Stability of prostaglandin E₁ solutions stored in polypropylene syringes for continuous intravenous administration to newborns

Abstract

Objective

We aimed to monitor the physicochemical stability of prostaglandin E₁ (PGE₁) 1.5 and 15 µg/mL in 10% dextrose stored in polypropylene syringes.

Methods

We developed a liquid chromatography-high resolution mass spectrometry (LC-HRMS) method to detect and quantify levels of PGE₁. Method selectivity was performed with a mixture of PGE₁ and its degradation products. Forced degradation tests were performed to determine which degradation products were most likely to form. PGE₁ injection solutions in 10% dextrose were stored in unprotected and shielded-from-light polypropylene syringes in a climatic chamber. Samples were taken immediately after preparation (T₀) and after 24, 48, 72 and 168 hours for analysis. PGE₁ solutions were considered stable if ≥90.0% of the initial concentration was retained.

Results

The LC-HRMS method was validated in the range of 0.086-0.200µg/mL PGE₁ with trueness values between 98.2% and 100.3%, and repeatability and intermediate precision values of <2.2%and <4.7%, respectively. The quantification and detection limits of the method were 0.086 and 0.026µg/mL, respectively. PGE₁ and its degradation products were resolved chromatographically. PGE₁ injection solutions were≥90.0%stable after 48hours in unprotected from light (UPL) syringes. The solutions remained clear without precipitation, colour or pH modification and subvisible particles within the permitted levels. Prostaglandin A₁ was the sole degradation product observed.

Conclusions

A LC-HRMS method to evaluate PGE₁ stability in a 10% dextrose was developed and validated. PGE₁ 1.5 and 15µg/mL in 10% dextrose solution are stable for 48hours when stored at 30ºC in UPL polypropylene syringes.

Introduction

Intensive care units treat the most critical patients, some of whom take 10 or more different prescription drugs.^1–3 Thus, one nurse must prepare and administer drugs frequently, often more than 20 times daily.² In non-intensive care units, nurses spend around 20% of their time performing drug-related tasks, including drug preparation, administration and documentation.^{4 5} These activities could reach up to 40% of the daily nursing workload,⁶ and we expect this burden could be even higher in a neonatal intensive care unit (NICU).

In Europe and in the USA, some hospital pharmacies provide centralised intravenous additive services. This is not the case in our hospital. Instead, in accordance with our current standard operating procedure, all intravenous continuous-administration drugs are prepared daily by nurses in 50 mL syringes and administered over 24 hours. In most cases, less than half of the volume is administered after 24 hours. Unfortunately, the half-filled syringes must be discarded and new ones are prepared for the next 24 hours infusion period. This practice is time consuming and has important financial implications for the NICU and risk of infections remains a major concern.

Prostaglandin E₁ (PGE₁) is an expensive drug administered by intravenous continuous infusion. PGE₁ maintains the permeability of the ductus arteriosus temporarily in newborns with ductal-dependent congenital heart disease until the diagnosis is confirmed and/or surgery can be performed. In some patients, PGE₁ treatment can last several weeks, which is costly.

The Centers for Disease Control and Prevention⁷ recommend replacing continuously used intravenous administration sets no more often than at 96 hours intervals, but at least every 7 days based on the Cochrane reviews^{8 9} and randomised control studies.^10–12 In response to this recommendation, we planned to extend the PGE₁ infusion period beyond 24 hours.

The objectives of this study were twofold. First, we sought to develop and validate an analytical liquid chromatography-high resolution mass spectrometry (LC-HRMS) method for dosing PGE₁ in 10% dextrose and detect its degradation products. In newborns, 10% dextrose is favoured over normal saline solution for drug infusion preparation. Dextrose provides a caloric energy source critical for neonatal brain development. Second, we sought to assess the physicochemical stability of 1.5 and 15 µg/mL PGE₁ (range of concentrations used in our NICU) in 10% dextrose stored in polypropylene syringes exposed to, or protected from, light at 30°C (maximal temperature observed during the summer in our NICU) for 24, 48, 72 and 168 hours in order to simulate an intravenous continuous infusion of the drug.

