Abstract
Background:
Furosemide is a commonly used diuretic for the treatment of edema. The pharmacokinetics of furosemide in neonates as they mature remains poorly understood. Microsampling assays facilitate research in pediatric populations.
Results:
We developed and validated a liquid chromatography–tandem mass spectrometry method for the quantitation of furosemide in human whole blood with volumetric absorptive microsampling (VAMS) devices (10 μl). Furosemide was stable in human whole blood VAMS under the study's assay conditions. This work established stability for furosemide for 161 days when stored as dried microsamples at -78°C.
Conclusion:
This method is being applied for the quantitation of furosemide in whole blood VAMS in an ongoing prospective pediatric clinical study. Representative clinical data are reported.
Keywords: : furosemide, human whole blood, LC-MS/MS, pediatrics, pharmacokinetics, volumetric absorptive microsampling
Furosemide, a loop diuretic, is common in both adult and pediatric medicine. Use is particularly frequent in intensive care units, where furosemide is used to manage edema resulting from or complicating hepatic, renal, cardiac and respiratory failure [1–3]. Furosemide inhibits the NKCC cotransporter [4,5]. Inhibition of this electrolyte carrier in the thick ascending loop of Henle increases urinary sodium and water excretion and decreases extracellular water [6,7], thereby decreasing edema. However, furosemide diuresis also causes urinary wasting of sodium, potassium, chloride, phosphate, calcium and magnesium and has various extra-renal effects. In pediatric medicine, these undesired effects are thought to mediate associations between furosemide and various clinical harms, such as life-threatening electrolyte derangements, hypovolemia, nephrocalcinosis, metabolic bone disease and potentially irreversible hearing loss [8]. A greater understanding of furosemide pharmacokinetics (PK) and pharmacodynamics (PD) is needed to develop dosing strategies that optimize the risk–benefit profile of this commonly used drug. In pediatric populations, furosemide PK and PD are influenced by dynamic developmental changes in the processes underlying drug disposition, such as rapid changes in total body and extracellular water, and maturation of renal tubular secretion, the primary mechanism of furosemide elimination [9–15].
Furosemide use is common in preterm infants [2,16]. Among those developing severe bronchopulmonary dysplasia (sBPD), furosemide is the most frequently used pharmacotherapy [16]. Furosemide use in BPD has physiologic rationale and plausible benefit. BPD is caused by injury to the developing lung, interrupting airway, parenchymal, vascular and lymphatic development. Excess pulmonary blood flow, capillary leak and inadequate lymphatic fluid reabsorption contribute to pulmonary edema and impair lung function [17,18]. Furosemide may decrease pulmonary edema and improve lung function, thereby modifying respiratory disease course. The current knowledge on furosemide PK is neonates is limited [9–15] to younger preterm-born infants and does not generalize to this population as they mature beyond 36 weeks' postmenstrual age and develop BPD. Additional clinical pharmacology studies are needed to better define the PK and PD effects of furosemide in this frequently exposed population.
Furosemide concentrations are usually quantified in human plasma or serum by validated liquid chromatography–tandem mass spectrometry (LC-MS/MS) methods [19–21]. However, the large blood volumes required for these assays prevent the feasibility of conducting clinical pharmacology studies in infants with sBPD. Therefore, it is necessary to develop a small volume assay for measuring furosemide in infant blood samples for pharmacokinetic evaluations and further dosage optimization. Volumetric absorptive microsampling (VAMS) is an attractive opportunity for quantitative analysis of drugs. This is particularly true in pediatric critical care settings, where the collection of multiple blood samples of ≥1 ml over a short period of time is a barrier to the feasible conduct of clinical pharmacology studies, and ease and efficiency of use facilitates study conduct in an environment with various time-sensitive demands for patient care. VAMS devices have a sampling tip that absorbs the whole blood upon contact, either from blood collected in a tube or from the sampling site (finger or heel stick). This avoids the use of glass capillary tubes during an intermediary collection step, as recommended for dried blood spots (DBS), an alternative microsampling strategy. Additional benefits are that samples can be stored in their collection container and post-collection processing such as subpunching is obviated. VAMS devices allow the collection of fixed volume blood samples (10, 20 or 30 μl), helping to reduce the hematocrit effect associated with the quantitative analysis of drugs via DBS microsampling [22,23]. We previously developed and reported a LC-MS/MS furosemide assay in urine [24]. This is the first study to report a VAMS LC-MS/MS furosemide assay in human whole blood, and it has been evaluated for the analysis of representative samples from an ongoing pediatric prospective clinical study.
Materials & methods
Materials
Furosemide and furosemide-d5 were purchased from Toronto Research Chemicals (TRC, Toronto, Canada). Nanopure water from a Synergy UV-R system was used in preparing all solvents and for sample preparation. Formic acid (98%), DMSO, LC-MS-grade methanol, acetonitrile and isopropyl alcohol were purchased from EMD Millipore (MA, USA). VAMS devices (10 μl) were purchased from Neoteryx (CA, USA). Blank human whole blood was collected from healthy volunteers (employees) at the Children's Hospital of Philadelphia (Institutional Review Board [IRB] protocol no. 18-015852), which were used for all validation experiments.
Stock solutions, calibration standards & quality controls
Primary stock solutions of furosemide were prepared at a concentration of 1.0 mg/ml in DMSO for the preparation of calibration standards (STD) and quality control (QC) working solutions as two independent solutions in amber glass vials and stored at -20°C when not in use. From the stock solution, STDs 10.0, 25.0 and 50.0 μg/ml, were spiked in human whole blood, left at room temperature and gently inverted often for 15 min to equilibrate before preparing further dilutions. Additional STDs were prepared by serial dilution to the final concentrations of 0.05, 0.10, 0.25, 0.50, 1.0, 2.5 and 5.0 μg/ml. QC concentrations of 20 and 40 μg/ml were spiked in human whole blood, left at room temperature and gently inverted often to equilibrate for 15 min before continuing with further dilutions. QC samples were prepared in human whole blood at four concentrations: 0.05, 0.10, 20 and 40 μg/ml. The primary stock solution for the internal standard (IS) furosemide-d5 was prepared at 0.5 mg/ml in methanol: DMSO (50/50, v/v) in an amber glass vial and stored at -20°C when not in use. The final working solution for furosemide-d5 was prepared at a concentration of 5 ng/ml in acetonitrile for sample extraction and stored at 4°C.
