All samples and controls were analyzed in triplicates, except controls of narcotics, which were only measured in one parallel. Samples were analyzed immediately and 4 h after mixing. The time points were chosen to detect the immediate formation of precipitation whereas the 4‐h time point was chosen to study whether the potential precipitate formed increased or was dissolved covering situations with very slow infusion rates. All samples were prepared, stored, and analyzed at ambient temperature.
2.2. Analysis of potential particle formation (precipitation)
A number of well‐established methods was used to scrutinize for any sign of particle formation in the mixed samples. 5 , 6 , 15 The smallest capillaries in the body have a diameter of approximately 5 μm. Particles in this range are sub‐visible and cannot be detected by visual examination. Sub‐visible particle counts were analyzed using light obscuration (Accusizer Optical Particle Sizer with Syringe Injection Sampler, PSSNICOMP, Billerica, MA, USA). The total number of particles/ml was assessed. The particle counts were divided in particles/ml for particle sizes >0.5, >5, >10, and >25 μm. Particles with a size of 5 μm and above were of main concern. The two upper limits were included since these are used by the USP where large‐volume parenterals (i.e., infusions) should not contain more than 25 particles/ml larger than 10 μm and not more than 3 particles/ml larger than 25 μm. 16 In addition, a high number of very small particles may also be alarming since they can grow in size over time; therefore, an acceptance criteria in this study was set to contain no more than 2000 particles/ml larger than 0.5 μm. 15
Increased turbidity or haze could be a sign of microprecipitation. Turbidity was measured for all mixed samples and all controls using the portable 2100Qis Turbidimeter (Hach Lange GmbH). The acceptance criteria were that mixed samples should not deviate by more than 0.2–0.3 Formazine Nephelometry Units (FNU) from the turbidity of the unmixed control samples. 15
Since the drugs are salts of weak bases, their solubility will depend on pKa of the parent base and the pH of the sample. The fact that the electrolyte solution also contains a buffer can challenge the theoretical assessment of solubility; hence, pH measurements (Seven Compact pH‐meter, Mettler Toledo) of mixed samples compared to unmixed controls is a very useful tool. A pH‐shift of more than 1.0 pH unit in the mixed sample as compared to unmixed control should be regarded as alarming.
Large particles (>50 μm) can be captured by visual examination. Also, microprecipitates can be observed with the naked eye if one uses a focused Tyndall beam. Visual observation was performed in two ways, firstly, by shining a focused light beam (Schott KL 1600 LED, Germany) through the mixed samples, comparing them with the unmixed controls, and, secondly, by passing a 630–650 nm laser beam (P 3010 RoHS, Chongqing, China) through the samples and controls. The mixed sample should be free from visually observed particles and with no Tyndall effect (visible coherent laser line through the sample).
3. RESULTS
3.1. Compatibility with buffered electrolyte solution (Plasma‐Lyte 148)
In the case of midazolam, all analyzes indicated precipitation with values exceeding acceptance criteria. The detector of the particle counter was overloaded and could not detect each particle individually without dilution of the sample (which could not be performed due to the scope of the test as this would dissolve the particles). Precipitation was observed with the naked eye and high turbidity values were recorded. The pH in the mix shifted from approx. 3.5 in midazolam (unmixed) to around 5.1 for the mixing ratio 1 + 1.
Ketamine showed a low particle count for particles of all sizes immediately after mixing. However, 2 h (data not shown) and 4 h after mixing, the detector was overloaded for the 1 + 1 ratio, and single‐particle number could not be reported. However, there did not seem to be larger particles >5 μm in the mix. Elevated turbidity values (over the acceptance criteria) and clear signs of particles could be detected in the visual examination with Tyndall beam. The pH in the mixed samples was around 6.0 whereas that of the ketamine control was 4.6. In the ketamine mixing ratio of 1 + 3 with the buffered electrolyte, the total particle count was also elevated after 4 h and the turbidity exceeded the acceptance level in the 2 h sample (not shown) and 4 h after mixing. In this mixing ratio, the pH changed to 6.3.
