Abstract
The aim of this paper was to investigate the effects of formulation parameters on the physicochemical and pharmacokinetic (PK) behavior of amorphous printlets of lopinavir (LPV) manufactured by selective laser sintering 3D printing method (SLS). The formulation variables investigated were disintegrants (magnesium aluminum silicate at 5–10%, microcrystalline cellulose at 10–20%) and the polymer (Kollicoat® IR at 42–57%), while keeping printing parameters constant. Differential scanning calorimetry, X-ray powder diffraction, and Fourier-transform infrared analysis confirmed the transformation of the crystalline drug into an amorphous form. A direct correlation was found between the disintegrant concentration and dissolution. The dissolved drug ranged from 71.1 ± 5.7% to 99.3 ± 2.7% within 120 min. A comparative PK study in rabbits showed significant differences in the rate and extent of absorption between printlets and compressed tablets. The values for Tmax, Cmax, and AUC were 4 times faster, and 2.5 and 1.7 times higher in the printlets compared to the compressed tablets, respectively. In conclusion, the SLS printing method can be used to create an amorphous delivery system through a single continuous process.
Keywords: amorphous solid dispersion, dissolution, lopinavir, pharmacokinetics, printlets, selective laser sintering
Introduction
Solubility and dissolution are the major hurdle in oral absorption of solid dosage forms [1, 2]. Many drugs fail to reach clinical stage due to this reason even though the molecules have favorable safety and efficacy profiles [3, 4]. Various approaches reported in the literature for solubility and dissolution enhancement are particle size reduction, selection of suitable solid form such as amorphous form, chemical modification, and formulation approaches [5, 6]. Reducing particle size improve solubility and dissolution of poorly soluble drug. This approach is being used in many commercially available products like aprepitant and griseofulvin. However, it may not result in an increase in dissolution of extremely insoluble drugs [7, 8]. Salt formation may improve the solubility and dissolution of the drug [9, 10]. About 50% of drugs in clinical practice uses salt forms [11]. However, salt formation is not always possible especially for very weakly ionizable and neutral molecules. Disproportionation and polymorphism are other issues with salt formation approach [12, 13]. Using amorphous form of the drug is another approach which results in supersaturation, and increase in in vitro and in vivo dissolution rate [14]. Using amorphous drug is common approach to increase solubility, dissolution, and oral bioavailability of poorly soluble drugs. Turning the drug into the amorphous phase requires breaking crystalline lattice order so that molecules can undergo solvation. However, amorphous form is not stable compared to its crystalline counterpart. Amorphous drug can be stabilized by adding polymer to form amorphous solid dispersion (ASD). Various BCS class II and IV drugs are commercially available as ASDs. Notable among them is tacrolimus and itraconazole [15, 16]. Hot-melt extrusion and spray drying are the two of the most commonly used manufacturing methods adopted by the pharmaceutical industry for manufacturing ASDs. The ASD is presented as a tablet or capsule dosage form in multiple steps process after its formation [17, 18].
Lopinavir (LPV) is one of the antiretroviral molecules that acts as a protease inhibitor. It has high specificity towards HIV-1 protease and is used in the treatment of HIV-1 infection. LPV exhibits exceptionally low oral bioavailability around 25% due to extensive metabolism by hepatic oxidation (CYP3A isozyme), poor aqueous solubility, and dissolution [19]. It is often co-administered with other anti-HIV protease inhibitors, primarily in combination with ritonavir due to synergistic behavior when administered together [20]. Ritonavir is a strong inhibitor of CYP3A, which drastically reduces the hepatic oxidization of LPV resulting in increased plasma concentration of LPV, boosting its antiviral effectiveness [21]. Another approach of increasing bioavailability of LPV is converting into the ASD [14].
Selective laser sintering (SLS) is a 3D printing method that is reported for manufacturing various drug delivery systems [22–24]. In addition to infrared heat, it utilizes a low power laser beam to sinter and fuse the powdered ingredients together. Similar to other 3D printing methods that involves heat, it is proven to be capable of changing the solid phase of the drugs [19, 23–25]. Unlike hot-melt extrusion and spray drying methods that are lengthy and complex, ASD can be manufactured in a single step using SLS printing method. Potentially, SLS can be utilized in dispensing of medications at point-of-care facilities such as hospital and clinic [26]. This creates and opportunity for ASD products to be manufactured at point-of-care in an easy and timely manner. The ASD of LPV has not been reported using SLS method to the best of our knowledge. The focus of this work was to understand the effect of formulation variables on the amorphous printlet formulations of LPV manufactured by SLS method and study the effect of amorphous phase transformation on physicochemical and pharmacokinetic properties.
Materials and Methods
Materials
LPV (Ingore Pharmaceutical, China), Candurin® NXT Ruby Red (Merck, Darmstadt, Germany), Kollicoat® IR (BASF, Ludwigshafen, Germany), microcrystalline cellulose (MCC, JRS Pharma, Rosenberg Germany), acetonitrile, glacial acetic acid, methanol, formic acid, sodium dodecyl sulfate (Fisher Scientific, Asheville, NC), and magnesium aluminum silicate (MgAlSi, Spectrum Chemical MFG Corp, California USA) were used. Deuterated LPV (LPV-d5, Toronto Research Chemicals, Ontario, Canada), heparinized rabbit plasma, was purchased from BioChemed Services, Winchester, VA. In-house water (deionized water) obtained from Milli-Q Gradient A-10 (Millipore Corporation, Bedford, MA, USA) water purification system was used in all studies. All reagents were of analytical grade and used as received.
Methods
Printing of Printlets
Four different LPV formulations were prepared by changing formulation variables (MgAlSi, Kollicoat® IR, and MCC) while keeping process variables constant (Table I). The composition of the printlet formulation was as follows: LPV 25%, Candurin® NXT Ruby Red 3%, Kollicoat® IR 42–57%, MgAlSi 5–10%, MCC 10–20%. The ingredients were sieved through a #100 sieve and mixed for 5 min using a V-blender. The powder blend was then loaded into the SLS printer (Sintratec Kit, AG, Brugg, Switzerland) equipped with a 2.3-W blue diode laser (445 nm). 3D models of the printlets with a height of 3 mm and a diameter of 6 mm were created using Solidworks (Solidworks Inc., USA) and printed using the following parameters: chamber and surface temperatures of 80 and 100°C, laser scanning speed of 275 mm/s, and layer thickness of 0.15 mm during the printing process. A physical mixture (PM) and a placebo of F2 were also prepared for comparative physicochemical characterization.