Sample preparation
Following the instructions from Neoteryx, freshly prepared STD and QC samples prepared in human whole blood were loaded to VAMS devices (10 μl, fixed volume). Before loading the VAMS devices with whole blood, all STD and QC samples were gently mixed in an Eppendorf tube at an approximate speed of 6 on a Fisher Scientific Vortexer followed by manually inverting each tube at least 3–4 times to ensure a homogenous solution. STDs and QCs loaded onto VAMS devices were placed in a plastic 96-well drying rack provided by Neoteryx and allowed to dry for 1 h at room temperature unassisted (Figure 1). After drying, the dried absorption tips on the plastic holder were removed and placed into the corresponding wells of a clean 96-well plate for sample extraction. Samples were reconstituted with 10 μl of nanopure water, vortexed (1000 r.p.m. for 2 min), and then incubated (37°C for 10 min). The IS solution (990 μl, 5 ng/ml of d5-furosemide in acetonitrile) was added to the samples, except double-blank samples, which received 990 μl of 100% acetonitrile. The 96-well plate was sonicated (15 min), vortexed (15 min at 500 r.p.m.) and centrifuged (30 minutes at 3220x g at 4°C). Each supernatant was transferred (∼650 μl) to clean wells of a 96-well plate with a multichannel pipette for analysis on the ultra HPLC (UHPLC)-MS/MS instrument.
Figure 1. . Simple schematic for furosemide sample preparation in human whole blood volumetric absorptive microsampling for UHPLC-MS/MS analysis.
IS: Internal standard; RT: Room temperature; cps: Counts per second.
UHPLC-MS/MS analysis
UHPLC-MS/MS analysis for furosemide was performed with an Exion UHLPC coupled to a 6500+ QTrap mass spectrometer (AB Sciex, MA, USA) using furosemide-d5 as an internal standard with conditions optimized for furosemide urine assay [24]. All mass spectrometry conditions are listed in Table 1. Nitrogen gas from a PEAK Scientific generator (Inchinnan, UK) was used for the collision, curtain and nebulizer gases.
Table 1. . Mass spectrometry parameters for quantitation and analysis of furosemide and furosemide-d5 in whole blood volumetric absorptive microsampling samples.
| Analyte | Precursor ion | Product ion | DP† | EP† | CE† | CXP† |
|---|---|---|---|---|---|---|
| Furosemide (quantitative) | 329.0 | 77.9 | ‐55 | ‐10 | ‐40 | ‐9 |
| Furosemide (qualitative) | 329.0 | 125.8 | ‐55 | ‐10 | ‐44 | ‐13 |
| Furosemide-d5 | 334.0 | 206.0 | ‐55 | ‐10 | ‐31 | ‐15 |
Measured in volts. Ionization mode: negative electrospray; source temperature: 450°C; interface heater: on; collision gas: 10 psi; curtain gas: 20 psi; GS1: 65 psi; GS2: 65 psi; and IonSpray voltage: 4500 V.
CE: Collision energy; CXP: Collision cell exit potential; DP: Declustering potential; EP: Entrance potential.
Total runtime for each sample was 4.5 minutes with a flow rate of 0.500 ml/min and injections of 2 μl per supernatant. A Waters (Milford, MA) Acquity HSS C18 column (2.6 μm, 100 Å, 2.1 x 100 mm) was utilized to achieve chromatographic separation. All UHPLC conditions and related solvents are shown in Table 2. Analyst software (version 1.6.3) provided by AB Sciex was used for all data acquisition and processing.
Table 2. . UHPLC gradient and solvents for analysis of furosemide in the whole blood volumetric absorptive microsampling samples.
| Time (min) | Mobile phase A (%) | Mobile phase B (%) | Divert valve |
|---|---|---|---|
| 0.00 | 98.0 | 2.0 | on |
| 0.50 | 98.0 | 2.0 | off |
| 0.70 | 60.0 | 40.0 | |
| 1.50 | 2.0 | 98.0 | |
| 2.50 | 2.0 | 98.0 | |
| 3.00 | 98.0 | 2.0 | |
| 4.00 | 98.0 | 2.0 | stop |
| 4.50 | 98.0 | 2.0 |
Mobile phase A: 0.1% formic acid in water; mobile phase B: 0.1% formic acid in acetonitrile; weak wash: 1:1 (v/v) water/ acetonitrile with 0.1% formic acid; and strong wash: 2:2:2:2:1 (v/v/v/v/v) water/acetonitrile/methanol/isopropanol/phosphoric acid.
Validation
This assay was validated for the quantitation of furosemide in human whole blood VAMS devices in a 3-day validation study. All validation experiments were performed following guidance from the US FDA for bioanalytical validation [25]. FDA guidance recommends performing the following assessments: linearity, sensitivity, accuracy, precision, selectivity/specificity, matrix effect, recovery, hematocrit (HCT) effect, dilution integrity and stability.
Linearity & sensitivity
Assay linearity was assessed within the calibration range of 0.05–50 μg/ml, and linearity was evaluated by visual inspection of the calibration plot, the residuals plot and percent relative error (%RE) of the back-calculated concentration plot as described previously [23]. The standardized residual plot and back-calculated %RE plot were within the ±15% limits of the expected concentrations of furosemide. Whole blood (∼20 ml) was collected daily (with lithium heparin as the anticoagulant) from healthy volunteers (IRB no. 18-015852), who were not receiving furosemide, to prepare STD and QC working solutions. The lowest limit of quantitation (LLOQ) for this assay was 0.05 μg/ml. A signal-to-noise ratio at the LLOQ was used to determine assay sensitivity.
