How surfactants, pH and a suspending agent change drug leaching from enteric ASD particles in acidic vehicles

Background and motivation

Amorphous formulations are a common way to raise the solubility of poorly water-soluble drugs: the amorphous form dissolves to a supersaturated solution, which can raise bioavailability, but the same supersaturation drives crystallization, so polymers are added to slow it and the products are called amorphous solid dispersions (ASDs) (Kawakami 2009, ref 1; Newman 2012, ref 2; Hancock 2000, ref 3; refs 4–6). This paper looks at a less-studied situation: ASDs made with enteric polymers, which are insoluble at low pH, held in an acidic liquid.

That situation arises in practice in two ways. Suspensions of enteric ASD particles in an acidic vehicle are used to dose animals in pharmacokinetic, pharmacodynamic and toxicology studies and in early clinical trials, where simple formulations are preferred; the vehicle usually contains a suspending agent to raise viscosity and a surfactant to wet the powder (Byrn 2010, ref 7; Chen 2014, ref 8). And any oral dose of an enteric ASD meets gastric acid before it reaches the intestine. In both cases the polymer stays largely undissolved, but the authors point out that a small amount of drug can still dissolve and may form a supersaturated solution, so crystallization must be avoided both in the suspended particles and in the solution around them C1·C2.

Excipients complicate this. They can change both how much supersaturation develops in the medium and how readily the drug nucleates (refs 9–11). The group's earlier work on celecoxib had linked surfactant effects on nucleation induction times to the crystallization of suspensions of celecoxib ASD particles (Chen 2015, ref 12) C3. What was not known was how excipients in the vehicle change the dissolution of drug from the suspended particles, and hence the concentration the drug reaches. The authors frame the problem as two opposing effects: more drug leaching means more supersaturation and a higher crystallization tendency, while more polymer leaching "could retard" crystallization by inhibiting nucleation in solution, so both should be measured C4. Even very low polymer concentrations can act as crystallization inhibitors (Abbou Oucherif 2013, ref 15) C13.

The paper therefore sets out to measure how polymer choice, surfactant type, pH and a suspending agent affect the leaching of drug and polymer from ASD suspensions. Lopinavir (LPV) was chosen as the model drug because of its low aqueous solubility and low crystallization tendency: it lets leaching be followed without crystallization happening at the same time C5. The polymers are hydroxypropylmethylcellulose acetate succinate (HPMCAS, grades LF, MF and HF) and hydroxypropylmethylcellulose phthalate (HPMCP); the surfactants are sodium dodecyl sulfate (SDS) and sodium taurocholate (STC), both anionic, and poloxamer 188 and polysorbate 80, both nonionic; and the suspending agent is polyvinylpyrrolidone/vinyl acetate (PVPVA, Kollidon VA 64) (Figure 1) M1–M10.

Background concepts

Leaching from a suspension. In this paper leaching means drug or polymer dissolving out of ASD particles that themselves stay suspended and largely undissolved. At pH 3 the enteric polymers are mostly unionized and "nominally insoluble", so the particles persist and only a small fraction of each component passes into solution C8. Within the 3 h of each experiment drug and polymer leaching "typically reached a plateau", and the paper compares systems by the average concentration between 60 and 180 min P4.

Crystalline and amorphous solubility. The crystalline solubility is the equilibrium concentration over drug crystals; the amorphous solubility is the higher concentration reached over the pure amorphous drug. A solution between the two is supersaturated with respect to the crystal and can in principle crystallize, but it cannot form a separate amorphous drug phase. For LPV in pH 3 citrate buffer the paper gives a crystalline solubility of 2.5 ± 0.2 µg/mL and an amorphous solubility of about 16.6 µg/mL (read from Figure 8a) R5.a·R6. The authors also draw on the observation that the components of a mixture lower each other's aqueous solubility: this is known for organic liquid mixtures (Banerjee 1984, ref 20) and "seems to hold true" for amorphous mixtures such as ritonavir with darunavir (ref 21), felodipine with miglyol (ref 22) and ritonavir with lopinavir (Trasi 2015, ref 23) C33. For a drug intimately mixed with a polymer, the drug's solubility "might be expected" to depend on its amorphous solubility, the drug loading and the drug–polymer interactions C34.