Methods

Preparation of standard solutions

PGE₁ (>99.0% purity) and prostaglandin E₂ (≥98.0% purity) came from The Council of Europe (Strasbourg, France). The other PGE₁ degradation products: prostaglandin A₁ (PGA₁) (≥97.0% purity), 8-iso-PGA₁ (≥97.0% purity), prostaglandin B₁ (PGB₁) (≥98.0% purity), 8-iso-PGE₁ (≥98.0% purity), 11 β-PGE₁ (≥95.0% purity), (15R)-PGE₁ (≥95.0% purity), 15-keto-PGE₁ (≥98.0% purity), 13,14-dihydro-PGE₁ (≥98.0% purity) and 5,6-trans-PGE₂ (≥98.0% purity) came from Cayman Chemical (Ann Arbor, Michigan, USA). Stock solutions of PGE₁ and its degradation products were prepared in HPLC-grade ethanol (lot K47696783613; Merck, Darmstadt, Germany) at 100 µg/mL. Solutions were pooled and diluted in a mobile phase solution A (ultrapure water and formic acid; 99.9:0.1, v/v) and solution B (acetonitrile and formic acid; 99.9:0.1, v/v) (solution A and solution B; 76.0:24.0, v/v) at 1 µg/mL mobile phase.

Calibration standard solutions (CSs) and validation standard solutions (VSs) were prepared from a PGE₁ stock solution at 4000 µg/mL in HPLC-grade ethanol: five CSs (range, 0.25–4.00 µg/mL) by dilution in HPLC-grade ethanol and three VSs by diluting PGE₁ in 10% dextrose (lot 15287416; B. Braun, Crissier, Switzerland) (range: 0.50–2.00 µg/mL) followed by 1/10 dilution of each standard solution in the mobile phase (solution A and solution B; 76.0:24.0, v/v). For the stability study, the commercial injectable drug Prostin VR 500 µg/mL (lot J67635) (Pfizer, Zurich, Switzerland) was diluted in 10% dextrose to obtain PGE₁ solutions at 1.5 and 15 µg/mL.

Separation and detection methods

Chromatographic separation was performed using an LC system equipped with an autosampler (HTC PAL, CTC Analytics, Zwingen, Switzerland), a quaternary pump (Accelera 1250 Pump, Thermo Fisher Scientific, Reinach, Switzerland) and a thermostatted column compartment, on a C₁₈ column (100 Å 3.5 µm 2.1×100 mm, reference WAT200650, Waters, Milford, Massachusetts, USA); injection volume 10 µl; Gradient elution over 23 min from 76% mobile phase solution A to 100% solution B at a flow rate of 300 µl/min. Samples were stored at 4°C during the analysis. Ultra pure water (lot 1077551) and acetonitrile (lot 1018611) came from Biosolve BV (Dieuze, France). Formic acid (lot BCBP4740V) came from Sigma-Aldrich (Steinheim, Germany).

We used a Q-Exactive mass spectrometer (Q Exactive Benchtop, Thermo Fisher Scientific, Reinach, Switzerland) operating with an electrospray ionisation source set in the negative mode (sheath gas flow rate: 40, auxiliary gas flow rate: 2, spray voltage: 4 kV, capillary temperature: 360°C and S-lens level: 80). Resolution was set at 70 000 full-width at half-maximum, the Automatic Gain Control was set to 3⁶ charges and the maximal injection time was set to 100 ms. Interval of the total ion current (TIC) chromatogram was 50–500 mass-to-charge ratio (m/z) with an extraction window ≤10 ppm. The main signal m/z obtained in the TIC was used to generate the extracted ion chromatogram. LC-HRMS control, data acquisition and peak integration were performed with Xcalibur 2.2 (Thermo Fisher Scientific, Reinach, Switzerland).

Method selectivity was assessed using a mixture of PGE₁ powder and its degradation products, as documented in the European Pharmacopeia¹³ and in previous studies.^14–19 Forced degradation tests were additionally performed by exposing PGE₁ for 3 hours under strongly acidic (HCl 0.1 M, 25°C), softly basic (NH₄OH 5%, 25°C), strongly basic (NaOH 0.1 M, 25°C), oxidative (H₂O₂ 15%, 25°C), thermal (H₂O, 80°C) and photolytic (H₂O, UV light, 30°C) conditions. HCl 1 M (lot HC392097), NH₄OH 25% (lot K44577032) and NaOH 1 M (lot B0024695) were obtained from Merck (Darmstadt, Germany) and H₂O₂ 30% (lot BCBJ8339B) were obtained from Sigma-Aldrich (Steinheim, Switzerland).