Accuracy & precision
Four QC levels of 0.05, 0.10, 20 and 40 μg/ml for furosemide prepared in human whole blood employing the VAMS devices (n = 6) were evaluated by intra- and interday studies. A single-day analytical validation run (n = 6 at each QC level), was used to calculate the intraday accuracies and precisions for furosemide. Interday accuracies and precisions were calculated based on the 3-day validation runs (n = 18 at each QC level). For all validation studies, acceptable accuracies should be within ± 15% (± 20% for the LLOQ), and precisions should be ≤15% (≤20% for LLOQ) of the theoretical value.
Dilution integrity & carryover
Dilution integrity was assessed at a 20-fold dilution from 100 μg/ml of furosemide in human whole blood VAMS (n = 6). Sample extraction was performed following the validated sample preparation method. Single-blank extracts (prepared from blank whole blood VAMS devices) were utilized to prepare dilutions of the extracted samples (100 μg/ml of furosemide in human whole blood VAMS), then mixed well for analysis. Assessment of carryover was evaluated by placing two double-blank samples after the upper limit of quantitation (ULOQ). The peak area of furosemide and furosemide-d5 in the double-blank samples should not exceed >20% of the LLOQ sample.
Selectivity & specificity
Selectivity and specificity were assessed in six individual lots and a pooled lot of human whole blood (equal volumes of the six individual lots mixed together). VAMS devices were loaded as double blanks (n = 1), as well as spiked LLOQ samples (n = 3) in the six individual lots and one pooled lot of blood. Loaded VAMS samples were extracted utilizing the validated method to confirm the absence of interference peaks at the retention times for furosemide and furosemide-d5, as well as cross-interference between both.
Recovery & matrix effect
Recovery was assessed by comparing the peak areas of sample extracts (containing furosemide and furosemide-d5) with double-blank extracts (containing no furosemide or furosemide-d5), which were spiked with the same concentrations of furosemide and furosemide-d5 at all four QC concentrations (n = 6). Matrix effect was assessed at the four QC concentrations (n = 6). VAMS devices were loaded and extracted as double-blank samples, then spiked at all QC concentrations into the final extracts. Peak areas of the postspiked extracts for furosemide and furosemide-d5 were compared with the peak areas of neat samples (n = 6, no matrix) prepared in acetonitrile at the same concentrations.
HCT effect
We assessed for an HCT effect in our furosemide in human whole blood VAMS assay, using low (23.5 %), normal (42.0 %) and high (61.2 %) HCT levels. Blood was collected from a healthy volunteer (IRB no. 18-015852) with a HCT level of 42.0%. We then prepared low and high HCT levels using an aliquot of the normal HCT blood sample, adding or removing the necessary quantity of plasma to achieve the desired HCT level for analysis. An aliquot of each prepared HCT level was then measured at the Translational Core Lab at Children's Hospital of Philadelphia using a Sysmex xt2000 hematology analyzer (Kobe, Japan) to determine HCT values. Whole blood was spiked at 0.100, 20.0 and 40.0 μg/ml of furosemide in each HCT level. VAMS (10 μl; n = 6) were loaded with the human whole blood samples and analyzed following the validated method. Furosemide concentrations were measured against a freshly prepared set of calibration curves and QCs in blood with HCT of 42.0% to determine the recovery of furosemide and the effect of HCT on VAMS quantitation.
Stability
The stability of furosemide was evaluated by loading VAMS devices (n = 6) with the LLOQ, low, medium and high QC concentrations in human whole blood (except for the 24-h stability experiment, which was evaluated at the low and high QC levels), and drying unassisted for 1 h at room temperature. Once drying time was complete, the QC samples were stored at designated temperatures and evaluated for different lengths of time. Stability assessments were performed for up to 20 h (room temperature), 1 week (7 days; 40°C, room temperature, 4, -20 and -78°C) and 5 months (161 days; -78°C). After the specified time, samples were set at room temperature for approximately 1 h to thaw completely before sample processing. Stability QCs were measured against freshly prepared calibration curves and QCs in whole blood using VAMS devices. Autosampler stability was assessed by reanalyzing sample extracts stored in the autosampler of the instrument for 96 h at 10°C.
Clinical samples
This pediatric prospective study aims to characterize furosemide pharmacokinetics in infants with established sBPD. Whole blood VAMS samples (10 μl) are obtained by heel-stick at five timepoints between furosemide dose administrations. The IRB at the Children's Hospital of Philadelphia approved the protocol (IRB no. 20-017983), and informed consent was obtained from the parents of all study participants. This validated method was evaluated for clinical sample analysis of three pediatric subjects. VAMS samples were collected from the subjects’ heel by first placing a heel warmer at the site of collection, cleaning the heel with an isopropyl wipe, allowing the site to dry, then using a lancet on the desired incision site. The first drop of blood was wiped away, then duplicate VAMS tips (10 μl; n = 2) were loaded according to the manufacturer's instructions at each time point. The first sample was collected 30 min before the dose of furosemide was given to the subjects, with the following four samples collected over a period of ~8 h before the next dose was given ~12 h later. VAMS samples were dried at room temperature for a minimum of 1 h and then transported and stored in a -78°C freezer until sample analysis.
Results
Validation
Linearity & sensitivity
Calibration curves (n = 2 per analysis) were linear and reproducible over the calibration range of 0.05–50 μg/ml for furosemide in human whole blood using VAMS devices for the 3-day validation studies. The 3-day validation study employed a linear regression of 1/x2 weighting. The slope, y-intercept and correlation coefficient (r2) values (mean ± SD) were (3.02 ± 0.104) × 10-4, (1.67 ± 1.23) × 10-3 and 0.997 ± 0.000889, respectively. Calibration standards from the 3-day validation study had accuracies and percent coefficient of variation (%CV) of 88.6–107%, and 2.65–9.92, respectively. Representative chromatograms of the double blank, single blank and LLOQ in whole blood VAMS are shown in Figure 2. The LLOQ for this assay is 0.05 μg/ml with a signal-to-noise ratio >270.