Neither fentanyl nor potassium chloride precipitated when mixed with the buffered electrolyte solution (Table 4).
TABLE 4.
Results from analyses of precipitation after mixing fentanyl (F), ketamine (K), midazolam (M), sterile water (SW), and potassium chloride (KCl) with the two buffered electrolyte solutions in different combinations and mixing ratios (bold font indicates values outside acceptance criteria) (average ± SD; n = 3)
| Drug | Mix ratio | Particles/ml | Turbidity (FNU) | pH | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ≥0.5 μm | ≥5 μm | ≥10 μm | ≥25 μm | ||||||||||
| 0 h | 4 h | 0 h | 4 h | 0 h | 4 h | 0 h | 4 h | 0 h | 4 h | 0 h | 4 h | ||
| Fentanyl (F) | Control | 263 a | 99 a | 1 a | 1 a | 0 a | 1 a | 0 a | 0 a | 0.12 a | 0.07 a | 6.02 a | 6.05 a |
| Ketamine (K) | Control | 330 a | 260 a | 5 a | 4 a | 4 a | 4 a | 2 a | 2 a | 0.16 a | 0.23 a | 4.63 a | 4.59 a |
| Midazolam (M) | Control | 204 ± 32 | 218 ± 80 | 4 ± 2 | 4 ± 4 | 1 ± 0 | 2 ± 3 | 0 ± 1 | 1 ± 2 | 0.12 ± 0.03 | 0.23 ± 0.16 | 3.47 ± 0.21 | 3.52 ± 0.27 |
| Plasma‐Lyte 148 (P) | Control | 274 ± 29 | 196 ± 84 | 11 ± 2 | 10 ± 4 | 2 ± 1 | 4 ± 2 | 0 ± 0 | 0 ± 1 | 0.21 ± 0.08 | 0.16 ± 0.01 | 7.02 ± 0.07 | 6.85 ± 0.14 |
| Plasma‐Lyte 148–5%‐Glucose (PG) | Control | 193 ± 64 | 120 ± 37 | 3 ± 2 | 1 ± 0 | 1 ± 1 | 0 ± 1 | 0 ± 0 | 0 ± 0 | 0.13 ± 0.02 | 0.12 ± 0.03 | 5.22 ± 0.04 | 5.21 ± 0.05 |
| Potassium chloride (KCl) | Control | 221 ± 20 | 165 ± 1 | 1 ± 1 | 6 ± 4 | 1 ± 1 | 3 ± 3 | 1 ± 1 | 1 ± 1 | 0.07 ± 0.01 | 0.08 ± 0.00 | 5.84 ± 0.00 | 5.82 ± 0.01 |
| Potassium chloride (KCl) | Control | 221 ± 20 | 165 ± 1 | 1 ± 1 | 6 ± 4 | 1 ± 1 | 3 ± 3 | 1 ± 1 | 1 ± 1 | 0.07 ± 0.01 | 0.08 ± 0.00 | 5.84 ± 0.00 | 5.82 ± 0.01 |
| Drug (s) + Plasma‐Lyte 148 | |||||||||||||
| F + P | 1 + 1 | 185 ± 6 | 356 ± 53 | 2 ± 1 | 5 ± 2 | 1 ± 1 | 2 ± 2 | 0 ± 0 | 0 ± 1 | 0.07 ± 0.02 | 0.09 ± 0.02 | 6.78 ± 0.06 | 6.81 ± 0.01 |
| 1 + 20 | 147 ± 29 | 187 ± 51 | 6 ± 3 | 11 ± 5 | 2 ± 1 | 4 ± 2 | 0 ± 0 | 0 ± 0 | 0.11 ± 0.03 | 0.11 ± 0.01 | 7.04 ± 0.01 | 7.06 ± 0.01 | |
