A highly complex interacting system of parameters associated with absorption, distribution, metabolism, and excretion (ADME) determines the pharmacokinetic profile of orally administered drugs in relevant test species or in man. The absorption component of this cascade, a necessary but not sufficient step for drug bioavailability, can be assessed in the context of Fick's First law, where the flux (J) of a drug through the gastrointestinal wall depends on the permeability coefficient (P) of the gastrointestinal barrier for the drug and the drug concentration (C) in the gastrointestinal lumen (assuming sink conditions):
This relation is the basis to classify drugs in the biopharmaceutical classification system (BCS) according to permeability and solubility criteria.1
For poorly water-soluble drugs (BCS class II and IV), the maximum achievable intraluminal drug concentration may limit absorption. As described by Lipinski and others, the introduction of high-throughput screening and combinatorial chemistry in drug development has resulted in a shift such that more new chemical entities suffer from limited aqueous solubility and/or poor dissolution properties.2–4 Therefore, various approaches in drug formulation development have been directed at increasing the dissolution rate and improving drug solubilization in the gastrointestinal tract.5 It should be recognized, however, that the intraluminal concentration of a drug is not necessarily limited by its solubility in gastrointestinal fluids. Drugs may be in solution at a concentration above their saturation solubility, that is, in a state of supersaturation. The degree of supersaturation can be expressed by the supersaturation ratio S:
with Ceq representing the equilibrium solubility (saturation). An alternative way of expressing supersaturation is the relative supersaturation index σ, defined as
A solution is defined as unsaturated, saturated or supersaturated based on the following relationships: S < 1 (σ < 0), S = 1 (σ = 0), or S > 1 (σ > 0), respectively.
As the chemical potential of a supersaturated system is increased compared to the equilibrium condition (saturation), a supersaturated drug solution is thermodynamically unstable and has the tendency to return to the equilibrium state (lowest chemical potential) by drug precipitation. If, however, a supersaturated drug solution exists in a metastable state in the gastrointestinal lumen for a time period sufficient for absorption, the increased intraluminal drug concentration can result in an enhanced flux across the intestinal wall.
The potential impact of supersaturation on the transport of drugs across biological membranes was first recognized by Higuchi.6 Since then, the role of supersaturation has been most extensively studied in the field of transdermal drug delivery where the link between the saturation state of the drug substance in an applied vehicle and subsequent absorption is fairly well understood. By extension, drugs suffering from solubility-limited oral bioavailability may also profit from the generation of supersaturation in the gastrointestinal lumen.
This review focuses on the concept of supersaturation in the gastrointestinal tract as a strategy to enhance the intestinal absorption of poorly water-soluble drugs. We will discuss the approach of both generating and maintaining supersaturation and illustrate this with examples of supersaturating drug delivery systems. Next, the principles behind precipitation delay by excipients will be discussed more in detail. Finally, various issues regarding the in vivo relevance of intraluminal supersaturation and the biorelevance of supersaturation assays will be addressed.
THE SPRING AND PARACHUTE APPROACH
To exploit supersaturation as a strategy to improve intestinal absorption of poorly water-soluble drugs, two essential steps need to be considered: generation and maintenance of the metastable supersaturated state. Guzmán et al.7, 8 described this concept by using the term ‘spring and parachute approach’, as illustrated in Figure 1.
A thermodynamically unstable, supersaturated solution of a drug can only be generated starting from a higher energy form of the drug (as compared to the crystalline powder), that is, the “spring.” A number of formulation options may induce the generation of supersaturated solutions in the gastrointestinal lumen, including the delivery of drugs in solution (e.g., cosolvent systems, lipid-based formulations) and the delivery of high-energy solid forms that provide an accelerated dissolution and/or a higher apparent solubility (e.g., amorphous forms, crystalline salt forms, co-crystals and the like). It should be noted that the term “apparent solubility” describes the apparent equilibrium between drug in solution and a solid whose structure is not in the most stable state.9, 10 Apparent solubilities should not be confused with the “true” or equilibrium solubility, which describes the thermodynamic equilibrium between drug in solution and the most stable solid form.
Once supersaturation has been induced, drug molecules have the tendency to precipitate, through processes which may be kinetically or thermodynamically controlled. In order to benefit from the supersaturated state, the increased concentrations have to be maintained for a time period sufficient for absorption. This may require a temporary inhibition of precipitation through the use of pharmaceutical excipients or by other components that interfere with nucleation and/or crystal growth, that is, the “parachutes” or precipitation inhibitors. Throughout this review, the terms “precipitation inhibition” and “precipitation inhibitors” will refer to the delay or temporary inhibition in drug precipitation from supersaturated solutions.
Studying supersaturation in vitro requires the determination of drug concentrations in a test medium as a function of time after introducing a “spring” form of the drug. This “spring” might be a high-energy formulation of the drug or a highly concentrated solution in an organic solvent (cosolvent induced supersaturation). Weakly basic drugs can also be dissolved in an acidic medium (pH 1–2), thereby simulating dissolution in the acidic environment of the stomach before entering the intestine with increased pH (pH-shift method). It is crucial to choose the volume of test medium and the applied dose in such a way that complete dissolution of the dose exceeds the solubility of the drug in the test medium. The application of nonsink conditions in supersaturation dissolution testing is the major difference compared to conventional dissolution testing which is conducted under sink conditions and focuses on release kinetics. As supersaturation in vivo may be affected by intraluminal conditions, test media relevant for gastric and/or intestinal fluids are preferred, including USP simulated gastric or intestinal fluid (SGF/SIF), fasted or fed state simulated gastric fluid (FaSSGF, FeSSGF)11, 12 and fasted or fed state simulated intestinal fluid (FaSSIF, FeSSIF),13 which were recently revised (FaSSIF-v2, FeSSIF-v2).12 To investigate the role of excipients on the stabilization of supersaturation, they can be included either in the formulation or in the test medium. Stirring of the test medium ensures homogeneity and attempts to simulate the hydrodynamics in the intestinal tract. In order to obtain reliable drug concentrations, separation of dissolved versus undissolved molecules is crucial. Especially when working with nanoparticles, which are being investigated for their potential to generate supersaturation (see below), this separation step should be robust and requires the use of ultracentrifugation or filters with pore sizes in the nanometer range.14, 15
To measure the degree of supersaturation, the obtained concentrations need to be compared to the equilibrium solubility in exactly the same test medium (see Eq. 2). The following factors not only affect the degree of supersaturation but also drug solubility: composition of the test medium, presence of excipients, alterations in composition as a function of time and the influence of the cosolvent. Therefore, thermodynamic solubility assessment starting from the crystalline powder in the test medium (shake-flask method) is required in supersaturation studies.
SUPERSATURATING DRUG DELIVERY SYSTEMS (SDDS)
A variety of formulation approaches for poorly water-soluble drugs may induce intraluminal supersaturation. Table 1 reports a number of examples of drug delivery systems that induce supersaturation in vitro and have been investigated for their absorption enhancing capacity in vivo. They include solubilized formulations as well as physically and chemically modified high-energy solid forms.