Figure 2. . Representative chromatograms.
(A) Double blank, (B) single blank, and (C) lowest limit of quantitation (0.05 μg/ml) for furosemide (left) and furosemide-d5 (right) measured in human whole blood volumetric absorptive microsampling.
cps: Counts per second.
Accuracy & precision
Results for the intra- and inter-day accuracies and precisions for furosemide in whole blood VAMS (n = 6) at the four QC concentrations of 0.05, 0.10, 20 and 40 μg/ml are shown in Table 3. Intra- and inter-day accuracies ranged from 94.7 to 111% with %CV ranging from 3.35 to 11.5 across all QC concentrations.
Table 3. . Summary of validation outcomes: intra- and inter-day accuracies and precisions for furosemide in human whole blood volumetric absorptive microsampling.
| Parameter | Day | n | 3-day validation summary | |||
|---|---|---|---|---|---|---|
| LLOQ | LQC | MQC | HQC | |||
| Furosemide (μg/ml) | 0.050 | 0.100 | 20.0 | 40.0 | ||
| Mean ± SD | 1 | 6 | 0.0556 ± 0.00352 | 0.100 ± 0.0823 | 21.1 ± 1.46 | 39.2 ± 4.12 |
| Accuracy (%) | 1 | 6 | 111 | 100 | 106 | 98.0 |
| %CV | 1 | 6 | 6.33 | 7.45 | 6.92 | 10.5 |
| Mean ± SD | 2 | 6 | 0.0523 ± 0.00507 | 0.102 ± 0.0742 | 21.8 ± 0.731 | 42.6 ± 2.16 |
| Accuracy (%) | 2 | 6 | 105 | 102 | 109 | 106 |
| %CV | 2 | 6 | 9.70 | 7.29 | 3.35 | 5.08 |
| Mean ± SD | 3 | 6 | 0.0496 ± 0.00569 | 0.0947 ± 0.0108 | 21.0 ± 1.28 | 45.1 ± 2.64 |
| Accuracy (%) | 3 | 6 | 99.1 | 94.7 | 105 | 113 |
| %CV | 3 | 6 | 11.5 | 11.4 | 6.11 | 5.86 |
| Mean ± SD | 1, 2, 3 | 18 | 0.0525 ± 0.00521 | 0.0987 ± 0.00872 | 21.3 ± 1.19 | 42.3 ± 3.82 |
| Accuracy (%) | 1, 2, 3 | 18 | 105 | 98.7 | 107 | 106 |
| %CV | 1, 2, 3 | 18 | 9.92 | 8.83 | 5.58 | 9.04 |
%CV: Percent coefficient of variation; HQV: Higher quality control; LLOQ: Lowest limit of quantitation; LQC: Lower quality control; MQC: Middle quality control.
Dilution integrity & carryover
Twenty-fold dilution integrity was assessed by loading VAMS (n = 6) with 100 μg/ml of furosemide in human whole blood and diluting the sample extracts with single-blank whole blood extracts from VAMS. Results from the dilution integrity experiment proved to be within an acceptable range (107 ± 6.73%) for this assay. Minimal carryover was observed in the single-blank injections with <20% of furosemide in comparison to the LLOQ samples through all validation experiments.
Selectivity & specificity
Six individual lots and a pooled lot of human whole blood were loaded as double-blank samples onto VAMS devices (n = 1). Once completely dry, samples were extracted following the optimized method. There were no interference peaks at the retention times of either furosemide and furosemide-d5. The six individual lots and the pooled lot of whole blood were spiked with furosemide at the LLOQ concentration (n = 3). These samples were extracted following the validated method. The LLOQ samples were measured against freshly prepared STD and QC VAMS samples. Accuracies for furosemide at the LLOQ concentration were within an acceptable range (± 20%) between 84.9 and 109% for all individual and pooled lots.
Recovery & matrix effect
The recovery of furosemide from VAMS was evaluated at all four QC concentrations (n = 6) and were within 23.9–25.0% (Table 4). Recovery remained consistent across all QC concentrations. The matrix effect in the six individual lots and one pooled lot of human whole blood at all QC concentrations was minimal when calculating with both the analyte peak area and the IS-corrected (peak area ratio) values. Values for matrix effect calculated with peak area ranged from 98.7 to 107% and, when calculated with the peak area ratio (IS correction), ranged from 96.2 to 111%. These results show that whole blood VAMS samples across the four QC concentrations had minimal matrix effect.
Table 4. . Recovery and matrix effect of furosemide from human whole blood volumetric absorptive microsampling devices at each quality control concentration (n = 6).
