Investigation of incompatibilities of the parenteral drugs etacrynic acid and theophylline with other commonly used intravenous drugs in a paediatric cardiological intensive care unit

Chemicals and reagents

The drugs as well as 50 mL polypropylene syringes (Original Perfusor^® B. Braun Melsungen AG or BD™ Medical™ Plastipak™) and sterile 0.9 % saline solution (B. Braun Melsungen AG or Fresenius Kabi) were kindly provided by the hospital pharmacy of the UKSH Campus Kiel. Table 1 gives the drug formulations, pH of the solution (after reconstitution), batches and concentrations used. At the time of testing all solutions were within the stated expiry time. LC-MS grade water was obtained from VWR (83,645.320), acetonitrile gradient grade 99.9 % from honeywell (34,851), acetic acid from Sigma-Aldrich (1.00066) and sodium dihydrogen phosphate from Carl Roth (K300.2).

Study design

In cooperation with the paediatric cardiological intensive care unit (pcICU) of the Clinic for Congenital Heart Defects and Paediatric Cardiology and the hospital pharmacy of the UKSH Campus Kiel, frequently used drugs were identified. Since both newborns and young adults are treated in the pcICU, the drugs are used in a wide range of doses or concentrations. It is quite possible that concentrations vary by a factor of 10 depending on the age of the patient. In addition, it is conceivable that the infusion rates of two drugs are different. When two or more drugs are infused through the same lumen of an infusion catheter, a different infusion rate will result in different concentrations of the drugs. To simulate different infusion rates in compatibility studies, 10:1 and 1:10 mixtures are often investigated in addition to the 1:1 mixture [24], 25]. However, this was not considered appropriate for our study due to the wide range of drug concentrations on the pcICU. Instead, we decided to use a mixing scheme as shown in Figure 2. For each drug, a maximum and a minimum concentration was defined. To represent a worst-case scenario, the drugs were then mixed together in different concentrations as follows: maximum concentration drug A- maximum concentration drug B (a); maximum concentration drug A- minimum concentration drug B (b); minimum concentration drug A- maximum concentration drug B (c); minimum concentration drug A- minimum concentration drug B (d). Drugs, which are used in the pcICU only in one concentration (so-called fix concentration), were mixed only with the maximum (e) or minimum (f) concentration of the other drug.

Figure 2:

Mixing scheme. Resulting combinations are a, b, c, d, e, f.

Each combination was prepared in duplicates. A contact time between the two drugs of 4 h was defined as the clinically relevant time period, as this is considered to be the maximum likely contact time of drugs when administered simultaneously through the same lumen of a catheter [24], 26], 27]. If no relevant changes occurred within these 4 h, the combination was considered to be compatible. However, the samples were measured over a total period of 96 h in order to provide more evidence for the data obtained.

Literature and database research

Firstly, published compatibility data were searched. Therefore, the online databases King^® Guide (last updated 03/2020), ASHP^® Injectable drug information™ (last updated 06/2020) and stabilis (last updated 02/2023) as well as the prescribing information were used. In addition, the literature database PubMed^® was searched for compatibility data using the keywords compatibility or incompatibility or Y-site or stability or coinfusion and name of the drug.

Sample preparation

The drugs were reconstituted, if necessary, with 0.9 % saline (Reomax with 5 % glucose solution provided) according to the manufacturer’s instructions and then further diluted with 0.9 % saline. These drug solutions were mixed in a 1:1 ratio in 50 mL syringes as described in the mixing scheme (Figure 2). The filled syringes were sealed with a combi-stopper (B. Braun Melsungen) and stored at room temperature and under laboratory light. Immediately after mixing, as well as after 2 h, 4 h, 8 h, 24 h, 48 h and 96 h, 2 mL of solution were taken for pH determination and turbidimetric measurement. At each measurement point, the samples were also checked for visual conspicuities (e.g. particles, discolouration, production of gas and turbidity) in front of a white and black background using an adapted pharmacopoeia method (Ph. Eur. 10.0: 2.9.20 Particulate Contamination: Visible Particles) [28]. In addition, 0.8 mL of sample solution was taken directly after mixing and after 2 h, 4 h, 24 h and 96 h, aliquoted to 0.2 mL and frozen in liquid nitrogen. The 0.2 mL aliquots were stored at −80 °C until high performance liquid chromatography (HPLC) measurements. Prior to HPLC measurement, the samples were freshly thawed, diluted with 0.9 % saline solution if necessary and measured in triplicate.

pH measurement

1 mL sample solution was placed in a test tube and the pH was measured after a latency time of 10 min with a Hanna pH metre (HI 2211 + HI 1053), which was calibrated daily with ready-to-use buffer solutions (Carl Roth, A517.1, P713.1, 8086.4).