Method was validated by the accuracy profile method in accordance with the Société Française des Sciences et Techniques Pharmaceutiques requirements.^{20 21} Acceptance limits were set at ±10%.

A calibration curve was constructed with five CSs. Three VSs were used in the presence of matrix. CSs and VSs were prepared in duplicate and triplicate, respectively, and analysed in four independent acquisition series. Statistics and method validation were performed with Enoval (Arlenda, Liège, Belgium). Trueness was expressed as a percentage recovery for each VS. Repeatability and intermediate precision were expressed as a percentage of the relative SD of the theoretical concentrations. Sample stability at 4°C and −20°C was assessed by reanalysing one series of CSs stored at these temperatures for 9 hours (total time for one run series) and 1 week, respectively.

Stability assay

PGE₁ stability was assessed over 7 days. The 1.5 and 15 µg/mL PGE₁ in 10% dextrose were each stored in two different conditions: (1) polypropylene syringes unprotected from light (UPL-syringes) (BD Perfusion 50 mL, reference 300136, Becton-Dickinson, Allschwill, Switzerland) and (2) light-shielded polypropylene syringes (LS-syringes). The LS-syringes were obtained from two different companies, designated Manufacturer 1 (LS1-syringe) (BD Perfusion 50 mL Syringe, reference 300138, Becton-Dickinson, Allschwill, Switzerland) and Manufacturer 2 (LS2-syringe) (B. Braun Original Perfusor 50 mL Syringe, reference 8728861 F-06, B. Braun, Melsungen, Germany). The LS2-syringe was necessary due to problems occurring during analysis with the LS1-syringe. At each time point, three syringes for each concentration and each storage condition (shielded and unshielded from light) were pulled out and each analysed in duplicate.

The prepared syringes were stored in a climatic chamber (day-light (fluorescent technology light lamp), 30°C±2°C, relative humidity 65%±5%) (Climatic Test Cabinet RUMED series, Rubarth Apparate, Laatzen, Germany) over the full duration of the study. One millilitre of PGE₁ solution was collected in each syringe immediately after preparation (T₀) and at 24, 48, 72 and 168 hours. Samples were stored at −20°C until analysis.²² Immediately before analysis, the 15 µg/mL PGE₁ samples were diluted 10-fold with 10% dextrose injection solution and then 10-fold with mobile phase (solution A, solution B and HPLC-grade ethanol; 68.0:22.0:10.0, v/v). The 1.5 µg/mL PGE₁ samples were directly diluted 10-fold with the same mobile phase.

Data analysis

The mean PGE₁ concentration at 24, 48, 72 and 168 hours was measured and reported as a percentage of initial PGE₁ concentration. PGE₁ solutions were considered stable when ≥90.0% of the initial concentration was retained. Differential PGE₁ stability between solutions stored in UPL-syringes and LS-syringes was compared at each time point and assessed by the Wilcoxon signed-rank test using Stata V.12.1 software (StataCorp LP, College Station, TX, USA). Indicative amount of PGE₁ degradation products at each time point was expressed as a percentage of PGE₁ concentration (degradation product peak area/PGE₁ peak area×100). The pH and subvisible particles (≥10 and ≥25 µm) of PGE₁ solutions in UPL-syringes or LS-syringes were measured at each time point with a pH-metre (Mettler-Toledo metre SevenMulti, Mettler-Toledo, Greifensee, Switzerland) calibrated with a pH 4.01 reference standard (Technical Buffer Solution pH 4.01, Mettler-Toledo, Analytical, Schwerzenbach, Switzerland, lot 1A195H) and a liquid particle counting system (HIAC Royco model 9703, Hach Ultra Analytics, Geneva, Switzerland) combined with PharmSpec 2.2 software (Hach Ultra Analytics, Geneva, Switzerland) for data acquisition, respectively.

