Why an enteric polymer dissolves slowly, and how turning it into a salt fixes it: HPMCP, buffers and a miconazole ASD

Background and motivation

Enteric polymers carry carboxylic acid groups. They stay insoluble in the stomach and dissolve once the pH of the intestine passes a threshold, which is why they were first used as tablet coatings. The coatings behave well in the standard laboratory test, 50 mM phosphate buffer at pH 6.8, where they break up quickly. In people they often do not: scintigraphy studies found a lag before enteric-coated tablets disintegrate in the small intestine (Catteau et al., 1994; Lehmann et al., 1997; Wilding et al., 1992). The explanation offered in the literature, and adopted by the authors, is that intestinal fluid has a much lower buffer capacity than 50 mM phosphate. When the polymer ionizes it releases protons, a weak buffer cannot remove them fast enough, the pH right at the polymer surface drops, and dissolution slows. A lower pH near a dissolving enteric film was measured by Harianawala et al. (2002) and agrees with models of Ozturk et al. (1988) C7.

Over the last decade the same polymers, notably hydroxypropylmethyl cellulose acetate succinate (HPMCAS) and hydroxypropylmethyl cellulose phthalate (HPMCP), have become carriers for amorphous solid dispersions (ASDs), where the drug is dispersed at the molecular level in the polymer. An ASD differs from a coating in two ways that matter here. First, the polymer has to dissolve, not merely rupture, to free the drug. Second, the drug and polymer are mixed molecule by molecule, so the drug can change how the polymer dissolves and the polymer can change how the drug dissolves; the authors' group had found that different drugs slow enteric polymers to different extents (Hiew et al., 2022; Saboo et al., 2020; Yang et al., 2021) C9·C12. The effect of buffer capacity on enteric-polymer ASDs had not been widely studied C8.

What actually limits the dissolution of an acidic polymer is disputed C11. Reiser's percolation model (Yeh et al., 1992) treats the gel layer as a network of ionizable sites and makes the hopping of solvent species, including the base cation, from one site to the next the slow step; dissolution starts only when a critical number of sites per chain is ionized. Willson's model (Tsiartas et al., 1997) lets chains leave only from the outer surface, at a rate set by their degree of ionization. Both fitted low-molecular-weight polymers whose chains are not entangled. Nguyen and Fogler (2005) modelled an entangled enteric polymer and made the buffer concentration and pKa, relative to the surface and bulk pH, central, together with hydrodynamics and chain disentanglement. The percolation picture predicts that the size of the cation matters; the proton-carrier picture predicts that the buffer's pKa and concentration matter.

The paper sets out to separate these factors for HPMCP grade HP-50, and then to use what it finds. The plan has three steps: dissolve HP-50 in buffers that differ in cation size, buffer pKa and concentration; make HP-50 salts with sodium and eight amine or ammonium bases to test "pre-ionization" and the role of the counterion; and compare drug release from ASDs of the lipophilic weak base miconazole made with the protonated polymer and with two salts. The stated practical aim is a polymer that gives good release at higher drug loadings, a recurring difficulty with ASDs C1.

Background concepts

HPMCP and its salts

HPMCP is cellulose carrying methyl, hydroxypropyl and phthalyl substituents (Figure 3). The phthalyl group is a half-ester of phthalic acid, so each one carries a free carboxylic acid. The paper uses 1.61 mmol of phthalic acid per gram of HP-50 as the basis for adding one equivalent of base M1. The reported pKa of HPMCP is about 4.2 (Spitael and Kinget, 1979; Davis et al., 1986), so at pH 6.8 the polymer in solution is extensively ionized and soluble C23. A polymer salt, written HP-50-X, has these carboxylic acids already neutralized, the carboxylate paired with a cation X: an ammonium ion from an amine (R₁R₂R₃NH⁺), tetrabutylammonium or sodium (Figure 3). The authors call this "pre-ionization".

