Release mechanisms of PLGA-based drug delivery systems: A review

Abstract

Poly(lactic-co-glycolic acid) (PLGA)-based microparticles and implants are of continuously increasing importance as parenteral controlled drug delivery systems. However, the underlying drug release mechanisms are often not understood, rendering product optimization difficult: The effects of formulation and processing parameters on drug release can be surprising. Also, upscaling and troubleshooting during production at industrial scale can be highly cumbersome. This can be attributed to the complexity of the physicochemical processes, which can be involved in the control of drug release. Generally, tri-phasic drug release patterns are observed: An initial burst release is followed by a zero order release phase and a final, again, rapid release phase. The relative importance of the different phases can strongly depend on the: (i) composition (e.g., type & amount of drug and polymer), geometry and dimensions of the system, (ii) manufacturing procedure, and (iii) conditions in the surrounding environment (e.g., bulk fluid versus human tissue). Water penetration into the system, drug dissolution, limited solubility effects, drug diffusion through an “intact polymeric matrix” (polymer phase) and/or through water filled pores, pore closure due to local PLGA swelling, osmotic effects, polymer degradation, local drops in micro-pH, autocatalytic effects, substantial swelling of the entire system as well as other phenomena can be of importance. This article aims at giving an overview on the current knowledge in this field. Please note that it is hypotheses-driven, thus, general conclusions should be seen with caution. Also, each drug delivery system should be considered on a case-by-case basis. This article also aims at raising awareness on two aspects, which are often neglected: (i) Substantial system swelling is likely the root cause for the onset of the third drug release phase in many systems. (ii) In the case of microparticles, only looking at drug release from ensembles (hundreds of thousands/millions) of particles can be misleading.

Graphical abstract

1. Introduction

Poly(lactic-co-glycolic acid) (PLGA)-based drug delivery systems are of steadily increasing practical importance, because they offer a variety of advantages, including the following: (i) The systems are completely biodegradable. Thus, the removal of empty remnants upon drug exhaust is avoided. (ii) PLGA is degraded into its monomers: lactic acid and glycolic acid. Hence, the biocompatibility of the systems is generally good. (iii) The resulting drug release kinetics can be controlled during flexible periods of time, ranging from a few hours up to several months (Mauduit et al., 1993; Jiang et al., 2002; Chen et al., 2017; Park et al., 2019). (iv) Various manufacturing technologies can be used to prepare PLGA-based drug delivery systems (Wischke and Schwendeman, 2008; Zhang et al., 2020), such as emulsification - solvent extraction/evaporation (O'Donnell and McGinity, 1997, Yang et al., 2000, Berkland et al., 2003, Fu et al., 2003, Kim and Park, 2004, Yeo and Park, 2004., Freitas et al., 2005, Katou et al., 2008, Park et al., 2021b, Zhang et al., 2021), spray-drying (Wan et al., 2013; Wan and Yang, 2016; Arrighi et al., 2019; Shi et al., 2020; Melnik et al., 2023; Zhang et al., 2025a), ink-jet injection (Sato et al., 2023), electrospraying/spinning (Bohr et al., 2012; Liu et al., 2018; Wang et al., 2019; Wang et al., 2025), compression (Maturavongsadit et al., 2021), hot melt extrusion (Gosau and Mueller, 2010, Cossé et al., 2017. Koshari et al., 2022, Lehner et al., 2024), injection molding (McConville et al., 2015), 3D printing (Serris et al., 2020), and casting (Lehner et al., 2021). Pre-formed and in-situ forming systems can be used (Giteau et al., 2008; Sun et al., 2017). Also, the geometries and dimensions of the devices are flexible, with spherical microparticles and cylindrical implants being most frequently used. Since decades a large variety of PLGA-based controlled drug delivery systems is available on the market and successfully applied in daily practice for the benefit of the patients (e.g., Nkanga et al., 2020).

However, unfortunately, the optimization of PLGA-based controlled drug delivery systems is often highly cumbersome, because the underlying drug release mechanisms are often not well understood. Frequently, unexpected tendencies can be observed when varying formulation or processing parameters, or when going from in vitro to in vivo studies. Consequently, time-consuming and cost-intensive series of trial-and-error experiments are required. This is particularly true when long release periods are targeted. Also, upscaling and troubleshooting during industrial production can be highly challenging, if the systems are treated like “black boxes”. The limited understanding of how PLGA-based delivery systems control drug release, can generally be attributed to the complexity of the involved physico-chemical and biological processes (Siepmann and Goepferich, 2001; Raman et al., 2005; Fredenberg, 2011; Fredenberg et al., 2011a; Fredenberg et al., 2011b). To give just some examples, the following phenomena might play a role: wetting of the dosage form, desorption of drug from the system's surface, penetration of water into the device (via pores/channels and/or through an “intact polymer network”/dense polymer phase), drug dissolution (Siepmann and Siepmann, 2013), diffusion of dissolved drug molecules/ions through water-filled pores/channels and/or a dense polymer phase (Fick, 1855; Crank, 1975; Fan and Singh, 1989; Cussler, 1984), ester hydrolysis (Wang et al., 1990; Vert et al., 1994; Batycky et al., 1997; Li, 1999), polymer swelling (e.g., Gasmi et al., 2015; Bode et al., 2019; Rapier et al., 2021), pore closure (Kang and Schwendeman, 2007), local drops in micro-pH resulting in autocatalytic effects (ester hydrolysis being catalyzed by hydronium ions) (Fu et al., 2000, Liu et al., 2012, Versypt et al., 2013, Schaedlich et al., 2014. Hong et al., 2022), accelerated PLGA degradation due to base-catalyzed hydrolysis (Quan et al., 2023), osmotic effects due to the presence of water soluble drugs and/or water-soluble PLGA degradation products (Brunner et al., 1999), plasticizing effects of water on PLGA (Blasi et al., 2005), fusion of highly swollen microparticles into larger lumps (Klose et al., 2010), drug-polymer interactions (e.g., plasticizing effects, electrostatic attraction/repulsion, van der Waals forces) (Crotts et al., 1997; Cleland et al., 2001; Blasi et al., 2007), and phase separation and glassy-to-rubbery state transitions (Park et al., 2021a). Depending on the type of drug delivery system (e.g., its qualitative and quantitative composition, manufacturing procedure, geometry, dimensions, inner & outer morphology/structure) the relative importance of these phenomena can fundamentally vary. Often, one or just a few of these processes are “release rate controlling” or “dominant”: This means that a multitude of processes takes place, but most of them are not “decisive” for the control of the drug release rate. For example, if several processes occur in a sequence during drug transport and one of them is much slower than the other processes, only this slow process is determining the overall transport rate. In these cases, modifying the slow process allows adjusting desired drug release rates, whereas modifying a rapid process does not. Or, from a different perspective: The resulting release rate is highly sensitive to changes affecting the dominant process, while the system is robust (“forgiving”) with respect to changes affecting a rapid process.

A very important aspect, which should never be forgotten, is the fact that the key properties of PLGA-based drug delivery systems fundamentally change over time, e.g., the polymer molecular weight decreases and its hydrophilicity increases. These changes often alter the relative importance of the involved phenomena and different “phases of drug release” can be observed. For this reason, even a thorough characterization of the drug delivery system, which is limited to measurements conducted only before exposure to the release medium, is generally insufficient to elucidate the underlying drug release mechanisms: The dynamic changes of the system's key properties occurring during drug release should also be monitored.

In most cases, tri-phasic drug release patterns are observed from PLGA-based dosage forms (e.g., Luan and Bodmeier, 2006; Arrighi et al., 2019; Sharifi et al., 2020). However, the relative importance of the three phases can substantially differ. In certain systems, one or two phases might be negligible. In these cases, drug release appears to be only mono-phasic or bi-phasic. Fig. 1 exemplarily shows the tri-phasic release kinetics of ketoprofen from PLGA-based microparticles, prepared by an oil-in-water (O/W) emulsification - solvent extraction/evaporation technique. During day 1, the release rate is high: This is the first release phase, the so-called “burst release”. During the second release phase (which lasts for about 5 d in this example), the release rate is about constant: In other words, zero order release kinetics are provided. In the third release phase, drug release is again rapid and leads to complete exhaust (100 % release).

Fig. 1.

Typical, tri-phasic drug release patterns from PLGA-based delivery systems: The example shows ketoprofen release from PLGA microparticles loaded with 1.9 % drug into phosphate buffer pH 7.4. Adapted from Gasmi et al. (2015) with permission.

PLGA-based drug delivery systems undergo bulk erosion: Since the maximum dimensions of PLGA-based drug delivery systems are in the order of centimeters (and not for instance of thousands of kilometers), they undergo so-called “bulk erosion” (Goepferich, 1996, Goepferich, 1997; von Burkersroda et al., 2002). This means that the rate at which water penetrates into the systems upon contact with aqueous media (e.g., patient tissue) is much higher than the rate at which the ester bonds are hydrolyzed (Keles et al., 2015): The entire dosage form is rapidly wetted (e.g., within hours or a day) and polymer degradation takes place throughout the device. It must be pointed out that only limited amounts of water enter the polymer phase at this stage. In contrast, in so-called “surface-eroding” systems (based on other polymers, e.g. polyanhydrides), water penetration into the system is much slower than polymer hydrolysis (Goepferich and Tessmar, 2002; von Burkersroda et al., 2002). Consequently, the water cannot penetrate deep into the dosage forms, and polymer degradation is limited to surface-near regions only. However, it should not be forgotten that “pure bulk erosion” and “pure surface erosion” are two extremes. In real life, bulk eroding systems might also exhibit features of surface eroding systems and vice-versa. An example of this type of “hybrid behavior” is discussed in section 4 in more detail. Please note that in this article the term “degradation” refers to polymer chain cleavage, whereas the term “erosion” refers to the mass loss of the system.

