Stability of Compounded Digoxin Solution 0.05 mg/mL for Injection

Abstract

Background: In 2015, commercial pediatric digoxin injection 0.05 mg/mL was discontinued, leaving only one adult concentration (0.25 mg/mL) for injection on the Canadian market. No published studies have documented the chemical stability over a long period of time of a diluted solution of digoxin for injection. Objective: The aim of this study was to assess the chemical stability of 2 digoxin injection formulations 0.05 mg/mL diluted in 2 vehicles stored at 5°C or a 25°C. Methods: The compounded solution of digoxin 0.05 mg/mL for injection was prepared with digoxin 0.25 mg/mL after dilution in 2 different vehicles, normal saline, and a compounding of the commercial vehicle. Half of the compounding products were stored in 2 mL transparent glass vials at 25°C and the other half at 5°C. Chemical stability was evaluated by HPLC–UV analysis on days 0, 14, 30, 60, 90, 120, 150, 180 for each temperature conditions. In addition, samples were tested for organoleptic change, presence of particular matter as well as sterility. Results: For all tested preparations, the concentration of digoxin remained above 90.0% of the initial concentration throughout the 180-day study. Furthermore, no organoleptic change was observed; particulate matter assessment was in acceptable range; and sterility specifications were met. Conclusions: Digoxin 0.05 mg/mL obtained with a dilution of digoxin 0.25 mg/mL by normal saline or a copy of the commercial vehicle remained stable for at least 180 days at 5°C and 25°C.

Introduction

Digoxin, an old cardiotonic glycoside, is a drug used to treat some types of cardiac pathologies. ¹ In pediatrics, digoxin is mostly prescribed for the treatment of mild to moderate heart failure and atrial fibrillation.^2,3 In Canada, this medication is commercially available in 3 different forms: oral tablet, oral liquid, and injection. Before 2015, the injection was available in 2 different concentrations, 0.05 mg/mL for pediatric use and 0.25 mg/mL for adult. Pediatric digoxin injection has been discontinued in Canada in 2015, leaving solely the more concentrated injection on the market. ⁴ According to Bédard and Goyer, administering the adult formulation in young children is challenging and potentially hazardous as the required volume to inject is difficult to measure especially in low-weight infants. ⁴ Digoxin has a narrow therapeutic index, enforcing the need for high quality formulations. Digoxin is a high alert medication as per the Institute of Safe Medication Practice.^2,5

This problem led the team study to compound an injectable digoxin 0.05 mg/mL solution from the commercially available 0.25 mg/mL injection to facilitate the administration to children and thus reducing the risk of error. The objective of this study was to evaluate the stability of an injectable digoxin 0.05 mg/mL formulation diluted in normal saline or in a copy of the commercial vehicle stored at 5°C or at 25°C in transparent glass vials for up to 180 days.

Methods

Compounded Preparation

Preparation of a commercial vehicle copy

A compounding of the commercial vehicle was made by using the same excipients reported in the product monograph and the recipe given by Sandoz Canada (Boucherville, QC, Canada). The excipient list in the 0.05 mg/mL formulation and the 0.25 mg/mL formulation are the same. Accurately weighted citric acid monohydrate 0.08%w/w (0.0875 g; Galenova, Saint-Hyacinthe, QC, Canada, Lot 12495-4136) and sodium phosphate dibasic NaHPO₄ 0.3%w/w powders (0.3 g; Galenova, Lot 12999-4042) were solubilized in sterile water for injection (49.6 g; Baxter, Mississauga, ON, Canada, Lot: 47-123-DK). After complete solubilization, accurately weighed ethanol 10%w/w (10 g; Alcools de Commerce, Boucherville, QC, Canada, Lot: 018950) and propylene glycol 40%w/w (40 g; Galenova, Lot: 13874-4363) were added under continuous stirring using a magnetic stirrer until complete dispersion. Then, the compounded vehicle was filtered using a 0.45 μm nylon filter (Millipore, Etobicoke, ON, Canada).