Turbidimetry

To examine the sample solutions for non-visible changes 0.8 mL of the sample solution was placed in a semi-micro cuvette (Brand^®, 759115) and measured in triplicates at 350 nm, 410 and 530 nm in a photometer (Varian Cary^® 50 UV-Vis Spectrophotometer) based on the recommendations from Bardin et al. [29]. Changes in absorption at these wavelengths can indicate chemical changes, discolouration or turbidity of the solution.

HPLC and validation

For the combinations evaluated as physically compatible, the drug content was determined, with the exception of potassium chloride, glucose and sodium bicarbonate. For this purpose, an HPLC method was developed that allowed the quantification of all selected drugs. The HPLC method follows a method already described in Jirschitzka et al. [30].

The HPLC system consisted of a degasser (Biotech Degasi^® Classic), Waters™ 1,525 binary HPLC pump, Waters™ 717plus autosampler, column oven (Techlab GmbH) and a Waters™ 2,487 Dual λ absorbance detector (254 nm, 270 nm). An Agilent^® Poroshell 120EC-C18, 2.7 μm, 3.0 mm × 50 mm column with Phenomenex^® SecurityGuard™ Ultra cartridge, UHPLC C18 3.0 mm inclusive guard holder was used. The flow rate was set to 0.8 mL/min. A gradient was used. Mobile phase A consisted of a 10 mM sodiumphosphate buffer pH 3.2 and mobile phase B consisted of acetonitrile/sodiumphosphate buffer pH 3.2 60:40. The gradient started at 95 % A for 3 min, then changed to 80 % A within 2 min to continue to 40 % A until 16 min. Within another 2 min the mobile phase A decreased further to 30 % to remained there for 1 min. Then, the gradient changed within 1 min–95 % A and re-equilibration started for 5 min. The entire gradient, including re-equilibration, ran for 25 min. The column oven was set to 20 °C. Waters™ Breeze^TM-software (version 3.3) was used for data acquisition and processing.

The HPLC method was validated according to ICH guideline [31]. Therefore the linearity of the method was determined by a five-point calibration and the precision and reproducibility were checked measuring the maximum, middle and minimum concentrations of the range in triplicates. To validate the freezing process, the maximum and minimum concentrations were measured both before and after freezing and thawing in triplicates.

Limit values for assessing drug combinations as incompatible

In the literature, different limit values for the methods usually used in compatibility studies are discussed to decide whether a combination can be considered compatible or incompatible. In particular, there are different views on the maximum acceptable deviation of the pH value. For example, in some compatibility studies, a pH deviation of more than 0.2 compared to the beginning is already considered as a sign of incompatibility [32], 33]. In other studies, however, deviations of the pH value of maximum two are also accepted [34]. As a limit value for turbidimetric measurements, absorbance changes of 0.01 at 530 nm or 0.04 at lower wavelengths are often defined [25], 35]. One study even accepted changes in absorbance of 0.1 [26]. However, since a visually perceptible discolouration is already observable from an absorbance change at 410 nm of approximately 0.015, a maximum absorbance change of 0.04 as a limit value for 410 nm seemed unsuitable to us. Therefore, we have decided to use a lower limit value for the absorbance at 410 nm. To describe the chemical compatibility of intravenous drugs, a residual content of at least 90 % is generally considered acceptable [29]. In summary, in this study, drug combinations were considered as physicochemical compatible if no changes occurred seen with the unaided eye within 4 h, the pH did not change by more than 1, the absorbance did not change by more than 0.04 at 350 nm and 0.01 at 410 and 530 nm, and drug content did not decrease by more than 10 %.

LC-MS

To identify degradation products, LC-MS and LC-MS/MS experiments were carried out for conspicuous combinations according to a method described in Jirschitzka et al. [30]. LC-MS measurements were performed with an Agilent 1260 HPLC system (G1379B 1260 μ-degasser, G1312B bin pump, G1367E 1260 HiP ALS + G1330B, G1316A 1260 TCC, G4212B 1260 diode array detector) with gradient elution (A: 0.2 % acetic acid in water–B: 0.2 % acetic acid in acetonitrile) coupled to a Bruker Amazon SL mass spectrometer (Bruker Daltonik, Bremen, Germany) with an ion trap and MSn facility. The gradient started with a flow rate of 1 mL/min at 95 % A for 2 min, then continued to 65 % A in 0.5 min and further to 20 % A in 3.5 min. The gradient changed to 3 % A in 0.5 min and remained there for 1.5 min. Finally, the gradient changed to 95 % A in 0.5 min and re-equilibration started at an increased flow rate of 1.3 mL/min for 1 min. The MS was switched on from min two to min 9.2 in positive mode. The column oven was set to 25 °C. A Phenomenex^® Luna^® 3 μm C18(2) 100 Å, 100 × 4.6 mm column inclusive Phenomenex^® SecurityGuard™ Ultra cartridge, UHPLC C18 3.0 mm with guard holder was used. Hystar 3.2 SR two software was used to control the HPLC, Bruker Trap Control 7.0 software was used to control the MS and the MS data were analysed with Bruker Data Analysis 4.0.