When considering drug transport through a PLGA-based drug delivery system, different aspects are of key importance, including the type/grade of PLGA, the physical states of the polymer and drug as well as the inner and outer morphology/structure of the dosage form (e.g., porosity). These features are described in some more detail in the following.

Different types/grades of PLGA can be used to adjust desired drug release kinetics for a given type of drug and drug dose: In practice, often the polymer molecular weight, “lactic acid:glycolic acid” ratio, and type of end groups (-COOH or ester) are varied (e.g., Janoria and Mitra, 2007; Mylonaki et al., 2018). Garner et al. (2015) proposed procedures to determine these key parameters of PLGAs. For interested readers: Kinam Park, a world-wide leading pioneer in the field of PLGA-based drug delivery systems, recently published a very helpful review article on advanced characterization methods for PLGA polymers in final drug products (Park, 2025). Higher polymer molecular weight PLGAs are more hydrophobic (because the relative importance of the hydrophilic end groups is limited) and more intensively entangled. Higher contents in lactic acid units lead to slower polymer degradation, because the methyl group sterically hinders the attack of water initiating ester bond cleavage (lactic acid and glycolic acid only differ in the presence/absence of a methyl group). Carboxyl end groups are much more hydrophilic than ester end groups and can be ionized. In addition, also the sequence of the monomers along the PLGA chains can be altered, e.g., they can be randomly distributed or arranged in more or less large blocks built from the same monomer (Washington et al., 2017; Washington et al., 2018). However, in practice the monomer distribution is rarely altered to adjust desired drug release kinetics, probably because of the limited commercial offer of PLGAs differing in this respect. The type/grade of PLGA can significantly affect the relative importance of the involved physico-chemical phenomena controlling drug release. Furthermore, desired release kinetics might also be obtained by blending different PLGAs (e.g., Wang et al., 2014; Gu and Burgess, 2015; Ravivarapu et al., 2000). Furthermore, the presence of residual monomers and/or oligomers from the polymer synthesis (e.g., acting as plasticizers for PLGA) might play a role, as well as the width of the polymer molecular weight distribution.

Physical state of the polymer: PLGAs used in controlled drug delivery systems are generally amorphous. The polymer is often in the glassy state in the dry dosage form: Its glass transition temperature (Tg) is above storage temperature. However, once administered into the patient's body, limited amounts of water rapidly wet the entire system, and water acts as a plasticizer for PLGA (Blasi et al., 2005). Consequently, the Tg often rapidly falls below 37 °C. Thus, the system undergoes a glassy-to-rubbery phase transition. In the rubbery state, the mobility of the polymer chains is much higher than in the glassy state. Hence, also drug molecules become more mobile. Nevertheless, most drugs cannot be expected to rapidly diffuse through an intact PLGA network/dense polymer phase containing only a few percent of water. An interesting recent study applying molecular dynamics simulations (Zhang et al., 2025b) suggests that the diffusion coefficients of water and small oligomers is 100 to 1000-times higher in water-filled pores compared to the polymer phase. Similar differences can be expected for most drugs.

PLGA and water: 1 or 2 phases? PLGAs have a certain capacity to “host” water within the polymeric networks, depending on their hydrophilicity/hydrophobicity. The latter is for instance affected by the type of end groups (-COOH versus ester) and polymer molecular weight (determining the relative importance of the hydrophobic polymer backbones versus the hydrophilic end groups). These are “1 phase systems”. However, if the amount of water exceeds the amount which can be “hosted” in the PLGA phase, two separate phases will exist: a polymer phase containing “some” water, and a water phase containing “some” PLGA. Between these phases, “phase boundaries” exist (e.g., the walls of water-filled pores or channels). Recently, very interesting molecular dynamics simulations have been reported by Zhang et al. (2025b), investigating such systems: They considered PLGAs with a 50:50 lactic acid:glycolic acid ratio, carboxylic end groups and number averaged molecular weights (Mn) ranging from 1 to 100 kDa. The theoretical water content was varied between 0 and 32 % (w/w). Importantly, according to these simulations, only up to 1 % water can be expected to be located within the polymer phase (occupying the free volume between the polymer chains and forming small clusters of fewer than 10 water molecules) in the case of 20 kDa PLGA (at 45 °C). Above 1 % water content, the water molecules are expected to tend to aggregate to “significantly larger water clusters”. Above 8 % water, the existence of “larger separate water phases” is suggested, as illustrated in Fig. 2. The polymer phase still likely contains only about 1 % water. These theoretical simulations are consistent with experimental studies reported earlier by Blasi et al. (2005) including Differential Scanning Calorimetry (DSC) and Karl Fisher measurements: PLGA samples were exposed to bulk water or high relative humidity. Irrespective of the experimental conditions, the water decreased the glass transition temperature (Tg) of the polymer by about 15 °C. They concluded that only a fraction of the absorbed water is responsible for the plasticizing effect, whereas the “additional” water is present as “clusters”.

Fig. 2.

Molecular dynamics simulations of the distribution of polymer chains (grey and half transparent), water (blue) and oligomers (purple) in PLGA (50:50, 20 kDa) containing 0 to 32 % water (w/w) at 45 °C. Adapted from Zhang et al. (2025) with permission. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

So, even at limited water contents in PLGA-based drug delivery systems, phase separation is likely occurring at the nanoscale, and water-filled nanopores might be formed, potentially contributing to the control of drug release. According to the molecular dynamics simulations of Zhang et al. (2025b), the aqueous phase is expected to be “more or less” free of polymer. Importantly, above 16 % water content in the case of 20 kDa PLGA (at 45 °C), the simulations suggest the presence of continuous water phases (Fig. 2). Such continuous water-filled channel networks can be expected to play a key role for the control of drug release. Drugs are much more mobile when dissolved in an aqueous phase compared to a dense PLGA phase. It is noteworthy that also Small Angle X-ray Scattering (SAXS) measurements were consistent with the hypothesis that increasing water contents in PLGA lead to nano-phase separation into an aqueous phase and a polymer phase (Zhang et al., 2025b). Please note that the estimated values of the above described “threshold” water contents (e.g., 1, 8 and 16 %) should be view with some caution: They have been determined by molecular dynamics simulations for a specific grade of PLGA (50:50 lactic acid:glycolic acid ratio, -COOH end groups, 20 kDa Mn) at 45 °C. Blasi et al. (2005) suggested that the PLGA grade they investigated (50:50 lactic acid:glycolic acid, -COOH end groups, 30 kDa Mw; Resomer RG503H, Boehringer Ingelheim, Germany) took up about 2.6 % water at 37 °C without phase separation: Adding up to 2.6 % water led to a continuous decrease in the glass transition temperature (Tg). Thus, the water molecules can be expected to be in intimate contact with the PLGA chains. Importantly, the Tg they observed with PLGA containing 20 % water was similar: This suggests that in both cases, about 2.6 % water are located within the polymer phase (and plasticize the PLGA), whereas the excess (e.g., 17.4 % in the case of 20 % total water content) is located in a separate aqueous phase.

It has to be pointed out that, yet up to now, the importance of phase separation in PLGA-based drug delivery systems (especially at the nano-scale) is not fully understood. The potential existence of nano-pores and channels is generally not even considered. This can at least in part be attributed to the fact that they are difficult to detect (e.g., artifacts are easily created during sample preparation for SEM analysis during drying and sputter coating). This is unfortunate, because the existence of a more or less continuous network of pores/channels in the nanometer range can be expected to substantially affect the resulting drug release kinetics.

Physical state of the drug: In addition to the physical state of the polymer, the physical state of the drug in the dosage form can be of key importance for the underlying drug release mechanisms (e.g., Zhang and Bodmeier, 2022). The drug can be dissolved (molecularly dispersed) in the system, or distributed in the form of tiny particles throughout the device (“non-dissolved”). These tiny particles be crystalline (molecules/ions arranged in a “perfect order”) or amorphous (“chaotic” arrangement). The drug can also partially be dissolved and partially be present in the form of amorphous and/or crystalline particles. Importantly, only dissolved drug molecules or ions can diffuse, drug particles are too large for this type of transport mechanism. The chemical structures of the drug and polymer as well as the type of preparation technique of the delivery system can strongly affect the physical states of the compounds.

Environmental conditions: It should not be underestimated to which degree the conditions in the surrounding of a PLGA-based delivery system can potentially alter the observed drug release kinetics (Johansen et al., 1998; Klose et al., 2011; Stein et al., 2018). This is true in vitro as well as in vivo (Doty et al., 2017b). For example, it has been shown that drug release can depend on the osmotic pressure of the release medium (Faisant et al., 2006), the pH in the surrounding environment (Faisant et al., 2006; Doty et al., 2017a), the volume of bulk fluid the system is exposed to as well as its renewal rate (Klose et al., 2011; Bassand et al., 2022a), the degree of agitation of the release medium (Schoubben et al., 2012; Delplace et al., 2012), the potential presence of a hydrogel around the system (mimicking living tissue) (Maeder et al., 1997; Kozak et al., 2021; Bassand et al., 2022b), and the temperature (Zolnik et al., 2006; Faisant et al., 2006). In vivo, the resulting release kinetics can also be affected by inflammatory reactions at the site of administration (Shive and Anderson, 1997), the formation of a fibrotic capsule (Sastre et al., 2007), mechanical stress experienced by the delivery system, and the presence of enzymes, especially esterases (Zolnik and Burgess, 2008), to mention just a few factors. Ideally, the sensitivity of drug release to variations in these conditions is known. Such information can for instance be very helpful to better understand differences in drug release behavior in vitro versus in vivo, and be used to improve the therapeutic efficacy and safety of the treatment for the patient.