Surface pH and proton carriers

When an acid dissolves into a medium whose pH is above its pKa, each molecule that ionizes on entering solution frees a proton at the interface. The protons can leave in three ways: by diffusing across the aqueous boundary layer, by meeting hydroxyl ions diffusing in from the bulk, or by binding to a buffer species (a proton carrier) that diffuses away and releases the proton in the bulk. Ionization is fast but diffusion takes time, so a pH gradient forms, lower at the surface than in the bulk. Its size depends on hydrodynamics, on the solute's pKa and solubility, and on the buffer's pKa and concentration (Mooney et al., 1981; Higuchi et al., 1958; Hens et al., 2022) C24. For HP-50 films in a phosphate buffer of similar buffer capacity to the one used here, the surface pH was reported to be about 0.4 units below the bulk (Harianawala et al., 2002) R18. A good proton carrier must be able to take up a proton at the lower surface pH and give it up at the bulk pH, so its pKa should lie close to both C26.

Buffer capacity

Buffer capacity is the amount of acid or base a medium can absorb per unit change in pH, reported here in mM·ΔpH⁻¹. The 50 mM phosphate buffer used as the default medium has a measured capacity of 29 mM·ΔpH⁻¹; diluted to 5 mM it has 3 mM·ΔpH⁻¹ M16·M17. Intestinal fluid varies within and between people and has a buffer capacity well below that of 50 mM phosphate (Vinarov et al., 2021; Hens et al., 2017; Kalantzi et al., 2006) C32.

How a polymer dissolves

Polymer dissolution involves water entering the solid, the chains becoming mobile and disentangling (reptation), and the freed chains diffusing across the boundary layer into solution (Hunek and Cussler, 2002). For an enteric polymer an ionization step is added: ionized chains take up more water, which plasticizes them, forms a gel layer and lets them disentangle C10. For neutral water-soluble polymers above the entanglement molecular weight, water ingress or disentanglement is usually the slow step. Nguyen and Fogler found that an acidic polymer can be limited either by mass transport or by disentanglement, depending on hydrodynamics and on the concentration of proton-carrying buffer species; in the mass-transport regime the H⁺ concentration at the interface decides the polymer's solubility there and hence its flux C25. The solvent fraction in the gel layer in turn sets how fast chains disentangle (Narasimhan and Peppas, 1996) C30.

Surface-normalized (intrinsic) dissolution

All dissolution tests in the paper use a rotating disk: 100 mg of powder is compressed into an 8 mm die so that only one face of 0.5 cm² touches the medium, and the die rotates at 100 rpm in 100 mL of buffer at 37 °C P10. Because the area and the hydrodynamics are fixed, release can be expressed as a rate per unit area, in mg·min⁻¹·cm⁻². The paper does not print how the rates were fitted (time window, mass basis) X5.

Congruent release from an ASD

An ASD releases its drug well when drug and polymer leave the solid together, at the same fraction of their content per unit time ("congruent release"). When the polymer leaves faster than the drug, the drug is left behind at the surface and release stalls. Many ASDs release congruently at low drug loading but fail as the loading of a lipophilic drug rises (Yang et al., 2021; Indulkar et al., 2019) C38. When the released drug exceeds its amorphous solubility, it can separate as a drug-rich phase that makes the medium turbid.

Methods in detail

The study has four parts, run on the same rotating-disk set-up (Section 3.9) P10. The method flow below shows how they connect.

Part 1: which property of the medium controls HP-50 dissolution

Problem. If the percolation model were right, the size of the base cation should matter; if proton transport is limiting, the buffer's pKa and concentration should matter. What they did. HP-50 compacts were dissolved in three sets of pH 6.8 media. (a) Phosphate buffer at equal molarity (50 mM) with five different cations: K, NH₄, Cs, Na and tetrabutylammonium (Bu₄N), whose sizes run from 182 pm (ammonium, from a Van der Waals volume of 25.28 ų assuming a sphere) through 227 pm (Na) and 343 pm (Cs) to 415 pm (tetrabutylammonium, 300.36 ų) M18–M22·E1. The hydrated radius is similar for K, Cs and ammonium and largest for Na (SI Table S1). The figure legend gives the cation concentrations as 70, 41, 72, 72 and 74 mM, so the cations were not at equal concentration even though the phosphate was X11. (b) Hydrochloric acid–amine buffers at 50 mM made from BIS-TRIS, Tris and TEA, compared with the 50 mM sodium phosphate reference (75 mM Na); more buffers are in the SI M23–M26. (c) Phosphate at 50 and 5 mM, with measured buffer capacities of 29 and 3 mM·ΔpH⁻¹ M16·M17. Note by the AI writer. The same compact geometry is used for the polymer, its salts and its ASDs, which makes the rates of all four parts directly comparable with one another.