| Extraction conditions | Area ratio at furosemide concentration (μg/ml) | |||
|---|---|---|---|---|
| 0.05 | 0.1 | 20 | 40 | |
| Regular extraction (n = 6) | 0.0127 | 0.0241 | 4.71 | 8.69 |
| 0.0150 | 0.0256 | 4.37 | 8.95 | |
| 0.0146 | 0.0240 | 4.40 | 7.99 | |
| 0.0125 | 0.0226 | 4.40 | 8.18 | |
| 0.0125 | 0.0225 | 4.57 | 7.34 | |
| 0.0131 | 0.0210 | 4.42 | 9.81 | |
| Mean | 0.0134 | 0.0233 | 4.48 | 8.49 |
| Post extraction (n = 6) | 0.0506 | 0.0938 | 16.5 | 36.2 |
| 0.0501 | 0.0944 | 19.1 | 37.9 | |
| 0.0496 | 0.0928 | 19.4 | 38.1 | |
| 0.0503 | 0.0848 | 18.9 | 38.7 | |
| 0.0519 | 0.0873 | 18.0 | 30.7 | |
| 0.0515 | 0.0889 | 19.3 | 34.2 | |
| Mean | 0.0507 | 0.0903 | 18.5 | 36.0 |
| Neat (n = 6) | 0.0430 | 0.0963 | 17.3 | 38.4 |
| 0.0478 | 0.0942 | 16.3 | 37.6 | |
| 0.0459 | 0.0934 | 14.7 | 39.6 | |
| 0.0426 | 0.0843 | 20.0 | 35.8 | |
| 0.0498 | 0.0884 | 18.9 | 37.4 | |
| 0.0454 | 0.0868 | 17.8 | 35.6 | |
| Mean | 0.0458 | 0.0906 | 17.5 | 37.4 |
| Recovery (%) | 26.4 | 25.8 | 24.2 | 23.6 |
| Matrix effect (%) | 111 | 100 | 106 | 96.2 |
| Coefficient of variation matrix effect (%) | 1.9 | 4.3 | 6.4 | 8.2 |
HCT effect
The effect of HCT on the quantitation of furosemide (0.10, 20.0 and 40.0 μg/ml) in whole blood VAMS was evaluated in vitro, and the results are shown in Table 5. On the basis of the calibration curves prepared at normal HCT levels (42.0%), the recovery of furosemide was higher at the low HCT level (23.5%) resulting in accuracy greater than 115%. At normal HCT levels, the accuracies ranged from 97.3 to 105%. At higher HCT levels (61.2%), furosemide had accuracies ranging from 78.6 to 86%. The furosemide concentration differences were statistically significant at all three QC concentrations at low versus normal HCT levels (p < 0.01), normal versus high HCT levels (p < 0.03), and low versus high HCT levels (p < 0.005).
Table 5. . Hematocrit effect on the quantitation of furosemide in human whole blood volumetric absorptive microsampling devices at the low, medium and high quality control concentrations (n = 6) in low, normal and high hematocrit levels.
| Furosemide (μg/ml) | Accuracy (%) ± %CV | ||
|---|---|---|---|
| Low HCT (23.5%) | Normal HCT (42.0%) | High HCT (61.2%) | |
| 0.100 | 168 ± 14.5 | 98.2 ± 7.42 | 78.6 ± 14.5 |
| 20.0 | 126 ± 6.64 | 105 ± 7.36 | 86 ± 4.61 |
| 40.0 | 121 ± 3.55 | 97.3 ± 11.3 | 80.1 ± 7.97 |
CV: Coefficient of variation; HCT: Hematocrit.
Stability
Stability assessments of furosemide (Table 6) were conducted using loaded VAMS devices at the LLOQ, low, medium and high QC concentrations (n = 6), except for the 20-h room temperature stability experiment, which was only evaluated at the low and high QC concentrations (n = 6). Stability studies were evaluated under the following conditions: 20 h at room temperature; 1 week at 40°C, room temperature, 4°C, -20°C and -78°C; 5 months (161 days) at -78°C; and autosampler stability. QC samples at room temperature remained stable for 20 h with accuracies ± precisions of 103 ± 11.4% (low) and 85.6 ± 12.0% (high). Results for 1 week stability were only stable at -78°C with accuracies and %CV ranging from 92.3 to 103% and 7.27 to 12.8, respectively. QC samples stored for 5 months (161 days) at -78°C were stable with accuracies and %CV ranging from 83.3 to 112% and 2.49 to 8.27, respectively. Extracts stored in the autosampler for 24 h were also stable with accuracies ranging from 91.0 to 113% with %CV ranging from 3.59 to 13.4.
Table 6. . Stability of furosemide (n = 6) in human whole blood volumetric absorptive microsampling samples at different storage conditions.
| Furosemide (μg/ml) | Accuracy (%) of furosemide concentrations ± %CV | ||
|---|---|---|---|
| 20 h at RT | 1 week at 40°C | 1 week at RT | |
| 0.100 | 103 ± 11.4 | 71.6 ± 6.80 | 88.6 ± 14.7 |
| 20.0 | ND | 69.7 ± 16.6 | 74.8 ± 15.9 |
| 40.0 | 85.6 ± 12.0 | 72.1 ± 8.03 | 57.3 ± 20.0 |
| Furosemide (μg/ml) | 1 week at 4°C | 1 week at -20°C | 5 months at -78°C |
|---|---|---|---|
| 0.100 | 84.8 ± 9.47 | 97.6 ± 10.8 | 91.7 ± 6.88 |
| 20.0 | 78.0 ± 10.5 | 90.5 ± 9.37 | 112 ± 2.49 |
| 40.0 | 56.8 ± 9.92 | 70.7 ± 19.2 | 105 ± 8.27 |
%CV: Percent coefficient of variation; ND; Not determined; RT: Room temperature.
Clinical samples
This method was used for furosemide analysis in VAMS samples from three infants after a nasogastric or nasoduodenal dose of furosemide (Table 7). Representative chromatograms for the predose sample and the sample collected 1.5 h after the first dose from subject 1 are shown in Figure 3. The first sample was collected before the dose of furosemide, and VAMS samples for each subject were below the limit of quantitation. The final sample in subject 1 was also below the limit of quantitation. All three subjects showed a similar trend, where the third VAMS sample reaches the Cmax, and the samples that follow progressively decrease in concentration (Figure 4).