Table 1. Selected Examples of Supersaturating Drug Delivery Systems
Spring Form (Type of Formulation)
Parachute (Precipitation Inhibitors)
In Vivo Performance (PK)
AUC, area under the curve (plasma concentration–time profile); Cmax, maximum concentration (plasma concentration–time profile); EPAS, evaporative precipitation into aqueous solution; HPC, hydroxypropyl cellulose; HPMC, hydroxypropyl methylcellulose; P407, poloxamer 407 (Pluronic® F127); PK, pharmacokinetics; PVA, polyvinylalcohol; PVP, polyvinylpyrrolidone; S-cosolvent, supersaturating cosolvent formulation; SDS, sodium dodecyl sulfate; SEDDS, selfemulsifying drug delivery system; SFL, spray freezing into liquid; S-SEDDS, supersaturating SEDDS; TPGS, d-α-tocopheryl polyethylene glycol 1000 succinate.
HPMC + HPMC phtalate (carrier)
Bioavailability ↑ in case of reduced gastric activity (vs. physical mixture, in rabbits)
Circumventing the dissolution step by administering the drug as a solution is a commonly used approach to enhance the intestinal absorption of poorly water-soluble drugs. Hydrophobic drugs are solubilized in the formulation in a mixture of hydrophilic cosolvents, water-soluble and -insoluble surfactants, complexing agents (e.g., cyclodextrins) and/or oils.16, 17 Combining different ratios of these components results in a variety of solubilized formulations. An important class are the lipid-based formulations, consisting of pure oils (triglycerides or mixed mono- and diglycerides) or a combination of oils with cosolvents and/or surfactants (oil-in-water emulsions or self-(micro)emulsifying drug delivery systems (S(M)EDDS)).18–21
Dilution and dispersion of a solubilized formulation in the gastrointestinal lumen typically forms a complex mixture of colloidal species consisting of both formulation components and endogenous species (including bile salts and phospholipids). Moreover, the composition of this medium will fluctuate with time, especially if products of lipid digestion are involved. The capacity of the gastrointestinal fluid to keep the drug in solution dictates to a large extent the success of these delivery systems. In case the solubilization capacity is insufficient, a metastable supersaturated state is generated, potentially resulting in drug precipitation. Therefore, an in vitro evaluation of the utility of drug solutions as a formulation requires the assessment of the rate and extent of precipitation in conditions simulating the dispersion and processing of the formulation (i.e., lipid digestion) in biorelevant media.22, 23 In general, it is observed that the likelihood of immediate drug precipitation upon dispersion is increased for formulations that contain relatively higher amounts of water miscible surfactants or cosolvents. The solubilizing capacity of cosolvents strongly decreases upon dilution as a consequence of the extended Hildebrand equation (logarithmic relation between cosolvent concentration and drug solubility).24, 25
Traditionally, attempts to avoid drug precipitation from solubilized formulations focus on maximizing solubilization in the gastrointestinal media formed upon dispersion. However, this approach does not always yield an improved oral bioavailability. Sometimes, it might not be possible to adjust a formulation in such a way that the solubilizing capacity upon dispersion is sufficient to keep the complete dose of a poorly water-soluble drug in solution. In addition, solubilization in the gastrointestinal tract does not necessarily result in an improved absorption: uptake across the intestinal barrier is limited to the free fraction of drug molecules in the intermicellar phase, which is in equilibrium with the fraction of drug molecules solubilized in the mixture of colloidal species (emulsified oil, mixed micelles, etc.).26, 27 In the case of strongly lipophilic drugs and excipients, the free drug concentration can be very low resulting in a poor systemic exposure, despite solubilization.28
When the limits of the solubilization approach are reached, it might still be possible to profit from the high intraluminal concentrations initially generated from solubilized formulations by maintaining the supersaturated state of the drug; addition of precipitation inhibitors may be necessary. Most evaluation procedures for solubilizing formulations do not strictly distinguish between solubilization and maintaining supersaturation as the mechanism responsible for limiting precipitation; therefore, it is possible that supersaturation plays a role in the absorption process of various solubilizing dosage forms, albeit unintended. Gao and Morozowich29 successfully employed the supersaturation strategy to formulate a number of poorly water-soluble drugs in so-called supersaturable SEDDS (S-SEDDS). In contrast to conventional SEDDS, which attempt to completely solubilize the drug by emulsification, S-SEDDS aim at generating high, supersaturated free drug concentrations in the gastrointestinal tract and reducing the precipitation rate to allow for sufficient absorption. To achieve this goal, S-SEDDS contain a reduced amount of surfactant and a precipitation inhibitor, usually hydroxpropyl methylcellulose (HPMC). This approach was originally evaluated using the antitumor agent paclitaxel as a model drug.28 Due to its extremely low aqueous solubility (<1 µg/mL) and its interaction with the efflux carrier P-glycoprotein, the development of a successful oral formulation is challenging. The solubilized formulation Taxol®, containing paclitaxel dissolved in ethanol and high concentrations of the surfactant Cremophor® EL, is used for intravenous administration but failed to yield therapeutic paclitaxel plasma concentrations upon oral administration in rats; presumably, this can be attributed to a low free concentration of paclitaxel, solubilized by high surfactant levels. Gao et al. developed a SEDDS containing a reduced amount of Cremophor® EL. Dilution of the SEDDS in simulated gastric fluid resulted in supersaturated concentrations of paclitaxel and subsequent precipitation. The precipitation rate could be decreased by including HPMC (5%) in the formulation; oral administration of this S-SEDDS in rats resulted in a significantly enhanced systemic exposure of paclitaxel (AUC 4.7-fold higher as compared to administration of Taxol®). Similar S-SEDDS were able to improve the absorption of the poorly water-soluble drug candidates PNU-91325 (for which a supersaturable cosolvent formulation was also developed) and AMG 517.30, 31 These results indicate that S-SEDDS, containing a reduced amount of surfactant and a precipitation inhibitor, are a valuable formulation option for the intestinal delivery of poorly water-soluble drugs by generating and maintaining increased free drug concentrations above the saturation solubility. Similar to the development of S-SEDDS, other solubilized dosage forms might be adjusted into supersaturating formulations.
High-Energy and/or Rapidly Dissolving Solid Forms
In addition to the delivery of drugs in solution, a number of formulation strategies that deliver drugs in a solid form can induce supersaturation in the gastrointestinal lumen. A prerequisite for the generation of the thermodynamically unstable state is that the drugs should be administered as a high energy or otherwise rapidly dissolving form.32 Drug particle engineering techniques (e.g., milling, cogrinding, solvent evaporation, melting, freezing techniques, crystal engineering, etc.) aim at generating high-energy or rapidly dissolving solid forms of drugs by altering the morphology, particle size and/or wettability.33 Less stable polymorphs or amorphous solids require less energy to dissolve, resulting in higher apparent solubilities and increased dissolution rates.34, 35 Co-crystals and crystalline salt forms may provide improved solubility and dissolution properties without being thermodynamically unstable.36 Particle size reduction and improved wettability may increase dissolution rates by enhancing the surface area available for dissolution. In addition to an increased surface area, nanoparticles may also provide an enhanced apparent solubility:37 according to a modified version of the Kelvin equation or the Ostwald-Freundlich equation, the apparent solubility of small particles (especially those below 200 nm) increases as a function of the high curvature of these particles.38, 39
A number of studies have reported the creation of supersaturation in vitro from solid form formulations that rely on one or more of these principles (see Tab. 1). These examples include solid dispersions, nanoparticles, coground mixtures, the use of inorganic matrices as carrier, crystalline salt forms and prodrugs of higher aqueous solubility.