Part 2: polymer salts and how they were characterized

Preparation. For the amine salts, 1.61 mmol of amine (one equivalent to the phthalic acid in 1.0 g of HP-50) was added to a clear solution of 1.0 g HP-50 in methanol/dichloromethane (10:90 v/v), stirred for 30–60 min at room temperature and dried by rotary evaporation at 40 °C; tris base needed extra methanol to dissolve. HP-50-Na was made by adding a sodium base (sodium methoxide is listed among the materials) and the tetrabutylammonium salt by salt metathesis; both procedures are in the SI P1·M2·M10. The eight amine or ammonium salts are coded PAMP (2-amino-2-methyl-1-propanol), PTHAM (tris base), PMP (morpholine), PBTP (BIS-TRIS propane), PDIP (N,N-diisopropylethylamine, DIPEA), PBTM (BIS-TRIS), PTEA (triethylamine) and PTBA (tetrabutylammonium). Their precursors differ widely: calculated pKa 6.5 (BIS-TRIS) to 10.8 (DIPEA), logP −4.7 (BIS-TRIS propane) to 2.1 (DIPEA), zero to six hydroxyl groups, and Van der Waals volumes of 90 to 300 ų (Table 1) M3–M10. This spread is what allows the authors to ask whether any property of the counterion matters.

Characterization. Sodium in HP-50-Na was measured by ICP-OES after drying at 80 °C under vacuum; the ammonium content of the amine salts by ¹H NMR; salt formation by NMR (SI Figure S9); the carbonyl region by transmission FTIR of spin-coated films; thermal stability by TGA at 10 °C/min to 500 °C; and the glass transition both by modulated DSC (onset, second heating, scanned to 140 °C) and by DMA of powder in compression (peak of tan δ, 2 °C/min, 1 Hz) P4–P9. Water uptake was measured by weighing 200 mg of dried powder in an open vial at 100% relative humidity and 37 °C for up to 96 h P3.

Part 3: how much ionization is needed

To separate "salt" from "ionized", the authors made BIS-TRIS salts with 20, 50, 80, 90, 95 and 100 mol% of base and measured their dissolution rates, with HP-50 itself (in sodium phosphate) as the 0% point (Figure 9) M8·R13. The axis is the amount of base added; the actual ionization of the partial salts was not measured X9.

Part 4: miconazole ASDs

Drug. Miconazole is a lipophilic, weakly basic antifungal, BCS class II, with pKa 6.5 and an amorphous solubility of about 5 µg/mL (both from Yang et al., 2021) M11. Preparation. ASDs at 20% drug loading (DL) were made with HP-50, HP-50-Na and PTEA by dissolving drug and polymer in dichloromethane/methanol (50:50 v/v), evaporating at 50 °C, drying 48 h under high vacuum and cryo-milling; PTEA ASDs can also be made in one pot by adding triethylamine to the HP-50 solution before the drug P2. Test. 100 mg of ASD powder (20 mg of drug) was compressed and dissolved as above. Miconazole was measured by HPLC after diluting each sample with water and methanol to a clear solution, so the reported drug release counts all drug in the medium, dissolved or in a drug-rich phase; the polymer was measured with a phenol–sulfuric acid colorimetric assay P11·P12·X15.

Experimental analysis

The buffer species matters far more than the cation

In 50 mM phosphate with five different cations, HP-50 release was linear in time and reached between about 22.6% (tetrabutylammonium) and about 35.7% (ammonium) after 60 min (Figure 1) R1. The authors conclude that cation size "has little impact", with tetrabutylammonium "slightly slower" C13. Note by the AI writer: sodium (about 28.4%) also sits clearly below potassium, ammonium and cesium, although the sodium ion is not the largest; the paper does not comment on this.

Changing the buffer species had a much larger effect (Figure 2) R2:

Table A. HP-50 release in buffers of different species (read from Figure 2, approximate; n = 3). pKa values: BIS-TRIS, Tris and TEA calculated (Table 1); phosphate 7.2 as stated in Section 5.1 R2·M16·M23–M25.