Table I.
Formulation Compositions and Printing Parameters
Formulation | MgAlSi (%) | MCC (%) | Kollicoat® IR (%) | Candurin® NXT Ruby Red (%) | Lopinavir (%) | Surface temp (°C) | Chamber temp (°C) | Laser scanning speed (mm/s) |
---|---|---|---|---|---|---|---|---|
F1 | 5 | 20 | 47 | 3 | 25 | 100 | 80 | 275 |
F2 | 10 | 20 | 42 | 3 | 25 | 100 | 80 | 275 |
F3 | 5 | 10 | 57 | 3 | 25 | 100 | 80 | 275 |
F4 | 10 | 10 | 52 | 3 | 25 | 100 | 80 | 275 |
Preparation of Amorphous Lopinavir
Solid-phase conversion of LPV from crystalline to amorphous was achieved by melt and quench method. Five grams of the crystalline drug was placed on a hotplate until the drug completely melted. The melted drug was immediately quenched in liquid nitrogen to ensure rapid solidification. The sample was immediately placed in a desiccator to prevent water sorption followed by milling in a porcelain mortar and pestle. The yield was more than 90%. The amorphous form of the melted LPV was confirmed by Fourier-transformed infrared (FT-IR), differential scanning calorimetry (DSC), and X-ray powder diffractometer (XRPD).
Weight, Diameter, Thickness, and Disintegration Time
The weight of ten printlets were determined and the average weights were calculated. The diameter and thickness of the printlets were measured using a digital caliper. The disintegration time of six printlets of each formulation was determined in water at 37 ± 1°C using a tablet disintegration apparatus (Vankel Varian VK-100, NC, USA). The time required for complete disintegration of the printlets was recorded.
Hardness
Hardness of the printlets were determined using a texture analyzer (TA.XT Plus, Stable Micro Systems, Surrey, U.K.). Texture analyzer was fitted with a 50-kg load cell, while hardness values were recorded using compression mode with 5-mm compression distance, 5-mm/s pretest speed, 1-mm/s test speed, and 10-mm/s post-test speed. Printlets were placed on the stage radially, and distance between loadcell and stage was set to be 1.5 cm while trigger force was set to be 1 N. Measurements were done in triplicates.
Scanning Electron Microscopy
Surface images of the LPV printlets were taken with a scanning electron microscope (SEM, JSM-7500F, JEOL, Tokyo, Japan) as previously described [26]. Briefly, the printlets were coated with 5-nm thickness of carbon layer with a sputter coater under high vacuum (argon gas pressure 0.01 mbar), and high voltage (40 mV). The morphology of the printlets was captured at 15-mm working distance, 5-kV accelerated voltage, and 20-μA emission current.
Hyperspectroscopy
Cryogenically cooled spectral camera with mercury-cadmium-telluride sensor with resolution of 384 × 288 pixels and pixel size of 24 × 24 μm was used to capture images. Camera was controlled via a camera link interface and the Via-Spec SWIR control systems. The samples were loaded on a metal sample holder and placed on the movable stage of Via-Spec II illuminated with halogen line light source. Pushbroom configuration was used by the instrument to construct a chemical image in which the sample moves underneath an imaging spectrograph to acquire simultaneous spectral measurements from a series of adjacent spatial positions. The distance between the lens and the object was adjusted and focused on the sample to get better image in the reflectance mode. The data were collected at an integration time of 7.005 ms, frame rate of 75 MHz, and scan axis speed of 0.2 in. s−1. Reference images containing pure white and dark images were obtained prior to sample measurement. The data acquisition software used was Middleton Spectral Vision and the data analysis software was Prediktera Evince™ (Prediktera AB, Umea, Sweden) [26].
Vibration Spectroscopy
FTIR spectra of samples were collected using a modular Nicolet™ iS™ 50 system (Thermo Fisher Scientific, Austin, TX). Data collection parameters were absorbance mode, wavenumber range 400–4000 cm−1, data resolution 8 cm−1, and 100 scans. OMNIC software, version 9.0 (Thermo Fisher Scientific, Austin, TX), was used to capture and analyze the spectra.
Differential Scanning Calorimetry
DSC patterns of crystalline and amorphous LPV, placebo, MCC, PM, and printlets were collected using Q2000 instruments (TA Instruments Co., New Castle, DE, USA). Approximately 1–2 mg samples were weighed into sealed aluminum pans. The temperature-scanning rate was 10°C/min and scanned up to 300°C for DSC measurement. Nitrogen gas was purged at a pressure of 20 psi and 50 mL/min flow rate to provide inert atmosphere during the measurement.
X-ray Powder Diffraction
XRPD patterns of the samples were collected using Bruker D2 Phaser SSD 160 Diffractometer (Bruker AXS, Madison, WI) as previously described [27]. Briefly, a powder sample equivalent to 400 mg was evenly spread on the sample holder used for data collection. Samples were scanned over 2θ range of 5–30° with a step size of 0.0343 at 4 s per step (750 total steps). Samples were rotated at 15 rpm/min to get average diffractogram of the sample. The XRPD data collection and analysis were performed using Diffrac.EVA Suite version V4.2.1, and processed using File Exchange 5.0 (Bruker AXS, Madison, WI).
Dissolution
Dissolution of LPV printlets (n = 3) was performed using a USP-II apparatus (Model 708-DS with 850-DS autosampler, Agilent Technologies, CA, USA). Printlets were placed in 900 mL of 0.35% sodium dodecyl sulfate (SDS) solution. The paddle speed was fixed at 75 rpm and dissolution tests were conducted at 37 ± 0.5°C. A volume of 1 mL sample was collected at 5, 10, 15, 30, 60, 90, and 120 min. The percentage of LPV dissolved from the printlets was determined using the HPLC method as described in later section of the manuscript.
