International Journal of Drug Development and Research

  • ISSN: 0975-9344
  • Journal h-index: 38
  • Journal CiteScore: 46.78
  • Journal Impact Factor: 1.95
  • Average acceptance to publication time (5-7 days)
  • Average article processing time (30-45 days) Less than 5 volumes 30 days
    8 - 9 volumes 40 days
    10 and more volumes 45 days
20+ Million Readerbase
Indexed In
  • Genamics JournalSeek
  • China National Knowledge Infrastructure (CNKI)
  • CiteFactor
  • Scimago
  • Directory of Research Journal Indexing (DRJI)
  • OCLC- WorldCat
  • Publons
  • MIAR
  • University Grants Commission
  • Euro Pub
  • Google Scholar
  • Secret Search Engine Labs
  • ResearchGate
Share This Page

Research Article - (2022) Volume 14, Issue 11

Exposing Profound Screening Potential of Ethanol-Based Dissolution Media in the Development of Oral-Modified Dosage Forms

Berginc Katja*, Huzjak Tilen, Lavric Olivera, Buzeti Uros, Jordan Vanja and Veluscek Dejan
Lek d.d. Verovskova ulica 57 1526 Ljubljana, Slovenia
*Correspondence: Berginc Katja, Lek d.d. Verovskova ulica 57 1526 Ljubljana, Slovenia, Email:

Received: 11-Nov-2022, Manuscript No. IJDDR -22-13165; Editor assigned: 15-Nov-2022, Pre QC No. IJDDR -22-13165; Reviewed: 21-Nov-2022, QC No. IJDDR -22-13165; Revised: 25-Nov-2022, Manuscript No. IJDDR -22-13165; Published: 29-Nov-2022, DOI: 10.36648-0975-9344-14.11-979


Characterization of drug release from modified-release products in the presence of alcohol is required to evaluate the possibility of dose dumping due to safety assurance reasons. A generic product containing BCS IV drug and HPMCAS polymer was formulated as amorphous solid dispersion with hot-melt extrusion having the same qualitative and quantitative composition as the reference product. Although formulations were bioequivalent, they exerted different vulnerability towards ethanol, and the root cause was sought to establish appropriate control during product life cycle. To identify critical process parameters (CPPs) and critical material attributes (CMAs) according to the Quality by Design, a systematic screening of excipient properties and process parameters was performed. The indepth evaluation revealed that dose dumping could be avoided with appropriate selection of specification limits for binder viscosity and particle size, specific surface area of the glidant and adjustments in the milling and tableting parameters. Due to the effective prevention of dose dumping, additional changes in qualitative or quantitative composition of the generic product were not required; thus subsequent bioequivalence study was not warranted.


Keywords:Dose-Dumping; Critical Material Attributes; Critical Process Parameters; Microfluxtm; Bioequivalence; Hot-Melt Extrusion


The control of drug release from modified release (MR) products mainly depends on the properties of polymer in the matrix or coating. Ingestion of MR product with alcohol can lead to alcohol induced dose dumping (AIDD), which refers to abrupt release of entire dose or a significant fraction thereof in a short period of time [1,2]. Regulatory considerations to prevent AIDD are reflected in the existing guidelines, which are not yet fully harmonized worldwide regarding the media that should be used (0.1 M HCl in US versus media used for routine testing in EU), the concentration of ethanol to be added (up to 40 % in US versus up to 20 % in EU) or the duration of the dissolution test (up to 2 h in US versus not defined for EU) [1, 2]. Although it is desired to have rugged MR formulation with drug release being insensitive to the ethanol, this is not always possible [1]. Therefore, in vivo studies and/or risk assessments evaluating patient safety are justified. From generic perspective, to-be-marketed MR formulation should exert the same or superior resistance to ethanol (i.e. provide the same or slower release rate) compared to the originator's product [1].

In the assessment of formulation vulnerability towards ethanol one can expect ethanol causing changes in solubility, wettability, swelling, and mechanical properties of drug, excipients, and/ or formulation [5]. Generally, ethanol increases solubility of hydrophobic drugs and in various ways affects performance of matrix forming excipients, which ultimately affect susceptibility of formulation to the media uptake, the onset of polymer swelling, and finally to deterioration of matrix/coating integrity and its resistance to erosion. On the other hand, the incoming material's porosity and compactness may also affect susceptibility of the surface to dissolution media and thus dose dumping [5]. All listed factors ultimately lead to altered drug release in the presence of alcohol [1,3-5].

While it could be argued that in vitro dissolution tests for assessment of AIDD may not capture in vivo bio-relevancy and are thus not predictive of in vivo situation owing to short-lived persistence of high ethanol concentrations in the stomach [3, 4], valuable information can be gathered with alcohol dissolution experiments regarding formulation performance, appropriateness of excipient selection [6], excipient critical material attributes (CMA), and critical process parameters (CPP). For generic industry such data may assist in the reverse-engineering process during development of bioequivalent products and in the Quality by Design strategy to establish appropriate control for the selected product.