Relatively simple in vitro characterization methods: In order to get deeper insight into the underlying release mechanisms playing a dominant role in a specific PLGA-based delivery system, rather simple experimental characterization techniques can often be very helpful. The additional workload can be kept relatively limited. For example, at each sampling time point of the drug release studies, a photo of the delivery system can be taken (e.g., by optical microscopy or macroscopy). Also pictures taken with a standard smartphone might be very telling, e.g., to monitor changes in the systems' size and geometry (e.g., swelling, erosion, disintegration, fusion of microparticles into lumps). Furthermore, the pH of the release medium can often very easily be measured (in withdrawn samples or the remaining bulk fluid). This can potentially reveal very interesting phenomena, such as a massive release of water-soluble short chain acids (generated due to PLGA degradation). In addition, if the pH of the release medium substantially drops in vitro due to a “non-optimal” protocol of bulk fluid renewal (e.g., no sampling during long week-ends or holidays), this can artificially accelerate polymer chain cleavage (since ester bond hydrolysis is catalyzed by hydronium ions). Thus, the observed in vitro release kinetics might be very different from the in vivo release behavior. As long as the system is mechanically sufficiently robust, wet weight measurements can also offer valuable insight into the underlying drug release mechanisms. However, caution must be paid that withdrawing and weighing the samples do not damage the delivery system. The key advantage of taking pictures, measuring the pH and/or wet weight of samples, is that these techniques are not destructive. So, the “added work load” to the drug release experiment (which is conducted anyway) is generally limited. In the following, also some destructive analyses are described, which generate a higher additional work load, but the obtained information can be worth the effort.

Scanning Electron Microscopy (SEM) pictures of surfaces and cross-sections of the devices before and after exposure to the release medium can be very helpful to get an idea of the outer and inner structure of the systems. However, great care must be taken when drawing conclusions from SEM pictures obtained after exposure to aqueous media: The samples are generally dried prior to analysis, and this drying step can fundamentally alter the structure of the systems. For example, upon drying highly swollen surface-near regions, they are generally visualized as porous, shriveled surfaces. Also, sample cutting might lead to artefact creation (e.g., tiny pores might be closed), and the observed structures might be misleading.

Measuring the dry mass of the delivery system (in addition to its wet mass) allows quantifying the erosion of the system as well as its water content (generally being equal to the difference “wet minus dry mass”). The required equipment is not particularly expensive and often already available in many laboratories: an oven and a balance.

Gel permeation chromatography (GPC) allows monitoring the decrease in polymer molecular weight over time, which can be a highly valuable information for a better understanding of the release mechanisms. However, the interpretation of the data is not always straightforward: Generally, the entire delivery system is analyzed and the results reflect the whole device. If PLGA degradation is not homogeneous throughout the system (e.g., due to autocatalytic effects at its center), caution must be paid when drawing conclusions. It should also not be forgotten that water-soluble short chain acids diffuse out of the system and are no more included the subsequent samples studied by GPC. In addition to the average polymer molecular weights, also the latter's distribution should be looked at. Furthermore, when comparing absolute values reported in different studies, caution should be paid, because different analytical methods might lead to somewhat different results.

Differential Scanning Calorimetry (DSC) and X-ray diffraction analyses of the drug delivery system before exposure to the release medium are generally relatively easy to perform (provided the equipment is available or accessible), and can be very helpful to better understand the physical states of the drug and polymer. For instance, the latter's glass transition temperature (Tg) can be very telling, e.g., revealing potential plasticizing effects of the drug.

In addition, the sampling protocol of the in vitro drug release studies is generally easy to vary, e.g., the volume of release medium the sample is exposed to, the volume of the withdrawn samples, partial versus complete bulk fluid renewal, as well as the sampling frequency. In case the observed drug release patterns strongly depend on the sampling protocol, this can be very helpful information and great caution should be paid when drawing conclusions from the overserved in vitro drug release kinetics: The (more or less) arbitrarily chosen experimental conditions affect them, not only the drug delivery system. Furthermore, the composition of the release medium (e.g., its osmolarity, type of ions, ionic strength) might be altered to get an idea of the sensitivity of drug release to such changes (e.g., Doty et al., 2017a).

In addition, more complex experiments can be performed, for instance aiming at the monitoring of potential changes in the pH within the dosage form during drug release. Also, potential effects of the presence of a surrounding hydrogel (mimicking living tissue) and of mechanical stress experienced by the dosage form during drug release (e.g., mimicking muscle contractions) might be monitored (Bassand et al., 2022b). In the case of PLGA-based microparticles, drug release studies from single microparticles can be extremely helpful. However, they can be highly challenging to perform in the case of very small particles. In these cases, larger microparticles might serve as surrogates. But caution should be paid when drawing conclusions, since the dimensions of the system can potentially alter the relative importance of the involved phenomena (e.g., of autocatalytic effects).

2. Drug release from single vs. multiple unit dosage forms

When studying the release mechanisms from multiple unit dosage forms (e.g., microparticles), drug release should not only be measured from ensembles of systems, but also (if possible) from single dosage forms (Borgquist et al., 2004). From a clinical point of view, the release kinetics from the ensembles of microparticles are most important. However, when only looking at the sum of the release rates from hundreds of thousands/millions of individual particles, key mechanistic aspects might not be visible. Fig. 3 illustrates a theoretical (extreme) example: In Fig. 3A, the release kinetics from several single microparticles are shown: Each microparticle exhibits pulsatile drug release kinetics, the lag-time being different for each particle. In this example, the lag-times are homogeneously distributed over the observation period. Thus, when summing up the amounts of drug released from hundreds of thousands/millions of microparticles of this type of microparticles (or in other words: when looking at the release kinetics from an ensemble of microparticles), a constant release rate is observed (zero order kinetics, as illustrated in Fig. 3B). An analysis based only on the constant release rate observed with an ensemble of microparticles can easily be misleading. Please note that a considerable variability of the drug release patterns at the single microparticle level is not necessarily a concern for the patient: In case of good batch-to-batch reproducibility, the standard deviations observed with ensembles of microparticles can be low. Thus, system performance in vivo can be reliable.

Fig. 3.

Schematic presentation of the drug release kinetics from: A) several single microparticles exhibiting pulsatile release with different lag-times, and B) an ensemble of hundreds of thousands/millions of microparticles of the same type.

On the other hand, measuring drug release only from single microparticles might also be misleading: For example, potential “particle-particle” interactions cannot be detected. If, for instance, at later time points (upon substantial system swelling), single microparticles fuse together to form larger “lumps”, this fundamentally changes the conditions for subsequent drug release: The length of the diffusion pathways to be overcome by the drug (and water-soluble short chain acids) increases, potentially altering drug release (and the importance of acidic micro-environments).

Please note that throughout this article, schematic presentations illustrate specific geometries, namely spheres and cylinders. These geometries have been arbitrarily chosen in the different schemes. The illustrated phenomena can generally be expected to occur in devices of different geometry.

In the following, the drug release mechanisms, which might control the three drug release phases observed with PLGA-based delivery systems, are discussed.

3. First phase: burst release

At early time points (generally during the first day), rapid drug release is observed from many PLGA-based drug delivery systems. This is the so-called “burst release phase”. The importance of this “burst” very much depends on the type of system and can have different root causes. Fig. 4 illustrates some of the possible underlying mechanisms:

• (i)Parts of the drug might only “loosely” be adsorbed onto the system's surface. Upon contact with the release medium, the drug molecules/ions are rapidly desorbed and, thus, released. In case of highly porous systems, in which the pore network has immediately direct access to the surface, this type of release mechanism might be relatively important, because of the considerable total surface of the system.
• (ii)The drug is distributed in the form of tiny particles throughout the dosage form, e.g., in the form of small crystals. Some of these crystals have direct access to the surface of the device and will, thus, be exposed to the release medium right from the beginning. These drug particles will rapidly dissolve and the drug quickly be released. Fig. 5 shows an example for this type of drug delivery systems: Diprophylline release from ensembles of PLGA microparticles into phosphate buffer pH 7.4 is illustrated. The release from three microparticle batches is shown, which differ in their mean particle size (indicated in the diagram +/− standard deviation). Importantly, the burst release is much more pronounced in the case of the smaller microparticles compared to larger ones (e.g., red versus yellow curve). This can be explained by the higher probability for a drug particle to be in direct contact with the system's surface in smaller microparticles: As illustrated in the cartoon on the right-hand side of Fig. 5, the number of drug particles with direct surface access increases significantly when decreasing the system's size, while keeping the microparticles' composition and inner & outer structure the same.
• (iii)The drug is distributed in the form of tiny particles throughout the dosage form and parts of these particles are connected via channels/pores (of sufficient size to allow for rapid water penetration). If one of the particles or pores/channels of such a network has direct access to the system's surface from the beginning, water rapidly enters the network and comes into contact with the drug particles of this network. Consequently, these particles quickly dissolve, and the dissolved drug molecules/ions rapidly diffuse through the interconnected, water-filled pore/channel network.

Fig. 4.

Schematic presentation of possible root causes for the initial burst release phase. Crosses represent individual drug molecules/ions, rectangles drug particles (e.g., crystals). The “brackets” around certain crosses indicate that the respective drug molecules/ions are highly mobile.

Fig. 5.

Diprophylline release from ensembles of PLGA microparticles in phosphate buffer pH 7.4: Impact of the mean particle size (indicated in the diagram +/− standard deviation). The burst release increases in importance with decreasing microparticle size. A possible explanation is illustrated in the cartoon on the right-hand side: The probability that drug crystals are in direct contact with the release medium from the beginning increases with decreasing microparticle size. Adapted from Tamani et al. (2019a) with permission.