| K + P | 1 + 1 | 224 ± 76 | OL b | 3 ± 1 | 2 ± 1 | 2 ± 1 | 1 ± 2 | 1 ± 0 | 0 ± 1 | 0.45 ± 0.09 | 0.85 ± 0.07 | 6.05 ± 0.01 | 6.04 ± 0.03 |
| 1 + 3 | 538 ± 43 | 1594 ± 1316 | 1 ± 1 | 3 ± 1 | 0 ± 1 | 1 ± 1 | 0 ± 0 | 0 ± 0 | 0.26 ± 0.05 | 0.38 ± 0.10 | 6.30 ± 0.03 | 6.30 ± 0.03 | |
| M + P | 1 + 1 | OL b | OL b | OL b | OL b | OL b | OL b | 636 ± 41 | 273 ± 34 | 313 ± 11.0 | 122 ± 42.6 | 5.09 ± 0.19 | 5.10 ± 0.22 |
| 1 + 19 | 423 ± 202 | 293 ± 202 | 7 ± 6 | 2 ± 6 | 4 ± 2 | 0 ± 1 | 0 ± 1 | 0 ± 0 | 0.25 ± 0.20 | 0.08 ± 0.01 | 6.18 ± 0.02 | 6.21 ± 0.02 | |
| KCl + P | 1 + 1 | 155 ± 50 | 290 ± 25 | 3 ± 2 | 12 ± 6 | 2 ± 1 | 5 ± 4 | 0 ± 0 | 1 ± 2 | 0.10 ± 0.04 | 0.11 ± 0.01 | 6.87 ± 0.05 | 6.60 ± 0.36 |
| 1 + 12 | 264 ± 95 | 272 ± 94 | 11 ± 9 | 14 ± 5 | 3 ± 2 | 6 ± 2 | 0 ± 0 | 1 ± 1 | 0.10 ± 0.01 | 0.11 ± 0.02 | 7.00 ± 0.01 | 7.00 ± 0.03 | |
| M + SW + P | 1 + 2 + 3 | 853 ± 217 | 943 ± 450 | 57 ± 12 | 56 ± 16 | 31 ± 7 | 30 ± 8 | 5 ± 3 | 5 ± 1 | 1.41 ± 0.45 | 13.5 ± 15.1 | 5.33 ± 0.28 | 5.57 ± 0.14 |
| F + KCl + K + M + P | 3 + 1 + 24 + 16 + 55 | 3883 ± 528 | OL b | 32 ± 7 | 86 ± 26 | 15 ± 3 | 35 ± 12 | 2 ± 1 | 3 ± 2 | 0.70 ± 0.17 | 1.00 ± 0.10 | 5.63 ± 0.03 | 5.64 ± 0.01 |
| F + KCl + K + SW + P | 3 + 1 + 24 + 16 + 55 | 496 ± 209 | OL b | 2 ± 1 | 3 ± 1 | 0 ± 1 | 1 ± 1 | 0 ± 0 | 0 ± 0 | 0.33 ± 0.07 | 0.48 ± 0.09 | 6.24 ± 0.02 | 6.26 ± 0.03 |
| Drug(s) + Plasma‐Lyte 148 ‐ 5% | |||||||||||||
| M + PD | 1 + 1 | 516 ± 98 | 884 ± 455 | 60 ± 85 | 49 ± 76 | 18 ± 26 | 43 ± 72 | 0 ± 0 | 0 ± 0 | 1.99 ± 1.40 | 79.9 ± 53.1 | 4.77 ± 0.09 | 4.74 ± 0.09 |
| 1 + 19 | 120 ± 33 | 127 ± 29 | 0 ± 1 | 1 ± 2 | 0 ± 0 | 1 ± 1 | 0 ± 0 | 0 ± 1 | 0.08 ± 0.01 | 0.11 ± 0.03 | 5.23 ± 0.03 | 5.25 ± 0.02 | |
| M + SW + PD | 1 + 2 + 3 | 220 ± 25 | 265 ± 52 | 5 ± 3 | 3 ± 1 | 1 ± 1 | 1 ± 0 | 0 ± 0 | 0 ± 0 | 0.25 ± 0.20 | 0.08 ± 0.01 | 5.07 ± 0.03 | 5.06 ± 0.02 |
| F + KCl + K + M + PD | 3 + 1 + 24 + 16 + 55 | 1016 ± 86 | OL a | 4 ± 1 | 3 ± 1 | 2 ± 2 | 1 ± 0 | 0 ± 0 | 0 ± 0 | 0.54 ± 0.14 | 0.76 ± 0.21 | 5.07 ± 0.02 | 5.06 ± 0.02 |
Over the detector limit (OL), >9000 particles/mL.