The aim of this study is to present a case of MR generic product prepared with hot-melt extrusion that contains poorly soluble BCS IV model drug and hydroxypropyl methyl cellulose acetate succinate (HPMCAS) polymer as release-rate controlling excipient. The model drug exerts pH-dependent solubility (it is soluble in acidic environment), and very low saturated solubility in the physiologically relevant range. A pilot study was performed to establish in vitro-in vivo correlation between in vitro dissolution under non-sink conditions and in vivo results. With the established correlation, bioequivalent prototype on production scale was produced but dose dumping tests indicated that generic product would fail to pass AIDD regulatory requirements before entering the market. Therefore, an in-depth evaluation was triggered to identify potential CPPs and CMAs responsible for higher vulnerability of generic formulation towards alcohol and to establish control strategy in accordance with Quality by design to assure constant product quality/safety attributes through product life cycle without additional modifications in the qualitative or quantitative composition, since such alterations would necessitate additional bioequivalence study.

Materials and Methods


Excipients for extragranular phase were purchased from DuPont (microcrystalline cellulose - MCC), Ashland (hydroxypropylcellulose - HPC), DuPont (croscarmellose sodium -CCS ), Grace GmbH (silicon dioxide) and Peter Greven (magensium stearate). HPMCAS was purchased from Ashland. Silicon dioxide was from Grace GMBH. Salts for dissolution media and solvents were of analytical grade (HCl, octanol, NaCl, NaH2PO4, NaOH) and were sourced from Merck KGaA, Germany, ethanol 96 % was sourced from Győr Distillery & Refinery Co., Hungary. SIF powder for FaSSIF media was from Biorelevant, United Kingdom.

Drug quantification was performed with UV on MicrofluxTM apparatus at 256 – 276 nm. HPLC quantification was performed for dissolution experiments; the method used was validated.



FaSSIF buffer solution pH 6.5 was prepared by dissolving 6.186 g of NaCl and 3.438 g of NaH2PO in 800 mL of distilled water, adjusting pH to 6.5 using 1 M NaOH solution. Next, 2.24 g of SIF powder and distilled water to 1000 mL volume was added and stirred for 2 hours. Dissolution-permeability measurements were performed on MicrofluxTM apparatus, Pion Inc. Tablets were pealed, crushed and sieved. Fraction of 250 – 400 μm of crushed sample containing 2.5 mg of model drug was added into micro-dissolution vessels containing 10.0 mL of FaSSIF pH 6.5 at 37 ± 0.5 °C. Revolution was put to 500 rpm for the first 3 minutes to homogeneously disperse the extrudate, followed by adjustment to 100 rpm for the rest of the experiment. 10.0 mL of 1-octanol at 37 ± 0.5 °C was slowly added into the vessels and UV probes were immersed in 1-octanol for drug quantification with the in-situ Rainbow UV-VIS spectrophotometer, Pion Inc for the next 8 hours. Quantification was done using standardization curve (R2 > 0.98) correlating drug concentration and AUC of 2nd derivative of the measured absorbance in the 256 – 276 nm wavelength interval.

Extrusion process

Model drug and HPMCAS were weighted in ratio 1 : 3 and blended with Cube Mixer, Erweka GmbH. Blend was sieved using Frewitt oscillating sieve 1.5 mm and feed in 18 mm twin screw extruder. Extrudate strands were pelletized, milled with hammer mill, blended with extragranular excipients and used for tableting.


Milled extrudate and external excipients (silicon dioxide, HPMCAS, HPC, MCC, CCS, and magnesium stearate) in 2: 1 ratio were blended with Cube Mixer and compressed with single press tableting machine.

In Vitro Dissolution

Dissolution tests were performed on Agilent apparatus using paddle and 50 rpm. Samples were withdrawn at predetermined time points every 15 min up to 2 h. Sampled aliquots were filtered through 0.45 μm PVDF filters discarding the first 5 mL of withdrawn sample. Withdrawn volume of dissolution media (10 mL) was not replaced.

Drug Release from Extrudate Samples

The amount of extrudate corresponding to one, two, or three doses were weighed, and 50 mL of testing media was added. 50 mL falcon tubes with extrudate and media were placed into shaker at 37 °C and 650 rpm. Samples were withdrawn, filtered through 0.2 μm MSVGN 2250 filter plates from Corning and the filtrate was diluted with dissolution media before HPLC analysis.

Saturated Solubility of Model Drug

Model drug (50 mg) was weighed into 15 mL tubes. After adding 10 mL of dissolution media, eprouvettes were placed into shaker at 37 °C and 650 rpm. Samples were withdrawn, filtered through 0.2 μm MSVGN 2250 filter plates from Corning and the filtrate was diluted with dissolution media before HPLC analysis.

In vivo bioequivalence study

A comparative bioavailability study was conducted under fast conditions to investigate the in vivo behavior and compare the bioavailability of the two test formulations to that of the reference product. The protocol for the study was approved by the relevant Independent Ethics Committee and was conducted in accordance with Good Clinical Practice and Declaration of Helsinki. Plasma concentrations of the active substance were measured by a validated, sensitive and specific LC/MS/MS assay. The pharmacokinetic parameters AUC (area under the plasma concentration time curve) and Cmax (peak plasma concentration) were calculated using a non-compartmental approach.