The group of Steven Schwendeman performs ground-breaking research in this field. For example, they showed that PLGA swelling can rather rapidly close “surface pores”. Fig. 6 exemplarily illustrates SEM pictures of octreotide-loaded PLGA microparticles after different exposure times to the release medium. Clearly, the microparticles' surfaces were highly porous at t = 0. But most of these pores were closed within 24 h. Once they are closed, this type of rapid drug transport stops and the release rate at the end of the burst release phase decreases. The group of Anders Axelsson investigated the mechanism of pore closure of acid capped PLGAs in more detail: as a function of the pH of the surrounding environment (Fredenberg et al., 2011c). They found that at pH 7.4, pore closure is likely driven by swelling of the polymer network: The -COO⁻ end groups are negatively charged at this pH and, thus, sufficiently hydrophilic and mobile. In contrast, pore closure at pH 3.0 is probably driven by polymer–polymer interactions: The -COOH end groups are protonated at this pH and, thus, more hydrophobic. The pore closure is driven by the mutual attraction of hydrophobic polymer surfaces (being separated by water in a pore). The release of the surface-bound water increases the entropy of the system. This type of pore closure might for instance be of importance within PLGA-based drug delivery systems with local acidic microenvironments.

Fig. 6.

Pore closure can end the burst release phase: SEM pictures of surfaces of octreotide-loaded PLGA microparticles after different exposure times to the release medium. Adapted from Wang et al. (2002) with permission.

It has to be pointed out that for PLGA implants, the “surface area to volume ratio” is much lower compared to PLGA microparticles. If the implants' surface is initially smooth and non-porous, the “burst effect” is often very limited, if visible at all (especially when only looking at relative drug release rates). The top row in Fig. 7 shows some examples: Dexamethasone release is shown from PLGA- (or PLA-) based implants (prepared by hot melt extrusion) upon exposure to phosphate buffer pH 7.4. The initial drug loading was varied as indicated in the diagrams. Clearly, in all cases, the initial burst release (during the first day) was negligible. This can be explained by the fact that no noteworthy drug amounts had direct surface access at t = 0 (referred to the 100 % reference values).

Fig. 7.

Drug release (top row) and swelling (bottom row) of dexamethasone-loaded PLGA- or PLA-based implants prepared by hot melt extrusion upon exposure to phosphate buffer pH 7.4. The initial drug loading is indicated in the diagrams, the type of polymer at the top. Adapted from Bode et al. (2019) with permission.

Please note that in the case of in-situ forming implants, the root cause for the initial burst release is different, as briefly discussed in section 6.4.

In many cases, the importance of the observed “burst release” is too high, resulting in potentially toxic drug concentrations. Various strategies have been proposed to try to limit drug release at early time points (e.g., Yamaguchi et al., 2002; Ahmed et al., 2008; Beig et al., 2022).

4. Second phase: zero order release

The underlying mechanisms controlling drug release during the second release phase are probably the less well understood so far, compared to the first and third release phases. This is at least partially due to the lack of experimental data on the dynamic changes in the systems' key properties during this phase. Also, depending on the type of drug delivery system, the root cause(s) for the about constant drug release rate might be very different. In the following, hypotheses are described, which are likely true for a range of systems. However, experimental data allowing to evaluate their validity is not very strong up to date. It would be interesting to study this phase in more detail in the future, e.g., applying advanced characterization techniques (e.g., Yang et al., 2017) to monitor the changes in the inner and outer structure of the systems with high spatial resolution, ideally including the nanometer range.

Fig. 8 schematically shows a PLGA-based drug delivery system, which is exposed to an aqueous medium (e.g., bulk fluid or human tissue). As discussed above, limited amounts of water rapidly wet the entire system, leading to a water concentration of a “few percent” throughout the polymer phase. Consequently, ester bond cleavage starts all over the system (“bulk erosion”). However, polymer chains in surface-near regions, which are in direct contact with the surrounding aqueous phase, are exposed to much higher water concentrations (⁓100 % water). Hence, the degradation rate in these surface-near regions can be expected to be much higher compared to the rest of the drug delivery system. This has key consequences, including the following:

• (i)Each ester bond cleavage creates two new hydrophilic end groups: an -OH group and a -COOH group. Thus, the polymer network becomes rapidly more and more hydrophilic upon degradation.
• (ii)The length of the macromolecules decreases, resulting in a decreased degree of entanglement. When the PLGA chains are less entangled, the polymer network is mechanically less stable.
• (iii)More and more water-soluble degradation products are generated, creating a steadily increasing osmotic pressure, attracting water from the surrounding environment.

Fig. 8.

Schematic presentation of the dynamic changes in surface-near regions in a degrading PLGA-based drug delivery system. Due to the high concentration of water in the surrounding bulk fluid, polymer chain cleavage is accelerated in these regions, leading to substantial local swelling: A highly swollen “PLGA corona” is formed.

At a certain time point, “critical” conditions are reached in the surface-near regions: The PLGA network is so hydrophilic and mechanically fragile that it cannot withstand the “arrival” of substantial amounts of water, which are driven into the system by the generated osmotic pressure. Thus, the surface-near regions undergo substantial local swelling: They are transformed into highly swollen PLGA gels. Please note that more rapid polymer degradation at the surface compared to the bulk is characteristic for “surface eroding” drug delivery systems. As discussed above, PLGA-based devices are considered to be “bulk eroding”, since hydrolytic ester bond cleavage is much slower than water penetration into the system. However, “hybrid effects” can be observed, e.g. a predominantly “bulk eroding” system might also exhibit certain features of a “surface eroding” system (and vice versa). Overall, “bulk erosion” remains dominant for PLGA-based drug delivery systems.

Importantly, these fundamental changes in the structure of surface-near regions can be expected to strongly affect drug transport in these zones. Fig. 9 schematically illustrates 4 types of systems, differing in the physical state of the drug and potential presence of a partially or fully interconnected pore/channel network. Please note that also combinations of these different “types” of systems are possible. Furthermore, also other phenomena might be decisive (e.g., drug saturation effects): The illustrated examples are not exhaustive.

• a)The cartoon in Fig. 9A shows a PLGA-based drug delivery system, in which the drug is initially molecularly dispersed throughout the device. The crosses represent individual drug molecules or ions. The mobility of the drug in the polymer phase is very low, even if a few percent of water are present and the PLGA is in the rubbery state. However, drug molecules/ions located in surface-near regions become much more mobile once these regions undergo substantial swelling (are transformed into a highly swollen PLGA gel, as discussed above). Consequently, these drug molecules/ions rather rapidly diffuse out of the system (due to concentration gradients). In addition, since the highly swollen outer polymer layer contains considerable amounts of water, the “not yet swollen PLGA layer” right underneath is exposed to higher water concentrations than PLGA located in deeper regions. Hence, ester bond cleavage in this “subsequent” polymer layer is also accelerated and a “swelling front” very slowly moves inwards. Since the water concentration in the highly swollen layer can be expected to be about constant, this “swelling front” likely moves at a constant rate. If the drug delivery system has the geometry of a thin film, this leads to zero order release kinetics (the surface area does not change over time and the front moves at constant rate). In the case of other geometries (e.g., spheres or cylinders), the surface of the moving swelling front decreases over time, which should lead to decreasing drug release rates. However, often second release phases are limited in duration, and this effect might not be noteworthy in these short time periods. Furthermore, continuous polymer degradation in the “not yet highly swollen core” of the drug delivery system can lead to increased drug mobility in the “core”, potentially compensating the effect of the decrease in surface area of the swelling front.
• b)The scheme in Fig. 9B schematically illustrates a PLGA-based drug delivery system, in which the drug is initially distributed in the form of tiny particles throughout the device. Also in this case, the above described substantial swelling of surface-near regions can be expected. If a drug particle (e.g., a crystal) is located in such a region, once its surrounding undergoes the transition from a dense polymer phase to a highly swollen PLGA gel, it is exposed to high amounts of water and rapidly dissolves. The dissolved (individual) drug molecules/ions rather rapidly diffuse out through the highly swollen PLGA gel (due to concentration gradients). The reason why this leads to about constant drug release rates is explained in the following section, together with the next case.
• c)The scheme in Fig. 9C is similar to the one Fig. 9B, but in this case, the drug particles are partially connected via channels/pores. Once such a network of particles/channels/pores gets into contact with a highly swollen surface-near region, drug release can be expected to be rather rapid for all particles being part of the network. Thus, this type of drug release mechanism can be expected to lead to occasional “sudden” drug release events, concerning only parts of the total drug loading. Interestingly, this type of drug release behavior has been reported for different types of PLGA-based microparticles. Fig. 10A shows an example (Tamani et al., 2019a). Diprophylline from single PLGA microparticles into phosphate buffer pH 7.4 is shown. Each curve corresponds to a specific microparticle. As can be seen, each microparticle releases the drug “in its own way”. The red flashes indicate occasional “sudden” partial drug release events. SEM pictures of cross-sections and X-ray diffraction patterns of the microparticles revealed that at least parts of the drug were distributed in the form of tiny crystals all over the system before exposure to the release medium. Many of them were likely interconnected via pore/channels. Thus, the inner structure probably corresponds to the scheme in Fig. 9C. Once an interconnected particle/pore/channel network gets into contact with a highly swollen surface-near region, all of its drug is probably rather rapidly released. Importantly, each microparticle has its own inner structure, in particular drug crystal distribution. So, each microparticle “releases its drug in its own way”. Some of them release parts of their drug loads at certain time points, whereas others do not release any drug in this example during the first week (Fig. 10A). The latter behavior can probably be explained by the fact that in these specific microparticles no drug particles were located in surface-near regions undergoing substantial swelling.