Bioavailability Assessment
This study was conducted to compare the pharmacokinetic profile and bioavailability of the printlet and compressed tablet formulations of LPV. The printlet LPV formulation F2 and a compositionally identical formulation of compressed tablets were employed in the study. Four albino New Zealand white rabbits (n = 4, 2 males and 2 females) were utilized for this study, with the rabbits’ weight ranging from 2.5 to 4 kg at the start of the study. Prior to conducting the study, approval from the Institutional Animal Care and Use Committee (IACUC) of Texas A&M University (AUP #2020–0046) was obtained. The printlets or compressed tablets, equivalent to 40 mg LPV, were administered orally to the animals using a pill popper, following a parallel study design with a 21-day washout period between the two studies. Blood samples (1 mL) were collected via the ear vein at 0, 0.25, 0.5, 1, 2, 3, 4, 6, 8, 12, 24, 36, 48, and 72 h postdosing. The collected blood was placed in heparinized tubes. Plasma was obtained by centrifuging the tubes at 13,300 rpm at 4°C for 15 min and then stored at − 80°C until the time of analysis.
Plasma samples were extracted using the protein precipitation method with methanol. In an Eppendorf tube, 300 μL of plasma was mixed with 50 μL of the internal standard (LPV-d5, 500 ng/mL), 850 μL of methanol, and 100 μL of 0.1% formic acid in water. The samples were vortexed for 1 min and then centrifuged at 13,300 rpm at 4°C for 15 min.
High-Performance Liquid Chromatography
High-performance liquid chromatography (HPLC) method was developed to quantify the drug in the samples. HPLC equipment was (HP 1260 series, Agilent Technologies, Wilmington, DE, USA) equipped with a quaternary pump, autosampler, and UV/Vis detector. A mobile phase consisting of acetonitrile and 0.05 M (pH 3) glacial acetic acid (65:35) was pumped at a flow rate of 1 mL/min through Econosphere 5 μm C18 column 250 × 4.6 mm (Alltech, Deerfield, IL, USA). The volume of the sample injection was 45 μL. The eluent was detected at a wavelength of 210 nm and retention time for LPV was 6.3 min. Data were collected and analyzed using OpenLab software (Agilent Technologies, Wilmington, DE, US).
Ultra-Performance Liquid Chromatography-Mass Spectrometry
A validated ultra-high performance liquid chromatography (UPLC) combined with mass spectrometry (MS) method was used for quantitative determination of LPV in the plasma. The method was validated as per FDA’s bioanalytical method validation guidance [28]. UPLC analysis was performed on a Waters Acquity® UPLC system (Waters Corporation, Milford, MA, USA) equipped with quaternary pump, PDA, and QDa detectors. Separation of plasma samples was performed using Poroshell 120 EC-C18 (4.6 × 50 mm, 2.7 μm) (Agilent, Santa Clara, CA) maintained at 45 °C with mobile phase 0.1% formic acid in water and acetonitrile (70:30 V/V) flowing at 0.5 mL/min. Injection volume of 45 μL was injected with 3-min run time. Electrospray positive ionization (ESI +) mode with 0.8 kV capillary voltage and 15 V collision energy were used to detect LPV and LPV-d5 at 629.3 and 637.3 Dalton, respectively. The retention times of LPV and LPV-d5 were 2.01 and 1.992 min, respectively, and their peaks were well separated from the baseline. Calibration curve was linear over 2.5 to 250 ng/ml range.
Statistics
Data were presented as mean and standard deviation (mean ± SD). Pharmacokinetic parameters such as area under the curve (AUC), time to reach maximum plasma concentration (Tmax), and maximum plasma concentration (Cmax) were calculated using PKSolver add-in for Excel software [29]. Non-compartmental extravascular model was used, and AUC was calculated using linear trapezoidal method. For bioequivalence calculations, geometric mean of logarithmically transformed Cmax and AUC, and 90% confidence interval (CI) were calculated using Excel software. In vivo drug dissolved (% drug absorbed) were calculated by Wagner-Nelson method [30]. Statistical significance between means were determined by Student’s t-test.
Results and Disscussion
Printing of Formulations
During the printing process, the application of infrared and laser energy to the powder mixture can induce drug dissolution and/or melting in a polymer matrix [31]. The polymer matrix decreases the conversion of amorphous drugs into thermodynamically stable crystalline forms by elevating the drug’s glass transition temperature [32]. In our previous publication, we reported a melting point of 124°C for LPV [14]. To ensure the complete melting of the drug, surface and chamber temperatures of 100 °C and 80 °C, respectively, along with a laser speed of 275 mm/s (laser energy) were selected. These parameters were chosen while maintaining a constant drug percentage in all formulations (25%). The laser scanning also imparts energy to melt the drug and subsequently entrap it in the polymer matrix, forming an ASD [19]. A laser speed of 275 mm/s was selected to print printlets of good quality with sufficient mechanical strength, without subjecting them to excessive laser energy. For the polymer, Kollicoat® IR was chosen due to its fast disintegration, flowability properties, spherical particle shape and size distribution, and pH-independent release profile [19, 23, 33]. The concentration of Kollicoat® IR varied from 42 to 57%. MgAlSi was added to enhance disintegration alongside MCC; it was found to improve disintegration times. Additionally, Candurin® NXT Ruby Red was added to all formulations at 3% (w/w) to enhance the absorption of laser energy, facilitating the sintering of the otherwise white-colored powders [23].
Diameter, Height, Weight, Hardness, Disintegration Time, and Assay
The tablets exhibited diameters, heights, and weights ranging from 5.65 to 5.79 mm, 3.58 to 3.79 mm, and 38.92 to 41.43 mg, respectively (Table II). Weight variations between formulations were < 6.5%, which can be attributed to minor differences in the printlet dimensions. These differences in printlet size arise from the selection of printlets from different positions on the printing stage. Notably, the printlets closer to the powder reservoir tended to have larger dimensions than those situated in the middle of the printing area. Disintegration time for the printlets demonstrated a direct correlation with disintegrant concentration. Among the formulations, F2 (103 ± 7 s) and F3 (147 ± 12 s) exhibited the fastest and slowest disintegration times, respectively. Hardness displayed an inverse correlation with disintegrant concentration. This could be elucidated by the varying polymer concentrations in the printlets. F3, with the highest polymer concentration (57%) due to lower disintegrant concentration in the formulation, exhibited the highest hardness. Conversely, F2, with high disintegrant concentration and the lowest polymer concentration (42%), displayed the lowest mechanical strength. Finally, the drug content in the printlets ranged from 93.6 to 96%. Furthermore, the disintegration time data demonstrated statistically significant (p < 0.05) differences between F1 and 2, as well as F3 and F4 (p < 0.05), while no statistically significant differences (p > 0.05) were observed for diameter, height, weight, hardness, and assay data.