All subjects were provided with an ICF in their language of preference for review. Prior to initiation of study procedures, the ICF was verbally reviewed with subjects by qualified staff, allowing sufficient time for review of the information provided and to answer any questions subjects had. Subsequently, subjects were informed of any developments or changes to procedures that could have influenced their continued participation in the study. Subjects did not anticipate any direct benefits from participation in this research study, with the exception that they received a health evaluation. Subjects who participated in this study were compensated for their time; however, they were not offered any incentives.


Dissolution under Non-Sink Conditions to Establish In Vitro – In Vivo Correlation

Evaluation of drug release with compendial dissolution apparatus using 900 mL FaSSIF pH 6.5 as release media shows precipitation under non-sink conditions as shown in Figure 1. Since saturated solubility is attained within 15 min and high variability (RSD > 10 %) is evident throughout the test, dissolution in FaSSIF was found unreliable and not sufficiently rugged to guide prototype development. During drug release and due to extremely low solubility, phase transformation of dissolved drug is imminent and several different nano-species (i.e. nanodroplets, LLPS, crystalline seeds) were identified [7, 8], which cannot be completely separated from dissolved drug with currently available filter techniques. Data obtained with compendial dissolution approach is therefore highly variable. Also, with the concomitant occurrence of several nano-species and dynamic equilibria among them, questions regarding which specie is absorbed and which constitutes drug depo that drives the absorption remain unanswered. Thus, instead of monitoring amount dissolved in the aqueous phase, the amount distributed into organic phase was evaluated by using MicrofluxTM (Figure 2), and an excellent correlation between in vitro accumulation of dissolved drug in 1-octanol and in vivo pilot study results (Table 1) was established. The pilot study was performed using two test products (Test product 1 and Test product 2), which differed in the extruder configuration but have the same qualitative and quantitative composition as reference product. Based on favorable in vivo results seen with Test product 2, selected extruder configuration was used to prepare Test product 3 on scale-up on as to-bemarketed product. However, due to processability issues during scale-up extrusion, silicon dioxide was added into the extruding mass to avoid process stoppages. Nevertheless, biorelevant dissolution method in the MicrofluxTM indicated that this change would not impact formulation bio-performance.

Test product Cmax AUCt-¥
1 126 % (117 – 136 %) 118 % (112 – 124 %)
2 115 % (106 – 125 %) 111 % (105 – 118 %)
Table 1 .  In vivo results (average value and 95% CI) for pilot study under fast conditions for generic product.

Figure 1: Typical dissolution profile of one tablet (n = 3) in 900 mL of FaSSIF pH 6.5 media, Apparatus 2, 100 rpm.


Figure 2: Permeation of dissolved drug from aqueous to organic phase – figure shows drug concentration in the organic phase.

Selection of the Most Discriminatory Dose- Dumping Dissolution Conditions

Test product 2 was bioequivalent (Table 1). It also conformed to requirements regarding dose dumping; the release from generic Test product 2 was similar or slower/lower than the one of reference product used in the in vivo pilot study. Interestingly, the release was the highest in media with 20 % ethanol, not as one would expect in the media with the highest ethanol concentration of 40 % (Figure 3).


Figure 3: Drug release form Test product 2 and Reference product in media required to demonstrate the absence of dose-dumping. Graphs show average values of n = 3 repetitions for every sample with corresponding standard deviations.

However, when Test product 3 was screened against reference procured for the pivotal bioequivalence study, dose-dumping for the generic formulation in media with 20 % of ethanol was recorded (Figure 4). Conditions with 10 % ethanol were thus not evaluated.


Figure 4: Drug release form Test product 3 and Reference product in media required to demonstrate the absence of dose-dumping. Graphs show average values of n = 3 repetitions for every sample with corresponding standard deviations.

Based on evaluations in media for AIDD, 20 % of ethanol was selected as the most discriminatory dissolution condition that is sensitive to certain formulation and/or process parameters, which trigger dose dumping. Also, we observed that the reference product itself does not perform in a uniform and rugged manner, since reference lot used in pilot study and the one intended for pivotal study did not provide similar release in 20 % ethanol media.

Drug Substance Solubility in Proposed Ethanolic Media

Saturated solubility of model BCS IV drug in media used during dose dumping was evaluated (Table 2) and a theoretical number of doses that could completely dissolve were calculated. Based on presented results, drug solubility increases with the increasing amount of ethanol and in all cases, solubility is sufficient to allow for complete release. Furthermore, sink conditions are assured across all media (i.e. at least 3 doses would dissolve in 900 mL), thus, complete release 900 mL of each media is possible.

Time (h) 0.1 M HCl 0.1 M HCl + 10% Et 0.1 M HCl + 20% Et 0.1 M HCl + 40% Et
Conc (mg/mL) number of doses Conc (mg/mL) number of doses Conc (mg/mL) number of doses Conc (mg/mL) number of doses
0.5 0.6 5 2.2 19 3.3 29 5.5 50
1 1.0 9 2.2 19 4.6 42 5.4 50
2 1.0 9 2.2 20 4.7 42 5.4 50
4 1.0 9 2.2 20 4.7 42 5.4 50
6 1.0 9 2.2 20 4.8 43 5.3 50
Et – ethanol Table 2 . Solubility of drug substance in dose dumping media. Table shows average solubility and the number of doses that could completely dissolve in 900 mL.