Table 7. . Measured concentrations of furosemide in volumetric absorptive microsampling samples collected via heel-prick from three pediatric subjects .
| Subject 1 | Subject 3 | Subject 6 | |||
|---|---|---|---|---|---|
| Dosing weight (kg): 3.8 | Dosing weight (kg): 4.7 | Dosing weight (kg): 4.5 | |||
| Dose: 7.6 mg (2 mg/kg) Route: gastric |
Dose: 4.7 mg (1 mg/kg) Route: gastric |
Dose: 9 mg (2 mg/kg) Route: duodenal |
|||
| Sample ID (time after dose) | Furosemide concentration (μg/ml) | Sample ID (time after dose) | Furosemide concentration (μg/ml) | Sample ID (time after dose) | Furosemide concentration (μg/ml) |
| VAMS 1 (predose) | BLQ | VAMS 1 (predose) | BLQ | VAMS 1 (predose) | BLQ |
| VAMS 2 (16 min) | 0.584 | VAMS 2 (16 min) | 0.0513 | VAMS 2 (15 min) | 0.718 |
| VAMS 3 (93 min) | 0.876 | VAMS 3 (91 min) | 0.592 | VAMS 3 (89 min) | 0.987 |
| VAMS 4 (240 min) | 0.153 | VAMS 4 (239 min) | 0.334 | VAMS 4 (240 min) | 0.522 |
| VAMS 5 (480 min) | BLQ | VAMS 5 (479 min) | 0.148 | VAMS 5 (482 min) | 0.118 |
BLQ: Below the limit of quantitation; VAMS: Volumetric absorptive microsampling.
Figure 3. . Representative chromatograms.
Furosemide (left), and furosemide-d5 (right) in whole blood volumetric absorptive microsampling collected from a heel-stick from subject 1 (dose: 2 mg/kg). (A) Predose. (B) 1.5 h after first dose.
cps: Counts per second.
Figure 4. . The whole blood volumetric absorptive microsampling pharmacokinetic profile of furosemide in pediatric patients.
Subject 1 (dose: 2 mg/kg gastric oral solution), subject 3 (dose: 1 mg/kg gastric oral solution), and subject 6 (dose: 2 mg/kg duodenal oral solution). Predose samples (all subjects) and the final sample in subject 1 are not shown because concentrations were below the limit of quantitation.
Discussion
We developed and validated a furosemide assay with blood microsampling, allowing for the feasible conduct of a pharmacokinetic study in a critically ill infant population, hospitalized preterm-born infants with severe bronchopulmonary dysplasia (sBPD). A major impediment to clinical pharmacology research in this population is the need to restrict phlebotomy losses because of low total circulating blood volumes. Here, we describe a furosemide assay with 10 μl of human whole blood using VAMS devices. Drops of blood obtained through heel-prick, a routine procedure in neonatal intensive care units, suffice to obtain this volume. This small sampling volume allows for a PK study with a rich sampling strategy, an approach that would not be possible with traditional sample volumes. Data from the application of this assay in the first three enrolled infants are displayed in Table 7, showing plausible concentration-time patterns and supporting the utility of this assay in meeting our research objectives.
Although furosemide has been used in the management of neonatal lung disease for more than 40 years and is commonly used in modern practice, important PK uncertainties remain. A key knowledge gap is the characterization of how furosemide PK changes during the first year of life in preterm born infants. Severe BPD is a common and consequential morbidity of preterm birth, and furosemide is the most common pharmacotherapeutic exposure [16]. However, the neonatal furosemide PK literature is limited to a few small studies [9–15] that do not generalize to older preterm infants with sBPD because of dynamic maturation of the processes responsible for drug disposition between preterm birth and diagnosis at 36 weeks' postmenstrual age [11,12]. Most significant is the maturation of renal clearance, which drives both furosemide elimination and PD effect. Renal clearance first increases gradually as glomerular filtration rises after birth and then more rapidly after 32 weeks postmenstrual age, when tubular secretion matures and plays a more prominent role in furosemide urinary excretion [11]. Absorption (gut maturity), distribution (decreasing extracellular volume) and metabolism (increased glucuronidation) are also susceptible to maturational changes compelling sBPD specific study [26]. Second, enteral furosemide administration is routine in infant populations. Bioavailability data have only been collected in a few neonates, with a range of 20–106% reported [11,13]. Standard clinical dosing presumes equal 50% bioavailability for both gastric and duodenal use despite the lack of data to support bioequivalence among these enteral administration routes. Lastly, several other covariates with strong plausibility for influencing furosemide PK, including PMA, renal function, hypoalbuminemia and feeding status, vary notably within and between hospitalized infants [11,26,27]. An appropriate dosage regimen may benefit from considering readily available clinical covariates to guide individualized treatments. Therefore, it is necessary to develop a small volume assay for measuring furosemide in infant blood samples for pharmacokinetic evaluations and dose optimization. Although furosemide urinary excretion rates are the most accurate effective drug concentration for evaluating diuretic effect, blood levels are important in considering off-target harms such as ototoxicity and to estimate PK parameters such as volume of distribution and blood clearance. Furosemide quantitation in whole blood and urine allows a more comprehensive PK characterization. Both will be obtained in an ongoing clinical study.
Our assay has limitations. First, we noted an HCT effect, with an overestimation of furosemide concentrations at low hematocrit and an underestimation at high hematocrit. This is notable because a key benefit of VAMS over DBS is mitigating the HCT effect associated with the latter [22,23]. In DBS, HCT influences viscosity and the extent to which a blood spot spreads, thereby influencing the amount of blood volume and analyte present in a fixed-sized subpunch from the DBS paper substrate. VAMS, by sampling a consistent blood volume, can help overcome this. However, additional factors can lead to a HCT effect despite consistent volume sampling. Furosemide is highly protein-bound [10], and we speculate that greater binding and lower extraction recovery at higher HCT values may have led to our pattern of results: underestimation at high levels and overestimation at low levels (Table 5). Of interest, this pattern is opposite the direction of bias expected from a HCT effect observed with DBS, where lower HCT and viscosity samples are expected to have greater spread, resulting in less blood volume and an underestimation of analyte. Indeed, an opposing direction of bias at low HCT values was noted in a furosemide DBS assay [28]. Prior work has emphasized the importance of high extraction recovery in minimizing this volume-independent HCT effect [29]. Indeed, a related limitation of our assay was low recovery values (Table 4). We accepted these recovery values as we noted high sensitivity and signal saturation, our reported extraction approach provided cleaner samples and our extraction recovery was consistent across our desired concentration range. These limitations are particularly relevant when applying the assay to determine absolute furosemide concentrations across populations with a wide range of hematocrit values, as may be the case, for example, if used for therapeutic drug monitoring. There is no clinical indication for therapeutic drug monitoring of furosemide in our study population. This limitation is less critical in the context of our study objectives because our PK study relies primarily on within-subject changes in furosemide concentrations over a single dose without meaningful differences in the patient's hematocrit during this short period. Further, HCT is maintained within a relative narrow range in our study population, rarely varying more than 10% between patients. In contrast, our assay scrutinized a 38% HCT range. Nonetheless, given this HCT effect of furosemide quantitation we are collecting levels in our patients and evaluating it as a covariate in modeling. Lastly, although our stability data at -78°C were acceptable, we did not meet accepted standards at -20°C. In response, samples were directly transferred to a -78 C freezer and maintained in this storage condition until quantitation. We speculate that additional differences between our -20°C and -78°C freezers, such as differences in humidity or the frequency with which they were opened as part of routine laboratory activity may have played a role.