Table II.
Physical Properties of All Formulations. Diameter, Height, Weight, Disintegration time, Harness, and Assay Values. Data Presented as Mean ± Standard Deviation, n ≥ 3
Diameter (mm) | Height (mm) | Weight (mg) | Disintegration Time (s) | Hardness (N) | Assay (%) | |
---|---|---|---|---|---|---|
F1 | 5.8 ± 0.2 | 3.6 ± 0.1 | 39.0 ± 1.9 | 113 ± 5 | 6.1 ± 0.1 | 94.0 ± 4.2 |
F2 | 5.6 ± 0.2 | 3.6 ± 0.1 | 38.9 ± 2.8 | 103 ± 7 | 5.8 ± 0.1 | 96.0 + 3.8 |
F3 | 5.7 ± 0.1 | 3.8 ± 0.2 | 41.4 ± 2.5 | 147 ± 12 | 7.3 ± 0.1 | 93.6 ± 5.1 |
F4 | 5.7 ± 0.2 | 3.6 ± 0.1 | 40.9 ± 1.2 | 134 ± 17 | 7.1 ± 0.1 | 92.5 ± 6.8 |
Scanning Electron Microscopy
In the powder mixture, the individual components of the formulation can be identified based on our previous paper [19]. LPV appeared as irregularly shaped crystals, Kollicoat® IR as near-perfect spherical particles, Candurin® NXT Ruby Red as thin flake-shaped particles adhering to the surface of Kollicoat® IR particles, and MgAlSi particles recognized as rectangular-shaped particles. Candurin® NXT Ruby Red and MgAlSi can be differentiated by their particle size. The particle size range of Candurin® NXT Ruby Red and MgAlSi was approximately 20 μm and 130 μm, respectively (Fig. 1). The flake-like particles adhering to the polymer were Candurin® NXT Ruby Red due to its small size. MCC was observed as long rod-shaped particles. The printlets’ surfaces exhibited slightly deformed Kollicoat® IR particles that had fused with the formulation components. The deformation of Kollicoat® IR was a result of the polymer softening due to laser scanning. Furthermore, Candurin® NXT Ruby Red, MCC, and MgAlSi particles can be identified within the melted matrix. However, the drug could not be identified, possibly due to the melting of LPV and its subsequent fusion with the polymer to form ASD, owing to the drug’s lower melting point compared to the polymer. Candurin® NXT Ruby Red, MCC, and MgAlSi did not melt and did not undergo significant deformation under the process conditions. Nonetheless, varying degrees of melting/fusion were observed among formulations F3 and F2, as evident from the surface morphology and pore size of the printlets. This could possibly be attributed to the different polymer and disintegrant concentrations between each formulation, resulting in these distinctions. The higher polymer concentration in the F3 formulation led to denser printlets with smaller pores compared to the F2 formulation.
Fig. 1.
SEM images for Printlets F1 (a), F2 (b), F3 (c), and PM (d)
Near-Infrared Hyperspectral Imaging
Hyperspectral imaging, also known as chemical imaging, is used for the qualitative and quantitative estimation of formulation components or their changes during manufacturing. These measurements are nondestructive, in contrast to other analytical methods where tablets are dissolved in a suitable dissolution medium before analysis. The qualitative spatial distribution of individual components in the samples was demonstrated using near-infrared hyperspectral imaging. The hyperspectral image data were mathematically treated with mean centering and multiplicative scatter correction to generate principal component analysis (PCA) concentration images [34]. Crystalline LPV can be visualized as orangish-yellow color pixels, while Kollicoat® IR appeared as light blue pixels. Additionally, MCC appeared as dark blue color pixels (Fig. 2). The color distribution of pixels in the physical mixture was noticeably different from that of the placebo sample. The placebo formulation had blue pixels, whereas the physical mixture sample had lighter teal-colored pixels interspersed with yellow pixels. The yellow pixels represented the drug, which was absent in the placebo sample. The difference in color pixels between pure LPV and LPV in the printlets and physical mixture could be attributed to the dilution of LPV in the PM and printlet samples with other components. The printlet exhibited yellow pixels similar to the PM, indicating the presence of LPV. This observation further suggested that during printing and powder spreading steps, the individual components of the physical mixture were deposited homogeneously on the powder bed, without any segregation of components. The distribution of LPV in the printlet followed a pattern similar to the LPV distribution in the physical mixture. However, a significant difference existed between the two samples. The color pixels representing the LPV distribution in the physical mixture were more localized and had well-defined, brighter yellow areas compared to the printlet sample. The printlet sample had wider, less bright yellow spots that covered a larger area on the printlet’s surface. This difference could be explained by the melting of LPV during the printing process to create an ASD through fusion and intimate mixing with the polymer and other formulation components. The melting of LPV caused localized pockets of LPV to dissolve into the matrix, reducing its concentration in the area and resulting in wider, less bright yellow pixels. To confirm this hypothesis, additional testings were performed using X-ray diffraction and differential scanning calorimetry that support this hypothesis (described in subsequent sections).
Fig. 2.
Near-infrared hyperspectral images generated using PCA model. Printlet F2 (a), PM (b), placebo (c), Kollicoat IR (d), Crystalline LPV (e), and microcrystalline cellulose (f)
Fourier Transformed Infrared
The spectrum of the crystalline LPV was compared to the amorphous LPV sample (Fig. 3). Crystalline LPV showed major bands at 3395 (N–H stretching), 3030–3069 (CH– aromatic), 2962 (aliphatic-CH stretching), 1735 (C = O), 1656 (C = O stretching, present in the amide group), 1507–1610 (C–C aromatic), 1443 (C = C stretching, aromatic), 1345 (C–N), 1302 and 1164 (C–O–C, asymmetrical and symmetrical stretching, respectively), and 395 cm−1 (N–H stretching) [35, 36]. PM showed additive spectrum encompassing bands of crystalline drug and excipients. Bands characteristic to the crystalline drug have been observed at 3395 and 1735 cm−1. This bands observed to be of lower intensity in PM formulation. Reduction of intensity was due to the dilution of crystalline drug with other formulation components. In addition to changes in bands intensity, the amorphous LPV spectrum showed broader bands in the 3100–3500 cm−1 region, lower intensity bands at 1735 cm−1, and 1580–1712 cm−1 region, and bands the 1420–1580 cm−1 regions to merge into a single peak at 1500 cm−1. These observations can be related to transformation of crystalline LPV to amorphous form of the drug [37]. Furthermore, peak at 1735 cm−1 disappeared in the printlets which may be attributed crystalline to amorphous transformation of the drug during printing.