Across buffer species the release at 60 min spans about 25-fold, in the order of buffer pKa. The authors note that the slow buffers have high pKa and therefore low buffer capacity at pH 6.8, and the fast buffers a pKa close to the dissolution pH; rate "appear[s] to trend" with these properties, and there is "clearly no correlation" with cation size C14. The buffer capacities of the amine buffers are not printed in the main text, so this part of the trend cannot be checked from it X12.

The salts are ionized and much more hygroscopic

Elemental analysis gave 3.38–3.52 wt% sodium in HP-50-Na, which the authors convert to a molar sodium content of 91–95%, "indicating nearly complete ionization"; the ammonium content of the amine salts was "around 100%" by ¹H NMR R3·R4·C15.Card check: the 91–95% is reproduced by dividing the sodium per gram of sample by the 1.61 mmol phthalic acid per gram of HP-50; expressed per gram of the salt itself, the same sodium content would mean about 94–98% neutralization. The paper does not state its basis (note X10). The FTIR spectrum of HP-50-Na shows, beside the carbonyl band near 1720 cm⁻¹, a strong new band near 1600 cm⁻¹ that the authors assign to carboxylate (Figure 4) R6·C16. HP-50-Na dissolves in plain water, giving an NMR spectrum, whereas HP-50 gives no NMR signal after addition to water R16.

Water uptake at 100% RH and 37 °C (Figure 7) separates the polymers clearly R11:

Table B. Water uptake (read from Figure 7) set beside the dissolution rate in 50 mM phosphate (read from Figure 6); both approximate, n = 3. Putting the two figures in one table is the AI writer's choice R11·R10.

The salts were still gaining water at 96 h, so these are not equilibrium values X13. The authors observe that the BIS-TRIS salt, with its hydrophilic cation, took up more water than the DIPEA and tetrabutylammonium salts, "although the differences were smaller than might be anticipated" C19.

Every salt dissolves about three times faster than HP-50

In 50 mM phosphate, HP-50-Na dissolved almost completely by 80 min (about 99%), while HP-50 had released about 39% (Figure 5) R9. The abstract and results call this "~ 3-fold faster", the discussion "more than twice as fast" C3. Figure 6 extends the comparison to all nine salts: PTBA about 3.09, PBTM about 2.93, PAMP about 2.87, PMP about 2.86, PTHAM about 2.85, HP-50-Na about 2.80, PTEA about 2.53, PDIP about 2.39 and PBTP about 2.35 mg·min⁻¹·cm⁻², against about 0.92 for HP-50 R10. The authors see "no notable difference" among the counterions, "regardless of amine molecular weight or lipophilicity" C18. Neither the pKa of the base (which spans more than four units) nor its size or lipophilicity tracks the rate. The authors contrast this with the strong effect of the same bases when used as buffers for HP-50, in particular triethylamine: the PTEA salt dissolves at about 2.53, whereas HP-50 released only about 1.8% in 60 min in TEA buffer (the direct comparison is in SI Figure S8) C34·R2.Card check: the three slowest salts (PTEA, PDIP, PBTP, about 2.35–2.53) lie about 18–24% below the fastest; the SD bars overlap and no statistical test is reported (note X7).

The salt is insensitive to buffer capacity; the protonated polymer is not

Lowering the phosphate concentration from 50 to 5 mM, which lowered the measured buffer capacity from 29 to 3 mM·ΔpH⁻¹, slowed HP-50 from about 0.92 to about 0.14 mg·min⁻¹·cm⁻², the "6-fold reduction" of the text (Figure 8) R12·C4. HP-50-Na went from about 2.80 to about 2.45, which the authors describe as "the same" C4.Card check: the salt's rate is about 12% lower at the low buffer capacity, within the overlapping SD bars (note X6). This is the central experiment of the paper. It fits the proton-transport picture: if protons generated at the surface limit the protonated polymer, a weaker buffer should slow it, and a polymer that generates no protons should not care.

Rate rises with neutralization, most steeply between 80 and 100%

Table C. Dissolution rate of partially neutralized BIS-TRIS salts (read from Figure 9, approximate; n = 3) R13.