Apparent supersaturation. The paper's central quantity is the ratio of the LPV concentration leached from the ASD to the crystalline solubility in a medium containing the same excipients at the same temperature (Methods, "Apparent Supersaturation"; Figure 9 caption):

Here $C$ is the LPV concentration leached from the dispersion and $C_{\mathrm{e}\mathrm{q}}$ the equilibrium crystalline LPV solubility in a medium of the same composition. In plain terms: the ratio says how many times above its crystal solubility the drug sits in that particular medium. Because micelles and polymers in solution also hold drug, the authors call the ratio apparent: a concentration ratio in such complex media "may or may not" represent the thermodynamic driving force for crystallization (Raina 2015, ref 13) E1·C7.

Critical micelle concentration and micellar solubilization. Above its critical micelle concentration (CMC) a surfactant forms micelles that can take up a poorly soluble drug, so the drug's apparent solubility rises with the surfactant concentration above the CMC. The authors determine each CMC this way: they measure crystalline LPV solubility against surfactant concentration and take the CMC where the solubility starts to rise sharply P3. In pH 3 buffer with LPV present they find 0.050% for SDS, 1.40% for poloxamer 188, 0.020% for polysorbate 80 and 0.73% (w/v) for STC R1. For the leaching work each surfactant was used at 10% above its CMC, so that some micellar surfactant, and hence some solubilization, is present C6·P4.

Polymer–surfactant complexes. Surfactants interact with dissolved polymers and can begin to aggregate on the polymer chain at a critical aggregation concentration (CAC) that is usually well below the CMC; for SDS the CAC can be more than two orders of magnitude lower (Kwak 1998, ref 24; Goddard 1976 and 1977, refs 25, 26). Such complexes form between nonionic polymers and ionic surfactants, ionic polymers and nonionic surfactants, and oppositely charged pairs, and can solubilize solutes below the CMC; their solubilizing power can be lower or higher than that of the micelles alone, depending on the system (Ammar 1976, ref 30) C41. This literature becomes the authors' main tool for explaining the effects of HPMCP, HPMCAS and PVPVA on surfactant performance.

Enteric polymers and pKa. HPMCAS grades are made by substituting HPMC with different amounts of acetate and succinate groups; HPMCP carries phthalate groups (Figure 1) M2–M5·C10. The grades differ in pKa: the paper states that HF has a higher pKa than MF and LF, and that LF has a lower pKa than HF (Ashland handbook, ref 16), so at pH 5 LF is more ionized than HF; no pKa values are printed C9·C27. LPV itself, the authors state, does not ionize between pH 3 and 5 C24.

Methods in detail

The study is one experimental design applied repeatedly: suspend a fixed amount of ASD in a vehicle, follow drug and polymer in solution for 3 h, and compare the plateau with the crystalline solubility of LPV in a vehicle of the same composition. What varies between experiments is the polymer in the particles, the surfactant in the vehicle, the pH and the presence of PVPVA A1–A4.

The dispersions

LPV (10 mg/mL) and polymer (20 mg/mL) were dissolved together in acetone and spray dried on a ProCepT Microspray Dryer (0.6 mm bifluid nozzle, 5 mL/min, outlet temperature 55 ± 5 °C, atomization air 2.5 L/min), then dried overnight in a vacuum oven at room temperature P1. The paper prints the drug loading as "33.3% (w LPV /w polymer)", but the solution concentrations give 33.3% LPV in the dry solids, which is also how the text later speaks of "33% drug loading"; the card reads it that way X1. Four dispersions were made, with HPMCAS LF, MF, HF and HPMCP, plus the four polymers spray dried alone as controls A1. Their surface-weighted mean particle sizes, D[3,2], measured by laser diffraction in 1% Methocel A4M in pH 3 buffer, were 40.9 (LF), 37.0 (MF), 24.6 (HF) and 14.2 µm (HPMCP) for the dispersions and 29.2, 29.9, 34.6 and 24.1 µm for the pure polymers P2·M2–M5. The paper reports no solid-state characterization of the dispersions (no X-ray diffraction, thermal analysis or spectroscopy), so their miscibility is inferred later rather than measured C35.