Results

Accuracy profile

Method validation was determined based on the CS and VS raw data of the four series analysed. Weighted linear regression (1/x) was the best model for validation of the method. Recovery, repeatability and intermediate precision intervals were in the limits of 98.2%–100.3%, 1.9%–2.2% and 2.7%–4.7%, respectively. The accuracy profile showed that the method was valid over a concentration range of 0.086–0.200 µg/mL. The calibration curve was linear over the concentrations of 0.025, 0.05, 0.1, 0.2 and 0.4 µg/mL with a coefficient of determination (r²) equal to 0.9966 and a risk profile guaranteeing that 95.0% of the measurements performed in this range will be within the acceptance limits of ±10%. PGE₁ quantification and detection limits were 0.086 and 0.026 µg/mL, respectively.

Selectivity

PGE₁ and its degradation products were identified according to their m/z and retention time (R_t) values. The R_ts of the products in the mixture varied from 10.82 min (8-iso-PGE₁) to 17.95 min PGB₁, with the PGE₁ peak R_t occurring at 11.45 min (table 1). PGE₁ and the other products were separated successfully, excepted for 15-epi-PGE₁. For each product, peak resolution (R_S) was >1.5, showing that PGE₁ degradation products did not interfere with the peak of PGE₁ itself. However, PGE₁ and 15-epi-PGE₁ were not completely separated (R_S=1.35) despite our method optimisation.

Table 1.

Liquid chromatography-high resolution mass spectrometry analysis of PGE₁ solution and PGE₁ degradation products

Compound	Mass-to-charge ratio (m/z)			Chromatographic retention
	Theoretical	Measured	Difference (ppm)	Time (min)	Relative to PGE1
PGE1	353.23225	353.23172	−1.50	11.45	1.00
8-iso-PGE1		353.23166	−1.67	10.82	0.94
11β-PGE1		353.23157	−1.93	11.97	1.05
(15R)-PGE1		353.23169	−1.59	11.16	0.97
PGA1	335.22169	335.22122	−1.40	17.59	1.54
8-iso-PGA1		335.22141	−0.84	16.46	1.44
PGB1		335.22128	−1.22	17.95	1.57
PGE2	351.21660	351.21603	−1.62	10.94	0.96
5,6-trans-PGE2		351.21625	−1.00	11.90	1.04
15-keto-PGE1		351.21594	−1.88	13.45	1.17
13,14-dihydro-PGE1	355.24790	355.24750	−1.13	12.63	1.10

PGA₁, prostaglandin A₁; PGB₁, prostaglandin B₁; PGE₁, prostaglandin E₁.

Forced degradation tests

PGE₁ was extensively degraded under basic, thermal and oxidative conditions. PGA₁ was the main degradation product observed in forced degradation tests (figures 1 and 2). Under basic or high thermal conditions, PGA₁ isomerises to PGB₁; epimers of PGE₁ and PGA₁ were also observed, namely 8-epi-PGE₁, 11-epi-PGE₁ and 15-epi-PGE₁ (M−H⁺ m/z 353.23225, R_t 10.82, 11.97 and 11.16 min, respectively) and 8-epi-PGA₁ (M−H⁺ m/z 335.22169, R_t16.46 min). Under oxidative conditions, 15-keto-PGE₁ (M−H⁺ m/z 351.21594, R_t 13.45) was observed in minute amounts. UV treatment did not affect degradation. For each product analysed, the main peaks observed in the TIC corresponded to the ion [M−H⁺] (table 1).

Figure 1.

Chemical structures of prostaglandin E₁ degradation products observed in accelerated degradation studies.

Figure 2.

Prostaglandin E₁(PGE₁) accelerated degradation assay. (A) PGE₁ in 10% dextrose injection solution. (B) Strongly acidic condition (HCl 0.1 M, 25°C). (C) Softly basic condition (NH₄OH 5%, 25°C). (D) Strongly basic condition (NaOH 0.1 M, 25°C). (E) Oxidative condition (H₂O₂ 15%, 25°C). (F) Thermal condition (H₂O, 80°C). (G) Photolytic condition (H₂O, UV-light, 30°C). All treatments were performed for 3 hours.