  Since the “sudden” partial drug release events are randomly occurring (because the drug crystals are randomly distributed throughout the system), summing up all single partial release events from hundreds of thousands/millions of microparticles leads to about constant drug release rates (please also see Fig. 3 and the discussion in section 2). The corresponding release curves observed with ensembles of microparticles of this type are shown in Fig. 5 (three ensembles are shown, differing in their mean particle size). As can be seen, during the second release phase, the release rate is indeed about constant in all cases. This is a good example for a system, for which the interpretation of the drug release patterns observed only from ensembles of microparticles can be misleading when trying to understand the underlying drug release mechanisms. Fig. 10B and C illustrate the swelling behavior and dynamic changes in the morphology of single microparticles of this type upon exposure to the release medium. Clearly, during the first week, overall system swelling is limited. Looking at the microparticles' surfaces observed by optical microscopy on day 0 and day 3, it can be seen that the sharpness of the interface “microparticle – release medium” seems to decrease with time. The interface also appears to become less regular, which might serve as an indication for the swelling of surface-near regions.

• d)The cartoon in Fig. 9D schematically illustrates a PLGA-based drug delivery system, in which the drug is distributed in the form of tiny particles (e.g., crystals) throughout the device and a continuous network of pores/channels interconnects them. The channels/pores are sufficiently large to allow for water penetration into the core. Once surface-near regions undergo substantial swelling, water can access the particle/pore/channel network. At least parts of the drug particles dissolve and diffuse through the water-filled pore/channel network and the highly swollen surface-near regions into the release medium. So, there are two resistances for drug transport: one within the “core”, and one in the highly swollen “corona”. If the particle/pore/channel network in the core contains significant amounts of water (leading to high drug mobility), the mass transport resistance in the swollen hydrogel corona is likely dominant. The thickness of this “highly swollen PLGA gel corona” might be about constant during a certain time period (since the swelling front advances very slowly and continuously erodes at the interface “hydrogel – release medium”). If drug dissolution and diffusion in the (sufficiently large) pores/channels in the ‘not yet swollen” core is rapid, the interconnected network of the core is saturated with drug. Combined with perfect sink conditions in the release medium, this can be expected to result in zero order release kinetics (similar to a reservoir system with a constant activity source (Siepmann and Siepmann, 2012). An example for a delivery system, which might follow this release mechanism is shown in Fig. 11 (Tamani et al., 2019b). PLGA-based microparticles loaded with 6–7 % caffeine were prepared by an oil-in-water (O/W) emulsion solvent extraction/evaporation technique. Resomer RG 504H was used (50:50 lactic acid: glycolic acid, acid end groups; Evonik, Germany). X-ray diffraction and Scanning Electron Microscopy (SEM) revealed that the drug was distributed throughout the microparticles in the form of tiny crystals, which were highly interconnected via pores/channels. On the left hand-side of Fig. 11A a cross-section of a microparticle before exposure to the release medium is shown. Drug release measurements from ensembles of particles (differing in their mean size) exhibited a limited burst release on day 1, followed by a zero order release phase during about 1 week, irrespective of the particle size. Fig. 11D shows that the overall swelling of the microparticles was very limited during this time period (the diameter remained about constant). Importantly, drug release measurements from single microparticles (Fig. 11C) revealed that many of them released the drug at an about constant rate during the second release period, which might be explained by the above described mechanism (reservoir system with contact activity source, cartoon in Fig. 11A in the middle). Please note that in certain cases (e.g., the blue curve), the release patterns were less continuous, and showed some “steps”, which suggests that some of the drug might also be released according to the mechanisms illustrated in Figs. 9B or C. When comparing the caffeine release kinetics from the ensembles of microparticles differing in their mean size in Fig. 11B, it can be seen that the smaller microparticles release the drug at a higher relative rate during the second release phase than the larger particles. This can be attributed to the fact that their relative surface area is higher.

Fig. 9.

Schematic presentation of potential root causes for the 2nd drug release phase (exhibiting zero order kinetics): Four different types of systems are illustrated (A-D). The drug is initially dispersed in the PLGA matrix at the molecular level (dissolved) or in the form of tiny drug particles (e.g., crystals), which are optionally interconnected via channels/pores. Details are described in the text.

Fig. 10.

Behavior of single diprophylline-loaded PLGA microparticles upon exposure to phosphate buffer pH 7.4: A) Drug release, B) Swelling, and C) Morphological dynamic changes observed by optical microscopy (the same microparticle is shown in each row, its initial size is indicated on the left-hand side). In A) and B), each curve corresponds to a specific single microparticle. The same colour corresponds to the same microparticle. The red flashes indicate “sudden” partial drug release events. Adapted from Tamani et al. (2019a) with permission. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 11.

Caffeine-loaded PLGA-based microparticles prepared by an oil-in-water (O/W) emulsion solvent extraction/evaporation technique. The polymer was a 50:50 PLGA with acid end groups (Resomer RG 504H), the initial drug loading was about 6–7 %. A) Cross-section of a microparticle before exposure to the release medium and two cartoons illustrating the dominant mass transport mechanisms for the 2nd and 3rd release phases. B) Drug release from ensembles of microparticles (differing in their mean particle size +/− standard deviation). C) Drug release from single microparticles (each curve corresponds to a specific microparticle, the initial diameter is indicated in the diagram). D) Swelling of single microparticles (each curve corresponds to a specific microparticle, the initial diameter is indicated in the diagram). Adapted from Tamani et al. (2019b) with permission.

Please note that the swollen PLGA “corona” might also have a heterogenous inner structure, e.g. zones with higher polymer concentrations and zones with lower polymer concentrations. The latter might for example stem from pores, which were initially present at the system's surface (and contributed to the burst effect), but which were closed due to local PLGA swelling. The polymer density in these “former pore zones” can be expected to be lower compared to zones, which were initially dense PLGA phases. So, the swollen PLGA “corona” might also be seen as a “skin” through which the drug has to diffuse, and which has been formed upon pore closure (Park, 2025; Elkharraz et al., 2011). It would be interesting to study the structure of the PLGA “corona” or “skin” in more detail in the future (taking care that the applied methods do not create artifacts during sample preparation, e.g., for SEM analysis).

5. Third phase: again rapid release leading to complete drug exhaust

In many cases, the onset of substantial swelling of the entire drug delivery system can explain the beginning of the third release phase, which leads to complete drug exhaust. Fig. 12 schematically shows the involved key phenomena. For reasons of simplicity, the swelling of surface-near regions is not illustrated. Initially, the device is hydrophobic, even in the case of PLGA with acid end groups: The hydrophobicity of the polymer backbone dominates. The drug is dispersed within the polymer matrix: either molecularly (dissolved) and/or in the form of tiny particles (which can be crystalline and/or amorphous). Upon exposure to the release medium (or administration into the patient's body), only limited amounts of water penetrate into the system. But these amounts are sufficient to start ester bond cleavage throughout the device (“bulk erosion”), generating new hydrophilic end groups. As described in more detail in section 4, this has three major consequences: (i) The polymer phase becomes more and more hydrophilic. (ii) The PLGA network is more and more fragile (due to the decreasing molecular weight and degree of chain entanglement). (iii) More and more water-soluble degradation products (monomers and short oligomers) generate a steadily increasing osmotic pressure inside the system. Once a “critical stage” is reached, large amounts of water are driven into the system. In contrast to the second release phase, not only surface-near regions are affected, but the entire devices. As illustrated on the right-hand side of Fig. 12, this allows potentially remaining drug particles to dissolve (only dissolved drug being able to diffuse), and fundamentally increases the mobility of dissolved drug molecules/ions. Consequently, the resulting drug release rate increases, and this phase leads to complete drug exhaust. The optical microscopy pictures in Fig. 10C show examples of microparticles undergoing this fundamental transformation: in that case, between day 9 and 17. In the following, several examples are described, which have been reported in the literature: for different types of drugs, PLGA grades and delivery systems.

Fig. 12.

Schematic presentation of the “orchestrating role” of substantial entire system swelling for the onset of the third (again rapid) drug release phase: Once the polymer network becomes sufficiently hydrophilic & mechanically fragile, substantial amounts of water are driven into the system due to the osmotic pressure generated by water-soluble degradation products (and potentially the drug itself, if it is water soluble). Consequently, remaining non-dissolved drug particles can dissolve, and the mobility of the dissolved drug molecules/ions substantially increases.

Fig. 13A illustrates ketoprofen release from PLGA-based microparticles measured upon exposure to phosphate buffer pH 7.4 (containing 0.02 % Tween 80). The black symbols refer to the left y-axis and illustrate drug release. The open symbols refer to the right y-axis, showing the dynamic changes in the microparticles' diameter. The pictures in Fig. 13B show optical microscopy pictures of microparticles after 8.3 and 10 d exposure to the release medium, respectively. The ovals in the diagram highlight the onsets of substantial microparticle swelling and of the third release phase. As can be seen, substantial swelling of the entire systems is followed by the onset of the third drug release phase.

Fig. 13.

A) Ketoprofen release from (left y-axis) and swelling of PLGA microparticles (right y-axis), loaded with 0.6 % ketoprofen upon exposure to phosphate buffer pH 7.4 (containing 0.02 % Tween 80). B) Optical microscopy pictures of microparticles after 8.3 and 10 d exposure to the release medium. Adapted from Gasmi et al. (2015) with permission.

Diprophylline release from single PLGA microparticles in phosphate buffer pH 7.4 is shown in Fig. 10A. The green rectangle highlights the third drug release phase, which starts as soon as substantial entire system swelling sets on (Fig. 10B). Looking at the optical microscopy pictures in Fig. 10C, it becomes evident that the considerable system swelling starting after about 9 d in this case, fundamentally changes the conditions for drug release. Please note that all microparticles undergo this drastic change at about the same time: This is because the water content in the polymer phase can be expected to be about roughly the same throughout all systems, following rapid wetting after exposure to the release medium. Thus, the “critical conditions”, at which large amounts of water are osmotically driven into the system, are reached at about the same time point.