Conventional Solid Dispersions
Solid dispersions usually contain amorphous drug particles dispersed in a hydrophilic carrier matrix of one or more polymers (e.g., polyvinylpyrrolidone (PVP), polyethyleneglycols (PEG), polymethacrylates, cellulose derivatives, inulin, etc.) and/or (polymeric) surfactants (e.g., Inutec® SP1, Gelucire®, poloxamer 407, etc.).40–42 Through vitrification, specific drug–excipient interactions and/or reduced mobility, the matrix stabilizes, to a certain extent, the dispersed amorphous drug particles or domains. Solid dispersions aim at generating high and possibly supersaturated intraluminal concentrations of poorly water-soluble drugs by increasing their apparent solubility and/or dissolution rate. In the extreme case of a molecular dispersion (solid solution), release of drug molecules is fully dictated by dissolution of the hydrophilic carrier. In other cases, codissolution of the drug with the hydrophilic carrier is improved by several mechanisms, including the higher apparent solubility, the carrier-induced increase in wettability and the increased surface area available for dissolution of the dispersed amorphous drug particles.43 The dissolution characteristics of solid dispersions depend to a large extent on the physical state (ideally: amorphous), drug dispersivity (ideally: molecular dispersion) and particle size. Therefore, drug loading, matrix composition and preparation technique will dictate the initial degree of supersaturation.44 The duration of supersaturation will depend on the presence of codissolving matrix components that act as precipitation inhibitor (as discussed below).
The commercially available capsule-based Sporanox® formulation is a solid solution relying on the principal of supersaturation to enhance the intestinal absorption of the antifungal itraconazole, a weak base (pKa = 4) with an extremely low and pH dependent aqueous solubility (ca. 1 ng/mL in water, 6 µg/mL in 0.1 M HCl). This formulation comprises a molecular dispersion of itraconazole in an HPMC matrix and coated on inert sugar spheres. Dissolution of HPMC in media simulating the gastric environment releases supersaturated concentrations which are maintained for at least 4 h. HPMC is believed to prevent itraconazole from precipitation in the stomach and in the intestine,45 resulting in significant absorption (maximum fraction absorbed ca. 85%) and oral bioavailability (ca. 55%).
Yamashita et al.46 investigated the dissolution in acidic medium of solid dispersions containing the macrolide lactone tacrolimus in an amorphous state comparing three different polymers (HPMC, PVP, and PEG 6000) as the carrier. They observed rapid dissolution and the creation of supersaturated concentrations of tacrolimus which were up to 25-fold higher than the equilibrium solubility (2 µg/mL). While the polymer choice did not affect the maximum degree of supersaturation, only HPMC could fully inhibit precipitation for up to 24 h. Administration of the HPMC-based solid dispersion to beagle dogs resulted in a 10-fold increase in Cmax and AUC as compared to administration of the crystalline drug powder.
Conventional solid dispersions often contain microparticles. The use of nanoparticles may further enhance the capacity to generate supersaturation, as a result of the increased surface area and enhanced apparent solubility due to the high curvature of the particles (as discussed above). This was illustrated by Matteucci et al.,14 who studied itraconazole supersaturation in acidic medium from freeze dried nanoparticles stabilized by surface-located HPMC (prepared by controlled precipitation) versus conventional HPMC-based solid dispersions (prepared by solvent evaporation). The higher surface area of the nanoparticles resulted in a much faster development of supersaturation. For instance, after 20 min of dissolution, the supersaturation ratio amounted to 64 for a 1:1 itraconazole:HPMC nanoparticle formulation (surface area of 8100 cm2) versus 14 for a solid dispersion with the same composition (surface area of 704 cm2).
Nanostructured solid dispersions of tacrolimus were prepared using ultrarapid freezing by Overhoff et al.15 They compared the effect of three different stabilizers (sodium dodecyl sulfate (SDS), polyvinylalcohol (PVA), and poloxamer 407 (P407)) on the maximum degree and stability of tacrolimus supersaturation in both acidic and pH-shift conditions (simulating transfer from the stomach to the intestine). Interestingly, they also dosed the different formulations to rats, which allowed the development of a relation between in vitro supersaturation behavior and in vivo absorption. The best in vivo absorption of tacrolimus was observed upon administration of the formulation containing P407 as stabilizer (largest Cmax and AUC). In vitro, this formulation dissolved rapidly and showed the highest degree of supersaturation in acidic medium. However, P407 performed poorer compared to SDS in stabilizing the supersaturated state. Thus, the in vitro/in vivo correlation suggested rapid uptake of tacrolimus, making the maximum degree of supersaturation more important than the stability of this state.
As the amorphous form of a drug has the highest apparent solubility, it is not surprising that amorphous-based dosage forms are a popular formulation strategy for poorly water-soluble drugs. However, amorphous materials are thermodynamically unstable and avoiding recrystallization during storage is a major issue in formulation development. Vogt et al.47 evaluated the possibility to generate supersaturation starting from crystalline EMD 50733. While the coarse material resulted in a very poor dissolution rate and no detectable absorption in dogs, micronized powder (in the presence of lactose) did improve the dissolution rate and bioavailability (37%) but no supersaturation was observed. Jet-milling of the physical mixture of EMD 50733 and lactose resulted in a coground mixture (still containing the drug in its crystalline form) which was able to induce limited supersaturation (up to 2.8-fold) and further increase the bioavailability (55%). Upon inclusion of HPMC in the coground mixture, precipitation could be avoided for at least 3 h and the bioavailability increased to 68%. Similar results were obtained for felodipine but no supersaturation could be created for albendazole and danazole.48 While cogrinding can be useful to increase dissolution rates, the creation of supersaturation can be variable and might not be sufficient for high-dose poorly water-soluble drugs.
Inorganic Carriers for Delivery of Poorly Water-soluble Drugs
Recently, inorganic materials (e.g., silica) have been studied as carrier for the delivery of poorly water-soluble drugs.49, 50 In this respect, Mellaerts et al.51 developed an itraconazole formulation using ordered mesoporous silica (OMS) as a carrier. OMS exhibits a two-dimensionally ordered array of cylindrical pores oriented parallel to each other and separated by thin walls. Itraconazole can be molecularly dispersed in these pores up to a certain loading (ca. 30% by weight). The influx and competitive adsorption of water to the pore surfaces provides for a rapid release of itraconazole. It was shown that the release of the weak base itraconazole from OMS gave rise to supersaturation in simulated gastric fluid; a subsequent pH shift to simulated intestinal fluid caused only limited precipitation and supersaturated concentrations were maintained for at least 4 h.52 Although the initial degree of supersaturation in acidic medium was not as pronounced as compared to the earlier mentioned Sporanox® oral capsule formulation, less precipitation was observed after the pH increase. As a precipitation inhibitor (HPMC) was included in Sporanox® but not in the OMS formulation, the results were somewhat difficult to interpret. In comparison with the administration of crystalline powder, the oral bioavailability of itraconazole in rabbits and dogs was significantly enhanced upon administration of both OMS and Sporanox®.53
Crystalline Salt Forms
Crystalline salt forms of weak bases or acids often provide faster dissolution and a higher apparent solubility as compared to their unionized counterparts. Therefore, dissolving a salt form of a drug may induce supersaturation. An advantage of using a crystalline salt form instead of the amorphous phase is their better stability during storage. An interesting example of the potential use of crystalline salt forms to induce supersaturation and improve intestinal absorption has been reported by Guzmán et al.8 Celecoxib is a poorly water-soluble (ca. 1 µg/mL in water) anti-inflammatory drug which is administered orally at relatively high doses (100–400 mg). The bioavailability of the marketed formulation (Celebrex®), which contains the unionized form of the weak acid celecoxib, ranges from 22% to 40% in dogs. Celecoxib salts (e.g., sodium and sodium propylene glycol salts) provide an increased apparent solubility, but recrystallize as the free acid almost immediately after dissolution. However, the induced supersaturated state in SGF (>10-fold the equilibrium solubility of the free acid) could be maintained for at least 30 min by inclusion of a mixture of a surfactant (TPGS or Pluronic® F127) and hydroxypropyl cellulose (HPC) (2 mg/mL both) as precipitation inhibitors. The bioavailability of a formulation comprising the salt form (as “spring”) and precipitation inhibitors (as “parachutes”) amounted to more than 90%, compared to 30% for the formulation containing the free acid.