The rate rises over the whole range, gently up to 80% and then steeply: from 80 to 100% it almost doubles (about 1.68 to 2.93), the "almost 2-fold" step the authors emphasize C21. They explain it with a minimum degree of ionization that a hydrophobic ionizable polymer needs before it can dissolve (Heller et al., 1978), and with the critical ionization of about 83% that Hiew and Taylor estimated for HP-50 (Hiew et al., 2022), which "may explain" the steep part. Partially neutralized salts can ionize further in the medium but still need proton removal, "albeit to a lesser extent" than HP-50 C36·R19.

Salt ASDs release miconazole quickly and nearly together with the polymer

Table D. Miconazole and polymer release from 20% DL ASD compacts in 50 mM phosphate pH 6.8 (read from Figures 10 and 11, approximate; n = 3) R14·R15.

From the HP-50 ASD, release was "slow and incomplete": about 8.6% of the drug after 60 min and about 11.7% after 120 min, while the polymer reached about 23.3% and about 42.2%. The polymer left faster than the drug, which the authors call incongruent release R14·C22. The text gives this as "Only 10% ... after 120 min" in the results and "only ~ 10% ... in an hour" in the discussion; the figure lies between the two readings X2. The HP-50-Na ASD "dissolved very quickly": about 89.5% of the drug after 60 min, levelling at about 91%, with the medium turning turbid as the drug exceeded its amorphous solubility and formed a drug-rich phase. The PTEA ASD behaved similarly, at about 83.7% after 60 min and about 87% at the plateau R14·C22. Normalized drug release rates were about 0.22, 2.70 and 3.15 mg·min⁻¹·cm⁻² for HP-50, PTEA and HP-50-Na; the authors state that drug release from the HP-50-Na ASD was 14 times faster than from the HP-50 ASD (the PTEA ratio is about 12) R15·C5.

Three points need care when reading these numbers. First, the discussion says the salt ASDs "show complete drug release" within an hour, but the figures level off at about 87–92% X3. Second, the authors say the polymer from the HP-50-Na ASD released "at the same normalized rate as the drug", but the polymer fraction runs ahead of the drug fraction from 20 min onward and plateaus above 100% (about 107–112%), which points to an offset in the basis of the colorimetric polymer assay that the paper does not discuss X4. Third, the Figure 11 rates appear to be expressed per mass of ASD rather than per mass of drug: the drug release slope from Figure 10 divided by the 0.2 drug fraction reproduces them (for HP-50-Na, about 63% of 20 mg over 40 min on 0.5 cm² gives about 0.63 mg·min⁻¹·cm⁻² of drug, and dividing by 0.2 gives about 3.15) X5. Read that way, the HP-50-Na ASD dissolves at about the same rate as the neat HP-50-Na compact (about 2.80), whereas the HP-50 ASD (about 0.22) is far slower than neat HP-50 (about 0.92). Note by the AI writer: the same contrast shows in the polymer itself, which left the HP-50 ASD more slowly (about 42.2% at 120 min) than neat HP-50 compacts (about 60.3%) R14·R9; the paper does not comment on this.

Glass transitions and thermal stability

The glass transitions in Table 2 span a wide range R8. By DSC onset, the amine salts gave 54.9 °C (PTHAM), 58.1 (PMP), 68.2 (PAMP), 78.1 (PDIP), 84.0 (PTEA), 84.5 (PBTP) and 92.1 °C (PTBA), against 127.1 °C for HP-50; PBTM and HP-Na showed no DSC glass transition. By DMA (peak of tan δ) PBTM gave 135.7, HP-50 172.8 and HP-Na 231.3 °C. The 20% DL ASDs gave 83.4 (HP-50) and 77.2 °C (PTEA) by DSC and 142.0, 216.2 (HP-Na) and 116.7 °C by DMA, all below the neat polymers measured by the same method. The authors summarize the salts as ranging "from a low of 55 °C for the tris polymer salt to a high of 231 °C for the sodium salt" C17.Card check: this range mixes DSC onset and DMA peak values, which differ by about 46 °C for HP-50; the DSC scans stopped at 140 °C, below the DMA values of HP-Na and its ASD, which may be why no DSC transition was seen for them (note X8). TGA of HP-50-Na showed events at 174.2 °C (0.47% total weight loss), 267.7 °C (6.8%) and 327.2 °C (26.8%) R7.