Choosing the surfactant levels

To find each CMC, excess crystalline LPV was shaken for 48 h at 25 °C in 15 mL pH 3 buffer containing increasing amounts of surfactant; the samples were filtered through 0.45 µm cellulose acetate and analysed by HPLC P3. The same titration was repeated with an excess of HPMCAS HF or HPMCP for SDS and poloxamer 188 (Figure 4), and for SDS with 0.5% PVPVA plus HF (Figure 7b) R11·R12·R18. The leaching vehicles then held 0.055% SDS, 1.54% poloxamer 188, 0.022% polysorbate 80 or 0.81% STC (w/v), 10% above each CMC measured without polymer P4. These levels were set from pH 3 measurements only; the pH 5 experiments presumably used the same concentrations, and the SDS level of the PVPVA experiment is not stated X10.

Leaching experiments

One hundred milligrams of ASD (or of a pure spray-dried polymer) was weighed into a 20 mL vial with 10 mL of 50 mM citrate buffer at pH 3 or pH 5, containing the chosen surfactant and/or PVPVA, and stirred at 300 rpm at 25 °C in a water-jacketed block P4. Samples were taken between 5 and 180 min, in triplicate, through 0.45 µm cellulose acetate syringe filters P4. LPV was measured by reversed-phase HPLC at 210 nm (LOD 0.05 µg/mL, LOQ 0.17 µg/mL) P7. The cellulose polymers were measured by the phenol–sulfuric acid colorimetric method at 490 nm, with LOD and LOQ of 1.0 and 3.2 µg/mL for HPMCAS HF, 1.3 and 4.4 for MF, 1.5 and 5.1 for LF and 1.3 and 4.4 µg/mL for HPMCP P8. Poloxamer 188, STC and PVPVA interfered with the colorimetric assay at the concentrations used, so polymer leaching could not be measured in those vehicles P8. The suspensions were checked by polarized light microscopy, and no crystallization was reported during the experiments R21.

Note by the AI writer (not discussed in the paper): the printed reagent volumes of the colorimetric assay (2 mL of sample, 50 mL of phenol solution, 5 mL of concentrated sulfuric acid) look inconsistent, the phenol volume being 25 times the sample; the paper does not allow the volumes to be checked, so they should not be copied from it X13.

Solubility measurements

Crystalline LPV solubility in each leaching vehicle was measured with excess crystalline LPV, with and without an excess of polymer, after 48 h at 25 °C; the phases were separated by ultracentrifugation (35000 rpm, 30 min, 25 °C) before HPLC, with at least three replicates P5. These values make up Table I and serve as $C_{\mathrm{e}\mathrm{q}}$ R5. The amorphous solubility was measured only once, in pH 3 buffer: a 5 mg LPV film cast from methanol/dichloromethane onto a vial was stirred in 10 mL buffer for 3 h at 25 °C and the supernatant analysed after centrifugation, with polarized light microscopy before and after to exclude crystals P6·R6.

Note by the AI writer (not discussed in the paper): three different separations were used for the three key concentrations: 0.45 µm filtration for the leached concentration $C$, ultracentrifugation for $C_{\mathrm{e}\mathrm{q}}$, and centrifugation of unstated conditions for the amorphous solubility. The ratio $C/C_{\mathrm{e}\mathrm{q}}$ therefore mixes methods, which is one more reason to treat it, as the authors do, as apparent X8.

Statistics

One-way ANOVA with Tukey's test at p < 0.05 was used P9. Only one test result is reported in the text: LPV leaching from the three HPMCAS grades was "not statistically different" R4. Other statements that an effect was or was not "significant" carry no reported test X12.

Experimental analysis

The polymer sets the baseline: little polymer, a few µg/mL of drug

Figure 2 gives the baseline at pH 3 without additives. The pure spray-dried polymers dissolved very little: less than 100 µg/mL in every case, under 1% of the solids R2·C8. Averaged over 60–180 min, 15 ± 2 µg/mL of LF, 19 ± 2 of MF and 68 ± 11 of HF dissolved, and less than 5 µg/mL of HPMCP R2. The authors call the high solubility of HF "somewhat counterintuitive", since HF has the highest pKa, and suggest that the grades "may" contain different amounts of a small low-pH-soluble polymer fraction left by incomplete or variable substitution during manufacture C9·C10. From the LPV dispersions, polymer leaching fell to below 5 µg/mL for every grade R3. The authors conclude that "it is clear" HPMCAS leaching was "substantially suppressed" by incorporating LPV, while noting that these values lie close to the detection limit, with high experimental errors; for HPMCP the leaching was too low to tell C11·C12. Even this little polymer, they argue, is not negligible, since very low polymer concentrations can inhibit crystallization C13.