A special case of circumventing the dissolution issues of a poorly water-soluble drug is the administration of a prodrug with improved solubility and/or dissolution characteristics.54 If the permeability for the prodrug is decreased (e.g., by the addition of an ionizable promoiety), this approach requires conversion of the dissolved prodrug into the parent drug in the gastrointestinal tract. Examples include the phosphate ester prodrugs of poorly water-soluble drugs phenytoin or amprenavir, that is, fosphenytoin and fosamprenavir, respectively. The charged prodrug rapidly dissolves in the gastrointestinal lumen, but diffusion across biological membranes is limited meaning that dephosphorylation by intestinal alkaline phosphatase is required for intestinal uptake. Due to the lower solubility of the parent drug, a supersaturated solution may be generated, increasing the driving force for absorption.55, 56 Potential precipitation of the parent drug prior to intestinal uptake has been identified as one of the major pitfalls for the success of phosphate ester prodrugs.57 In Figure 2 amprenavir transport across Caco-2 monolayers from a saturated solution (suspension) is compared to the transport from a supersaturated amprenavir solution that was generated by dephosphorylation of fosamprenavir in human intestinal fluids. The intestinal flux after 60 min from the supersaturated solution was almost fivefold higher than the flux from the saturated suspension, indicating the large impact that supersaturation may have on transepithelial uptake of drugs.
The Special Case of Weak Bases: Supersaturation Induced by the Gastrointestinal pH Gradient
In most cases, supersaturation is induced from solubilized formulations or formulations that contain a high-energy state of the drug. However, for weakly basic drugs, even intake of the crystalline powder may result in supersaturation in the small intestine. Due to the pH gradient in the gastrointestinal lumen in fasted state conditions (pH 1.5–2 in the stomach vs. pH 5–8 in the intestine), the gastric solubility of weak bases (ionized form) typically exceeds their intestinal solubility (unionized form). Hence, after dissolution of poorly water-soluble weak bases in the stomach, transfer to the intestine may result in supersaturated concentrations and an increased flux across the intestinal mucosa. By simulating the gastrointestinal pH-shift during dissolution experiments, this behavior can be monitored. For instance, Kostewicz et al.58 evaluated the behavior of three weakly basic drugs (dipyridamole, BIBU 104 XX, and BIMT 17 BS) in an in vitro system simulating both the pH gradient between stomach and intestine and the presence of bile salts and phospholipids in the intestine. Upon transfer of a solution of the drug in an acidic medium (pH 2, simulating fasted state gastric conditions), supersaturated concentrations of the weak bases were observed in both fasted and fed state simulated intestinal fluids (FaSSIF and FeSSIF). Presumably, this mechanism plays an important role in the intestinal absorption of various poorly water-soluble weak bases.
The efficiency of this behavior depends on both dissolution of the drug in the stomach and metastable supersaturation in the intestine (in relation to the uptake rate). To stabilize the supersaturation upon transfer to the intestine, inclusion of precipitation inhibitors in the formulation, such as polymers, may be required.59 The ‘spontaneous’ supersaturation can be enhanced by formulation approaches that improve gastric dissolution. Typical examples are formulations of the poorly water-soluble weak base itraconazole (pKa = 4). The dissolution rate and solubility of crystalline itraconazole are insufficient, even in acidic medium (6 µg/mL in 0.1 M HCl), to allow for significant absorption. The previously mentioned Sporanox® and OMS-based formulations improve the gastric dissolution rate of itraconazole and create supersaturated concentrations already in the stomach.
As described above, absorption of poorly water-soluble weak bases relies on the acidic pH in the stomach. In case of elevated gastric pH—due to for instance hypochlorhydria or the concomitant intake of antacids or proton pump inhibitors—gastric dissolution of weakly basic drugs and subsequent supersaturation in the intestine will be impaired, resulting in insufficient and variable drug absorption. For instance, absorption of the weakly basic HIV protease inhibitor atazanavir was drastically impaired upon coadministration with the proton pump inhibitor lansoprazole.60 A reduction of itraconazole absorption from Sporanox® in AIDS patients is probably related to hypochlorydia.61, 62 In order to avoid these phenomena, pH-independent formulation strategies are preferable. In this respect, itraconazole release from OMS is promising, as supersaturated itraconazole concentrations are created not only in acidic medium, but also in FaSSIF at pH 6.5 (in contrast to itraconazole release from Sporanox®).52
Since transepithelial transport of weak bases will be more pronounced in the small intestine (preferential uptake of the unionized form), sufficient precipitation inhibition is required upon transfer of the supersaturated solution to the intestine. Therefore, one cannot rely on dissolution studies at constant acidic pH to predict the performance of formulations of weak bases in vivo.33 For instance, Six et al.63 observed a discrepancy between the results of in vitro dissolution tests in acidic medium and in vivo absorption for four solid dispersions of itraconazole: faster release and increased supersaturation in acidic medium correlated with lower bioavailability. Presumably, this effect can be explained by differences in recrystallization rate upon transfer to the small intestine (increased driving force for precipitation in case of higher supersaturation). Thus, it is crucial to simulate the gastrointestinal pH-shift during supersaturation dissolution testing of weak bases to evaluate whether supersaturation is maintained in the small intestine. An approach to direct supersaturation of weak bases to the intestine instead of the stomach was investigated by Miller et al.64 They developed solid dispersions of itraconazole in an enteric matrix based on Eudragit® L 100-55. Due to the pH-dependent dissolution of the matrix, itraconazole was only released in supersaturated concentrations after transfer to the intestine. Incorporation of Carbopol® 974P as a stabilizing agent helped to maintain the supersaturation.65
Supersaturating Versus Solubilizing Drug Delivery Systems: Conceptual Advantages
The examples presented so far show that supersaturating drug delivery systems rely in part on conventional principles to improve the intestinal absorption of poorly water-soluble drugs: generating increased intraluminal concentrations by administering high-energy dosage forms that circumvent or accelerate the dissolution step. As soon as the intraluminal concentrations exceed drug solubility in the gastrointestinal fluids, there is a risk for drug precipitation and reduced absorption. The major difference between solubilizing and supersaturating formulations lies in the way they try to avoid precipitation. Solubilizing formulations attempt to avoid the creation of a supersaturated state by increasing the solubilizing capacity of the gastrointestinal environment (thermodynamic approach); in contrast, supersaturating formulations aim at the creation of a metastable supersaturated state, stabilized by temporary inhibiting precipitation (kinetic approach).