Fig. 3.
FT-IR spectra of the crystalline and amorphous LPV, printlets, PM, placebo, and Kollicoat IR
Differential Scanning Calorimetry
The crystalline LPV showed two endothermic peaks at 117.3 and 122.5 °C, which represent the melting points of type IV non-solvated crystal forms [38]. These peaks completely disappeared in the amorphous LPV sample further confirming crystalline to amorphous transformation of the drug (Fig. 4). Kollicoat® IR exhibited a broad melting peak at 214.5 °C. Placebo exhibited only one endothermic peak around 214 °C that matched with Kollicoat® IR. This indicated that other components were either amorphous or did not melt. On the other hand, PM showed a broad melting endothermic peak of LPV and Kollicoat® IR at 117.7 and 214.3 °C, respectively. In contrast to PM, the printlets did not show melting peak of the drug that suggested the drug conversion to the amorphous form during the printing process. LPV transformation in the printlets can be attributed to melting and subsequent solubilization within the polymer, and leading to drug and polymer intimate mixing [23].
Fig. 4.
DSC thermogram of the crystalline and amorphous LPV, printlets, PM, placebo, and Kollicoat IR
X-ray Powder Diffractometry
XRPD patterns for the crystalline and amorphous forms of LPV are shown in Fig. 5. The crystalline form of LPV exhibited characteristic reflection peaks at 2θ of 7.3, 12, 12.6, 14.8, 16.4, 17.7, 18.5, 19.3, 20.5, and 21.7°. Placebo and PM exhibited additive diffractograms of individual components present in their composition. The XRPD pattern of placebo formulation exhibited a broad characteristic reflection peak at 2θ at 19.2° due to Kollicoat® IR [39]. MCC showed a reflection peak at 2θ of 15.5 and 22.1°, respectively, that matched with the previous findings [40, 41]. PM exhibited characteristic low-intensity reflection peaks of the drug due to dilution with the excipients. In addition to drug peaks, diffractogram of PM also showed broad, low-intensity characteristics halo reflection peaks of MCC and Kollicoat® IR at 22.2 and 19.3°, respectively. On the other hand, the characteristic reflection peaks of the drug completely disappeared in the printlets’ diffractograms and were similar to placebo formulation. It showed characteristics halo diffractogram of amorphous material further confirming the conversion of crystalline to amorphous LPV during printlets. The printlets showed varying intensities peak around 19.2°. Intensity of these peaks were changing between printlets due to varying Kollicoat® IR percentage. F3 had the highest intensity reflection peak at 19.2° followed by F4 and F1 while F2 had the lowest. Differences in peak intensities can be explained by the difference between concentration of polymer across formulations. F3 had the highest polymer percentage while F2 has the lowest, which correlated with the peak intensities observed. The DSC and FTIR data concurred with the XRPD results, which confirmed the crystalline conversion of LPV into amorphous form during printing.
Fig. 5.
XRPD diffractogram of the crystalline LPV, printlets, PM, and placebo
Dissolution
The recommended dissolution medium for LPV and ritonavir combination tablets and soft gelatin capsules, as per USP and FDA guidelines, is 60 mM polyoxyethylene 10 lauryl ether in water and 0.05 M polyoxyethylene 10 lauryl ether in phosphate buffer pH 6.8, respectively [42, 43]. Similarly, literature reports dissolution methods using sodium dodecyl sulfate in concentrations of 0.5 to 2.5% to dissolve LPV in Kaletra soft gel capsules [44, 45]. In our previously published work, we used a 0.5% sodium dodecyl sulfate concentration in the dissolution medium [19]. In this study, a 0.35% SDS concentration was employed to further enhance the discriminatory power of the method, thereby enabling differentiation between the formulations. The dissolution tolerance for LPV, as per USP, is not less than 85% in 90 min. The dissolved LPV after 2 h for F1, F2, F3, and F4 were 90.4 ± 3.8, 99.3 ± 2.7, 71.1 ± 5.7, and 85.4 ± 1.7%, respectively (Fig. 6). The only formulation that met the 85% dissolved drug limit in 90 min was F2 (99.2 ± 2.7%).
Fig. 6.
Dissolution profiles of the LPV-loaded printlets
As the process parameters were maintained constant for all formulations, differences in the dissolution profiles among the formulations are attributed to the formulation variables, namely MCC and MgAlSi, as well as the percentage of Kollicoat® IR. Both of these excipients function as disintegrants within the concentration ranges employed in the formulations [46, 47]. Disintegration times for the formulations were correlated with dissolution and can help explain the distinctions between the dissolution profiles. Formulations with the highest total concentration of disintegrants and the lowest polymer content (F2) exhibit the lowest disintegration time and the fastest dissolution profile among the formulations. Conversely, the F3 formulation with the lowest total disintegrant concentration and the highest polymer content displays the lowest extent of dissolution. During dissolution, it was observed that formulations with low disintegrant and high polymer concentrations (F3 and F4) did not fully disintegrate until around 60 and 90 min, respectively, into the dissolution process. These formulations remained partially intact in the dissolution medium before eventually disintegrating. This behavior can be explained by the lower disintegrant and higher polymer concentrations in these formulations, which elucidates the differences in the dissolution profiles.
To further assess the impact of composition on dissolution, the dissolution efficiency (DE) was calculated. DE represents the correlation between the area under the dissolution curve between specific time points, normalized by the total dissolution. The area under the dissolution curve can be calculated using the model-independent trapezoid method. The equations used for calculating DE and the corresponding values are provided as follows:
where t is the time point, yt is the percent dissolution at time point t, and y100 is the maximum dissolution value.