Stability study

PGE₁ solutions stored in UPL-syringes and LS1-syringe remained transparent without precipitation or colour change. No significant pH changes were observed in the PGE₁ solutions stored in UPL-syringes throughout the study, whereas PGE₁ solutions stored in LS1-syringe exhibited pH elevation (from 4.26, 95% interval of confidence (IC95%) 4.2 to 4.3) at T₀ to 6.25 (IC95% 6.1 to 6.4) at 168 hours; table 2). Follow-up complementary ultraviolet-visible (UV–VIS) spectrophotometry analysis of 10% dextrose stored in the UPL-syringe and LS1-syringe conducted at T₀, 24, 48, 72 and 168 hours with a UV–VIS spectrophotometer (Cary 50 Bio UV-Visible spectrophotometer, Agilent Technologies, Basel, Switzerland) equipped with scan application software (CaryWin UV Scan application, version 5.0.0.999, Agilent Technologies, Basel, Switzerland) revealed no changes in the UV–VIS spectrum in UPL-syringes across sampling times. In LS1-syringe, however, a peak at 245.7 nm (range: 244.1–247.1 nm) was observed at the 24 hours time point and later, but not at T₀. Because of these troubling pH and spectrophotometric results, we stopped using LS1-syringe and continued the physicochemical stability study of PGE₁ in 10% dextrose with another brand of light-shielded polypropylene syringes (LS2-syringe). No change in pH was observed in PGE₁ solutions stored in LS2-syringe, and the solutions remained totally transparent without precipitation or colour change throughout the study (table 2). Amounts of subvisible particles (≥10 and ≥25 µm) in UPL-syringes and LS2-syringes were in compliance with the European Pharmacopoeia requirements for injectable drugs throughout the study.

Table 2.

Mean pH of 1.5 and 15 µg/mL PGE₁ in 10% dextrose injection solution stored in UPL and LS1 and LS2 polypropylene syringes at 30°C

PGE1 concentration	Time	Syringe type
		UPL		LS1		LS2
		Mean	IC95%	Mean	IC95%	Mean	IC95%
1.5 µg/mL	T0	4.08	(4.0 to 4.1)	4.26	(4.2 to 4.3)	4.05	(4.0 to 4.1)
	24 hours	4.17	(4.1 to 4.2)	4.96	(4.6 to 5.3)	3.99	(3.9 to 4.0)
	48 hours	4.12	(4.1 to 4.1)	5.70	(5.1 to 6.3)	4.01	(4.0 to 4.0)
	72 hours	4.20	(4.2 to 4.2)	6.00	(5.5 to 6.5)	4.04	(4.0 to 4.1)
	168 hours	4.09	(4.0 to 4.1)	6.25	(6.1 to 6.4)	3.97	(3.9 to 4.0)
15 µg/mL	T0	4.10	(4.1 to 4.1)	4.27	(4.3 to 4.3)	4.01	(4.0 to 4.0)
	24 hours	4.18	(4.1 to 4.2)	4.93	(4.8 to 5.1)	4.00	(4.0 to 4.0)
	48 hours	4.17	(4.1 to 4.2)	5.58	(5.4 to 5.8)	4.03	(4.0 to 4.0)
	72 hours	4.20	(4.2 to 4.2)	5.93	(5.7 to 6.2)	4.04	(4.0 to 4.1)
	168 hours	4.17	(4.1 to 4.3)	6.22	(6.1 to 6.4)	4.03	(4.0 to 4.0)

LS, light-shielded; PGE₁, prostaglandin E₁; UPL, unprotected from light.

After 48 hours, PGE₁ concentrations were 96.1% (IC95% 92.4 to 99.8) and 94.9% (IC95% 90.8 to 99.0) in UPL-syringes and 91.4% (IC95% 87.7 to 95.1) and 96.7% (IC95% 92.2 to 101.2) in LS2-syringe for the 1.5 and 15 µg/mL PGE₁ solutions, respectively. The difference in PGE₁ concentration between the UPL-syringe and LS2-syringe was not statistically significant (p=0.07 (1.5 µg/ml) and p=0.60 (15 µg/ml)). After 72 hours, the PGE₁ concentrations were 90.9% (IC95% 88.9 to 92.9) and 92.8% (IC95% 87.7 to 97.9) in UPL-syringes and 90.0% (IC95% 85.8 to 94.2) and 91.9% (IC95% 85.4 to 98.4) in LS2-syringe for the 1.5 and 15 µg/mL PGE₁ solutions, respectively. The stability results for 1.5 and 15 µg/mL PGE₁ in 10% dextrose in UPL-syringes and LS2-syringe at each time points are presented in table 3.

Table 3.