The supersaturating approach has some conceptual advantages over solubilizing methods. First, enhancing the solubilizing capacity of the gastrointestinal environment can be limited for compounds with very poor aqueous solubility; higher intraluminal concentrations can be reached using supersaturating formulations. In addition, the solubilizing approach merely solubilizes drugs by their incorporation in colloidal species or complexing agents; the free drug fraction, in equilibrium with the solubilized fraction, is still limited by the poor aqueous solubility. Supersaturation, however, creates enhanced free drug concentrations, having a more pronounced effect on the uptake flux. Finally, supersaturating formulations do not need the incorporation of large amounts of solubilizing excipients in the vehicle, which may result in a reduced pill burden and/or a lower toxicity.
PRECIPITATION INHIBITION TO STABILIZE SUPERSATURATION
Supersaturation is a thermodynamically unstable condition and the driving force for precipitation. In order to take advantage of the creation of intraluminal supersaturation, this state should be stabilized for a time period allowing sufficient transepithelial transport by temporary inhibiting precipitation. This may require the intraluminal presence of the so-called precipitation inhibitors which are typically included in the SDDS. In this section, we will present an overview of excipient classes that have been investigated as precipitation inhibitors. In addition, we will summarize the theory behind drug precipitation and discuss the mechanisms underlying precipitation inhibition by pharmaceutical excipients. In order to make a rational selection of excipients to add as precipitation inhibitors, insight into the mechanisms of precipitation kinetics is crucial.
A supersaturated solution is characterized by an increased chemical potential (µ) compared to a stable, saturated solution (µeq). From a thermodynamic point of view, the difference in chemical potential (Δµ) is the driving force for drug precipitation (crystallization):
From the definition of chemical potential, it follows that:
where R is the gas constant, T is the temperature and a and aeq are the activity of the solute in a supersaturated and saturated state, respectively.66
Assuming no difference in the activity coefficients of the solute in the supersaturated and saturated state, the equation becomes:
where C is the drug concentration in the supersaturated solution, Ceq is the equilibrium solubility of the drug, and S is the supersaturation ratio as defined in Eq. (2).
Mechanistically, drug precipitation from a supersaturated solution essentially consists of two processes: nucleation and crystal growth. Starting from a supersaturated solution, dissolved molecules have to form small clusters/aggregates (nucleation), which can then grow to macroscopic crystals (crystal growth). Despite the fact that precipitation from a supersaturated solution is a thermodynamically favored process (decrease in Gibbs free energy), the nucleation step requires an activation energy (see Fig. 3). The increased Gibbs free energy of small clusters of molecules can be attributed to the high interfacial tension between the high-curvature clusters and the solvent. A critical cluster is defined as a cluster with critical radius r* with maximum interfacial energy. In case this activation energy is too high, no new crystals can be formed and a metastable, supersaturated solution arises. The range of supersaturated concentrations that can exist (for a certain time period) without the formation of new crystals is called the metastable zone. In many cases, stabilizing supersaturation by precipitation inhibition can be considered as increasing the width of the metastable zone.
The nucleation rate Jn, that is, the net production of critical clusters per unit of time and unit of bulk volume, is defined as:
where N0 is the number of molecules in a unit volume, v is the frequency of molecular transport at the nucleus-liquid interface, kb is the Boltzmann's constant, and ΔG* is the Gibbs free energy change for the formation of critical clusters. Assuming homogeneous nucleation (i.e., only nucleation in solution and not on surfaces) and spherical clusters, ΔG* equals:
and the nucleation rate Jn becomes:
where υ is the molecular volume of the crystallizing solute and γns is the interfacial energy per unit area between the cluster and the surrounding solvent. A more detailed discussion of these equations can be found elsewhere.66–69 Eq. (9) indicates the strong dependence of the nucleation rate on (1) the degree of supersaturation S and (2) the interfacial energy γns between the critical cluster and the solvent. It should be noted that nucleation in the gastrointestinal lumen will be facilitated by the presence of various surfaces and interfaces that may act as a catalyst for nucleation by decreasing the required activation energy. Despite being of practical importance, this heterogenous nucleation (nucleation on surfaces) is more difficult to model.66
Once the energy barrier for nucleation has been overcome, critical clusters can grow to macroscopic crystals. Crystal growth basically consists of two steps: diffusion of molecules from the supersaturated solution to the crystal interface and integration of the molecule into the crystal lattice (which is accompanied by desolvation). The net growth of each particle (increase of the radius r) is governed by:69
where D is the diffusion coefficient of the molecule, k+ is the surface integration factor, NA is the Avogadro constant, and (C − Ceq) is the difference between the bulk concentration and the concentration in the solution directly next to the cluster surface (which is assumed to be in equilibrium with the cluster). If r ≫ D/k+, the process is diffusion-controlled while if r ≪ D/k+, the process is controlled by the surface integration.
Precipitation inhibitors may act by a number of possible mechanisms that can be inferred from the theory of drug precipitation, including:
reducing the degree of supersaturation by increasing the solubility (decrease in both nucleation and crystal growth);
increasing the viscosity, resulting in a reduced molecular mobility (decreasing nucleation), and diffusion coefficient (decreasing crystal growth);
increasing the cluster–liquid interfacial energy (decreasing nucleation);
changing the adsorption layer at the crystal-medium interface by, for example, adsorbing onto the crystal surface thereby hindering crystal growth; this may be accompanied with crystal habit modifications;
changing the level of solvation at the crystal–liquid interface, thereby affecting the integration of drug molecules into the crystal.
Obviously, these mechanisms depend on properties of the inhibitor, the drug and the medium.
Pharmaceutical Excipients as Stabilizers of Supersaturation
Different classes of excipients have been investigated as precipitation inhibitors to be included in SDDS. The identification of potential precipitation inhibitors for a given drug requires the assessment of precipitation in presence and absence of the candidate inhibitors. In essence, precipitation is assessed by determining the induction time for (measurable) precipitation or monitoring the concentration–time profile upon generation of supersaturation. It should be taken into account that inhibitory effects will depend not only on the concentration of the excipient, but also on the initial degree of supersaturation.70, 71 From a mechanistic point of view, it is essential to distinguish between thermodynamic and kinetic precipitation inhibition. While excipients that inhibit precipitation by avoiding supersaturation are valuable for solubilizing formulations, SDDS require excipients that can delay precipitation in case of supersaturation. Thus, in addition to their effect on drug precipitation, excipients should always be evaluated for their impact on the thermodynamic solubility, as determined by means of the shake-flask method. In order to gain insight into the precise mechanisms of precipitation inhibition, additional experiments are required, including assessing the particle size distribution,70 evaluating the morphology of the precipitate,31, 72 discriminating between inhibitory effects on nucleation and crystal growth69 and/or investigating drug–excipient or crystal–excipient interactions (e.g., by spectroscopic techniques).69, 73
The potency of polymers to stabilize supersaturation has been reported quite often in literature.8, 14, 15, 28, 30–32, 39, 46, 72–77 Examples include cellulose derivatives (e.g., MC, HPC, HPMC), vinyl polymers (e.g., PVA, PVP, PVPVA), and ethylene polymers (e.g., PEG). As can be seen in Table 1, HPMC is often the excipient of choice to include in SDDS as a precipitation inhibitor. In case of solid dispersions or nanoparticles, HPMC (and other polymers) can function as both the carrier/stabilizer of the formulation and stabilizer of the supersaturated state upon release of the drug.