Only one parallel for the narcotics (local laboratory guidelines).
Finally, in the complex mix of five components mimicking the case, precipitation was observed, with a high total number of particles >0.5 μm immediately after mixing, which developed into detector overload after 4 h. The number of large particles (>10 μm) exceeded the acceptance limit after 4 h. The turbidity was high, and Tyndall effect could be observed. pH in the mix was around 5.6. When replacing midazolam with sterile water, fewer particles were observed, but still all analysis methods used in this study indicated that a precipitation occurred.
3.2. Compatibility with buffered electrolyte‐glucose solution (Plasma‐Lyte 148 with 5% glucose)
The buffered electrolyte‐glucose solution had a pH of approximately 5.2 whereas the one without had a pH of around 7.0. To elucidate which effect the pH would have on the precipitation on the complex regime, the analyses focused around midazolam.
First, in the two‐component mix consisting of equal parts of midazolam and the buffered electrolyte‐glucose solution the total number of particles >0.5 μm was slightly increased but most importantly the number of particles with a diameter >10 μm was over the acceptance level and showed an increasing trend over time (Table 4).
Replicating the full case by mixing fentanyl, ketamine, midazolam and potassium chloride with the glucose‐containing buffered electrolytes, particles developed with time and after 4 h, there was detector overload. Also, turbidity measurements indicated particle precipitation and the pH values were above acceptance criteria.
4. DISCUSSION
The physical compatibility of the buffered electrolyte solution and the buffered electrolyte‐glucose solution has never been studied in a complex mixture of several intravenous drugs that are often administered at PICUs simultaneously. Since the use of these buffered electrolyte products offers the advantage of avoiding hyperchloremic metabolic acidosis that tends to occur with the use of 0.9% NaCl, it is of benefit to study the impact its intravenous co‐administration might have on the physical stability of drugs given simultaneously via the same catheter. To our best knowledge, this study is the first to analyze the physical stability of a multi‐drug mixture with this type of products.
By replicating the case where precipitation occurred during co‐administration of fentanyl, ketamine, midazolam, and potassium chloride in the same i.v.‐catheter line as Plasma‐Lyte 148 using the mixing volumes arising from the infusion rates, our study confirmed that a precipitation was formed and identified the problematic drugs in the mix to be midazolam and ketamine.