Solubility of Extrudate With and Without Silicon Dioxide in Proposed Ethanolic Media

Since the only qualitative difference between Test product 2 and Test product 3 lies in the silicon dioxide addition into the extrudate to improve HME processability, drug release from extrudate sieved fraction 100 - 125 μm, was monitored. The amount of drug dissolved was tested at different initial extrudate amounts, corresponding to 1, 2 and 3 tablets being dissolved in 900 mL to evaluate if drug release is limited by the polymer solubility. Because drug solubility in 0.1 M HCl and ethanol is not limiting, any amount dissolved from the extrudate would be direct reflection of extrudate surface properties, wettability and or resistance of HPMCAS matrix to dissolution media (Figure 5).


Figure 5: Impact of starting amount of extrudate on the % of drug released in proposed ethanolic media for extrudate with and without intragranular silicon dioxide, fraction analyzed 100 – 125 μm.

Several surprising effects were noticed. Firstly, no difference in dissolution was recorded between both compositions, confirming that inclusion of silicon dioxide during extrusion is not causing dose dumping of the finished formulation (Figure 4). When dissolution media with 10 % - 40 % ethanol was used, complete release was recorded within 24 h, which is also in accordance with high drug saturated solubility results.

Based on presented results, HPMCAS in the extrudate does not prevent AIDD and a complete release would be expected. Namely, in 2-h time-frame, as prescribed for dose dumping tests, approximately 10 % (in 0.1 M HCl), 50 % (in 10% ethanol in 0.1 M HCl), and complete release (in 20% and 40% ethanol in 0.1 M HCl) would be anticipated according to results in Table 3. The discrepancy in drug release in the corresponding media between pure extrudate and finished dosage form (regardless of silicon dioxide) can thus only stem from down-stream processing after hot-melt extrusion process.

without SiO2 with SiO2
Time (h) 0% Et 10% Et 20% Et 40% Et 0% Et 10% Et 20% Et 40% Et
1 11 34 95 102 13 38 93 104
2 10 54 100 98 12 58 99 101
4 12 75 100 100 15 80 100 99
6 15 87 99 98 17 91 100 97
  Table 3 . The % of drug dissolved in each proposed ethanolic media for extrudate with and without silicon dioxide. Data corresponding to 1 tablet being dissolved in 900 mL of individual media are shown. The amount remaining at the end of experiment was quantified and the % based on the initial mass is reported.

Mechanical Properties of Extrudate and Compactness

Tablets were compressed from both extrudates (i.e., with and without silicon dioxide) and dissolution in 20 % ethanol media was performed. Dissolution was compared between both compositions for tablets with similar tablet hardness and/or similar disintegration (Figure 6 and Table 4). Dissolution results show that samples are ranked according to disintegration time within each composition – the shorter disintegration, the faster drug release. However, tablets that contain silicon dioxide show significantly shorter disintegration time than samples without it in spite of having similar tablet hardness (i.e. sample A is faster than D, sample B is faster than E). Overall, dissolution data and physical properties of analyzed tablets indicate that mechanical properties between both compositions are different.

Sample Tablet hardness (in kp) Disintegration time (in min)
A – with SiO2 8.9 7-8
B – with SiO2 13.9 13-17
C – with SiO2 12.8 18-19
D – without SiO2 11.0 12-14
E – without SiO2 13.5 19-20
Table 4 . Average tablet hardness and range of disintegration time for tablets prepared from extrudate with and without silicon dioxide.

Figure 6: Impact of tablet hardness and disintegration time on drug release in 20 % ethanol in 0.1 M HCl (Apparatus 2, 50 rpm, 900 mL of 20 % ethanol in 0.1 M HCl) for tablets prepared from extrudate with and without silicon dioxide. Dissolution profiles were generated on n = 3 tablets for each sample. Data are shown as average values with corresponding standard deviations.

Impact of Extrudate Particle Size

After extrusion extrudate is milled twice, first through 2.0 mm mesh, followed by the second step of milling through 0.5 mm mesh. The impact of different milling speeds on D50 values of extrudate were evaluated for both compositions (Figure 7). Also, dissolution in 20 % ethanol was evaluated to test, if D50 also affects drug release from tablet cores.


Figure 7: Left - Impact of extrudate particle size on drug release in 20 % ethanol in 0.1M HCl (Apparatus 2, 50 rpm). Tablet cores were analyzed in n = 3 repetitions, graph shows average values with corresponding standard deviations. Right – impact of milling speed on extrudate particle size depending on the presence of silicon dioxide in the extrudate (SiO2).c

The milling study indicates that under the same milling speeds significantly bigger extrudate particles are obtained when silicon dioxide is present in the extrudate (see Appendix), indicating different mechanical properties. Furthermore, bigger extrudate particles also give faster and more variable dissolution in 20 % ethanol.