Prior clinical and PK studies for furosemide have been performed in human plasma or serum [19–21]. However, there are no published microsampling methods using whole blood VAMS devices to quantitate furosemide. A prior assay using DBS has been reported. In addition to the contrasting hematocrit effect bias, important differences include greater sample volumes (30 vs 10 μl) and a lower upper limit of quantitation (1.25 vs 40 μg/ml) in the prior assay [29]. The results of the clinical samples included in this article are part of an ongoing, multi-year clinical study to estimate furosemide PK in infants with established severe BPD. We have included the results from three subjects to demonstrate the feasibility of sample collection, storage and analysis using the described approach. These data demonstrate the feasibility of conducting a PK study, but additional assay optimization is needed to evaluate formally the use of VAMS for absolute furosemide concentration measurements as part of clinical practice. Future analyses to determine the blood to plasma distribution of furosemide would also be useful because these would facilitate quantitative comparisons with previous literature reporting plasma furosemide levels.
Conclusion
An LC-MS/MS method for the quantitation of furosemide in 10 μl of human whole blood was developed and validated using VAMS devices. The calibration range (0.05–50 μg/ml) showed good accuracy and precision with values within ±15% (±20% for the LLOQ) of the theoretical values. Representative clinical samples from an ongoing pediatric PK study were assessed to determine the feasibility of this validated method. The collection a larger volumes of blood samples to investigate plasma pharmacokinetics of furosemide in critically ill infants is not feasible. This furosemide VAMS assay allows for a small volume collection of blood to investigate PK in infants.
Future perspective
The present study demonstrates that alternate sampling with microsampling devices is a valuable tool for pediatric clinical pharmacology research. The recent advancements in bioanalytical chemistry and enhanced sensitivity of modern LC-MS instruments provide the needed capabilities to quantify low concentrations of furosemide from a small volume of clinical samples. VAMS provides several benefits to patients, including the collection of small volumes of blood and the potential for remote sampling from home.
Summary points.
Background
A volumetric absorptive microsampling (VAMS) assay for the quantitation of furosemide was developed, validated and evaluated for clinical samples from an ongoing pediatric pharmacokinetic research study.
Experimental
VAMS devices were extracted by a simple protein precipitation method and analyzed by liquid chromatography–tandem mass spectrometry for furosemide quantitation (0.05–50 μg/ml) as dried whole blood VAMS.
Results
Intra- and inter-day accuracies (94.7–111%) and precision (≤11.5%) based on a 3-day validation study were within acceptable range following the US FDA bioanalytical guidelines.
The stability of furosemide was acceptable for 20 h at room temperature and 161 days at -78°C in human whole blood VAMS samples.
Conclusion & future perspective
This validated method accurately and precisely quantifies furosemide in dried human whole blood VAMS devices, which has been applied to clinical samples from an ongoing pediatric prospective pharmacokinetic study.
Footnotes
Financial & competing interests disclosure
This research is supported by the Eunice Kennedy Shriver National Institute of Child Health & Human Development of the National Institutes of Health under award no. K23HD101651 (to NA Bamat). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
No writing assistance was utilized in the production of this manuscript.
Disclaimer
The funding sources had no role in the study design; the collection, analysis, and interpretation of the data; the writing of the report or the decision to submit for publication.
Ethical conduct of research
Blank human whole blood was collected from healthy volunteers (employees) at the Children's Hospital of Philadelphia (Institutional Review Board protocol no. 18-015852 and 20-017983), which were used for all validation experiments. The authors state that they have obtained appropriate institutional review board approval or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In addition, for investigations involving human subjects, informed consent has been obtained from the participants involved.