The DE values for formulations F1, F2, F3, and F4 were 75.2%, 94.7%, 54.6%, and 64.1%, respectively, which correlated with the extent of dissolution after 2 h. The effect of each disintegrant on dissolution was calculated by dividing the difference of DE values of selected formulations with each other. The effect of MgAlSi was calculated by comparing formulations with varying MgAlSi and constant MCC concentrations (F3 vs F4 and F1 vs F2), and the effect of MCC was calculated similarly by comparing constant MgAlSi with varying MCC concentrations (F3 vs F1 and F4 vs F2). This comparison provides insight into the change in DE with respect to the each individual disintegrant. Formulations F3 and F4 both contained 10% MCC along with 5% and 10% MgAlSi, respectively; when comparing the DE of F3 and F4, a difference of 14.9% between DE of F3 vs F4 was observed. This suggests that a 5% change in MgAlSi concentration, when 10% MCC was present, resulted in 14.9% difference in dissolution. When comparing F1 and F2 (both with 20% MCC and, 5% and 10% MgAlSi, respectively), the difference between dissolution efficiencies appears to be even greater at 20.6%. This can be interpreted as the effect of MgAlSi on dissolution with higher MCC concentration was lower. Suggesting MCC has bigger impact on dissolution than MgAlSi when both are present. Finally, higher total disintegrant concentration leads to a faster disintegration time and greater dissolution.
Bioavailability Assessment
Pharmacokinetic profiles of the printlets (test product (T)) and compressed tablets (reference product (R)) of LPV were not superimposable. Major differences in pharmacokinetic parameters AUC, Tmax, and Cmax were observed between two formulations. Tmax, Cmax, and AUC of the printlets and the tablets were 1.0 ± 0.0 and 4.0 ± 0.0 h, 124.5 ± 3.7 and 49.3 ± 17.6 ng/ml, and 392.6 ± 151.3 and 225.3 ± 83 ng/ml·h, respectively (Fig. 7). The difference between Tmax and Cmax of two formulation were statistically significant (p < 0.05) while AUC difference between two treatment was not statistically significant due to higher values of standard deviation in the data (p > 0.05). Differences in pharmacokinetic parameters between two formulations can be explained by solid phase present in the tablets and the printlets. The printlets contained amorphous form of the drug while compressed tablets contained crystalline form of the drug. Amorphous form of LPV is previously shown to have more favorable physicochemical properties especially dissolution [19]. The printlets would increase in vivo dissolution and maintain supersaturation that results in faster absorption as indicated by shorter Tmax and higher absorption as reflected by Cmax and AUC compared to compressed tablets. Tmax, Cmax, and AUC were fourfold faster, and 2.5- and 1.7-fold higher in the printlets compared to the compressed tablets. Thus, the printlets is 1.7–2.5 times more bioavailable than the compressed tablets. Further, in vivo drug dissolved (absorbed) was calculated by Wagner-Nelson deconvolution method. Calculated in vivo dissolved drug was higher in the printlet compared to the compressed tablets (Fig. 8). To further prove that these two formulations were not bioequivalent, the pharmacokinetics parameters were logarithmically transformed followed by calculating geometric mean and 90% CI. 90% CI of R/T and T/R of AUC were 0.89–0.92 and 1.08–1.13, respectively. Similarly, 90% CI of R/T and T/R of Cmax were 0.71–0.89 and 1.10–1.39, respectively. T/R ratios of geometric mean of Cmax and AUC were 1.245 and 1.104, respectively. Similarly, R/T ratios of geometric mean of Cmax and AUC were 0.803 and 0.905, respectively. Two formulations are considered to be bioequivalent when 90% CI of both Cmax and AUC are within 90% CI. Cmax did not meet 90% CI criteria and, thus, two formulations were not bioequivalent as per FDA guideline [48].
Fig. 7.
Blood plasma levels of 3D printed PV formulation (F2) vs compressed LPV formulation
Fig. 8.
Estimated in vivo dissolved drug of 3D printed PV formulation (F2) vs compressed LPV formulation
Conclusion
Amorphous 3D printed LPV printlets were developed in a single step using SLS 3D printing method. Formulations’ variables have significant impact on physicochemical attributes such as hardness, disintegration time, and dissolution. Different disintegrant type and amounts were found to have an effect on the dissolution profiles of the printlets by effecting the disintegration behavior. As expected, reverse trend was observed for disintegration time and dissolution. Dissolution efficiency of printlets was calculated and differences were compared to each other. Spectral, thermal, and diffractogram data suggested crystalline drug transformed into amorphous phase. Due to amorphous phase of the drug being present in the printlets, bioavailability of the LPV was higher in the printlets compared to compressed tablets’ formulation. The drug was 1.7–2.5 times more bioavailable from the printlets compared to the compressed tablets. Thus the SLS method was used to prepare Lopinavir prinlets that had improvement in the rate and extent of bioavailability as compared to the compressed tablets.
Acknowledgements
We acknowledge the TAMU Materials Characterization Facility for allowing us to use SEM.
Funding
This work was partially supported by the National Institute of Health R56 grant #1R56HD106612 and 1R01HD112077.
Footnotes
Conflict of Interest The authors declare no competing interests.