Three studies investigating the intravenous compatibility of the same buffered electrolyte solution and the buffered electrolyte‐glucose solution with one‐drug‐at‐a‐time have been published. These studies reported concentration dependent compatibility between midazolam and the two buffered electrolytes, irrespective of whether glucose was present. 17 , 18 , 19 Hammond et al. 17 investigated physical compatibility by visual inspection and reported that precipitation formed immediately when mixing three parts of midazolam 3 mg/ml with two parts of the buffered electrolyte solution or the buffered electrolyte‐glucose solution. Dawson et al. 18 did not see any precipitation when investigating compatibility of midazolam 0.25 mg/ml with the buffered electrolyte solution. On the contrary, the manufacturer Baxter Medical reported that midazolam 1 mg/ml was compatible with the buffered electrolyte solution. 19 The concentration of midazolam in our case was 5 mg/ml, which caused precipitation and resonates well with the literature confirming that there is a concentration dependent solubility challenge of midazolam when the pH in the mix deviates from the pH of the drug product, a rather acidic pH of 3.5. It is well‐known that water solubility of midazolam is pH‐dependent exhibiting a drastic decrease with pH values exceeding 4. 20 , 21 Midazolam base has a pKa of 6.2 (Table 2), which means that when the pH of its environment, in this case the mixture with the buffered electrolyte (and for the case also other drugs) reaches the pKa‐value of midazolam, more than 50% of midazolam molecules will be deprotonated and less soluble. It is important to mention that the structure of midazolam changes with increased pH. Upon deprotonation a ring‐structure is formed in the molecule (Figure 1), which affects the physico‐chemical properties of midazolam from being water‐soluble at low pH to more lipid‐soluble with increased pH. 22 This explains the precipitation. Our analyses suggest that at a pH between 4.7–5.7 the dilution of midazolam 5 mg/ml is not sufficient, neither in 1 + 1 ratio (=2.5 mg/ml) nor in the simulated case 1 + 5 ratio (=0.83 mg/ml), to keep the drug in solution. This is in agreement with the low solubility of midazolam of 0.1 mg/ml at neutral pH (Table 2). 21 At a mixing ratio 1 + 19, the dilution will be higher (=0.25 mg/ml), which should still indicate too low solubility. This mixture shows a higher pH and is more influenced by the buffer of the electrolyte fluid. These factors will also have an impact on the degree of dissociation and solubility, illustrating how complex buffered systems are. This complicates quick theoretical estimation of mixed pH in a clinical setting.
FIGURE 1.
Schematic representation of pH dependent ring‐formation in midazolam structure
Dawson et al. studied the compatibility of ketamine 0.2 mg/ml (1 + 30) with the buffered electrolyte solution and concluded that the mix was physically compatible in the mixing ratio of 1 + 30. 18 Baxter Medical concluded that equal parts of ketamine 2 mg/ml and the buffered electrolyte solution to be physically compatible. 19 It should be noted that the ketamine concentration in the mixtures of both these studies were lower than in our studies (=6.5 μg/ml and 1 mg/ml, respectively), compared with our clinical case and study using 10 mg/ml for 1 + 1 (=5 mg/ml) and 1 + 3 (=2.5 mg/ml). Even though we could not find the exact solubility of ketamine base, only the theoretically predicted one of 46 μg/ml from DrugBank, it is reasonable to assume that it would be below 2.5 mg/ml. Again, ketamine is a weak base, and the solubility depends on the pH of the surroundings and the pKa of the drug. However, the buffering electrolytes will also have an impact on the ionization and solubility. As discussed earlier, the buffered electrolyte governs the pH of the mixture of ketamine (pH control of 4.6) and the buffered electrolyte solution (pH control approx. 7) and keeps it at 6–6.3. Since ketamine is a weak base with a pKa of 7.5 (the strongest base), a higher proportion of the drug will be deprotonated and can precipitate.
Fentanyl is also a weak base with a pKa on the basic side (8.77). It is used as the citrate salt in the product, which showed a pH of 6.0 in the unmixed control (Table 4). The solubility of fentanyl base was reported to be 0.74 mg/ml. However, fentanyl is a very potent opioid, and in our hospital, the clinical concentration used is 50 μg/ml. Hence, in our studies, the mixing ratios 1 + 1, 1 + 20 and 1 + 32 (the latter represents the case), all represented drug concentrations well below the solubility limit (25 μg/ml, 2.4 and 1.5 μg/ml, respectively), and the precipitation could therefore not be traced to fentanyl. Baxter Medical have concluded 10 μg/ml fentanyl, in mixing ratio 1 + 1, to be compatible with the buffered electrolyte solution, 19 and Dawson et al. 18 tested 30 μg/ml of fentanyl in mixing ratio of 1.2 + 1 and concluded it to be compatible with both the buffered electrolyte solution and the buffered electrolyte‐glucose solution. All concentrations were below solubility limits and confirmed our findings. It can be concluded that fentanyl is safe to co‐administer with either one of the two buffered electrolytes.