LPV leaching rose within 15 min and reached about 7 µg/mL for all three HPMCAS grades and about 5 µg/mL for HPMCP (Figure 2c) R4. The HPMCAS grades were "not statistically different", although HF polymer is about four times more soluble than LF or MF at pH 3 C14. Table A collects the baseline values.

Table A. Baseline at pH 3, 25 °C, 100 mg ASD in 10 mL. Pure-polymer values are text averages over 60–180 min; dispersion polymer values are the range read from Figure 2b; LPV leaching from the bars of Figure 8a; crystalline solubilities with HF, HPMCP and in buffer from Table I, with LF and MF read from Figure 8a (approximate) R2·R3·R19·R5·R6·M2–M5·P8.

Every suspension was supersaturated with respect to crystalline LPV, which the authors note carries a risk of solution crystallization; yet, "interestingly", the maximum LPV concentration from the HPMCAS dispersions stayed below the amorphous solubility of about 16.6 µg/mL C15·C16·R6. The authors interpret the suppressed solubility of both LPV and HPMCAS in the dispersions, compared with each pure form, as "consistent with a miscible solid dispersion" C35. That the three grades leach the same amount of LPV fits drug–polymer interactions that "would be expected to be similar" across grades; the lower leaching with HPMCP "might suggest" more favourable LPV–HPMCP interactions, or a higher water content in the spray-dried particles C36·C37. Neither interactions nor water content were measured C37.

Note by the AI writer (not discussed in the paper): the polymer values from the dispersions should be read qualitatively. The LF and MF values lie below the stated limits of detection, the HF values near the limit of quantitation X7. Other figures disagree with the "below 5 µg/mL" figure: Figure 3d plots about 6.3–7.5 µg/mL of HPMCP leached from LPV–HPMCP between 30 and 120 min, above the pure HPMCP values of Figure 2a, and Figure 6b plots about 6–11 µg/mL of LF from LPV–LF at pH 3, against about 0.05–1.15 in Figure 2b X6.

Surfactants raise leaching in step with crystalline solubility

Figure 3 shows the surfactant series for the HF and HPMCP dispersions at pH 3. With LPV–HF, all four surfactants increased LPV leaching. STC raised it most, to about 25 µg/mL, which the authors describe as a 3.5-fold increase; poloxamer 188 raised it only "marginally" R7·C17. With LPV–HPMCP, only polysorbate 80 and STC raised LPV leaching, SDS and poloxamer 188 had "no significant impact", and for each surfactant the leaching was lower than with HF R9·C18.

The crystalline solubility in the same media (Table I) moved the same way. Without polymer, the surfactants changed LPV solubility little (2.2–3.3 µg/mL against 2.5), which the authors find "consistent with micellar solubilization" just above the CMC R5.a·C19. With polymer present the effects grew: with HF, solubility rose from 1.9 to 3.0 (poloxamer 188), 3.3 (SDS), 4.5 (polysorbate 80) and 7.2 µg/mL (STC); with HPMCP from 2.7 to 2.7, 3.1, 3.2 and 6.0 µg/mL R5.b·R5.c. With STC and either polymer, solubility doubled or more relative to STC alone (3.0 µg/mL), and the authors count this as enhanced solubilization by a surfactant–polymer pair R5·C19·C43. Figure 8 sets leaching and crystalline solubility side by side for every system, and the authors conclude that the two "mirror each other very well": the extent of drug leaching "appears to be increased based on the solubilization power of surfactant" R19·C39. Table B shows the numbers and the resulting ratio.

Table B. Surfactant series at pH 3 (µg/mL). Leaching read from Figures 8b and 8c and ratios from Figure 9 (approximate); Ceq from Table I R19·R5.b·R5.c·R20·X2.

The ratio $C/C_{\mathrm{e}\mathrm{q}}$ is the paper's key result. It stayed at about 3.25–3.9 for every HF system and about 1.55–1.8 for every HPMCP system (Figure 9) R20. The authors conclude that for ASDs made with a given polymer the ratio "remained the same", so the surfactants changed the apparent solubility and the leaching of LPV but "did not appear to significantly impact" its supersaturation, assuming $C/C_{\mathrm{e}\mathrm{q}}$ is a reasonable approximation; in the conclusions this becomes supersaturation that "remained approximately constant for a given lopinavir polymer system" C40. The supersaturation "does appear to depend" on the polymer, with HPMCAS-HF suspensions about twice as supersaturated as HPMCP suspensions at the same loading C47·R20.