Discussion

The authors' explanation, step by step

The discussion (Section 5.1) assembles the results into one causal chain. The chain is an interpretation of rate data; the surface pH, the proton flux and the hydration of the gel layer were not measured in this paper X16.

Within this chain, the buffer results are "largely rationalized" by pKa C26. BIS-TRIS (pKa 6.5) and phosphate (pKa 7.2) can take up protons at the lower surface pH and release them at the bulk pH. Assuming the surface pH is about 0.4 units below the bulk, as reported by Harianawala et al. (2002), BIS-TRIS has more conjugate base available at the surface than phosphate and so carries protons away more effectively. Tris (8.1) and TEA (10.2) are mostly protonated at both surface and bulk pH and therefore "cannot act as proton carriers, leading to slower polymer dissolution rates" C27·R18. Note by the AI writer: the pKa values used in this argument are calculated (Chemicalize) for the amines, and the 0.4-unit shift is borrowed from films; neither was measured for these compacts X16.

Why the percolation model is set aside

Earlier work had linked the dissolution rate of acidic polymers to the size of the base cation: Arcus (1986) found the rate falling non-linearly with the specific volume of the cation, "presumably" because larger cations diffuse more slowly, and Reiser used this in support of the percolation model C28. Here neither the buffer cation (Figure 1) nor the salt counterion (Figure 6) showed a size trend, which "indicates that a percolation model of dissolution is likely not appropriate" for HP-50 C29. The authors instead take Nguyen and Fogler's view, that the polymer's solubility at the interface, set by the surface pH and so by the proton-carrier concentration, is what matters, as "a better explanation" C29. They add that the surface pH also sets how much the gel layer hydrates, and that the solvent fraction in the gel sets the disentanglement rate (Narasimhan and Peppas, 1996) C30.

What pre-ionization does

Pre-ionization removes the generation of protons at the surface altogether. The authors list three observations that "confirm" the role of proton generation and removal: HP-50-Na dissolves more than twice as fast as HP-50 in 50 mM phosphate; its rate is not affected by buffer capacity, whereas HP-50 slows 6-fold; and it dissolves in water, where HP-50 does not C31·R9·R12·R16. They also point to hydration: all the salts take up far more water than HP-50, which aids water ingress, raises the solvent fraction of the gel layer and speeds disentanglement, so "the improved hydration extent accompanying salt formation is likely a key contributor" C35. They call it "somewhat surprising" that the counterion does not matter when the sodium salt takes up clearly more water than the ammonium salts C35. Note by the AI writer: this is the weakest link of the chain. Among the salts, water uptake does not rank the dissolution rates (HP-50-Na took up the most water but dissolved at about 2.80, PTBA took up less and dissolved at about 3.09), so whatever hydration does, it is not what separates one salt from another; the large step is between the protonated polymer and any salt X17.

The ASD result and drug loading

For the ASDs, the authors place the result in the context of a familiar problem: many ASDs release well at low drug loading but fail as the loading of a lipophilic drug rises, as the 20% DL miconazole–HP-50 ASD illustrates C38. They propose that the higher hydration of the polymer salts "helps to mitigate the deleterious impact of a lipophilic drug on the water solvent fraction in the gel layer" ("It is likely") C40. Because low-potency drugs need high drug loadings, they see "the potential advantage of polymer salts to enable higher drug loadings in ASD formulations without compromised release", and they expect that at higher loadings the differences in hydration between salts "may play a role" in when release fails, which they leave to a future study C39·C41.

What it suggests for formulation

Stated by the authors. A polymer salt in an ASD "may offer advantages" in dissolution rate and in robustness of that rate to the conditions of the intestine, whose buffer capacity is variable and much lower than that of 50 mM phosphate C33·C32. Salt formation makes the dissolution "more robust to variations in media composition, in particular buffer capacity" C43. The wide choice of pharmaceutically acceptable counterions makes this "a relatively unexplored approach" to adjust polymer properties and widen the ASD formulation and processing space C44. Neutralized enteric polymers are not new as such: they have been used to make aqueous enteric coatings, and partial neutralization with a double coating was proposed to shorten the lag of enteric coatings, mostly for methacrylic polymers; their use to improve ASD dissolution is what the authors call under-explored C37.

