Evaluating hyaluronic acid dermal fillers: A critique of current characterization methods

Abstract Soft‐tissue augmentation has gained much popularity in recent years. Hyaluronic acid (HA) based dermal fillers; a non‐permanent injectable device, can restore volume loss, fill fine lines and wrinkles and add curves and contours. HA based dermal fillers entered the non‐surgical treatment market in the late 1990s, however there is a lack of data and literature comparing the range of products and detailing the complexities of these products and how it relates to tissue performance. Measuring the physico‐chemical properties of these dermal fillers provide key parameters to predict their performance after injection into the body. This article reviews the currently reported methods and parameters used to characterize dermal fillers. The review of these methods and data from the literature provides a useful guide to clinicians and injectors in selecting the optimal product suitable for the needs of each patient.


| INTRODUCTION
The human face ages as a product of complex microscopic and macroscopic volumetric changes. These changes are a product of the resorption of bony structures, gravity, subcutaneous fat redistribution, and skin damage. 1 Dermal fillers are deployed to augment the face to meet the aesthetic concept of beauty, dictating that certain curves, contours, dimensions, and ratios are fulfilled in order to produce a conventionally attractive face, or to restore volumetric dimensions and hence youth in the aging face. 2 Facial rejuvenation using soft tissue biodegradable fillers-a non-permanent injectable device-provide an affordable and relatively safe procedure compared to permanent surgical cosmetic procedures.
The past decade has witnessed a dramatic rise in social media and online influencers. This, in turn, has contributed to the rise in popularity of aesthetic medicine. A recent survey suggests that the nonsurgical treatment market in the UK could be worth in excess of £3 billion within the next 5 years. 3 Hyaluronic acid (HA) fillers are regarded as a class III medical device in the UK and not a medicine, due to the lack of an active pharmaceutical ingredient, where there is a pharmacological effect.
Unlike medicines, medical devices such as dermal fillers have no legal requirement to provide safety and efficacy data, or how they perform in comparison to other market brands. 4 In Europe manufacturers are expected to complete a certificate of conformity, in line with the Medicine Device Regulations general safety and performance requirements. This is mainly linked to manufacturing and risk assessment. As the UK is no longer part of Europe, the certification of devices is changing, and CE will become UKCA. 5 The FDA require more detailed information linking to safety and effectiveness of the product, therefore there is more rigor applied to approving these products for use. 6 As HA fillers are not medicines, claims can be made with little to no evidence to substantiate them. The lack of regulation in the HA filler market, and the financial value of the industry, means there is a lack of objective guidance to clinicians when choosing HA fillers.
Understanding the physico-chemical properties of dermal fillers is important to make informed product selection. HA fillers differ from each other due to the different crosslinking technologies used, which aim at tuning the mechanical properties to the target tissue and the biological outcome after injection. During the injection process, gels are subjected to shear stress and vertical compression/elongation forces, which cause the filler to deform. Dermal fillers under low stress are gel-like materials, but flow under increasing shear stress, to different extents depending on their specific manufacturing conditions and composition. 7,8 Key questions are: can these products correctly mimic the soft tissue or bones they are replacing? Which measurable physicochemical parameters can be used to predict the long-term performance of fillers? How do stress, deformation, temperature and enzymatic degradation affect their properties over time?
While addressing these questions should be a long-term endeavor, the objective of this review is to present the state of research on cross-linked HA dermal fillers, focusing on current methods and parameters reported in the literature, to evaluate whether these measurements can be used as a valid measure of how HA fillers may perform in vivo.

| REVIEW METHODOLOGY
A literature search was conducted using Embase and Google Scholar to identify suitable research papers. Articles from the last 20 years were included. References from each article were reviewed to identify any papers of further interest. The aim was to identify papers reporting the physico-chemical and rheological properties of HA dermal fillers, specifically FDA-approved HA fillers and Revolax (a filler from the Korean market that is widely used globally as a less costly product). No information regarding Revolax was found in scientific publications, therefore the manufacturers device patent was accessed to provide product information. It was necessary to search for patent information for Teoxane RHA, Restylane OBT, and Juvaderm Vycross technologies to establish the degree of cross linking. This measurement often differs in research literature therefore manufacturers patent application was used to find this value.
Papers were analyzed and critiqued, to review the methodologies used to investigate these HA properties in seminal literature. The search used the following key words: hyaluronic acid, dermal fillers, soft tissue augmentation, rheology, hyaluronidase, cross-linked hyaluronic acid.