References
- 1.Wu C-j, You J-z, Wang X-j. Thermal decomposition mechanism of tenofovir disoproxil fumarate. J Therm Anal Calorim. 2018;132(1):471–82. 10.1007/s10973-017-6910-3. [DOI] [Google Scholar]
- 2.Kesisoglou F, Wu Y. Understanding the effect of API properties on bioavailability through absorption modeling. Aaps j. 2008;10(4):516–25. 10.1208/s12248-008-9061-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Savjani KT, Gajjar AK, Savjani JK. Drug solubility: importance and enhancement techniques. ISRN Pharm. 2012;2012: 195727. 10.5402/2012/195727. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hughes JP, Rees S, Kalindjian SB, Philpott KL. Principles of early drug discovery. Br J Pharmacol. 2011;162(6):1239–49. 10.1111/j.1476-5381.2010.01127.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Jain S, Patel N, Lin S. Solubility and dissolution enhancement strategies: current understanding and recent trends. Drug Dev Ind Pharm. 2015;41(6):875–87. 10.3109/03639045.2014.971027. [DOI] [PubMed] [Google Scholar]
- 6.Sareen S, Mathew G, Joseph L. Improvement in solubility of poor water-soluble drugs by solid dispersion. Int J Pharm Investig. 2012;2(1):12–7. 10.4103/2230-973x.96921. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Zhou Y, Du J, Wang L, Wang Y. Nanocrystals technology for improving bioavailability of poorly soluble drugs: a minireview. J Nanosci Nanotechnol. 2017;17(1):18–28. 10.1166/jnn.2017.13108. [DOI] [PubMed] [Google Scholar]
- 8.Chogale MM, Ghodake VN, Patravale VB. Performance parameters and characterizations of nanocrystals: a brief review. Pharmaceutics. 2016;8(3). 10.3390/pharmaceutics8030026. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Hamideh A, Rahman Z, Dharani S, Khuroo T, Mohamed EM, Nutan MTH, et al. Preparation and characterization of dicarboxylic acids salt of aripiprazole with enhanced physicochemical properties. Pharm Dev Technol. 2021;26(4):455–63. 10.1080/10837450.2021.1888978. [DOI] [PubMed] [Google Scholar]
- 10.Afrooz H, Mohamed EM, Barakh Ali SF, Dharani S, Nutan MTH, Khan MA, et al. Salt engineering of aripiprazole with polycarboxylic acids to improve physicochemical properties. AAPS PharmSciTech. 2021;22(1):31. 10.1208/s12249-020-01875-x. [DOI] [PubMed] [Google Scholar]
- 11.Gupta D, Bhatia D, Dave V, Sutariya V, Varghese Gupta S. Salts of therapeutic agents: chemical, physicochemical, and biological considerations. Molecules. 2018;23(7). 10.3390/molecules23071719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Rahman Z, Dharani S, Barakh Ali SF, Afrooz H, Reddy IK, Khan MA. Effect of processing parameters and controlled environment storage on the disproportionation and dissolution of extended-release capsule of phenytoin sodium. Int J Pharm. 2018;550(1–2):290–9. 10.1016/j.ijpharm.2018.07.042. [DOI] [PubMed] [Google Scholar]
- 13.Dharani S, Rahman Z, Barakh Ali SF, Afrooz H, Khan MA. Quantitative estimation of phenytoin sodium disproportionation in the formulations using vibration spectroscopies and multivariate methodologies. Int J Pharm. 2018;539(1–2):65–74. 10.1016/j.ijpharm.2018.01.005. [DOI] [PubMed] [Google Scholar]
- 14.Hamed R, Mohamed EM, Sediri K, Khan MA, Rahman Z. Development of stable amorphous solid dispersion and quantification of crystalline fraction of lopinavir by spectroscopic-chemometric methods. Int J Pharm. 2021;602: 120657. 10.1016/j.ijpharm.2021.120657. [DOI] [PubMed] [Google Scholar]
- 15.Bhujbal SV, Mitra B, Jain U, Gong Y, Agrawal A, Karki S, et al. Pharmaceutical amorphous solid dispersion: a review of manufacturing strategies. Acta Pharmaceutica Sinica B. 2021;11(8):2505–36. 10.1016/j.apsb.2021.05.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Zidan AS, Rahman Z, Sayeed V, Raw A, Yu L, Khan MA. Crystallinity evaluation of tacrolimus solid dispersions by chemometric analysis. Int J Pharm. 2012;423(2):341–50. 10.1016/j.ijpharm.2011.11.003. [DOI] [PubMed] [Google Scholar]
- 17.Chivate A, Garkal A, Dhas N, Mehta T. Hot-melt extrusion: an emerging technique for solubility enhancement of poorly water-soluble drugs. PDA J Pharm Sci Technol. 2021;75(4):357–73. 10.5731/pdajpst.2019.011403. [DOI] [PubMed] [Google Scholar]
- 18.Baumann JM, Adam MS, Wood JD. Engineering advances in spray drying for pharmaceuticals. Annu Rev Chem Biomol Eng. 2021;12:217–40. 10.1146/annurev-chembioeng-091720-034106. [DOI] [PubMed] [Google Scholar]
- 19.Hamed R, Mohamed EM, Rahman Z, Khan MA. 3D-printing of lopinavir printlets by selective laser sintering and quantification of crystalline fraction by XRPD-chemometric models. Int J Pharm. 2021;592: 120059. 10.1016/j.ijpharm.2020.120059. [DOI] [PubMed] [Google Scholar]
- 20.Sham HL, Kempf DJ, Molla A, Marsh KC, Kumar GN, Chen C-M, et al. ABT-378, a highly potent inhibitor of the human immunodeficiency virus protease. Antimicrob Agents Chemother. 1998;42(12):3218–24. 10.1128/AAC.42.12.3218. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Mangum EM, Graham KK. Lopinavir-ritonavir: a new protease inhibitor. Pharmacotherapy. 2001;21(11):1352–63. 10.1592/phco.21.17.1352.34419. [DOI] [PubMed] [Google Scholar]
- 22.Jamróz W, Kurek M, Łyszczarz E, Szafraniec J, Knapik-Kowalczuk J, Syrek K, et al. 3D printed orodispersible films with Aripiprazole. Int J Pharm. 2017;533(2):413–20. 10.1016/j.ijpharm.2017.05.052. [DOI] [PubMed] [Google Scholar]
- 23.Awad A, Fina F, Trenfield SJ, Patel P, Goyanes A, Gaisford S, et al. 3D printed pellets (miniprintlets): a novel, multi-drug, controlled release platform technology. Pharmaceutics. 2019;11(4). 10.3390/pharmaceutics11040148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Mohamed EM, Barakh Ali SF, Rahman Z, Dharani S, Ozkan T, Kuttolamadom MA, et al. Formulation optimization of selective laser sintering 3D-printed tablets of clindamycin palmitate hydrochloride by response surface methodology. AAPS PharmSciTech. 2020;21(6):232. 10.1208/s12249-020-01775-0. [DOI] [PubMed] [Google Scholar]
- 25.Trenfield SJ, Goyanes A, Telford R, Wilsdon D, Rowland M, Gaisford S, et al. 3D printed drug products: non-destructive dose verification using a rapid point-and-shoot approach. Int J Pharm. 2018;549(1–2):283–92. 10.1016/j.ijpharm.2018.08.002. [DOI] [PubMed] [Google Scholar]
- 26.Kayalar C, Rahman Z, Mohamed EM, Dharani S, Khuroo T, Helal N, Kuttolamadom MA, Khan MA, et al. Preparation and characterization of 3D-printed dose-flexible printlets of tenofovir disoproxil fumarate. AAPS PharmSciTech. 2023;24(6):171. 10.1208/s12249-023-02623-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Barakh Ali SF, Mohamed EM, Ozkan T, Kuttolamadom MA, Khan MA, Asadi A, et al. Understanding the effects of formulation and process variables on the printlets quality manufactured by selective laser sintering 3D printing. Int J Pharm. 2019;570: 118651. 10.1016/j.ijpharm.2019.118651. [DOI] [PubMed] [Google Scholar]
- 28.FDA: Guidance for industry – Bioanalytical method validation. https://www.fda.gov/files/drugs/published/Bioanalytical-Method-Validation-Guidance-for-Industry.pdf (2018). Accessed Oct 22 2022.