Note by the AI writer (not discussed in the paper). Two practical messages follow. First, the standard test medium flatters protonated enteric polymers: HP-50 lost 6-fold in rate when the buffer capacity fell from 29 to 3 mM·ΔpH⁻¹, so an HP-50 or similar ASD that releases well in 50 mM phosphate may not do so in low-capacity intestinal fluid; a low-capacity medium is a cheap screen. Second, the counterion barely changed the dissolution rate but moved the glass transition by tens of degrees (about 55 to 92 °C by DSC onset for the amine salts, 127 °C for HP-50), so the counterion could be chosen for processing temperature or stability. The other side of that freedom is that a salt that dissolves in plain water and takes up 25–42% water at 100% RH may lose the acid resistance that made the polymer enteric, and may be harder to keep stable in humid storage; neither was tested.

Limitations

Limits the authors state

The authors present the ASD work as "proof-of-concept studies" with one drug at one loading (20% DL), and state that the behaviour at higher drug loadings, where hydration differences between salts "may play a role", "will be evaluated in a future study" C41. Their mechanism is hedged in the discussion ("is likely", "appear to provide a better explanation", "may explain"), although the abstract presents the rate-limiting step as identified C2·C29·C36.

The AI writer's own judgement

• The mechanism is inferred, not measured. The surface pH, the proton flux and the hydration of the dissolving gel layer were not measured; the carrier argument uses calculated pKa values and a surface pH shift taken from earlier work on films. The rate data fit the proton-transport picture well, but they are the only evidence for it X16.
• Hydration does not rank the salts. The sodium salt took up the most water without dissolving faster than the amine salts, so hydration explains at most the step from protonated polymer to salt X17.
• Some statements are stronger than the figures. "Complete drug release" from the salt ASDs is about 87–92% in Figure 10; the polymer is "at the same normalized rate as the drug" only roughly, and its release exceeds 100%; "the same" rate for HP-50-Na at both buffer capacities is a 12% drop within the SD; "no notable difference" among the salts covers a spread of about 25% with no statistics X3·X4·X6·X7.
• Units and normalization are not stated. How the normalized rates were fitted is not printed, and the drug release rates of Figure 11 appear to be per mass of ASD X5. The "% ionization" of Figure 9 is the mol% of base added, not a measured ionization X9.
• Comparability. The buffer-cation series differed in cation concentration (41–74 mM); the buffer capacities of the amine buffers and the pH of the 5 mM phosphate are not printed; the T_g range mixes DSC onset and DMA peak values X11·X12·X8.
• Solid state and medium. No diffraction or polarized-light check of the ASDs or salts is shown; the drug-rich phase and turbidity are reported only for the HP-50-Na ASD, although the HP-50 ASD also exceeded the cited amorphous solubility of miconazole; drug release counts total drug, dissolved plus drug-rich phase X14·X15.
• Scope. One polymer grade, one drug at 20% DL, phosphate media at pH 6.8 without a gastric step, no biorelevant media, no stability study; the SI, which holds the salt preparations, the buffer compositions and several comparisons, was not available for this reading.

Conclusion

The paper makes a clear experimental case on three points. HP-50 dissolves at a rate that depends strongly on the buffer species (about 45% in 60 min in BIS-TRIS against about 2% in TEA) and on buffer capacity (6-fold slower at 3 than at 29 mM·ΔpH⁻¹), and hardly on the cation R1·R2·R12. Pre-ionized salts of HP-50 dissolve about three times faster, almost whatever the counterion, and the sodium salt barely notices the drop in buffer capacity R10·R12. And 20% miconazole ASDs made with the sodium or triethylamine salt release about 84–90% of the drug in an hour, against about 9% from the HP-50 ASD, with a normalized drug release rate about 12 to 14 times higher R14·R15. The explanation the authors give, that solubilization and hydration through ionization, limited by the removal of protons from the surface, is "likely" the slow step and that pre-ionization removes it, fits all of these results, but rests on rates rather than on measurements of the surface itself C2·C31·X16. For formulation, the work points to a lever that is easy to pull: neutralize the enteric polymer before making the ASD. Whether it holds at higher drug loadings, after a gastric step and in storage is the next thing to establish C39·C41.