Preparation of the diluted digoxin 0.25 mg/mL solutions

Two different diluents (copy of compounded commercial vehicle or NS [Baxter, Lot W5J03A0]) were used to dilute the 0.25 mg/mL solution of digoxin (Sandoz Canada, Lots FA6727 and EF7676) to obtain a target concentration of 0.05 mg/mL. A ratio of 1:4 (concentrated formulation: diluent) was used. The solution prepared from the copy of the commercial vehicle was filtered using a 0.22 µm nylon filter (Pall Canada, Mississauga, ON, Canada). The solutions were packaged in clear glass vials (fill volume 1 mL per 2 mL vial, Omega Laboratory, Montréal, QC, Canada). Each vial was closed using a rubber stopper and an aluminum crimp. Similarly, a total of 18 larger clear glass vials (fill volume 20 mL per 50 mL vial, Omega) was also prepared for each formulation.

Stability Study Design

Vials were stored under refrigeration (5°C ± 2°C) and in a stability chamber set at room temperature (25°C ± 2°C/60% ± 5% RH). No special care was taken to protect the product from light during storage. On each analysis day (0, 14, 30, 60, 90, 120, 150, 180), 3 vials (fill volume 1 mL) of each formulation were retrieved from each temperature condition and visually examined. The vials were vortexed for 20 seconds. A 200 µl sample was withdrawn for analysis by high performance liquid chromatography coupled with ultra-violet detection (HPLC-UV). At time 0 and after 180 days, 10 vials (fill volume 1 mL) were retrieved for each conditions to perform the sterility study. Particular matter was also assessed at predetermined time points (day 0, 90, and 180) using 1 larger vial (fill volume 20 mL) for each temperature condition.

Liquid Chromatography Method

Samples were analyzed using a HPLC system (Prominence UFLC, Shimadzu, Laval, QC, Canada) equipped with an LC-20AD binary pump, a DGU-20A5 solvent degasser, an SPD-M20A multiple wavelength photodiode array detector set at 220 nm, an SIL-20AC HT refrigerated autosampler set at 10°C and a CTO-20AC column oven set at 50°C. The stationary phase was an Eclipse XBD-C18 column (4.6 × 150 mm, 5 µm, P/N 993967-902, S/NUSKH037197, Agilent, Santa Clara, CA, USA). The mobile phase, a mixture of water: acetonitrile 68:32, at a flow rate of 1.1 mL/minute was used in an isocratic mode.

Calibration Curve

A stock solution of 1 mg/mL was prepared with digoxin bulk powder (AK Scientific, Union City, CA, USA, Lot TC34690) solubilized in methanol (Fisher scientific, Ottawa, ON, Canada, Lot 144689). Five standard solutions were obtained (10, 25, 50, 75, 100 µg/mL) by diluting samples of stock solution in a mixture of methanol: water (20:80 v/v). Freshly prepared standard solutions were injected in triplicate into the HPLC system to establish the calibration curve for each run of analyses.

Precision and Accuracy of the HPLC–UV Method

The precision was evaluated by calculating the coefficient of variation of the measured concentrations of the 5 standard solutions used to establish the calibration curve performed on the first day (n = 3). Similarly, these solutions were stored at 5°C and reanalyzed on 3 consecutive days to establish the interday variability. Accuracy of the method was evaluated using the backcalculated concentration using the calibration coefficients for each standard sample. The percent ratio of backcalculated concentration was divided by the real concentration of the standard sample.