Potassium chloride (KCl) has a very high solubility (>100 mg/ml, Table 2), and there was no reason to expect salting out or precipitation of KCl. Neither of the ions form poorly soluble salts or complexes with any of the other constituents, also not with any of the excipients from the various drug products (Table 2). Baxter Medical supplied information of compatibility of 0.5 mmol/ml potassium chloride with both buffered electrolytes, 19 also supporting the findings in the current study.
When replicating the case but replacing midazolam with sterile water in order to see which of the components, midazolam or ketamine, contributed to precipitation, it was clear that precipitation occurred also without midazolam for the buffered electrolyte. The particle levels were lower in the admixture with sterile water, due to lack of midazolam, emphasizing that ketamine alone also led to an increase in turbidity. It would be interesting to conduct the same study and replace also ketamine with sterile water, however, this was not done.
Comparing the results of the complex regime for the two buffered electrolytes with and without glucose, midazolam showed a higher degree of particle formation when mixed with the buffered electrolyte solution than with the buffered electrolyte‐glucose solution. Nevertheless, the latter also showed signs of particle formation and pH‐values of the mix that should be alarming. Still, the pH of the buffered electrolyte‐glucose solution (pH approx. 5.2) is more favorable for keeping midazolam in solution than the corresponding product without glucose (pH approx. 7). 21 The acidic pH‐range of heat‐sterilized glucose (5% Glucose approx. 3.5–6.5 23 ) is described to be a result of glucose decomposition into levulinic and formic acids at temperatures in the autoclave. 24 Since the glucose content is the only difference between the two buffered electrolyte solutions, this is the probable cause of the more acidic product pH in the glucose‐containing product. For midazolam, the acidic pH of the glucose‐containing product was beneficial. Nevertheless, our studies emphasize that a pH difference of approximately 2 pH units between two corresponding products is not trivial and switching between the two product types in a clinical scenario should cause extra attention if co‐administration with drugs comes into question.
The experimental setup in this study, the mixing volumes of drugs in tubes to simulate Y‐site co‐administration of i.v.‐drugs, does not replicate the true clinical scenario and our results should therefore be interpreted with this in mind. However, we have performed both visual and sub‐visual particle analysis and analyzed clinically relevant concentrations and mixing ratios combined with theoretical evaluations, which makes our conclusions sufficiently robust. This is in contrast to many, especially older published studies that only evaluate 1 + 1 mixing ratios and rely on visual examination alone. We maintain that when testing for drug compatibility, it is important to use several analytical methods and not only perform visual inspections since it is shown to be subjective and will not capture sub‐visual particles. Since drugs are given in ever‐changing infusions rates depending on the need of the patient it is advised to analyze mixed samples in at least three different mixing ratios. Given the variability in clinical practice where different buffered electrolyte solutions are used an extra safety precaution is to use in‐line filter, which could help to prevent infusion of precipitated particles into the bloodstream of the patient. 25 , 26 Last, but not least, it should be emphasized that the exact composition and pH of buffered electrolytes might be product specific, and caution should be taken if and when extrapolating the findings to other products.
5. CONCLUSION
Our case‐based analysis of a five component mixture identified midazolam 5 mg/ml and ketamine 10 mg/ml as the causative agents of the precipitation when co‐infused with the buffered electrolyte solution Plasma‐Lyte 148. Midazolam was found to be physically incompatible in a two‐component mix with the buffered electrolyte solution but also with the corresponding product containing 5% glucose and should not be co‐administered in the same i.v.‐catheter line. Ketamine also showed signs of incompatibility when mixed with the buffered electrolyte solution and co‐administration should also be avoided. Fentanyl 50 μg/ml and potassium chloride were found to be compatible with both buffered electrolytes.