Fig. 11C shows caffeine release from single PLGA microparticles upon exposure to phosphate buffer pH 7.4. Again, the green rectangle highlights the third drug release phase, which starts after about 1 week in this case. Looking at Fig. 11D, it can be seen that substantial entire system swelling preceded the onset of this final drug release phase. Thus, drug release in the 3rd phase is likely initiated by the penetration of considerable amounts of water into the system (as illustrated by the cartoon on the right-hand side of Fig. 11A).

Please note that there is no “universal” critical polymer molecular weight threshold value for all types of PLGA-based drug delivery systems, at which substantial entire system swelling sets on. This is because the length of the macromolecules is not the only factor, which plays an important role: The osmotic pressure within the device, the type of PLGA end groups (acids or esters) as well as the hydrophilicity/hydrophobicity and amount of drug present in the system, can also be decisive. Thus, it is not surprising that different “critical polymer molecular weights” have been reported in the literature for the beginning of the third release phase.

The “orchestrating role” of substantial entire system swelling for the control of drug release from PLGA-based delivery systems can of course also be questioned: Instead of swelling, for instance, polymer degradation might be the dominant process. However, at least for certain systems, experimental evidence shows that this is likely not the case (Bassand et al., 2022b). For example, Fig. 14A illustrates ibuprofen release from PLGA implants upon exposure to well agitated phosphate buffer pH 7.4 or upon inclusion into an agarose gel (which is exposed to well agitated phosphate buffer pH 7.4). The schemes on the right-hand side illustrate the two set-ups (Eppendorf tubes were used in both cases). Clearly, the burst release was negligible in both cases (because noteworthy amounts of drug did not have direct surface access at the beginning). Importantly, the second release phase was much shorter in the bulk fluid set-up: After about 5 d, ibuprofen release was accelerated. Importantly, substantial entire system swelling preceded the beginning of the third release phase, as illustrated by the open symbols in Fig. 14B. In contrast, when the implants were embedded within an agarose gel, the latter sterically hindered substantial system swelling, delaying its onset (likely up to day 9/10, filled symbols in Fig. 14B). This results in a delayed onset of the third release phase from the implants embedded within agarose gel (filled symbols in Fig. 14A, the 3rd phase did not yet start at the end of the observation period). The pictures on the right-hand side of Fig. 14B show implants exposed to the bulk fluid (top) or embedded within agarose gel (bottom) after 10 d. Clearly, ibuprofen release can be expected to be much faster in the highly swollen implants exposed to the bulk fluid compared to the much less swollen systems embedded in agarose gel. Importantly, the experimental set-up (agitated bulk fluid versus agarose gel) did not affect the polymer degradation rate, as can be seen in Fig. 14C: In both cases, the polymer molecular weight decreased exponentially from the beginning at a similar rate. If PLGA degradation (and not polymer system swelling) would play an “orchestrating role” for the control of drug release, drug release should be similar in the two set-ups.

Fig. 14.

Behavior of ibuprofen-loaded PLGA implants upon exposure to well agitated phosphate buffer pH 7.4 or upon inclusion into an agarose gel: A) Drug release, B) Implant swelling, C) Polymer degradation. The asterisk indicates that the average polymer molecular weight (Mw) was below 5 kDa. Adapted from Bassand et al. (2022b) with permission.

Fang et al. (2019) investigated donepezil-loaded PLGA microparticles in vitro and in vivo: upon s.c. administration to mice. They studied a 75:25 PLGA, with an average polymer molecular weight of 47 kDa (obtained from Jinan Daigang Biomaterial Company, Shandong, China). The microparticles were prepared using an oil-in-water emulsion - solvent evaporation technique. Fig. 15A illustrates the observed drug release patterns and Fig. 15B the experimentally measured water uptake kinetics of the systems in vitro (filled symbols) and in vivo (open symbols). As can be seen, in vitro the water uptake rate steeply increased after about 1 week (filled symbols). This time point coincides with the onset of the 3rd drug release phase in vitro. In contrast, in vivo, substantial water uptake was observed right from the beginning. This was consistent with the immediate onset of rapid drug release in vivo. The reason for the differences in the swelling kinetics in vitro versus in vivo are not fully understood.

Fig. 15.

In vitro and in vivo behavior of donepezil-loaded PLGA microparticles: A) Drug release, B) Water uptake. In vivo studies were conducted in mice (s.c. administration). Adapted from Fang et al. (2019) with permission.

A further example is illustrated in Fig. 16: The group of Diane Burgess (one of the leading teams working on PLGA-based drug delivery systems) studied dexamethasone-loaded PLGA microparticles prepared by an oil-in-water emulsion – solvent evaporation technique, which were embedded in a poly (vinyl alcohol) (PVA, Mw 30–70 kDa) gel (Gu et al., 2016). The PLGA was a 50:50 copolymer with acid end groups and a molecular weight of about 25 kDa, obtained from Boehringer-Ingelheim (Resomer RG503H). As can be seen in Fig. 16A, drug release was tri-phasic, the 3rd release phase started after about 9 d. Fig. 16B shows that substantial system swelling set on right before: The green, red and violet symbols refer to different microparticle concentrations in the PVA gel. The blue symbols illustrate the swelling behavior of pure PVA hydrogel.

Fig. 16.

A) Drug release from, and B) Swelling of dexamethasone-loaded PLGA microparticles embedded within a PVA hydrogel. The microparticles were prepared using an oil-in-water emulsion – solvent evaporation technique. The PLGA was a 50:50 copolymer with acid end groups (Resomer RG503H). Adapted from Gu et al. (2016) with permission.

6. Potential additional phenomena

A variety of additional phenomena might be involved in the control of drug release from PLGA-based drug delivery systems. For example, if the device is exposed to fluid flow and if the flowing liquid can access the system via sufficiently large pores/channels, convective mass transport likely also plays a role. Convection is much faster than diffusion: dissolved drug molecules/ions (and even particles) can be very rapidly transported in a flowing liquid. Furthermore, the addition of pore formers can alter the underlying drug release mechanisms. For example, the group of Roland Bodmeier (who published seminal articles in the field) reported on the addition of different amounts of sucrose to PLGA films, acting as porogen (Zhang and Bodmeier, 2024). In such systems significant time-dependent changes in the system's porosity due to porogen leaching can play a key role for the control of drug release. Also the addition of hydrophilic polymers to PLGA-based drug delivery systems can fundamentally alter the underlying drug release mechanisms, as suggested by altered shapes of the drug release curves (e.g., Hamoudi-Ben Yelles et al., 2017; Johnson et al., 2025). Furthermore, the use of polyethylene – poly(lactic-co-glycolic acid) (PEG-PLGA) copolymers can be used to adjust the resulting release kinetics, altering the conditions for drug transport (e.g., de Souza et al., 2021; Lehner et al., 2025). In addition, phenomena related to the systems' surface might be of key importance, especially in the case of PLGA-based nanoparticles, for which the relative surface area is fundamentally higher compared to microparticles and implants.

It was beyond the scope of this article to discuss such potential additional phenomena in more detail (the article would have become too long). In the following, only a few other examples are described in “limited” depth.

6.1. Autocatalysis

As discussed above, PLGA-based drug delivery systems are degrading in the bulk. Importantly, the hydrolysis of ester bonds generates shorter chain acids. Thus, acids are created throughout the device and protons (or more precisely: hydroniums ions) can accumulate within the system. Due to concentration gradients, the protons and water-soluble acids diffuse out into the surrounding environment, where they are neutralized. Also, bases from the release medium/environmental tissue can diffuse into the delivery system and neutralize the generated acids. However, diffusional mass transport through the drug delivery system is generally relatively slow, and the rate at which the acids are generated within the device can be higher than the rate at which they are neutralized. In these cases, the local pH within the dosage form can significantly decrease. For example, Ding and Schwendeman (2008) reported pH values as low as “3” in degrading PLGA microparticles, observed by Confocal Laser Scanning Microscopy. Also, Brunner et al. (1999) detected pH values <4.7, using a pH-sensitive spin probe and electron paramagnetic resonance measurements in degrading PLGA microparticles. The drop in local pH is generally most pronounced at the center of the delivery systems, because the diffusion pathways to the surface are the longest. Importantly, such local acidities catalyze further ester bond hydrolysis, and polymer degradation is accelerated. This phenomenon is also called “autocatalysis”: One of the products of the chemical reaction (an acid) catalyzes the reaction itself.

Fig. 17A schematically shows the acidification in differently sized PLGA microparticles during drug release: The effect is more pronounced in larger systems, because the diffusion pathways for the acids and bases are longer. Since polymer degradation is particularly accelerated at the center of the device, the latter can become highly porous with time. Fig. 17B illustrates the outer and inner morphology of PLGA microparticles, which have been exposed to phosphate buffer pH 7.4 for 7 d. Great caution must be paid, because the samples were dried prior to SEM analysis: The highly swollen PLGA networks changed their structure during sample preparation. Nevertheless, looking at the pictures, it becomes obvious that the cores of the larger microparticles have been more intensively degraded than the cores in smaller microparticles. Thus, drug mobility within the larger microparticles can be expected to increase more rapidly than in the smaller microparticles. This has important consequences for the resulting drug release kinetics: The release rates from (initially non-porous) microparticles with mean radii ranging from 7 to 53 μm were very similar, despite the increase in the length of the diffusion pathways by a factor of 7. It seems that this “diffusion pathway length” effect is compensated by the increasing drug mobility due to more pronounced autocatalysis in larger microparticles. In contrast, in the case of initially highly porous microparticles, the diffusion rates of acids and bases are much higher, thus, autocatalytic effects are less pronounced. This resulted in decreasing relative release rates with increasing system size in the case of lidocaine-loaded PLGA microparticles with radii varying between 7 and 54 μm (Klose et al., 2006). The creation of acidic micro-environments within the drug delivery system can also be detrimental for various drugs. For example, protein drugs generally lose their activity.

Fig. 17.