Specificity of the HPLC–UV Method

A stress degradation study was conducted to determine the specificity of the HPLC–UV method. Samples of the stock solution (0.5 mL) were mixed with water (0.5 mL), aqueous solution of 3% hydrogen peroxide (0.5 mL), 0.1 M HCl (0.5 mL), and 0.1 M NaOH (0.5 mL). These 4 solutions were incubated for 1 hour at 25°C. The acid solution (100 μL, pH 1.3) was neutralized with 0.1 M NaOH (50 μL) and diluted with methanol: water (20:80; v/v) (850 μL). The alkaline solution (100 μL, pH 12.7) was neutralized with hydrochloric acid 0.1 M (50 μL) and diluted with methanol: water (20:80; v/v) (850 μL). Water and peroxide solutions were diluted directly using methanol: water (20:80; v/v) (900 μL). The generation of pH altering degradation products was evaluated by submitting both compounded preparations to 70°C and evaluating the pH at time 0 and after 7 days.

Particular Matter

Particulate matter was evaluated using a particle counter LS-20 (Lighthouse Worldwide Solutions, Fremont, CA, USA). Each sample was analyzed by using a slightly modified method in USP chapter <788>. ⁶ In this study, the first 4 mL aliquot was rejected, and the measurement was performed on three 5 mL aliquots; the last 1 mL was discarded. A visual evaluation of the solutions was also performed using the naked eye against a white and black background.

Sterility

Sterility was evaluated at time 0 and after 180 days under each stability conditions using QT-Microsystem test kits (QI Medical, Grass Valley, CA, USA) as specified by USP <71>. ⁷ Ten vials (fill volume 1 mL) corresponding to 10% of the batch were used for each evaluation.

Statistical Analysis

Reported concentration of digoxine was calculated as follow: 2 analytical replicates were analyzed by HPLC for each of the 3 bottles; the average of both analytical replicates was first calculated for each bottle, then these 3 results were used to calculate the average concentration and its standard deviation. Results were then reported as percentage of the time 0 concentration. A linear regression model of the stability results was performed using Mathematica 12 (Wolfram Research, Champaign, IL, USA). The LinearModelFit function was first used to build all linear regression models from the time-concentration series. These linear models were then used to calculate the 2-tailed 95% confidence interval of the predicted concentration relative to time 0 at 180 days.

Definition of Stability

These specifications where used to determine stability.

• The product has the appearance of a clear colorless solution;

• The relative concentration of digoxine is not less than 90.0% of the concentration at time 0;

• USP <788> Particulate matter in injection specifications using a light obscuration technique for containers smaller than 100 mL, considering a fill volume of 20 mL: Not more than 300 particles per mL >10 µm/mL and not more than 30 particles per mL >25 µm. ⁶

• USP <71> Sterility tests: Absence of microbial growth. ⁷

Results

Assay Validation

Calibration of the method was performed by linear regression using the standard samples and resulted in an r² not <0.9999. As shown in Table 1, the coefficient of variation were within acceptable limits, ranging from 0.03% to 0.82% for the intraday assay and from 0.31% to 1.28% for the interday assay. The backcalculated concentration varied between 99.2% and 100.3% of the real concentration for the whole calibration range. Forced degradation test for 1 hour at 25°C/60% RH resulted in recoveries of 104% in water, 85% in acidic condition (0.1 M HCl), 59%, in alkaline condition (0.1 M NaOH) and 104% in oxidative condition (3% H₂O₂) . No overlap was observed between digoxin peak, excipients peaks, and degradation products peak in all conditions tested (Figure 1). All non-digoxin peaks eluted between 1 and 4 minutes. Digoxin peak purity index calculated was equal or superior to 0.9999 in all cases between 190 and 250 nm. No pH altering degradation products was observed for both compounded preparation submitted to 70°C during 7 days: pH varied from 7.43 to 7.46 and 6.12 to 6.28 for the copy of the commercial vehicle and the 0.9% saline, respectively.

Table 1.

Precision and accuracy.

Concentration (µg/mL)	Intraday CV (%)	Interday CV (%)	Accuracy (%)
10	0.82	0.53	99.2
25	0.06	1.26	100.3
50	0.03	1.04	100.2
75	0.03	0.31	99.9
100	0.04	1.28	100.0

CV = coefficient of variation.