- 29.Zhang Y, Huo M, Zhou J, Shaofei X. PKSolver: An add-in program for pharmacokinetic and pharmacodynamic data analysis in Microsoft Excel. Computer methods and programs in biomedicine. 2010;1;99(3):306–14. 10.1016/j.cmpb.2010.01.007 [DOI] [PubMed] [Google Scholar]
- 30.Wagner JG. Application of the Wagner-Nelson absorption method to the two-compartment open model. J Pharmacokinet Biopharm. 1974;2(6):469–86. [DOI] [PubMed] [Google Scholar]
- 31.Baghel S, Cathcart H, O’Reilly NJ. Polymeric amorphous solid dispersions: a review of amorphization, crystallization, stabilization, solid-state characterization, and aqueous solubilization of biopharmaceutical classification system class II drugs. J Pharm Sci. 2016;105(9):2527–44. 10.1016/j.xphs.2015.10.008. [DOI] [PubMed] [Google Scholar]
- 32.Browne E, Worku ZA, Healy AM. Physicochemical properties of poly-vinyl polymers and their influence on ketoprofen amorphous solid dispersion performance: a polymer selection case study. Pharmaceutics. 2020;12(5). 10.3390/pharmaceutics12050433. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Fina F, Goyanes A, Gaisford S, Basit AW. Selective laser sintering (SLS) 3D printing of medicines. Int J Pharm. 2017;529(1–2):285–93. 10.1016/j.ijpharm.2017.06.082. [DOI] [PubMed] [Google Scholar]
- 34.Dharani S, Barakh Ali SF, Afrooz H, Mohamed EM, Cook P, Khan MA, et al. Development of methamphetamine abuse-deterrent formulations using sucrose acetate isobutyrate. J Pharm Sci. 2020;109(3):1338–46. 10.1016/j.xphs.2019.12.003. [DOI] [PubMed] [Google Scholar]
- 35.Machado J, Baleanu D, Al-Zhrani AA, Alhamed YA, Zahid AH, Youssef TE. On similarities in infrared spectra of complex drugs. Romanian reports in Physics. 2014:382–93. [Google Scholar]
- 36.Patel G, Shelat P, Lalwani A. Statistical modeling, optimization and characterization of solid self-nanoemulsifying drug delivery system of lopinavir using design of experiment. Drug Deliv. 2016;23(8):3027–42. 10.3109/10717544.2016.1141260. [DOI] [PubMed] [Google Scholar]
- 37.Du S, Li WS, Wu YR, Fu Y, Yang C, Wang J. Comparison of the physical and thermodynamic stability of amorphous azelnidipine and its coamorphous phase with piperazine. RSC Adv. 2018;8(57):32756–64. 10.1039/C8RA05535A. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Reddy BP, Reddy KR, Reddy RR, Reddy DM, Reddy KSC. Polymorphs of lopinavir. Google Patents. 2013. [Google Scholar]
- 39.El-Badry M, Alanazi FK, Mahrous GM, Alsarra IA. Effects of Kollicoat IR® and hydroxypropyl-β-cyclodextrin on the dissolution rate of omeprazole from its microparticles and enteric-coated capsules. Pharm Dev Technol. 2010;15(5):500–10. 10.3109/10837450903300171. [DOI] [PubMed] [Google Scholar]
- 40.Das K, Ray D, Bandyopadhyay N, Sengupta S. Study of the properties of microcrystalline cellulose particles from different renewable resources by XRD, FTIR, nanoindentation, TGA and SEM. J Polym Environ. 2010;18(3):355–63. [Google Scholar]
- 41.Ramírez B Microcrystalline cellulose (MCC) analysis and quantitative phase analysis of ciprofloxacin/MCC mixtures by Rietveld XRD refinement with physically based background. Cellulose. 2018; v. 25(no. 5):pp. 2795–815–018 v.25 no.5. 10.1007/s10570-018-1761-z. [DOI] [Google Scholar]
- 42.USP43-NF38: lopinavir and ritonavir tablet monograph. Accessed 2022.
- 43.FDA: Kaletra (lopinavir/ritonavir) for oral use https://www.accessdata.fda.gov/drugsatfda_docs/label/2019/021251s058,021906s053lbl.pdf. Accessed Oct 22 2022.
- 44.Donato EM, Martins LA, Fröehlich PE, Bergold AM. Development and validation of dissolution test for lopinavir, a poorly water-soluble drug, in soft gel capsules, based on in vivo data. J Pharm Biomed Anal. 2008;47(3):547–52. [DOI] [PubMed] [Google Scholar]
- 45.Madgulkar AR, Bhalekar MR, Kadam AA. Improvement of oral bioavailability of lopinavir without co-administration of ritonavir using microspheres of thiolated xyloglucan. AAPS PharmSciTech. 2018;19(1):293–302. 10.1208/s12249-017-0834-x. [DOI] [PubMed] [Google Scholar]
- 46.Saigal N, Baboota S, Ahuja A, Ali J. Microcrystalline cellulose as a versatile excipient in drug research. J Young Pharmacists. 2009. p. 6. [Google Scholar]
- 47.Bhargava H, Shah D, Anaebonam A, Oza B. An evaluation of Smecta as a tablet disintegrant and dissolution aid. Drug Dev Ind Pharm. 1991;17(15):2093–102. [Google Scholar]
- 48.FDA: Guidance document- bioequivalence studies with pharmacokinetic endpoints for drugs submitted under an ANDA Guidance for industry. (2021). Accessed May 31 2022.