Although some polymers can increase the solubility of drugs,16, 67 this effect is usually limited and precipitation inhibition is most likely the result of direct interference of the polymer with nucleation and/or growth rate. Raghavan et al.72 investigated the crystallization of hydrocortisone acetate in presence and absence of HPMC. A concentration-dependent increase in induction time of crystallization in presence of HPMC (0.5–5%) was observed. The researchers suggested a mechanism based on hydrogen bonding between hydrocortisone acetate and HPMC, which contains (similar to other cellulosic polymers) a large number of hydroxyl functional groups. Hydrogen bonds between drug molecules and the polymer, which were confirmed by means of infrared spectroscopy,78 increase the activation energy for nucleation. In addition, the ability to form hydrogen bonds enables HPMC to adsorb onto the crystal surface; adsorbed polymer molecules hinder the incorporation of drug molecules into the crystal lattice and slow the crystal growth. The interference of HPMC with crystal growth was shown experimentally by the observed modifications in crystal habit.72 Gao et al.31 observed decelerated precipitation of AM 517 in the presence of HPMC. In agreement with the slowed precipitation, the particle size distribution profile, assessed 60 min after induction of supersaturation, showed significantly lower counts and was shifted to lower particles in presence of HPMC. Moreover, drug precipitates were identified as crystalline in absence of HPMC, but amorphous when HPMC was present. These data indicate a strong interference of HPMC with the crystallization process.
Recently, the effect of PVP on crystallization of bicalutamide was investigated in detail by Lindfors et al.69 PVP concentrations as low as 0.01% (w/w) significantly decreased the crystallization rate. Experiments were performed to discriminate between the effects of PVP on crystal nucleation and on crystal growth. It was shown that polymer adsorption to the growing crystal affected the surface integration kinetics and slowed the crystal growth rate. The nucleation rate was not affected, presumably due to the fact that PVP did not adsorb to individual molecules or very small particles.
While the capacity of surfactants to inhibit precipitation by completely solubilizing the drug (thermodynamic precipitation inhibition) is well known,79 surfactants may also delay precipitation from supersaturated solutions. Data on the stabilization of supersaturated solutions by surfactants is rather limited. When surfactants are added to a supersaturated solution at concentrations exceeding their critical micelle concentration (CMC), an increase in drug solubility will reduce the rate of nucleation and crystal growth by decreasing the degree of supersaturation (see Eqs. 9 and 10). Surfactants may improve the solvation of dissolved drug molecules, thereby increasing the activation energy required for desolvation during nucleation and crystal growth. The latter mechanism was postulated by Overhoff et al.15 to explain the ability of SDS to maintain supersaturation of tacrolimus at concentrations below 10% of the CMC. In addition, adsorption of SDS molecules to the surface of embryonic crystals of tacrolimus (with hydrophobic tails and anionic sulfate groups oriented towards the crystal and the surrounding water, respectively) will provide electrostatic repulsion that limits coalescence.15 Recently, Brewster et al.45 evaluated the capacity of three surfactants (TPGS, Tween® 20, and Cremophor® RH40) to stabilize supersaturation of itraconazole in acidic medium, induced by the cosolvent method. As reported in Figure 4, itraconazole immediately precipitated in absence of excipients; surfactants (2.5%) could maintain itraconazole supersaturation (concentrations exceeding the solubility in presence of the surfactants). In comparison with cyclodextrins, however, stabilization by the surfactants was limited to a rather short time period (5 min in case of Tween® 20 and Cremophor® RH40, 40 min in case of TPGS).
Similar to surfactants, cyclodextrins are well known for their solubilizing capability and capacity. Although their effect is usually attributed to the formation of inclusion complexes, cyclodextrins may also form noninclusion constructs, including cyclodextrin aggregation or solubilization related to surfactant-like properties.80, 81 In addition, a number of reports suggest the ability of cyclodextrins to delay drug crystallization from supersaturated solutions.82–86 In a recent report, Brewster et al.45 described the stabilization of supersaturated itraconazole solutions, generated by the cosolvent technique, by hydroxypropyl-β-cyclodextrin (HPβCD) and sulfobutylether-β-cyclodextrin (SBEβCD). Stable itraconazole solutions (36- and 8-fold supersaturated relative to the solubility under equilibrium conditions in the case of HPβCD and SBEβCD, respectively) were maintained for at least 4 h. This effect was superior to that generated by other excipients as illustrated in Figure 4.
A number of mechanisms have been suggested to explain the precipitation inhibition by cyclodextrins.45 Similar to surfactants, the solubilizing capacity of cyclodextrines may decrease the degree of supersaturation and thereby alter the precipitation kinetics in such a way the solution becomes metastable.84 Analogous to what has been reported for HPMC, cyclodextrins may also interact with drug molecules in solution or with growth sites on the crystal surface by means of hydrogen bonding. Similar to surfactants, the interaction with cyclodextrins may improve the solvation of dissolved drug molecules, thereby increasing the activation energy for desolvation during crystal growth. Finally, cyclodextrins have been suggested to act as a kosmotrope (i.e., a material that enhances the cohesive structure of water). According to Loftsson et al.,87 kosmotropes may stabilize supersaturated solutions while chaotropes (i.e., materials that reduce the water structure) may have the opposite effect. While solubilization through complexation by cyclodextrins may reduce the free drug concentration, possibly limiting the increase in transepithelial uptake, cyclodextrin-based stabilization of supersaturation is a noninclusion based phenomenon. As a consequence, free drug concentrations will be higher, resulting in a more pronounced effect on the uptake flux.
Screening for Precipitation Inhibitors
The increased attention for precipitation inhibition in the field of oral drug delivery is illustrated by a number of recent reports that describe the use of high-throughput screening to identify excipients that may delay precipitation upon induction of supersaturation. Vandecruys et al.88 described the screening of 10 excipients for their capacity to inhibit precipitation of 25 drug candidates and in a related study a group of 14 candidates was assessed using a similar technique.89 Supersaturation was created by a cosolvent/solvent quench based approach in which a solution of the drug in an organic solvent was added to the dissolution medium which included the excipient of interest (2.5%). Drug concentrations in presence and absence of the excipient were assessed as a function of time and compared (Tabs. 2 and 3). The collected data suggested that the presence of surfactants and cyclodextrins can result in strongly elevated concentrations (Tab. 2), but the duration of the effect is highly compound dependent (Tab. 3). Although concentrations were typically lower in the presence of polymers (Tab. 2), their stabilizing effect tended to last longer, but again this was very much compound related (Tab. 3). As potential excipient-induced increases in solubility were not assessed, the results could not discriminate between thermodynamic and kinetic inhibition of precipitation. As mentioned, the results also indicated that precipitation inhibition should be assessed on an individual drug–excipient basis. No significant correlations could be observed between the precipitation profiles and the physicochemical properties of the drug candidates. For one of the drug candidates, a number of simple dosage forms were prepared and evaluated in dogs. Excipients that provided the best precipitation inhibition in screening also gave the highest oral bioavailability.88 A similar screening approach was used by Janssens et al.90 for the identification of excipients to be included in a solid dispersion formulation of itraconazole.
Table 2. Excipient-Induced Increase in Concentration after Creation of Supersaturation in a Solvent Quench Assay for 39 Compounds*
The figures correspond to the ratio between drug concentrations after 5 min in presence and absence of the excipients. Adapted from Vandecruys et al.88 and Brewster et al.89
ND, not determined.