Impact of Hydroxy Propyl Cellulose Particle Size and Molecular Weight

Hydroxypropyl cellulose (HPC) is one of the excipients used in the extragranular phase necessary for tablet compressibility since it acts as binder. Dissolution data in 20 % ethanol in 0.1 M HCl confirms that particle size of this excipient is crucial for efficient prevention of AIDD and smaller HPC particles provide better resistance to ethanol for two evaluated grades (Figure 8). Besides particle size, molecular weight (at the same particle size of excipient) was also found to impact dose dumping (Figure 9).


Figure 8: Impact of HPC particle size on drug release in 20 % ethanol in 0.1 M HCl media (Apparatus 2, 50 rpm, 900 mL of 20 % ethanol in 0.1 M HCl). Graphs show average values of n = 3 repetitions for every sample with corresponding standard deviations. On the right side is table with PSD limits for used grades.


Figure 9: Impact of HPC molecular weight on drug release in 20 % ethanol in 0.1 M HCl media (Apparatus 2, 50 rpm, 900 mL of 20 % ethanol in 0.1 M HCl). Graph shows average values of n = 3 repetitions for every sample with corresponding standard deviations.

Impact of HPMCAS Quantity

HPMCAS is included in the extragranular phase to enable tableting by acting as binder. Omission of HPMCAS from extragranular phase leads to significant increase in drug release in 20 % ethanol in 0.1 M HCl, therefore this excipient is crucial to retaining tablet consistency and resistance to dose dumping (Figure 10).


Figure 10: Impact of HPMCAS quantity in extragranular phase on drug release in 20 % ethanol in 0.1 M HCl media (Apparatus 2, 50 rpm, 900 mL of 20 % ethanol in 0.1 M HCl). Graph shows average values of n = 3 repetitions for every sample with corresponding standard deviations.

Impact of Silicon Dioxide Specific Surface Area

Extragranular addition of silicon dioxide as glidant enables continuous tableting process. Release in the presence of ethanol was evaluated for tablets where the impact of specific surface area of silicon dioxide in the extragranular phase was varied. Based on Figure 11, the impact is profound and higher specific surface area is desired to decrease formulation vulnerability towards alcohol.


Figure 11: Impact of specific surface area of silicon dioxide in extragranular phase on drug release in 20 % ethanol in 0.1 M HCl media (Apparatus 2, 50 rpm, 900 mL of 20 % ethanol in 0.1 M HCl). Graphs show average values of n = 3 repetitions for every sample with corresponding standard deviations.

Drug Release from Reference Product

Several different lots of reference product were screened for drug release in 20 % ethanol in 0.1 M HCl media (Figure 12) and significant differences among lots were identified (see Appendix). Also, some references exerted significant variability (RSD > 10 % at every time-point in dissolution profile), while others gave consistent results (Lot 2 and Lot 4). Since the composition of reference product is constant, the most probable cause for such differences could be poorly controlled process parameters or overlooked critical material attributes in the extragranular phase.


Figure 12: Drug release in 20 % ethanol in 0.1 M HCl media (Apparatus 2, 50 rpm, 900 mL of 20 % ethanol in 0.1 M HCl) from different reference lot numbers. Graphs show average values of n = 3 repetitions for every sample with corresponding standard deviations.


Solubility of model BCS IV drug is pH dependent and in the physiologically relevant range limited solubility prevents the entire dose to dissolve in the compendial 900 mL of dissolution media. In ethanol however, drug solubility is sufficient to allow complete release. To improve therapeutic efficacy of the model drug, HPMCAS polymer is used as the carrier to obtain amorphous solid dispersion by hot-melt extrusion process. This polymer is also essential for prevention of drug crystallization on storage and stability, and it acts as inhibitor of precipitation during drug dissolution in vivo [9]. Also, HPMCAS polymer itself exerts pH dependent solubility and enables drug release in the environment with pH above 5.5 – 6.8, depending on the degree of substitution of succinoyl and acetyl groups [10, 11].