Autocatalysis in PLGA-based drug delivery systems: Hydrolytic ester bond cleavage generates new -COOH groups throughout the device. This can lead to local drops in the micro-pH. Since ester bond cleavage is catalyzed by protons (hydronium ions), this accelerates polymer degradation. A) Cartoon illustrating local drops in micro-pH in smaller and larger microparticles (white = neutral pH, deep red = highly acidic pH). B) SEM pictures of surfaces & cross-sections of smaller and larger PLGA microparticles after 7 d exposure to phosphate buffer pH 7.4. Adapted from Siepmann et al. (2005) with permission. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

6.2. Drug-PLGA interactions

The mobility of dissolved drug molecules/ions within a PLGA-based drug delivery system can strongly depend on the affinity of the drug to the polymer: to the backbone and/or the end groups of the macromolecules (e.g., -OH, -COOH, -COO⁻ in the case of PLGA) (Hirota et al., 2016). For example, attractive forces between positively charged drug ions and negatively charged -COO⁻ end groups can decrease the “apparent” drug mobility. Such interactions can be expected to occur within the polymer phase, at the latter's surfaces (including the walls of inner pores/channels) as well as within water-filled pores/channels containing “more or less” PLGA. It has for instance been reported that lidocaine release from PLGA-based microparticles prepared using an oil-in-water (O/W) solvent extraction/evaporation technique was much slower compared to the release of ibuprofen (protonated lidocaine cations being attracted by negatively charged -COO⁻ ions, but not ibuprofen molecules or anions) (Klose et al., 2008).

6.3. Saturation effects

Drug saturation effects can potentially very strongly affect the resulting drug release kinetics from PLGA-based delivery systems. It has to be pointed out that saturation might occur in the surrounding environment and/or within the dosage form. Fig. 18 schematically illustrates four possible cases. If saturation occurs inside the dosage form, the concentration gradients of the drug are limited by its solubility within the system. Importantly, only dissolved drug can diffuse. Non-dissolved drug particles “have to wait” until they can dissolve. Fig. 19 shows an example for such a delivery system: Dexamethasone release from PLGA-based microparticles is shown during the 3rd (final) drug release phase. The microparticles were prepared by an oil-in-water (O/W) emulsion solvent extraction/evaporation technique. The PLGA was a 50:50 copolymer with acid end groups (Resomer RG 504H, Evonik, Germany). The initial drug loading was varied from 2.4 to 61.9 %, as indicated in the diagram. The third release phase started shortly after the onset of substantial swelling of the entire microparticles (data not shown). In this system, despite the considerable amounts of water in the highly swollen PLGA gels, the whole amount of dexamethasone could not dissolve (due to its limited solubility). Thus, non-dissolved and dissolved dexamethasone co-existed, as illustrated in the cartoon in Fig. 18B. Independent of the initial drug loading, the highly swollen PLGA gels were saturated with drug. Sink conditions were provided in the surrounding release medium (except for the highest drug loading at late time points). This led to about constant drug release rates, as can be seen in Fig. 19 (drug saturation inside the system combined with sink conditions outside the device). Since the 100 % reference value for drug release increases with increasing drug loading, the slope of the corresponding relative release curves decreases.

Fig. 18.

Schematic presentation of potentially occurring drug saturation effects inside and outside a PLGA-based delivery system during drug release. Four cases can be distinguished, as illustrated (A-D). The cuboids represent drug particles (e.g., crystals), whereas the crosses represent dissolved (individual) drug molecules/ions. Adapted from Siepmann and Siepmann (2020) with permission.

Fig. 19.

Impact of the limited solubility of dexamethasone within PLGA-based microparticles on the drug release rate during the 3rd release phase upon exposure to phosphate buffer pH 7.4. Sink conditions were provided in the surrounding bulk fluid throughout the observation period (except for the highest drug loadings at very late time points, for which 40–50 % drug saturation was observed) Adapted from Gasmi et al. (2016) with permission.

A interesting study from the group of Karsten Maeder (who is one of the world-wide leading pioneers in the field) reports on nicardipine release from PLGA implants prepared by compression at 100 °C (Zlomke et al., 2019). They worked with a 50:50 PLGA with acid end groups (Resomer RG 502H, Evonik, Germany). Two drug loadings were investigated: 5 and 10 % (w/w). Fig. 20 shows the observed drug release kinetics. The third release phase set on after about 1 week in this case, shortly after the beginning of substantial entire system swelling (data not shown). Clearly, the relative release rate was higher for implants loaded with 5 % nicardipine compared to 10 % drug, which can be explained by the limited solubility of nicardipine within the system and the difference in the 100 % reference values.

Fig. 20.

Drug saturation effects within the dosage form: Nicardipine release from PLGA implants prepared by compression at 100 °C. The initial drug loading was 5 or 10 %, as indicated in the diagram. Adapted from Zlomke et al. (2019) with permission.

6.4. In-situ forming implants

In-situ forming implants are injected as liquid or semi-solid formulations into the patient's body, followed by rapid hardening/solidification (Kempe et al., 2008; Solorio et al., 2010; Kempe and Maeder, 2012; Solorio et al., 2012). Different triggering mechanisms can be used to induce implant formation, such as solvent exchange (Parent et al., 2013; Bode et al., 2018) or the increase in temperature from 20 to 37 °C upon administration (Kempe and Maeder, 2012). Compared to preformed implants, the initial burst release is generally much higher, because the systems are still liquid/semi-solid at early time points (thus, drug mobility is much higher and convective mass transport might also play a key role). Depending on the type of in-situ forming implant, the underlying mechanisms controlling device formation and subsequent drug release can substantially vary. Among other aspects, the inner and outer structures of the implants are of key importance. For example, implants with a “core – shell” structure might be formed. Also, continuous pore/channel networks might be created and be decisive for the control of drug release. Furthermore, the size and geometry of the hardened implants can substantially vary, determining the lengths of diffusion pathways and importance of potentially occurring autocatalytic effects. In-situ forming PLGA implants are not treated in this article, to keep its length “reasonable”.

7. Overview on release mechanisms and some more practical examples

Fig. 21 schematically summarizes the potential key mechanisms controlling drug release from PLGA-based delivery systems during the three phases. In brief, the burst release can be caused by drug molecules/ions/particles, which have direct surface access upon exposure to the release medium/administration to the patient. Pore closure due to local PLGA swelling might end this phase. Zero order release kinetics in the second phase might be caused by the fact that PLGA in surface-near regions is exposed to higher water concentrations and, thus, rather rapidly undergoes substantial swelling. Drug particles located in these regions dissolve, and drug molecules/ions become highly mobile and are rather rapidly released. The swelling front moves at a very low, about constant rate. Also, a “reservoir type”-like behavior (permeable shell combined with drug saturation in the core and sink conditions in the surrounding environment) might explain this release phase. The third release phase can be caused by the onset of substantial entire device swelling, increasing the mobility of all remaining drug. The different release phases can be more or less pronounced, as illustrated in the following examples.

Fig. 21.

Overview: Schematic presentation of the (potential) underlying drug release mechanisms in PLGA-based delivery systems. Each release phase is likely dominated by different phenomena. Several possible cases are illustrated for each phase. Please note that two or more “cases” might also be combined.

Fig. 7 shows the behavior of dexamethasone-loaded PLGA- or PLA-based implants prepared by hot melt extrusion upon exposure to phosphate buffer pH 7.4. Dug release is illustrated at the top, system swelling at the bottom. The initial drug loading was varied between 1 and 15 %, as indicated in the diagrams. The type of polymer is shown at the top. As discussed in section 3, no noteworthy burst release was observed in any case, the amounts of drug with direct surface access at the beginning was negligible with respect to the total drug loading (which is in part due to the much lower “surface to volume ratio” compared for instance to microparticles, and the preparation technique). Furthermore, the second release phases showed only very limited drug release, irrespective of the type of polymer and initial drug loading. This can be explained by the fact that - given the size of the implants - the relative amounts of drug present in surface-near regions are very small (and the absence of pores/channels with direct surface access). Depending on the type of polymer, the third drug release phase started after a few days (PLGA 50:50: RG 502H), about 10 d (PLGA 75:25: RG 752H) or roughly 6 weeks (PLA homopolymer: R 202H). Importantly, these differences in lag-time for drug release corresponded very well with the observed lag-times for the onsets of substantial system swelling, as can be seen in the bottom row of Fig. 7: As soon as the system is sufficiently hydrophilic, the polymeric networks sufficiently weak and the osmotic pressure high enough, large amounts of water penetrate into the implants (e.g., up to 1500 % referred to the initial implant weight in the case of PLGA 75:25: RG 752H). Consequently, remaining non-dissolved drug can dissolve, and the mobility of dissolved drug steeply increases in the highly swollen PLGA network (Fig. 21, 3^rd Phase, cartoon at the bottom). The observed differences in the lag-times in Fig. 7 for substantial entire implant swelling can be explained by the chemistry of the polymers: Due to sterical hindrance, hydrolytic ester bond cleavage is slowed down by the presence of the methyl group in lactic acid compared to glycolic acid. Thus, the critical conditions allowing for the penetration of substantial amounts of water into the implants are reached at later time points in systems containing more lactic acid and the observed lag-times increase in the following rank order: PLGA 50:50 < PLGA 75:25 < pure PLA.