Stability of 1.5 and 15 µg/mL PGE₁ in 10% dextrose injection solution stored in UPL and LS2 polypropylene syringes at 30°C

PGE1 concentration	Time	Percentage PGE1 remaining
		Syringe type
		UPL		LS2
		Mean	IC95%	Mean	IC95%	p Value
1.5 µg/mL	T0	100.0		100.0
	24 hours	95.7	(90.5 to 100.9)	95.9	(89.5 to 102.3)	0.92
	48 hours	96.1	(92.4 to 99.8)	91.4	(87.7 to 95.1)	0.07
	72 hours	90.9	(88.9 to 92.9)	90.0	(85.8 to 94.2)	0.46
	168 hours	85.8	(83.0 to 88.6)	84.1	(78.7 to 89.5)	0.25
15 µg/mL	T0	100.0		100.0
	24 hours	95.5	(91.3 to 99.7)	99.0	(93.9 to 104.1)	0.25
	48 hours	94.9	(90.8 to 99.0)	96.7	(92.2 to 101.2)	0.60
	72 hours	92.8	(87.7 to 97.9)	91.9	(85.4 to 98.4)	0.75
	168 hours	87.6	(83.5 to 91.7)	86.1	(80.7 to 91.5)	0.35

LS, light-shielded; PGE₁, prostaglandin E₁; UPL, unprotected from light.

PGA₁ was the sole PGE₁ degradation product observed during the stability study. Its levels were detectable at T₀ and increased over time in both PGE₁ formulations and in both types of syringes investigated, with the maximum value being 3.9% (IC95% 3.3 to 4.5) in 1.5 µg/mL PGE₁ solutions point in LS2-syringe at the 168 hours time (table 4).

Table 4.

Indicative amounts of PGE₁ degradation product PGA₁ over time in 10% dextrose injection solution stored in UPL and LS2 polypropylene syringes at 30°C

PGE1 concentration	Time	Percentage PGA1 detected
		Syringe type
		UPL		LS2
		Mean	IC95%	Mean	IC95%	p Value
1.5 µg/mL	T0	1.8	(1.8 to 1.8)	2.0	(1.9 to 2.1)	0.03
	24 hours	1.9	(1.9 to 1.9)	2.3	(2.2 to 2.4)	0.03
	48 hours	2.0	(1.8 to 2.2)	2.5	(2.2 to 2.8)	0.03
	72 hours	2.2	(2.1 to 2.3)	2.8	(2.5 to 3.1)	0.03
	168 hours	2.9	(2.6 to 3.2)	3.9	(3.3 to 4.5)	0.03
15 µg/mL	T0	1.8	(1.8 to 1.8)	1.9	(1.5 to 2.3)	0.46
	24 hours	2.0	(1.9 to 2.1)	1.9	(1.5 to 2.3)	0.75
	48 hours	2.2	(2.1 to 2.3)	2.1	(1.5 to 2.7)	0.75
	72 hours	2.5	(2.4 to 2.6)	2.4	(1.9 to 2.9)	0.46
	168 hours	3.7	(3.6 to 3.8)	3.5	(2.7 to 4.3)	0.46

LS, light-shielded; PGE₁, prostaglandin E₁; UPL, unprotected from light.

Discussion

Method validation

We developed and validated an LC-HRMS method that enables PGE₁ separation and dosing as well as the identification of PGE₁ degradation products. To our knowledge, this is the first LC-HRMS method developed and validated for the analysis of PGE₁ in 10% dextrose prepared at concentrations used in newborn patients.

Stability study

We observed noteworthy pH changes in PGE₁ solutions stored in LS1-syringe over time. Storage of 10% dextrose in LS1-syringe led to the appearance of a substance with an absorption peak at 245 nm present at the 24 hours time point and beyond. This substance was not detectable when the same solution was stored in UPL-syringes. The absence of this peak at T₀ suggests the possible presence of extractable or leachable impurities, such as plasticisers or dyes from the rubber stopper, or the plastic of the LS1-syringe itself, as previously reported.²³ We did not pursue resolution of the identification of the hypothetical substance(s), which would have been beyond the scope of the present study. Nevertheless, the manufacturer was informed about this phenomenon. The pH increase observed in samples stored in LS1-syringe could be a serious problem in unbuffered solutions of drugs that are stable only when maintained at a pH <6. Our observations are supported by a US FDA alert,²⁴ published while this study was underway, concerning the use of different sizes of Becton-Dickinson syringes for the storage of drugs and the risk of chemical interactions of the rubber stopper in certain lots of syringes. The LS1-syringe used in our study was not from any of the lots specified in the FDA alert. Nevertheless, given our findings, we chose to use another type of LS2-syringe for the rest of the stability study.