Figure 1.

Chromatograms of digoxin incubated for 1 hour at 25°C/60% RH observed during study period. (a) Digoxin standard solution 0.5 mg/mL, (b) neutral condition (water), (c) acidic condition (0.1 M HCl), (d) alkaline condition (0.1M NaOH), and (e) oxidative condition (0.1 M H₂O₂).

Stability Study

The appearance of all digoxin remained constant throughout the study for all samples, suggesting no visually detectable instability. As illustrated in Table 2, solutions of digoxin diluted to 0.05 mg/mL with the original vehicle or sodium chloride and stored in transparent glass vials retained more than 90.0% of the initial digoxin concentration during 180 days at 5°C and 25°C. The lower limit of the 95% confidence interval of the predicted concentration at 180 days was not less than 90.0% for all conditions and complied to the assay specification. Representative chromatogram at time 0 and 180 days are illustrated in Figure 2.

Table 2.

Concentration of digoxine relative to time 0. ^a

Time (days)	NS at 5°C (%)	NS at 25°C (%)	CVC at 5°C (%)	CVC at 25°C (%)
0 b	51.8 ± 0.1 µg/mL		55.1 ± 0.1 µg/mL
14	99.8 ± 0.1	99.7 ± 0.1	99.6 ± 0.2	99.5 ± 0.1
30	100.1 ± 0.1	100.1 ± 0.2	100.0 ± 0.2	100.0 ± 0.2
60	99.6 ± 0.2	99.7 ± 0.0	97.1 ± 0.2	99.5 ± 0.1
90	95.4 ± 0.2	94.6 ± 0.3	95.3 ± 0.1	95.2 ± 0.1
120	95.0 ± 0.3	94.0 ± 0.2	94.8 ± 0.0	94.6 ± 0.1
150	95.1 ± 0.3	94.1 ± 0.0	94.8 ± 0.0	94.6 ± 0.1
180	94.7 ± 0.2	93.5 ± 0.1	94.9 ± 0.1	94.6 ± 0.2
180 CI c	[93.1, 94.7]	[91.6, 93.6]	[92.9, 94.4]	[92.7, 94.4]

NS = normal saline; CVC = commercial vehicle copy.

^aReported as mean ± standard deviation.

^bTime 0 is reported as absolute concentration in µg/mL.

^cConfidence interval (95%) of the linear regression at 180 days.

Figure 2.

Chromatograms of digoxin formulations at time 0 and after 180 days at 25°C. (a) Digoxine in normal saline at time 0, (b) digoxine in commercial vehicle copy at time 0, (c) digoxine in normal saline after 180 days at 25°C, and (d) digoxine in commercial vehicle copy after 180 days at 25°C.

As illustrated in Table 3 the number of particles detected in all tested samples met the specification. No microbial growth was noted during the course of the study.

Table 3.

Particulate matter count.

		0 day	90 days	180 days
NS at 5°C	(>10 µm/mL)	1.9 ± 0.1	0.4 ± 0.3	2.9± 1.4
	(>25 µm/mL)	0.0±0.1	0.0 ± 0.0	0.3 ± 0.2
NS at 25°C	(>10 µm/mL)	1.6 ± 0.8	2.1 ± 1.1	1.0 ± 1.2
	(>25 µm/mL)	0.0 ± 0.1	0.0 ± 0.0	0.1 ± 0.1
CVC at 5°C	(>10 µm/mL)	2.7 ± 1.1	2.7 ± 1.5	3.2 ± 2.6
	(>25 µm/mL)	0.6 ± 0.3	0.6 ± 0.5	0.7 ± 0.3
CVC at 25°C	(>10 µm/mL)	1.4 ± 1.1	3.0 ± 1.4	5.0 ± 2.7
	(>25 µm/mL)	0.0 ± 0.1	0.4 ± 0.3	0.9 ± 0.6

NS = normal saline; CVC = commercial vehicle copy.