Note by the AI writer (not discussed in the paper): recomputing the ratios from the paper's own inputs (Figure 8 leaching over Table I solubility) reproduces nine of the ten bars of Figure 9 within plot reading. The exception is HPMCP with polysorbate 80: about 7.45 µg/mL leached over 3.2 µg/mL gives about 2.3, whereas Figure 9b shows about 1.55 X2·E1. If the inputs are right, the HPMCP series spans about 1.55–2.3 rather than 1.55–1.8, and the "constant ratio" for HPMCP is less tight than the figure suggests.

The polymers change how the surfactants work

Two surfactant–polymer pairs behaved unexpectedly, and the solubility titrations of Figure 4 are the authors' explanation. For SDS, LPV solubility began to rise at about 0.05% SDS both without polymer and with excess HF; with excess HPMCP the break moved to about 0.1% R11. The authors conclude that HF "did not alter" the SDS CMC while HPMCP increased it, so at the 0.055% used for leaching SDS "was probably actually below its CMC as a result of leached HPMCP", which accounts for its lack of effect on both LPV solubility and leaching with HPMCP C20·C21. For poloxamer 188, LPV solubility without polymer was flat up to about 1.4% and rose to about 6.2–6.7 µg/mL at 4%; with HF or HPMCP it stayed at about 0.6–3.3 µg/mL over 0.5–4.5% with no visible break R12. The authors read this as the polymers having "appeared to decrease" poloxamer's solubilizing power, and "this observation explains" why poloxamer 188 had little effect on leaching C22.

Note by the AI writer (not discussed in the paper): the titration curves and Table I give different solubilities for the same media. With HPMCP and no surfactant Figure 4c gives about 0.4 µg/mL against 2.7 ± 0.3 in Table I; with HF at 1.5% poloxamer 188 Figure 4e gives about 0.9–1.4 against 3.0 ± 0.1 (at 1.54%) X3. The two methods differ (0.45 µm filtration for the titrations, ultracentrifugation for Table I), and the Figure 9 ratios use Table I X3·P3·P5. By Table I, poloxamer 188 did raise LPV solubility with HF, from 1.9 to 3.0 µg/mL, and leaching from about 7.3 to about 10, increases not far from those of SDS; so the claim that the polymers suppress poloxamer's solubilizing power rests on the titration data alone X4. The Figure 4 caption also mentions an "impact of pH" that the panels do not show X3.

At pH 5 an ionic surfactant multiplies leaching

Raising the pH from 3 to 5 tests what a small degree of polymer ionization does (Figure 5). Without surfactant, pH made no significant difference to LPV leaching from LPV–HF (about 7 µg/mL at both), and polymer leaching was similar within experimental error R13.a–b·R14.c–d·C23. With SDS the picture changed: average LPV leaching rose from about 13 to about 36 µg/mL and HF leaching from about 21 to about 120 µg/mL R13·C24. The crystalline solubility of LPV with SDS and HF also rose, from 3.3 to 6.3 µg/mL, which the authors read as SDS having a higher solubilizing power at pH 5 R5.b·R5.e·C25. With STC leaching also increased at pH 5, by less (about 25 to 30 µg/mL); with polysorbate 80 and poloxamer 188 neither drug nor polymer leaching changed significantly, apart from a slight rise in polymer leaching with polysorbate 80 that did not affect drug leaching (the text points to supporting information that was not supplied) R15·R14·C26.

The HPMCAS LF dispersion, with the lower pKa and so more ionized at pH 5, was tested with and without SDS (Figure 6) C27. Without SDS, more LF leached at pH 5 (about 20–47 µg/mL against about 6–11) but LPV leaching did not change (about 6–8 µg/mL), comparable to the HF dispersion R16·C28. With SDS at pH 5, LPV rose to about 30 µg/mL at 15 min, settling to about 23–25, and LF to about 216–326 µg/mL R16. Table C summarizes the pH series.

Table C. pH 3 against pH 5 (µg/mL). The SDS–HF values are the text averages; other leaching values are ranges read from Figures 5 and 6 over 60–180 min (approximate); Ceq from Table I; "not determined" where poloxamer 188 or STC interfered with the polymer assay R13·R14·R15·R16·R5.b·R5.e·P8.