Literature research

Only two compatibility studies were found in PubMed^® for the drug combinations selected here. Firstly, Allen et al., who confirmed the physical compatibility between Eca and KCl [36], and secondly, Kershaw et al., who investigated the compatibility between Tph, Amp, Cef, KCl and Sul and rated them as visually compatible [37]. The three databases (King^® guide, ASHP^® Injectable drug information™ and stabilis) refer to these two studies when searching for compatibility data for Eca and Tph in combination with the other drugs. Compatibility data beyond this were not given in the databases. Unfortunately, precise information on compatibility with other drugs is also missing in the prescribing information. In the prescribing information for Reomax, at least, it is stated that Eca is incompatible with solutions with a pH below 5. By contrast, the prescribing information of afpred^® forte-THEO only states that no substances other than the diluting solution should be added in order to avoid incompatibilities. The prescribing information of the other drugs either mentions an acidic pH, which should be avoided (Lasix^®), gives no further information except that the respective drugs should be administered separately (Meropenem Hikma, Pantoprazol EVER Pharma, Paracetamol B. Braun) or states specific incompatibilities, but not for the drugs used in this study (Unacid^®, Cefazolin-saar^®). As a result of the literature research, the cross-table shown in Figure 3 became the basis for our compatibility study.

Figure 3:

Compatibility chart displayed as a cross-table based on the conducted literature research. Colour indicates the type of compatibility. beige: visually compatible, green: physically compatible, white: no information available.

Validation HPLC

An HPLC method was developed for the identification and quantification of the drugs. Figure 4 shows an overlayed representative chromatogram of all drugs. With the exception of Amp and Mero, all peaks are baseline separated. However, since Amp and Mero did not need to be identified at the same time, this was not problematic. Table 2 shows the results of the validation and the retention times. All drugs showed a linear relation between peak area and concentration over the whole concentration range. The coefficient of variation was always less than 2 %. The maximum and minimum recoveries for both direct measurement and frozen and thawed samples were between 97.86 and 101.99 %. This was considered acceptable for the quantification of the drugs. Therefore, the developed HPLC method was found to be appropriate.

Figure 4:

Overlayed HPLC-UV chromatogram of tested drugs.

Physicochemical compatibility

When examining the drug mixtures with the unaided eye in front of a black and white background according to Ph. Eur. 10.0: 2.9.20, no changes were noticed in the clinically relevant time period of 4 h. In Figure 5, the results of the compatibility study are presented in form of a cross-table. A total of 50 combinations of drugs were investigated, each in duplicates. Of these, 37 combinations were assessed as physicochemically compatible and 13 combinations were assessed as incompatible. Drug combinations assessed as incompatible include etacrynic acid + theophylline, etacrynic acid + ampicillin + sulbactam, etacrynic acid + cefazolin 100 mg/mL, theophylline + cefazolin and theophylline 4 mg/mL + meropenem 2 mg/mL.

Figure 5:

Compatibility table displayed as a cross-table. a,b,c,d,e,f represents the drug combinations according to Figure 2. The respective concentrations in mg/ml, unless otherwise stated, are entered in the columns or rows. Colour indicates the type of compatibility. light green: physically compatible, dark green: physicochemically compatible, red: incompatible, ^†relative concentration decreases by more than 3 % within 4 h, ^††appearance of an increasing degradation product, ^†††published data, concentration not included.

When determining the contents of the combinations of etacrynic acid with sodium bicarbonate, an additional increasing peak was observed in the chromatograms (see Figure 6). LC-MS and LC-MS/MS experiments were carried out to identify the unknown substance. A molecular ion (M+H⁺) with m/z 321 was found and allocated to the unknown substance. Together with the fragmentation pattern obtained from the LC-MS/MS experiment of m/z 321 ± 2 (see Figure 7), the substance was identified as 2-(2,3-dichloro-4-(2-(hydroxymethyl)butanoyl)phenoxy)acetic acid (OH-Eca).

Figure 6:

Overlayed chromatogram of etacrynic acid 0.8 mg/mL + sodium hydrogencarbonate 8.4 %. ?: unknown substance.

Figure 7:

MS/MS experiment (Bruker Amazon SL massspectrometer with ion trap and MSn facility) of m/z 321 ± 2 and presumed fragmentation of 2-(2,3-dichloro-4-(2-(hydroxymethyl)butanoyl)phenoxy)acetic acid.

A graphical representation of the results of the individual drug combinations over 96 h can be found in the supplementary materials.