Due to the opposing solubility characteristics of weakly basic model BCS IV drug and slightly acidic HPMCAS, and very low drug solubility in physiologically pH range, guiding product development with compendial dissolution in bio-relevant media such as FaSSIF pH 6.5 has limited potential due to the interplay of dissolution, supersaturation, and precipitation (Figure 1), which contribute to poor ruggedness of dissolution method. Therefore, dissolution tests to guide development of bioequivalent product and establish in vitro – in vivo correlation were executed with MicrofluxTM apparatus. This experimental set-up enables in situ monitoring of dissolution on one side in aqueous media with or without surfactants (i.e. such as SIF powder). Dissolved drug then permeates lipophilic membrane into the receiver compartment that contains acceptor media with high concentration of sodium lauryl sulphate, which assures sink conditions and prevents precipitation. However, for the purpose of this product, slight deviation from the envisioned use was selected. Namely, the use of lipophilic membrane was omitted and a combination of FaSSIF pH 6.5 and 1-octanol as receiver media above FaSSIF was used. Without lipophilic membrane an additional rate-limiting step (i.e. membrane permeation) was avoided and better discriminatory power was obtained. The amount of accumulated dissolved drug in 1-octanol was monitored to reflect changes of drug in plasma, while quantification in FaSSIF phase was not performed due to extensive precipitation into various species (i.e. crystalline substance, nanoparticles, LLPS) [7, 8]. With this method in vitro - in vivo correlation was established according to the outcome of pilot bioequivalence study on healthy human volunteers under fast conditions. Fast conditions were correlated because they were confirmed to be more discriminatory. Two samples of generic product with the same qualitative and quantitative composition as reference product but produced with two different extruder configurations were tested in vivo (Table 1) and Test product 2 was found bioequivalent. Based on in vivo results and in vitro profiles of dissolved drug in 1-octanol a correlation was established. Namely, dissolution method developed with MicrofluxTM correctly ranked Test product 1 and Test product 2 compared to reference product according to corresponding Cmax values. Thus, Test product 3 – final prototype sample with selected extruder configuration was manufactured on scale-up as the final formulation. However, due to processability challenges on production scale HME equipment, silicon dioxide had to be added into the extrudate of Test product 3 to prevent stoppages of the process. Regardless of this change, dissolution under nonsink conditions in MicrofluxTM confirmed that Test product 3 would be bioequivalent (Figure 2).

However, when submitted to AIDD in vitro dissolution testing Test product 3 failed to provide similar dissolution as reference, although Test product 2, which was manufactured on the same extruder configuration but without silicon dioxide (Test product 2) passed (Figures 3 and 4 ). Initially it was suspected that addition of silicon dioxide to extrudate of Test product 3 could be accountable for dose-dumping, since this change is the only difference among Test products 2 and 3. Namely, when silicon dioxide is added to extrudate, it can affect drug release as reported by Xia et al. [12]. Their study evaluated the impact of silicon dioxide positioning and noticed significant changes in itraconazole release in aqueous media when silicon dioxide was present in the extrudate, because it affected the performance of the polymer in solid dispersion by inducing polymer gelling and formation of a transition zone with higher viscosity. Also, it interacted with the drug in the solid dispersion, and it influenced wetting properties, which were observed as poorer deagglomeration of solid dispersion particles during dissolution. Slower penetration of dissolution media into the formulation, and diminished interactions between dissolution media and the drug were also identified as contributing factors to slower release. While their study noticed retardation of drug release, dissolution in FaSSIF pH 6.5 in our case remained unaffected. Differences were observed when dissolution in 20 % of ethanol was done, but they were not caused by inclusion of silicon dioxide, because release of model drug from both extrudates (with and without silicon dioxide) in ethanol was the same (Figure 5 and Table 3). Therefore, the root cause for dose-dumping had to be identified in extrudate properties, which alone or in combination with extragranular excipients and/or down-stream processing steps affect drug release.

Process steps after hot-melt extrusion include milling, blending with extragranular excipients, tableting and film-coating. Having extrudate with different mechanical properties could influence every of the above-mentioned processes. In order to retain qualitative and quantitative composition of bioequivalent Test product 3 and simultaneously prevent AIDD, Quality by Design principles were followed to identify possible process steps (i.e. CPPs) and extragranular excipients properties (i.e. CMAs) responsible for AIDD and establish appropriate product control to avoid jeopardizing patient safety.

Results confirmed that milling and tableting outcome depend on the presence of silicon dioxide. The milling study showed that under identical milling conditions (i.e. mesh size and milling speed), extrudate prepared with silicon dioxide yields bigger particles of amorphous dispersion, which also enable faster and more variable release in ethanol media (Figure 7). Furthermore, during tableting, tablet cores compressed from extrudate manufactured with silicon dioxide exerted shorter disintegration time than cores compressed with the same hardness from the extrudate that did not contain silicon dioxide (Figure 6) although in both cases identical extragranular mixture of excipients was added. Overall, both CPPs indicate that mechanical properties of the extrudate of Test product 3 were influenced by inclusion of silicon dioxide. When extruded, HPMCAS is known to have poor compressibility, because hot-melt extrusion with its heating and shear stress affects polymer orientation in a way that yields high extrudate particle density with increased dynamic hardness, making HPMCAS extrudate material less amenable for compression [13]. Therefore, any manipulations done to the extruded material that affects mechanical properties would be expected to affect processibility of the material such as milling and tableting. Also, bigger HPMCAS extrudate particles have less contact areas available for particle-particle interactions during tableting (i.e. bonding area), which favors inflow of dissolution media and thus drug release [14]. This presumption was also corroborated with AIDD results obtained for the tableting process.