The following excipients were used (concentration 2.5%, w/w): hydroxypropyl cellulose (HPC), hydroxypropyl methylcellulose (HPMC), PolyOx™ NF 100 k (PolyOx), polyvinylpyrrolidone covinyl acetate 64 (PVPVA), polyvinylpyrrolidone K30 (PVP K30), Cremophor® RH40 (RH40), Polysorbate 20 (Poly20), d-α-tocopheryl polyethylene glycol 1000 succinate (TPGS), hydroxypropyl-β-cyclodextrin (HPβCD) and polyethylene glycol 4000 (PEG4000).
Table 3. Stability of Excipient-Induced Increase in Concentration for 39 Model Compounds Based on a Solvent Quench Assay and as a Function of 10 Precipitation Inhibitors*
The figures correspond to the decrease in concentration over 120 min. Adapted from Vandecruys et al.88 and Brewster et al.89
ND, not determined.
The following excipients were used (concentration 2.5%, w/w): hydroxypropyl cellulose (HPC), hydroxypropyl methylcellulose (HPMC), PolyOx™ NF 100k (PolyOx), polyvinylpyrrolidone covinyl acetate 64 (PVPVA), polyvinylpyrrolidone K30 (PVP K30), Cremophor® RH40 (RH40), Polysorbate 20 (Poly20), d-α-tocopheryl polyethylene glycol 1000 succinate (TPGS), hydroxypropyl-β-cyclodextrin (HPβCD) and polyethylene glycol 4000 (PEG4000).
Dai et al.91 designed a high-throughput screening experiment to assess the precipitation kinetics of liquid formulations of a drug candidate (including different excipients) in biorelevant media (SIF, FaSSIF, and FeSSIF). Three formulations with distinct precipitation kinetics (fast, slow, and no precipitation) were selected and evaluated in a dog model. Among the three formulations, the fast precipitation formulation resulted in the lowest oral bioavailability.
In the examples on screening for precipitation inhibitors presented so far, supersaturated solutions were generated by using a solvent quench approach. An alternative approach is by using the amorphous drug substance. In this case, the extent and duration of supersaturation in the presence of the excipients of interest can potentially provide data which is more germane to solid oral dosage form development. In a recent study with itraconazole, solubilization of the crystalline material and the amorphous phase in the presence of cyclodextrins was completed.92 Data for HPβCD (20%) and dimethyl-β-cyclodextrin (DMβCD, 20%) are provided in Figure 5A and B, respectively. The data are interesting in that they demonstrate two principles in the formation and stabilization of supersaturated drug solutions. In the case of HPβCD, the complexation efficiency of the cyclodextrin is relatively low such that relatively high levels of supersaturation are generated (C/Ceq = 23.3 at day 1) which fall as a function of time (C/Ceq = 4.0 at day 32). In this instance, the HPβCD is acting more like a cellulosic/starch-based polymer. In the case of DMβCD, the complexation efficiency is significantly increased. The result is that the level of supersaturation is lower (C/Ceq = 3.8 at day 1) but the generated metastable solution is more stable over time (C/Ceq = 3.7 at day 32).
Predictive Selection of Precipitation Inhibitors
Currently, the selection of successful drug–excipient combinations for SDDS is mainly trial-and-error based. The available data on the physicochemical properties that govern the interactions between drugs and excipients are insufficient to enable a more rational or even predictive selection. In the previous paragraphs we cited a number of publications that report on the use of excipients as stabilizers of supersaturated solutions. However, it is important to realize that excipients may affect drug precipitation by a multitude of mechanisms. For instance, the mechanisms by which surfactants may alter precipitation kinetics include:
adsorption to the surface of a nucleus, thereby reducing the interfacial tension γns between this surface and the solvent; this results in an increased nucleation rate Jn (Eq. 9);
increasing the equilibrium solubility Ceq (at surfactant concentrations exceeding the CMC), thereby reducing the supersaturation ratio S; this results in reduced nucleation and crystal growth (Eqs. 9 and 10);
improving the solvation of dissolved molecules and, as a consequence, increasing the energy required for desolvation during crystal growth; this will decrease the surface integration factor k+ and reduce the crystal growth rate (Eq. 10);
reducing coalescence by electrostatic repulsion as a result of oriented adsorption of ionic surfactants to the surface of embryonic crystals; this will reduce crystal growth.
The relative contribution of different mechanisms will dictate the net effect of excipients on drug precipitation, which may vary from inhibition to stimulation. This will depend on properties of the compound, the surfactant and the environment (solvent, hydrodynamics, etc.) and requires more basic research.
SUPERSATURATION IN VIVO
The current research concerning SDDS mainly focuses on formulation development and in vitro evaluation of the generation and stabilization of supersaturation. While the in vivo assessment of intraluminal supersaturation might be possible by aspiration and analysis of gastrointestinal fluid samples after intake of a dosage form,93, 94 it has not been investigated to date. However, as is the case for other absorption enhancing strategies, including dissolution/solubility improvement,95 a number of questions can be raised regarding supersaturation in the gastrointestinal tract. To name a few:
Can SDDS significantly enhance intestinal absorption in vivo?
What is the relative importance of the two major parameters related to supersaturation, that is, the maximum degree and the stability of supersaturation, in enhancing intestinal absorption?
How does the gastrointestinal physiology (fluid composition, hydrodynamics, etc.) affect intraluminal supersaturation?
How can we improve the biorelevance of supersaturation testing in vitro?
In Vitro–In Vivo Correlations
Some of the examples presented in Table 1 indicate that SDDS are indeed capable of generating an improved pharmacokinetic profile as compared to conventional (nonsupersaturating) dosage forms. However, only a few studies provide data that allow one to correlate varying supersaturation profiles in vitro with intestinal absorption characteristics in vivo. Using crystalline salt forms combined with precipitation inhibitors, Guzmán et al.8 observed a significant improvement in the intestinal absorption of celecoxib in dogs. Both the AUC and the Cmax of the plasma concentration–time profiles obtained upon absorption of nine celecoxib formulations correlated with the AUC0–10 min parameter of the corresponding in vitro dissolution profiles. This suggests that rapid creation of high intraluminal concentrations dictates the absorption of celecoxib while maintaining the metastable zone of celecoxib supersaturation by precipitation inhibitors is only required for a limited time period. Similar conclusions were drawn by Overhoff et al.15 with respect to the absorption of tacrolimus in rats from nanostructured solid dispersions using different excipients. Comparing the pharmacokinetic profiles of four formulations, the one containing the surfactant P407 as stabilizer showed the greatest extent of absorption (Cmax, AUC). In vitro, this formulation generated the highest maximum degree of supersaturation, but performed least optimally in stabilizing the supersaturated state. Inhibiting precipitation for more than 2 h did not seem to enhance tacrolimus absorption.
For both celecoxib and tacrolimus, the initial production of a highly supersaturated system appears to be more important for absorption enhancement as compared to long-term stabilization of supersaturation. This is probably related to a fast transepithelial permeation of the drugs. For slowly permeating drugs, the required time during which supersaturation should be stabilized in order to obtain sufficient absorption may be longer. From a conceptual point of view, combining a high degree of initial supersaturation with long-term stabilization may be difficult to achieve, as higher degrees of supersaturation result in faster nucleation and crystal growth (see Eqs. 9 and 10). Therefore, knowledge about the relative importance of supersaturation parameters in improving intestinal absorption is crucial for the efficient development of SDDS.
Impact of Gastrointestinal Physiology on Supersaturation
It is obvious that the gastrointestinal physiology, including hydrodynamics and the composition of gastrointestinal fluids, may affect supersaturation. However, knowledge on these effects is limited.