Another example, illustrating the importance of the chemistry of the PLGA is illustrated in Fig. 22: Dexamethasone-loaded implants were prepared by melting and molding. The “lactic acid:glycolid acid” ratio was 50:50 or 75:25, the initial average polymer molecular weight (Mw) was varied from 38 to 92 kDa, the end groups were either acids or esters (as indicated in the figure legend). The drug loading was 9 % (w/w) in all cases. Dexamethasone release as well as the dynamic changes in the systems' wet mass and polymer molecular weight are shown upon exposure to phosphate buffer pH 7.4. As can be seen, in all cases, the initial burst release was very limited, due to the absence of noteworthy relative amounts of drug with direct surface access at the beginning (e.g., no surface pores were visible on SEM pictures, data not shown). Also in the second release phases, dexamethasone release was very limited in all cases. This is due to the macroscopic size of the implants (diameter and length were in the order of millimeters), the limited drug loading (9 %) and the absence of a continuous macroscopic network of pores/channels/particles with direct surface access allowing for rapid drug release. Thus, only minor amounts of drug became mobile upon swelling of surface-near regions during this period. Importantly, after a certain lag-time, the third release phase set on and led to complete drug exhaust, irrespective of the monomer ratio, initial polymer molecular weight and type of end groups. The observed lag-time significantly depended on the type/grade of PLGA, increasing in the following order: PLGA 50/50 A (38 kDa) < PLGA 50/50 A (92 kDa) < PLGA 75/25 A (70 kDa) < PLGA 50/50 E (50 kDa). This can be explained by the difference in the initial hydrophilicity of the systems and the initial length of the macromolecules. As discussed above, substantial swelling of the entire device sets on as soon as the system is sufficiently hydrophilic, the polymeric network sufficiently fragile and the osmotic pressure sufficiently high. If the initial polymer molecular weight is higher, it takes more time to get to this point, e.g., comparing PLGA 50/50 A (92 kDa) and PLGA 50/50 A (38 kDa) in Fig. 22A. If the macromolecules contain more lactic acid, ester bond cleavage is slower and the lag-time is prolonged, e.g., comparing PLGA 75/25 A (70 kDa) and PLGA 50/50 A (92 kDa) in Fig. 22A (overcompensating the polymer molecular weight effect in this case). If the end groups are esters (and not -COOH groups), the system is much more hydrophobic at the beginning. This reduces the amounts of water, which can wet the polymer phase, leading to slower PLGA degradation and prolonged lag-times, e.g. comparing PLGA 50/50 E (50 kDa) and PLGA 50/50 A (92 kDa) in Fig. 22A (also in this case, the effect of the polymer molecular weight is overcompensated). As can be seen, in Fig. 22B, the onset of the third release phases were preceded by the onset of substantial entire system swelling in all cases. This onset was closely related to the time point at which the “critical conditions” for substantial entire system swelling were reached, including a “certain” polymer molecular weight (Fig. 22C).

Fig. 22.

Impact of the PLGA chemistry: Dexamethasone-loaded PLGA implants prepared by melting & molding. A) Drug release, B) Dynamic changes in wet mass, and C) Dynamic changes in the polymer molecular weight (Mw) upon exposure to phosphate buffer pH 7.4. The initial average polymer molecular weight (Mw) and type of end groups (A: acid end groups, E: ester end groups) are indicated in the legend. Adapted from Wachowiak et al. (2023) with permission.

The impact of the temperature on the behavior of diprophylline-loaded PLGA microparticles is shown in Fig. 23. The release medium was phosphate buffer pH 7.4. The temperature was 37 °C or 20 °C, as indicated at the top. Please note the different time scales of the diagrams on the left- versus right-hand side. Fig. 23A illustrates drug release from ensembles of microparticles, differing in their mean size. The release kinetics observed at 37 °C were discussed in more detail in 4, 5. Clearly, at 20 °C drug release was substantially slower. After 100 d, the third release phase did not yet even start (except for maybe the largest microparticles at the very end of the observation period). This can be explained by the fact that at 20 °C, the ester hydrolysis rate is reduced (as can be seen in Fig. 23D), and the “critical” conditions initiating substantial entire system swelling are not yet reached (Fig. 23C, right-hand side). In contrast, at 37 °C, polymer degradation is faster (Fig. 23D), and the “critical” conditions are provided after about 9 d (Fig. 23C, left-hand side), initiating the onset of the third release phase (Fig. 23A, left-hand side). When looking at the diagrams on the right-hand side in Figs. 23A-C, it can be seen that the 2nd release phases at 20 °C can be explained by occasional, partial drug release events, which can be attributed to the fact that tiny drug crystals or networks of drug crystals/pores/channels become access to swollen surface-near regions (cartoons in Figs. 9B and C).

Fig. 23.

Impact of the temperature (indicated at the top) on the behavior of diprophylline-loaded PLGA microparticles upon exposure to phosphate buffer pH 7.4: A) Drug release from ensembles of microparticles (differing in their mean size, as indicated in the diagram +/− standard deviation). B) Drug release from single microparticles (occasional, partial drug release events are highlighted in red). C) Swelling behavior of single microparticles (the green rectangles mark the period with substantial entire system swelling). D) Decrease in the polymer molecular weight of ensembles of microparticles. Adapted from Tamani et al., 2019a, Tamani et al., 2021 with permission. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

8. Each microparticle releases in its own way

To again emphasize the fact that each microparticle is different and releases the drug “in its own way”, another example is illustrated in Fig. 24. It nicely demonstrates that this is true even if the drug loading, microparticle size and inner & outer system structures are similar: The variability of drug release from single microparticles can still be substantial. The drug release behavior of PLGA microparticles loaded with 36 or 48 % ibuprofen is shown (top row), as well as their swelling kinetics (bottom row). For a given drug loading, the same symbols refer to the same microparticles. The systems were exposed to well agitated phosphate buffer pH 7.4. The microparticles were prepared by an oil-in-water (O/W) solvent extraction/evaporation technique using either a classical “beaker” (results marked in blue) or a “microfluidics device” (results marked in green). As can be seen, substantial microparticle swelling started after about 3 d. Some of the microparticles released their entire drug loading well before, while others showed limited drug release until the substantial entire system swelling set on. This was true, irrespective of the initial drug loading and equipment used for microparticle preparation. SEM pictures of cross-sections of the microparticles as well as X-ray microcomputed tomography measurements revealed that continuous drug particle/pore/channel networks exist in these systems, and a more or less thin PLGA layer separates these interconnected networks from the release medium. Thus, the scheme in Fig. 9D likely best describes the microparticles' structure (considering a continuous, non-porous PLGA film at the surface before exposure to the release medium). Upon contact with water, the surface-near regions start to substantially swell. In case of microparticles with “weak points” (locations with very thin PLGA layers separating the release medium from the continuous drug particle/pore/channel network), the drug is rapidly released once this weak point swells or ruptures (due to mechanical stress caused by the convective flow in the well stirred release medium). If a microparticle does not have any “weak points” at the surface, it “has to wait” until substantial entire system swelling sets on after 3 d before drug release starts (Fig. 12).

Fig. 24.

Drug release (top row) and swelling (bottom row) of PLGA microparticles loaded with: A) 36 %, or B) 48 % ibuprofen upon exposure to well agitated phosphate buffer pH 7.4. The microparticles were prepared by an oil-in-water (O/W) solvent extraction/evaporation technique using two types of equipment: a classical “beaker” (blue curves/symbols) or a “microfluidics device” (green curves/symbols). The behavior of single microparticles is illustrated, their initial diameters are indicated on the right-hand side of the diagrams. The same symbols refer to the same microparticle in a column. Adapted from Lefol et al. (2025) with permission. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

9. Conclusions

Various phenomena are occurring when drugs are released from PLGA-based delivery systems. However, during each of the (typically observed) 3 phases of drug release, often only one process (or a combination of a few) is really dominant. Depending on the type of device (e.g., microparticle versus implant), qualitative and quantitative composition of the system (e.g., drug loading, type/grade of PLGA, type of drug), manufacturing method and (not to be forgotten!) environmental conditions, the relative importance of the involved phenomena can fundamentally vary. The general hypotheses described in this article can hopefully facilitate to better understand how a specific drug delivery system controls its release rate. This knowledge can be very useful for the optimization of new drug products, upscaling of manufacturing procedures and troubleshooting at industrial scale. Eventually, the therapeutic efficacies and the safety of the treatments for the patients can be improved.

Two aspects, which are often neglected/underestimated in the field of PLGA-based drug delivery systems are the following: (i) Substantial entire device swelling likely plays an “orchestrating role” in numerous systems, initiating the onset of the final rapid drug release phase (allowing all drug to dissolve and become mobile). (ii) In the case of microparticles, only looking at the behavior of ensembles (hundreds of thousands or millions) of particles can be misleading.

However, a word of caution should be added: The article is highly hypotheses-driven and solid experimental evidence is not always available to back-up the proposed theories. This is particularly true for the potential root causes of the second release phase and the potential role & existence of nanopores/channels (which are very difficult to detect). It would be very interesting if these aspects would be studied in more detail in the future. Also, relatively little is yet known on the conditions encountered at the site of administration in the patient (e.g., importance of esterases), potential time-dependent changes of these conditions and their impact on drug release. A better understanding of these aspects could likely help improving the establishment of reliable in vitro – in vivo correlations.

Although the focus of this article is on PLGA-based drug delivery systems, the described release mechanisms and underlying basic principles are likely also applicable to other types of biodegradable polymers (e.g., some examples with pure PLA-based systems are given).

CRediT authorship contribution statement

J. Siepmann: Writing – review & editing, Writing – original draft, Visualization, Resources, Investigation, Funding acquisition, Conceptualization. F. Siepmann: Writing – review & editing, Writing – original draft, Visualization, Resources, Investigation, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Juergen Siepmann reports financial support was provided by Interreg (2 Seas and FWVL). Juergen Siepmann reports financial support was provided by European Regional Development Fund. Florence Siepmann reports financial support was provided by Interreg (2 Seas and FWVL). Florence Siepmann reports financial support was provided by European Regional Development Fund. Given his role as Editor-in-Chief of the journal and as Guest Editor of the special issue this article is part of, Juergen Siepmann had no involvement in the peer review of this article and had no access to information regarding its peer review. Full responsibility for the editorial process for this article was delegated to another journal editor. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

This is a review article.