On the other hand, AIDD outcome can also be controlled by adjusting the properties and/or quantities of excipients selected for the extragranular phase (CMAs). We have shown that AIDD release depends on the properties of binder in the extragranular phase. Namely, molecular weight and particle size of hydroxypropyl cellulose (HPC), when carefully selected, prevent AIDD (Figures 8 and 9). This is in accordance with Johnson et al [15] study that evaluated the impact of ionic strength on HPC swelling, gelation, and disintegration time of compacts prepared from HPC and model drug. They confirmed that smaller particle size and lower molecular weight HPC resist premature disaggregation and the loss of HPC matrix integrity more efficiently, therefore enabling slower drug release. This is also in line with presented results obtained in 20 % ethanol evaluation in this paper, where ionic strength changes with addition of ethanol, which could explain our observations. Additionally, Johnson et al. [15] found that the relationship between drug release and ionic strength follows biphasic patterns (i.e. slower release at low and high ionic strength but faster release in medium ionic strength), which also explains why the fastest release in our study was observed with medium ethanol concentration of 20 % (and not at 40 %). Besides HPC, the investigated formulation in our case also contains HPMCAS in the powder form in the extragranular phase to assist in improving binding capacity between extrudate particles during tableting. The importance of HPMCAS role in the extragranular phase was demonstrated by significant dose-dumping when it was omitted from the formulation, therefore, the quantities of this excipient were not changed (Test product 2 and 3 retain the same quantity and quality – grade). Decreasing vulnerability of formulation towards alcohol with this excipient is logical, because HPMCAS is insoluble in acidic media and in this manner assists in retaining tablet integrity during dissolution as observed in patent application by Kanamaru et al. [16]. Their study corroborates that the presence of HPMCAS as well as particle size contribute significantly to mechanical resistance of sustained-release tablets under 200 rpm at pH 2.0. Silicon dioxide in the extragranular phase performs as glidant and it was also investigated as potential excipient whose specific surface area could impact AIDD release. We confirmed that the grade with bigger specific surface is better in preventing of AIDD.

Overall, several key points were identified that with carefully establish control could be applied in the production of generic product and whose overall effect could contribute to lowering sensitivity of generic formulation towards AIDD. Interestingly, when different reference lot products sourced from different markets were subjected to AIDD testing in 20 % ethanol (Figure 12) the same issue was identified. Some of tested originator lots showed dose-dumping effect indicating the absence of suitable control of CPPs/CMAs. Since the regulatory requirement states that dose-dumping of generic product must be compared to the dissolution profile of reference batch used in the bioequivalence study, superior quality pertaining to safety aspect is assured for the generic product presented in this paper.


This study presents the commonly overlooked potential of AIDD dissolution tests as one of crucial methods capable of explaining the complex interaction of CQAs, CMAs and CPPs in development of safe and effective final pharmaceutical dosage form. Also, it reveals gaps between the regulatory requirements and actual market situation. Namely, AIDD testing is a one-time in vitro study, performed on one reference batch. However, a systematic screening of different reference lots revealed poor resistance of some reference batches to the ethanol assuming the lack of systematic CPPs/CMAs control by the originator.

Since the reference batch procured for the bioequivalence study exhibited low drug release in 20 % ethanol, generic product was systematically designed to assure low drug release in a consistent manner regardless of the lot-to-lot variability.


Impact of SiO2 on particle size distribution (PSD)

In order to assess, if the presence of SiO2 in the formulation significantly influences the PSD of the granulate after milling; we have analyzed the data from Figure 7. Namely, the D50 represents the median of the PSD, therefore the significant difference in the D50 values with and without SiO2 show that there is a significant shift between the distributions with respect to the presence of the SiO2. There is no evidence that the difference of D50 values obtained from the granulates which differed only in the presence of the SiO2 is not normally distributed (see Figure 13 and Table 5). Hence, we use the two-sided paired t-test to compare the impact of the presence of the SiO2 on the D50. The significance level was set on 0.05 (Table 6). The paired t-test result (see below) Table 7 shows that there is a significant difference between D50 values between granulates with and without the SiO2. The addition of the SiO2 apparently significantly shifts the PSD to the right (particles are larger).

A2 Prob > A2
Anderson-Darling 0,2838914 0,5050
Table 5 . Anderson-Darling normality test for the difference of D50(SiO2-noSiO2).
Mean 30
Std Dev 16,329932
Std Err Mean 8,1649658
Upper 95% Mean 55,984565
Lower 95% Mean 4,0154347
N 4
 Table 6. Summary statistics of the difference of D50(SiO2-noSiO2).  
Hypothesized Value 0
Actual Estimate 30
DF 3
Std Dev 16,3299
Test Statistic 3,6742
Prob > |t| 0,0349*
Prob > t 0,0174*
Prob < t 0,9826
  Table 7. Paired t-test for D50(SiO2-noSiO2). 

Figure 13: Distribution, fitted normal density curve and Q-Q plot of the difference of D50(SiO2-noSiO2).

Differences between reference lots in 20% ethanol in 0.1 M HCl media

The comparison of reference lots in 20 % ethanol in 0.1 M HCl media was performed using f2 statistics on their dissolution profiles. The dissolution profiles of two lots are considered equivalent if the f2 statistics exceeds 50. The dissolution profiles will be treated as significantly different if the 90 % confidence interval (CI) of the f2 lies below the value 50 (Table 8).

For the purpose of the study, the 90% CI of f2 was calculated using the bootstrap technique with 10.000 samples. The biascorrected, accelerated CI type was used.

According to the Table 9 the Lot 6 is significantly different from all other lots and the lot 5 is a borderline case regarding the comparison to lot 1 and to lot 2.