The authors summarize the pH result as "a small amount of polymer ionization significantly enhanced LPV leaching in solutions containing an ionic surfactant" C24. They do not claim to know why. Both HPMCAS and SDS carry negative charge, so repulsion "might be anticipated"; the mechanism is "currently unclear", and they "can only speculate" that ionized chains become less coiled, which "may lead to" more contact between SDS and the less polar, uncharged groups of HPMCAS and to solubilizing aggregates C49. The degree of ionization, the polymer conformation and the aggregates were not measured C24·C49. They also note that surfactant–polymer complexation "can also enhance" polymer solubility, which matches the large rise in HPMCAS leaching with SDS C44.

Note by the AI writer (not discussed in the paper): the paper reports no $C/C_{\mathrm{e}\mathrm{q}}$ at pH 5. Calculated from the Figure 5 plateau readings and the pH 5 row of Table I, the ratio for LPV–HF is about 4.8 without surfactant, 5.5 with SDS, 4.0 with poloxamer 188, 2.7 with polysorbate 80 and 3.7 with STC X5. At pH 5 the ratio is therefore not constant, and with SDS it is the highest in the study X5.

SDS together with PVPVA gives the largest effect

PVPVA, added at 0.5% w/v as a suspending agent and crystallization inhibitor, raised LPV leaching from LPV–HF only a little on its own, from about 7 to about 9 µg/mL (60–180 min averages; Figure 7a) R17. With SDS added as well, leaching rose to about 50 µg/mL, "more than a five-fold increase" R17·C29. The crystalline LPV solubility with HF, PVPVA and SDS was 11.4 ± 0.8 µg/mL against 3.3 ± 0.2 with HF and SDS, so experiments with crystalline LPV "suggested" a considerably larger solubilizing capacity of SDS when PVPVA is present R5.d·R5.b·C30. In the titration with PVPVA and HF, LPV solubility rose steadily with SDS, to about 58–64 µg/mL at 0.3%, and the authors place the onset, a CAC, below 0.01% SDS, against a CMC of 0.05% without PVPVA R18·C42. They conclude that the SDS–PVPVA complex "clearly" has a higher solubilizing capacity for LPV than SDS alone, and that PVPVA and SDS "may be able to form a complex" more effective than either component C43·C31. In the conclusions this combination "was found to have the most profound effect" on LPV solubilization C50.

Note by the AI writer (not discussed in the paper): in Figure 7b the points at 0 and 0.01% SDS are similar (about 1.0–2.0 µg/mL) and the first clear rise is at 0.025%, so the plotted data place the onset between about 0.01 and 0.025% rather than clearly below 0.01% X9. The ratio for this system, about 47.7 over 11.4, is about 4.2, close to the HF series of Figure 9 X5.

Discussion

One picture: leaching follows solubilization

The authors pull the results into a single picture. The enteric polymer keeps the particles intact in acid, so only a small amount of drug and polymer leaches C8·C15. In a miscible dispersion the drug and the polymer hold each other back, which places the leached LPV between its crystalline and amorphous solubility, at a level that depends on the polymer C35·C16·C14. Anything that solubilizes crystalline LPV in the vehicle then raises leaching in proportion, leaving the apparent supersaturation about where the polymer put it C39·C40. The special cases are explained by how the surfactant interacts with the polymer in solution: HPMCP lifting the SDS CMC, both polymers blunting poloxamer 188, PVPVA and SDS forming a complex, and slight HPMCAS ionization at pH 5 enhancing SDS solubilization C21·C22·C31·C24·C49.

Two parts of this picture are directly measured: the parallel rise of leaching and crystalline solubility across media (Figure 8), and the near-constant ratio at pH 3 (Figure 9) R19·R20. The rest is interpretation, with the authors' own hedges: the miscibility reading ("consistent with"), the interaction readings ("would be expected", "might suggest"), the CMC shift ("probably"), the complex ("may be able to form") and the ionization mechanism ("We can only speculate") C35·C36·C37·C21·C31·C49.

What it means for formulation, as the authors state it

The authors draw three practical points. Reducing leaching "is likely to be critical" for drugs that crystallize faster than LPV, where supersaturation must be kept low, so the selection of polymer "is likely important" in controlling drug leaching and suspension stability C38. Because HPMCP gave a lower supersaturation than HPMCAS-HF, "it might be anticipated" that HPMCP suspensions carry a lower crystallization risk; this was not tested C48. And since the supersaturation persists when solubilizing surfactants are present, "it would clearly be important" to understand the effect of each surfactant on crystallization kinetics, given that surfactants can either promote or inhibit crystallization (Chen 2015, refs 12 and 31) C45. They expect the findings to "aid in the development of formulations with adequate stability to crystallization" C51.