AUTHOR CONTRIBUTIONS
N.N., K.N‐H, and I.T. conceptualized the study. N.N., K.N‐H., and I.T. performed methodology. V.N. and N.N performed formal analysis. N.N. performed writing—original draft preparation. K.N‐H., V.N., and I.T performed writing—review and editing. K.N‐H., N.N., and I.T. supervised the study. K.N‐H. and I.T involved in project administration. K.N‐H performed funding acquisition. All authors have read and agreed to the published version of the manuscript.
FUNDING INFORMATION
This research was funded by South‐Eastern Norway Regional Health Authority (grant number 2018096).
CONFLICT OF INTEREST
The authors declare no conflicts of interest. The funding authority had no involvement in study design, collection, analysis, and interpretation of data, and writing manuscript or decision to submit manuscript for publication.
ACKNOWLEDGMENTS
This study was funded by South‐Eastern Norway Regional Health Authority (grant number 2018096). We would like to extend our gratitude to the Hospital Pharmacy Enterprise South Eastern Norway and all nurses and physicians at the pediatric intensive care unit at Oslo University Hospital for continuous support. Many thanks also to Tove Larsen and Ivar Grove (engineers at University of Oslo) for support in the laboratory.
Nilsson N, Nguyen V, Nezvalova‐Henriksen K, Tho I. Exploring a case of incompatibility in a complex regimen containing Plasma‐Lyte 148 in the pediatric intensive care. Pediatr Anesth. 2023;33:211‐218. doi: 10.1111/pan.14598
Section Editor: Francis Veyckemans
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
REFERENCES
- 1. Lehr AR, Rached‐d'Astous S, Barrowman N, et al. Balanced versus unbalanced fluid in critically ill children: systematic review and meta‐analysis*. Pediatr Crit Care Med. 2022;23(3):181‐191. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Weinberg L, Collins N, Van Mourik K, Tan CO, Bellomo R. Plasma‐Lyte 148: a clinical review. World J Crit Care Med. 2016;5:235‐250. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Arya V, Kavitha M, Mittal K, Gehlawat V. Plasmalyte versus normal saline as resuscitation fluid in children: a randomized controlled trial. J Pediatric Crit Care. 2021;8(3):134‐138. [Google Scholar]
- 4. Curran JD, Major P, Tang K, et al. Comparison of balanced crystalloid solutions: a systematic review and meta‐analysis of randomized controlled trials. Crit Care Explor. 2021;3(5):e0398. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Nezvalova‐Henriksen K, Nilsson N, Østerberg CT, Staven Berge V, Tho I. Y‐site physical compatibility of Numeta G13E with drugs frequently used at neonatal intensive care. Pharmaceutics. 2020;12(7):677. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Nilsson N, Storesund I, Tho I, Nezvalova‐Henriksen K. Co‐administration of drugs with parenteral nutrition in the neonatal intensive care unit—physical compatibility between three components. Eur J Pediatr. 2022;181(7):2685‐2693. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Jack T, Brent BE, Boehne M, et al. Analysis of particulate contaminations of infusion solutions in a pediatric intensive care unit. Intensive Care Med. 2010;36(4):707‐711. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Puntis JW, Wilkins KM, Ball PA, Rushton DI, Booth IW. Hazards of parenteral treatment: do particles count? Arch Dis Child. 1992;67(12):1475‐1477. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Lehr HA, Brunner J, Rangoonwala R, Kirkpatrick CJ. Particulate matter contamination of intravenous antibiotics aggravates loss of functional capillary density in postischemic striated muscle. Am J Respir Crit Care Med. 2002;165(4):514‐520. [DOI] [PubMed] [Google Scholar]