Lot 2 Lot 3 Lot 4 Lot 5 Lot 6
Lot 1 89.99 45.63 57.06 42.90 25.12
Lot 2 46.59 58.27 43.78 25.54
Lot 3 63.18 84.99 35.17
Lot 4 57.63 30.04
Lot 5 36.81
Table 8. f2 values.
Lot 2 Lot 3 Lot 4 Lot 5 Lot 6
Lot 1 (75.30, 93.65) (28.01, 77.43) (49.61, 63.38) (37.80, 51.32) (21.46, 28.85)
Lot 2 (28.78, 77.69) (51.94, 62.93) (39.30, 52.22) (21.96, 29.16)
Lot 3 (39.62, 93.55) (77.12, 96.92) (24.97, 54.89)
Lot 4 (49.83, 71.98) (25.64, 34.60)
Lot 5 (30.55, 44.30)
  Table 9 . BCa f2 90% confidence intervals.


  1. Friebe TP, Asgarzadeh F, Gray A, Hughes K, Hebestreit JP et al. (2015) Regulatory Considerations for Alcohol-Induced Dose Dumping of Oral Modified-Release Formulations. Pharm Tech 38: 40-46.
  2. Google Scholar

  3. Cvijic S, Aleksic I, Ibric S, Parojcic J (2018) Assessing the risk of alcohol-induced dose dumping from sustained-release oral dosage forms: in vitro-in silico approach. Pharm Dev Technol 23: 921-932.
  4. Indexed at, Google Scholar, Crossref

  5. Henderson LS, Tenero DM, Campanile AM, Baidoo CA, Danoff TM (2007) Ethanol does not alter the pharmacokinetic profile of the controlled-release formulation of carvedilol. J Clin Pharmacol 47: 1358-1365.
  6. Indexed at, Google Scholar, Crossref

  7. D’Souza S, Mayock S, Salt A (2017) A review of in vivo and in vitro aspects of alcohol-induced dose dumping. AAPS Open 3: 5-9.
  8. Google Scholar

  9. Jedinger N, Khinasta J, Roblegg E (2014) The design of controlled-release formulations resistant to alcohol-induced dose dumping - A review. Eur J Pharm Biopharm 87: 217-226.
  10. Indexed at, Google Scholar, Crossref

  11. Gujjar CZ, Rallabandi BC, Gannu R, Deulkar VS (2016) Development and Optimization of a Novel Prolonged Release Formulation to Resist Alcohol-Induced Dose Dumping. AAPS Pharm Sci Tech 17: 350-357.
  12. Indexed at, Google Scholar, Crossref

  13. Ilevbare G, Taylor L (2013) Liquid-Liquid Phase Separation in Highly Supersaturated Aqueous Solutions of Poorly Water-Soluble Drugs: Implications for Solubility Enhancing Formulations. Cryst Growth Des 13: 1497-1509.
  14. Google Scholar

  15. Indulkar AS, Gao Z, Raina SA, Zhang GGY, Taylor LS (2016) Exploiting the Phenomenon of Liquid-Liquid Phase Separation for Enhanced and Sustained Membrane Transport of a Poorly Water-Soluble Drug. Mol Pharma 13: 2059-2069.
  16. Indexed at, Google Scholar, Crossref

  17. Friesen DT, Shanker R, Crew M, Smithey DT, Curatolo WJ et al. (2018) Hydroxypropyl Methylcellulose Acetate Succinate-Based Spray-Dried Dispersions: An Overview. Mol Pharm 5: 1003-1019.
  18. Indexed at, Google Scholar, Crossref

  20. Shin-Etsu AQOAT® for amorphous solid dispersion - SE Tylose
  21. Xia Z, Yuan M, Deng Z, Ke X, Ci T (2017) Different effects of silica added internal or external on in vitro dissolution of indomethacin hot-melt extrudates. Int J Pharm 534: 272-278.
  22. Indexed at, Google Scholar, Crossref

  23. Iyer R, Hedge S, Zhang ZE, Dinunzio J, Singhal D (2013)The impact of hot melt extrusion and spray drying on mechanical properties and tableting indices of materials used in pharmaceutical development. J Pharm Sci 102: 3604-3613.
  24. Indexed at, Google Scholar, Crossref

  25. Mishra SM, Richter M, Mejia L, Sauer A (2022) Downstream processing of itraconazole: HPMCAS amorphous solid dispersion: from hot-melt extrudate to tablet using a Quality by Design approach. Pharmaceutics 14: 1429-1448.
  26. Indexed at, Google Scholar, Crossref

  27. Johnson JL, Holinej J, Williams MD (1993) Influence of ionic strength on matrix integrity and drug release from hydroxypropyl cellulose compacts. Int J Pharm 90: 151-159.
  28. Google Scholar

  29. 1498384591355793633-US20150366810A1 (

Citation: Katja B, Tilen H, Olivera L, et al. (2022) Exposing Profound Screening Potential of Ethanol-Based Dissolution Media in the Development of Oral-Modified Dosage Forms. Int J Drug Dev Res J, Vol. 14 No. 11: 979.