Note by the AI writer (not discussed in the paper): for someone formulating an acidic dosing vehicle, the useful reading is that a surfactant added for wetting is not a crystallization safeguard. At pH 3 it raised the dissolved drug concentration without lowering the apparent supersaturation, so any crystals that do form would grow from a richer solution. An anionic surfactant combined with a vehicle pH near the start of HPMCAS ionization, or SDS combined with PVPVA as a suspending agent, can multiply the dissolved drug several-fold; vehicle pH and the full excipient list therefore matter more than the polymer grade, which made no difference to LPV leaching here R13·R17·R4.

Limitations

Limits stated by the authors. They call the supersaturation "apparent" because a concentration ratio in micellar or polymer-containing media "may or may not" represent the thermodynamic driving force C7. Their constant-ratio conclusion holds "for the surfactant levels used in this study", and assumes that $C/C_{\mathrm{e}\mathrm{q}}$ is a reasonable approximation C40. The polymer concentrations leached from the dispersions were close to the limit of detection, so experimental errors were high C12. Crystallization was not an issue only because LPV, a slowly crystallizing compound, was "deliberately selected", which "is unlikely to be the situation for many other compounds" C46. The mechanism behind the pH 5 effect is "currently unclear" C49.

The AI writer's own judgement (1): the constant ratio is narrower than it sounds. The near-constant $C/C_{\mathrm{e}\mathrm{q}}$ is shown for one pH (3), one drug loading and five media per polymer, without a statistical test X5·R20. The HF ratios span about 3.25–3.9, and one HPMCP bar does not reproduce from its own inputs (about 1.55 plotted, about 2.3 recomputed) X2. At pH 5, ratios calculated from the paper's data range from about 2.7 to 5.5 X5. The abstract's "proportional" therefore holds within one polymer at pH 3, not across polymers or pH X5·C39.

The AI writer's own judgement (2): two solubility data sets disagree. The surfactant titrations (Figures 4 and 7b, filtration) give lower LPV solubilities than Table I (ultracentrifugation) in the same media, by up to about sevenfold for HPMCP without surfactant (about 0.4 against 2.7 µg/mL) X3. The claim that the polymers blunt poloxamer 188 rests on the titrations, whereas Table I shows poloxamer raising the solubility with HF X4. Leached concentrations, crystalline solubility and amorphous solubility were each separated differently X8.

The AI writer's own judgement (3): polymer leaching and mechanisms. Polymer leaching from the dispersions without surfactant lies below the detection limit or near the quantitation limit and differs between figures, so "substantially suppressed" is a qualitative finding X6·X7. HPMCP leaching with SDS and polysorbate 80, which the text calls marginally higher, is two- to threefold higher in Figure 3d X14. Miscibility, drug–polymer interactions, polymer ionization and the surfactant–polymer complexes are inferred, not measured C35·C37·C24·C31. The amorphous solubility was measured only in plain pH 3 buffer, so the leaching levels reached with surfactants or at pH 5 (up to about 50 µg/mL) have no amorphous reference X15. The study covers one drug at one loading, 25 °C and 3 h P4·A1.

Conclusion

The paper measures how much drug and polymer leach from enteric-polymer ASD particles held in acidic vehicles, and how vehicle additives change it. At pH 3, LPV reached about 7 µg/mL from HPMCAS dispersions of any grade and about 5 µg/mL from HPMCP, between the crystalline and amorphous solubility, while polymer leaching stayed below 5 µg/mL R4·R5·R6·R3. Surfactants raised leaching roughly in proportion to how much they raised the crystalline solubility, so the apparent supersaturation stayed near 3.3–3.9 with HPMCAS-HF and 1.6–1.8 with HPMCP R19·R20·C39·C40. Two vehicle changes multiplied leaching: SDS at pH 5, where slight polymer ionization is the authors' proposed but unexplained cause, and SDS with PVPVA, read as a solubilizing complex R13·R17·C24·C31. For a formulator, the lesson is that a vehicle surfactant changes how much drug is dissolved but not, at pH 3, how supersaturated it is, and that the crystallization consequences still have to be measured for each drug and vehicle C40·C45.