Discussion

Limited information is available in the literature regarding the stability of digoxin and its degradation pathways. The stability of digoxine in dextrose 5%, normal saline, lactated Ringer’s and a mixture of dextrose, sodium chloride, and potassium chloride solutions was evaluated for a duration of 48 hours. ⁸ The stability of digoxine admixtures for injection was also evaluated with milrinone and amrinone for a duration of 4 hours.^9,10 In all cases, no degradation was observed. The results presented in this study confirms the stability of digoxin aqueous solutions at 5°C and 25°C for a duration of 180 days. This result was not unexpected considering that the studied preparations were essentially based on a stable commercial preparation. Nonetheless, a stability study was required as potential instability included precipitation as well as chemical degradation which could both occur during the duration of the study. The risk of precipitation was higher for the normal saline solution as this vehicle differed from the commercial vehicle. Particle count remained acceptable for all tested conditions throughout the study confirming that these digoxin solutions were physically stable.

Digoxin is a compound containing a chain of 3 glycosidic units attached to a steroid-like structure. Although no information could be found regarding the rate of hydrolysis of digoxin, it is expected that these glycosidic units could be hydrolyzed resulting in a complex mixture of mono-, di-, and tri-glycosides as well as corresponding nul-, mono-, and di-glycosidic degraded versions of digoxin. Extensive degradation was observed in acidic and alkaline forced degradation conditions, this is compatible with a hydrolytic degradation pathway. Considering the complex nature of the degradation mixtures, no effort was made to identify the degradation products. The purpose of the stress degradation study was rather to confirm that no degradation compounds would interfere with the peak of interest during the HPLC analyses. No overlapping peaks were observed with the peak of interest in the stress degradation samples. Furthermore, the UV-spectrum of the peak of interest did not change, confirming that a single product eluted in the peak of interest.

Sterility was also evaluated at the beginning and, after the 180 days of the study. It should be noted that this microbial stability result is indicative of adequate manufacturing conditions when the studied formulations were prepared. It is not indicative of the microbial stability if these preparations were to be prepared elsewhere as manufacturing conditions will differ. In order to comply with NAPRA and Ordre des pharmaciens du Québec recommendations (OPQ 2014.01), it would be required to evaluate the sterile manufacturing environment and a given batch sterility in order to use a beyond use date of 180 days.

In addition to meet all the stability specifications at all tested time point, a regression analysis was performed to model the degradation rate of digoxin in the tested preparation. For all tested conditions, the lower end of the 95% confidence interval of the predicted concentration at 180 days also met the assay specification. This statistical model add confidence in the results and shows that the degradation trend is statistically not fast enough to result in an out of specification assay after 180 days.

An established 180 days chemical stability reduces the risk of waste and allows to prepare in advance a greater amount of compounded product, potentially avoiding emergency extemporaneous preparation.

Potential loss of drug product must be considered when performing filtration of digoxin solutions. Some authors reported no losses on any type of digoxin filter,^11,12 while others reported losses as high as 73% when filtration was performed within the first 20 minutes using 0.22 µm filters. ¹³ Saturation of the filters might eventually limit losses. In the present study, the large quantity filtered could explain why no losses were observed when solutions were prepared using the copy of the commercial vehicle. Solutions prepared using 0.9% saline were not filtered.

No special care was taken to protect the product from light during storage although digoxin is in the list of drugs sensitive to light.^14,15 The incubators used to perform the stability study have a glass front door and are in a storage room where light is usually off. Therefore, we recommend protecting diluted solutions of digoxin from light.

Conclusion

In conclusion, the 2 compounded digoxin 0.05 mg/mL solutions for injection were stable at least for 180 days when stored at refrigerated temperature and at room temperature conditions. As the compounding process is easier by using the normal saline as a diluent, it is more practical to use this formulation.