References
Papers of special note have been highlighted as: • of interest; •• of considerable interest
- 1.Feudtner C, Dai D, Hexem KR et al. Prevalence of polypharmacy exposure among hospitalized children in the United States. Arch. Pediatr. Adolesc. Med. 166(1), 9–16 (2012). [DOI] [PubMed] [Google Scholar]
- 2.Hsieh EM, Hornik CP, Clark RH et al. Medication use in the neonatal intensive care unit. Am. J. Perinatol. 31(9), 811–821 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Mccoy IE, Chertow GM, Chang TI. Patterns of diuretic use in the intensive care unit. PLOS One 14(5), e0217911 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Cotton R, Suarez S, Reese J. Unexpected extra-renal effects of loop diuretics in the preterm neonate. Acta Paediatr. 101(8), 835–845 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Orlov SN, Koltsova SV, Kapilevich LV et al. NKCC1 and NKCC2: the pathogenetic role of cation-chloride cotransporters in hypertension. Genes. Dis. 2(2), 186–196 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.O'Donovan BH, Bell EF. Effects of furosemide on body water compartments in infants with bronchopulmonary dysplasia. Pediatr Res 26(2), 121–124 (1989). [DOI] [PubMed] [Google Scholar]
- 7.Segar JL, Chemtob S, Bell EF. Changes in body water compartments with diuretic therapy in infants with chronic lung disease. Early Hum Dev 48(1-2), 99–107 (1997). [DOI] [PubMed] [Google Scholar]
- 8.Stewart A, Brion LP. Intravenous or enteral loop diuretics for preterm infants with (or developing) chronic lung disease. Cochrane Database Syst. Rev. , CD001453 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Aranda JV, Lambert C, Perez J et al. Metabolism and renal elimination of furosemide in the newborn infant. J. Pediatr. 101(5), 777–781 (1982). [DOI] [PubMed] [Google Scholar]
- 10.Aranda JV, Perez J, Sitar DS et al. Pharmacokinetic disposition and protein binding of furosemide in newborn infants. J. Pediatr. 93(3), 507–511 (1978). [DOI] [PubMed] [Google Scholar]
- 11.Mirochnick MH, Miceli JJ, Kramer PA et al. Furosemide pharmacokinetics in very low birth weight infants. J. Pediatr. 112(4), 653–657 (1988). [DOI] [PubMed] [Google Scholar]; •• Suggests that there are rapid developmental changes in furosemide clearance.
- 12.Mirochnick MH, Miceli JJ, Kramer PA et al. Renal response to furosemide in very low birth weight infants during chronic administration. Dev. Pharmacol. Ther. 15(1), 1–7 (1990). [DOI] [PubMed] [Google Scholar]
- 13.Peterson RG, Simmons MA, Rumack BH et al. Pharmacology of furosemide in the premature newborn infant. J. Pediatr. 97(1), 139–143 (1980). [DOI] [PubMed] [Google Scholar]
- 14.Tuck S, Morselli P, Broquaire M, Vert P. Plasma and urinary kinetics of furosemide in newborn infants. J. Pediatr. 103(3), 481–485 (1983). [DOI] [PubMed] [Google Scholar]
- 15.Vert P, Broquaire M, Legagneur M, Morselli PL. Pharmacokinetics of furosemide in neonates. Eur. J. Clin. Pharmacol. 22(1), 39–45 (1982). [DOI] [PubMed] [Google Scholar]
- 16.Bamat NA, Kirpalani H, Feudtner C et al. Medication use in infants with severe bronchopulmonary dysplasia admitted to United States children's hospitals. J. Perinatol. 39(9), 1291–1299 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]; • Describes prominent use of furosemide in contemporary clinical practice.
- 17.Brown ER, Stark A, Sosenko I et al. Bronchopulmonary dysplasia: possible relationship to pulmonary edema. J. Pediatr. 92(6), 982–984 (1978). [DOI] [PubMed] [Google Scholar]
- 18.Northway WH Jr, Rosan RC, Porter DY. Pulmonary disease following respirator therapy of hyaline-membrane disease. Bronchopulmonary dysplasia. N. Engl. J. Med. 276(7), 357–368 (1967). [DOI] [PubMed] [Google Scholar]
- 19.Heffron B, Taddei L, Benoit M, Negrusz A. Quantification of several acidic drugs in equine serum using LC-MS-MS. J. Anal. Toxicol. 37(8), 600–604 (2013). [DOI] [PubMed] [Google Scholar]
- 20.Ritscher S, Georges C, Wunder C, Wallemacq P, Persu A, Toennes SW. Assessment of adherence to diuretics and beta-blockers by serum drug monitoring in comparison to urine analysis. Blood Press. 29(5), 291–298 (2020). [DOI] [PubMed] [Google Scholar]
- 21.Ritscher S, Hoyer M, Wunder C et al. Evaluation of the dose-related concentration approach in therapeutic drug monitoring of diuretics and beta-blockers – drug classes with low adherence in antihypertensive therapy. Sci. Rep. 9(1), 15652 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Spooner N, Deniff P, Michielson L et al. A device for dried blood microsampling in quantitative bioanalysis: overcoming the issues associated blood hematocrit. Bioanalysis 7(6), 653–659 (2015). [DOI] [PubMed] [Google Scholar]; • Early description of volumetric absorptive microsampling.
- 23.Deniff P, Parry S, Dopson W, Spooner N. Quantitative bioanalysis of paracetamol in rats using volumetric absorptive microsampling (VAMS). J. Pharm. Biomed. 108, 61–69 (2015). [DOI] [PubMed] [Google Scholar]; • Early description of volumetric absorptive microsampling
- 24.Vedar C, Bamat NA, Zuppa AF, Reilly ME, Moorthy GS. Development, validation and implementation of an UHPLC-MS/MS method for the quantiation of furosemide in infant urine samples. Biomed. Chromatogr. 36(3), e5262 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.US FDA. Bioanalytical Method Validation: Guidance for Industry (2018). www.fda.gov/regulatory-information/search-fda-guidance-documents/bioanalytical-method-validation-guidance-industry
- 26.Pacifici GM. Clinical pharmacology of furosemide in neonates: a review. Pharmaceuticals (Basel) 6(9), 1094–1129 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Ponto LL, Schoenwald RD. Furosemide (frusemide). A pharmacokinetic/pharmacodynamic review (part I). Clin. Pharmacokinet. 18(5), 381–408 (1990) [DOI] [PubMed] [Google Scholar]
- 28.Sujan Kumar DP, Palavan C, Seshagiri Rao JVLN. Development and validation of a dried blood sport LC-MS/MS assay to quantify furosemide in human whole blood: an approach to improve the design of PK studies in pediatric population. IOSR-JPBS. 10(2), 65–71 (2015). [Google Scholar]
- 29.Mano Y, Kita K, Kusano K. Hematocrit-independent recovery is a key for bioanalysis using volumetric absorptive microsampling devices, Mitra™. Bioanalysis 7(15), 1821–29 (2015). [DOI] [PubMed] [Google Scholar]