Our results showed that 1.5 and 15 µg/mL PGE₁ solutions prepared by dilution in 10% dextrose were physically and chemically stable for 48 hours at 30°C in UPL-syringes. Mean PGE₁ concentrations (relative to initial concentrations) remained above 90.0% in LS2-syringe, though their IC95% extended below the acceptable limit of 90.0%²⁵ in the low PGE₁ concentration. Thus, even if the PGE₁ concentrations at 48 hours were not statistically different from the baseline concentrations at T₀ (p=0.07 (1.5 µg/ml) and p=0.60 (15 µg/ml)) when stored in UPL-syringes and LS2-syringe, we could not confirm the stability of our PGE₁ solutions in the LS2-syringe at 48 hours. At 72 hours, mean PGE₁ initial concentrations remained ≥90.0% in both UPL-syringes and LS2-syringe. However, for both PGE₁ concentrations, the IC95% extended below the acceptable limit of 90.0%.

PGA₁ was the sole PGE₁ degradation product observed and its concentration increased over time. Despite a significant difference in PGA₁ levels between UPL-syringes and LS2-syringe loaded with 1.5 µg/mL PGE₁, a similar difference between syringe types was not found with the higher concentration solution. It appears that PGE₁ degradation into PGA₁ was not catalysed by light. Although neonatal exposure to PGA₁ has not been assessed specifically in prior published studies, data about its toxicity are reassuring. PGA₁shows no signs of toxicity and exhibits vascular actions similar to PGE₁.^{16 26 27} Given that 15-epi-PGE₁ did not appear during the stability study, the partial resolution was a non-issue.

In addition to PGE₁ transformation into PGA₁, some loss of PGE₁ could be explained by the phenomenon of adsorption already known,²⁸ thus explaining the variation of PGE₁ in LS2-syringes at low concentration.

A number of PGE₁ stability studies have been reported (4, 20, 25 and 40 µg/mL PGE₁ in 0.9% sodium chloride stable for 15 days at 20°C,²⁹ 9.8 days at 25°C,²² 10 days at 20°C³⁰ and 6 weeks at 21°C,³¹ respectively). None of these studies were performed at 30°C with 1.5 or 15 µg/mL PGE₁ in 10% dextrose injection solution. PGE₁ is a weak acid with a pKa of 4.85 whose stability in solution may be sensitive to heat and pH. Maximum stability of PGE₁ in normal saline, assessed as a function of pH at room temperature, was observed in pH range of 6–7; considerable activity loss was observed outside this range.¹⁶ In newborns, dextrose is the solute of choice for intravenous drug preparations. At pHs between 3.5 and 6, dextrose can provide a caloric energy source critical for neonatal brain development without delivery of unneeded sodium. Our findings indicate that PGE₁ solution stability seems to depend more on pH and temperature than on concentration and/or light exposure.

Conclusion

We developed and validated a LC-HRMS method for the dosage of PGE₁ in a 10% dextrose at the concentrations used in our NICU. This method enabled detection of major PGE₁ degradation products. Solutions of PGE₁ diluted to 1.5 and 15 µg/mL concentrations in 10% dextrose injection solution were stable for 48 hours when stored at 30°C in polypropylene syringes, UPL sources.

What this paper adds.

What is already known on this subject

• In neonatal intensive care units, nurses spend lot of time performing drug-related tasks.

• Solutions of prostaglandin E₁ (PGE₁) diluted in 0.9% sodium chloride are stable for many days at room temperature.

• Centers for Disease Control and Prevention recommend replacing continuous intravenous administration sets no more often than at 96 hours.

What this study adds

• A liquid chromatography-high resolution mass spectrometry method for PGE₁ separation and dosing and the identification of its degradation products in 10% dextrose solution was validated.

• PGE₁ 1.5 and 15 µg/mL in 10% dextrose solution are stable for 48 hours at 30°C in polypropylene syringes exposed to light.

• pH of unbuffered solutions varied significantly depending of the type of syringe.