- 10. Hill SE, Heldman LS, Goo EDH, Whippo PE, Perkinson JC. Fatal microvascular pulmonary emboli from precipitation of a total nutrient admixture solution. JPEN J Parenter Enteral Nutr. 1996;20(1):81‐87. [DOI] [PubMed] [Google Scholar]
- 11. Bradley JS, Wassel RT, Lee L, Nambiar S. Intravenous ceftriaxone and calcium in the neonate: assessing the risk for cardiopulmonary adverse events. Pediatrics. 2009;123(4):e609‐e613. [DOI] [PubMed] [Google Scholar]
- 12. Fernández‐Peña A, Katsumiti A, De Basagoiti A, et al. Drug compatibility in neonatal intensive care units: gaps in knowledge and discordances. Eur J Pediatr. 2021;180(7):2305‐2313. [DOI] [PubMed] [Google Scholar]
- 13. Somers MJ, Mattoo TK, Wilkie L, eds. Maintenance Intravenous Fluid Therapy in Children. UptoDate; 2022. [Google Scholar]
- 14. KOBLE Kunnskapsbasert Oppslagsverk om Barns Legemidler. The Norwegian Medicines Manual for Health Personnel; 2022. Accessed August 5, 2022. https://koble.info/om_KOBLE [Google Scholar]
- 15. Staven V, Wang S, Grønlie I, Tho I. Development and evaluation of a test program for Y‐site compatibility testing of total parenteral nutrition and intravenous drugs. Nutr J. 2016;15(1):29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. United States Pharmacopeia and National Formulary (USP 47‐NF 3) . <788> PARTICULATE MATTER IN INJECTIONS. Accessed August 9, 2022. https://online.uspnf.com/uspnf/document/1_GUID‐BFC6D11B‐21C5‐494 E‐A0C3‐2EB88E2F297A_2_en‐US?source=Search%20Results&highlight=788
- 17. Hammond S, Wignell A, Cooling P, Barrett DA, Davies P. Plasma‐Lyte 148 and Plasma‐Lyte 148 + 5% glucose compatibility with commonly used critical care drugs. Intensive Care Med Exp. 2020;8(1):25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Dawson R, Wignell A, Cooling P, Barrett D, Vyas H, Davies P. Physico‐chemical stability of Plasma‐Lyte 148® and Plasma‐Lyte 148® + 5% glucose with eight common intravenous medications. Pediatric Anesthesia. 2019;29:186‐192. [DOI] [PubMed] [Google Scholar]
- 19. Baxter Healthcare . Y‐Site Compatibility of Intravenous Drugs with Plasmalyte 148. Baxter Healthcare; 2015. [Google Scholar]
- 20. Forman JK, Souney PF. Visual compatibility of midazolam hydrochloride with common preoperative injectable medications. Am J Hosp Pharm. 1987;44(10):2298‐2299. [PubMed] [Google Scholar]
- 21. Andersin R. Solubility and acid‐base behaviour of midazolam in media of different pH, studied by ultraviolet spectrophotometry with multicomponent software. J Pharm Biomed Anal. 1991;9(6):451‐455. [DOI] [PubMed] [Google Scholar]
- 22. Gerecke M. Chemical structure and properties of midazolam compared with other benzodiazepines. Br J Clin Pharmacol. 1983;16(Suppl 1):11s‐16s. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Medicines.org.uk 2022 . Glucose 5% Intravenous Infusion BP. Baxter Healthcare Ltd, Summary of Product Characteristics (SPC). Accessed October 10, 2022. https://www.medicines.org.uk/emc/product/1825
- 24. Barnett MI, Cosslett AG, Duffield JR, Evans DA, Hall SB, Williams DR. Parenteral nutrition. Pharmaceutical problems of compatibility and stability. Drug Saf. 1990;5(Suppl 1):101‐106. [DOI] [PubMed] [Google Scholar]
- 25. Van Boxtel T, Pittiruti M, Arkema A, et al. WoCoVA consensus on the clinical use of in‐line filtration during intravenous infusions: current evidence and recommendations for future research. J Vasc Access. 2022;23(2):179‐191. [DOI] [PubMed] [Google Scholar]
- 26. Jack T, Boehne M, Brent BE, et al. In‐line filtration reduces severe complications and length of stay on pediatric intensive care unit: a prospective, randomized, controlled trial. Intensive Care Med. 2012;38(6):1008‐1016. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.