As discussed above, the impact of pH-shifts in the gastrointestinal tract on supersaturation of weakly basic drugs is well known. In addition to its effect on solubility, the gastrointestinal pH may also affect the efficiency of precipitation inhibitors, for instance when hydrogen bonding between ionizable groups is involved. Parameters including the surface tension and the viscosity of gastrointestinal fluids may influence drug precipitation kinetics in supersaturated solutions (see the drug precipitation theory). Without doubt, the phenomenon of heterogeneous nucleation (nucleation on impurity surfaces) occurs in the gastrointestinal tract, presumably more extensively than during in vitro tests.
Little is known about the influence of endo- and exogeneous components such as bile salts, phospholipids and food digestion products on intraluminal supersaturation. As these components can alter the dissolution and solubility of drugs,95, 96 they will affect the degree of supersaturation, and, as a consequence, also the precipitation kinetics. In addition, specific interactions between gastrointestinal components and supersaturated drugs or growing drug crystals can be expected but have not been studied yet. For instance, bile salts and phospholipids may decelerate drug precipitation in a way similar to surface-active excipients. As mentioned previously, Kostewicz et al.58 evaluated the behavior of three weakly basic drugs in an in vitro system simulating both the gastrointestinal pH gradient and the presence of bile salts and phospholipids in the intestine by using biorelevant media (FaSSIF and FeSSIF). Upon transfer of a solution of the drug in acidic medium simulating fasted state gastric conditions, supersaturated concentrations of the weak bases were observed in both FaSSIF and FeSSIF. Although higher concentrations were reached in FeSSIF (mainly the result of a higher solubility), the extent of supersaturation was higher in FaSSIF (maximum supersaturation ratio between 3 and 8.4 in FaSSIF and between 1.4 and 2.7 in FeSSIF). Having said that, the supersaturated concentrations were maintained longer in FeSSIF. Whether this is only due to the lower supersaturation ratio and hence, the reduced driving force for precipitation, or to a specific stabilizing effect of the increased bile salt and phospholipid levels, is not clear. Dai et al.91 investigated the precipitation behavior of the drug candidate JNJ-25894934 from three different liquid formulations (fast, slow and not precipitating based on a screening assay) in phosphate buffer, FaSSIF and FeSSIF. The solubility of the drug was similar in FaSSIF and FeSSIF, but was 8-fold lower in the phosphate buffer. Striking differences in precipitation were observed between the different media, especially for the fast-precipitating formulation. While precipitation from this formulation was immediate and complete in phosphate buffer, a metastable zone containing about 20% and 80% dissolved drug, was maintained during 8 h in FaSSIF and FeSSIF, respectively. After 8 h, precipitation continued. As there was no difference in drug solubility in FaSSIF versus FeSSIF, enhanced concentrations in the metastable zone in FeSSIF clearly suggest specific precipitation–inhibiting interactions that are more effective in FeSSIF, presumably due to the increased bile salt/phospholipid concentrations. Interestingly, the in vitro precipitation profiles in FeSSIF correlated better with in vivo absorption than those in phosphate buffer and FaSSIF.
The same study also illustrated an impact of hydrodynamics on the precipitation kinetics by evaluating drug precipitation in a 96-well plate (screening assay) and in a USP dissolution setup. While the initial precipitation profiles were similar, more drug was precipitated after 24 h of incubation in the USP dissolution apparatus as compared to the 96-well plate. In addition to a larger quantity of compound used in the USP dissolution method, the faster agitation in this setup was suggested to explain these observations. It appears that the hydrodynamics in the gastrointestinal tract, which are difficult to simulate in vitro, affect drug precipitation in vivo. For instance, the presence of convection currents may enhance diffusion-controlled nucleation and crystal growth.
Biorelevant Supersaturation Testing
With regard to the impact of gastrointestinal physiology, performing supersaturation assays in biorelevant conditions will be important in obtaining a reliable prediction of intraluminal supersaturation in vivo. Despite the limited knowledge on the influence of the gastrointestinal environment on supersaturation, a number of simple recommendations can already be taken into account to improve the relevance of supersaturation testing.
Simulate the gastrointestinal pH-gradient. For obvious reasons, simulating the pH-shift is required when evaluating weakly basic drugs. In other cases, it is advisable to evaluate supersaturation not only in gastric but also in intestinal conditions, since for most drugs the small intestine is the primary site of transepithelial uptake. A pH-shift can be simulated by multi-compartmental dissolution techniques.58, 97 In many cases, a simple manual transfer of gastric medium into intestinal medium may be sufficient.52
Use biorelevant media to simulate gastric and intestinal conditions. Considering the potential effects of, for instance, bile salts and phospholipids on drug precipitation, it is advisable to include relevant concentrations of these components in the dissolution media used for supersaturation testing. For now, one can use the standard biorelevant media for dissolution testing, including FaSSGF, FeSSGF, FaSSIF(-v2), and FeSSIF(-v2)11–13 which have recently been reviewed by Dressman et al.98 However, there is an urgent need to identify crucial intraluminal factors that affect supersaturation in both fasted and fed state; this would allow the definition of more appropriate biorelevant media.
Assess transepithelial transport from supersaturated solutions. The final goal of inducing supersaturation is improving the flux of drug molecules across the gastrointestinal mucosa. In order to evaluate the potential impact of a supersaturation-based strategy on intestinal absorption, it is important to determine drug transport across an epithelial layer starting from supersaturated solutions or formulations that induce supersaturation. While this is a common approach in the field of topical drug delivery,99–101 the literature reporting this type of studies in the field of intestinal drug delivery is rather limited.52, 56, 57 Assessing transepithelial transport may also provide information on the relative importance of the supersaturation parameters (maximum degree and stabilization). For drug compounds exhibiting fast permeation, the transepithelial flux may create a kind of sink condition, limiting the importance of long-term stabilization. Mellaerts et al.52 studied the transport of itraconazole from the intestinal lumen into the mesenteric vein by means of the in situ perfusion technique in rats. The transport curve shown in Figure 6 indicates the increase in uptake of itraconazole upon switching the perfusion medium from a saturated solution (0–60 min) to a supersaturated solution generated by a solvent quench approach (60–120 min).
Conceptually, the generation and stabilization of intraluminal supersaturation provide an efficient solution for the growing problem of solubility-limited oral bioavailability. Hence, it is not surprising to see a rapidly increasing number of reports on supersaturating drug delivery systems in recent years. While various examples presented in this review suggest useful applications of SDDS, it is also clear that more basic research is required to fill a number of crucial gaps in our knowledge on intraluminal supersaturation. From a technological point of view, the current understanding of precipitation kinetics and precipitation inhibition is rather limited. Insight into the physicochemical characteristics that govern the interactions between drugs and excipients during nucleation and crystal growth would lead to a more focused selection and even prediction of effective precipitation inhibitors. From a biopharmaceutical point of view, the impact of gastrointestinal physiology on supersaturation is not well understood. The identification of intraluminal factors affecting the generation and stabilization of supersaturation is crucial for biorelevant evaluation of SDDS. In addition, more in vitro–in vivo correlations are required to assess in vitro supersaturation parameters relevant for in vivo absorption. Eventually, improved insight related to these issues will enable rational design of SDDS and the full exploration and exploitation of this promising oral delivery strategy for poorly water-soluble drugs.