| Crosslinking and degree of crosslinking
Upon dissolution in water, HA forms highly viscous solutions. These solutions are not suitable for use in dermal fillers as they will not stay in place at the injection site and also be rapidly degraded by the enzyme hyaluronidase, 11 resulting in a limited residence time.
In order to achieve the required stiffness and persistence, a number of modifications and processing steps must be carried out. Specifically, crosslinkers are used to connect HA polymer chains together to create a network (Figure 2), and transforming the viscous liquid into a gel. 12 The crosslinked dermal fillers available in the UK today are mostly crosslinked with 1,4-butanediol diglycidal ether (BDDE).
The crosslinking technology differs across brands and has an impact on the properties of the resulting gel. During the crosslinking process, the crosslinker becomes permanently bound to the HA chains via strong covalent bonds. The crosslinker can either bind at both ends, creating a diol that is stable and does not react further, or bind only at one end, resulting in a "pendant" end 13,14 (Figure 2), the ratio of which will affect the stiffness of the final gel.
The degree of crosslinking (CrD) 14 can be quantified by the ratio of the crosslinker molecules that form crosslinks to the number of HA disaccharides. In Figure 2 Additionally, exceeding this threshold could affect the biocompatibility of the product, which could lead to an immune response or an adverse reaction to the gel. There is, however, no established higher threshold value for CrD.
In the UK, medical devices such as dermal fillers are only required to have a CE marking, 15 an indicator of the product's compliance with EU legislation. Quality control of the product is not a requirement, this means that the presence of unreacted or residual crosslinker molecules can remain in the final product, which can be toxic in high concentrations, 16  Based on product information provided by manufacturers.

| Crosslinking technology
The proprietary crosslinking technology of the HA gels influence the gel properties in many ways. Dermal fillers discussed in this review have been manufactured using three different methods: Resilient Hyaluronic Acid (RHA ® ), 18 Vycross ® , 19 and Optimal Balance Technology (OBT ® ) 20 (Table 1).
Revolax's crosslinking technology involves the encapsulation of ursolic acid, 21 a natural wax found in fruit peels, within the HA polymer network. This technology claims to maintain a long-term durability without increasing the crosslinking density. RHA ® is the technology used by Teoxane, whereby the gels are stabilized by natural and chemical crosslinks to produce gels with long chains of HA. Their range uses 1.9%-4.0% BDDE, 18 which is relatively low compared to other brands, and they also have differing concentrations of HA. Vycross ® , used in the latest Juvéderm range, is formulated with a mixture of high molecular weight HA and low molecular weight HA, with a higher ratio of the latter, linked with BDDE at both ends. 19 Finally, OBT ® is the technology used by Restylane. Products in this range have the same HA concentration but achieve a range of gel firmness by varying the degree of crosslinking. 20

| Manufacturing process: fragmentation
The manufacturing process used to produce HA dermal fillers involves breaking down the initial crosslinked gel into smaller HA gel fragments or particles. 16 This process allows the gel to flow through a needle for injection under the skin. After fragmentation, the gels may still be too stiff and hence resistant to deformation and potentially difficult to inject. In order to overcome this, some manufacturers introduce uncrosslinked HA as a lubricant to reduce the strength of the gel during injection 16 (and therefore the force required for injection In practice, a small amount of the dermal filler is placed between two plates, and one of them rotated by a small angle, creating a shear stress (force per unit area) and inducing deformation (strain). G* is defined as the ratio of the shear stress to the shear strain (the ratio of the displacement to the height of the sample, or gap). For instance, a higher extent of crosslinking will require a greater stress to achieve the same displacement (or will be deformed less for the same stress applied), resulting in a higher modulus value, characteristic of a "stiffer" gel.
Numerous studies report dynamic frequency-sweep tests conducted in oscillatory mode, using parallel-plates 13,14,23,24 or cone-andplate geometries. 13,25,26 In this type of experiment, the frequency is typically varied over a few decades of frequency (typically 0.01-100 rad/s, often a narrower range). Most studies have reported measurements at temperatures of 25 C, 13,23,25,26 however, physiological temperature is more relevant to the clinical application and this may need to be taken into consideration for future work. 24,27 The ratio of G 00 and G 0 , or tan δ, reflects the relative magnitude of the viscous and elastic modulus. A predominantly elastic gel (low tan δ) deforms under the action of stress and recovers its shape after removal of the force (Figure 3), while a gel where viscosity dominates (usually at low frequencies, i.e., long times) deforms but also flow ( Figure 3). As a result, the phase angle δ is often linked to the capacity or otherwise of a product to migrate, and Revolax claimed that the low phase angle of their product relates to limited product migration from the site of injection. 28 Only in an "ideal" gel (permanent cross-links, monolithic gel) are  Table 2 shows a partial summary of this data, with G 0 values extracted at After injection, HA fillers are subjected not only to dynamic shear stresses due to facial movements, but also compressional stresses.
However, rheological measurements (which apply shear stress) dominate the dermal fillers literature, apart from a few exceptions. 25 To understand how a product will persist in the tissue when under muscular forces (such as when speaking, smiling or eating), both the dynamic shear stress and compression stress of the gel should be measured. 7 Both G 0 and E 0 (the Young or compression modulus) provide complementary information and must be within a suitable range to withstand these facial forces. 25

| Swelling
When in solution, HA chains expand due to their affinity with water, 22,16 ; when the polymer chains are crosslinked, this results in swelling of the gels. Swelling is an essential parameter for dermal fillers as it is directly related to how the filler will expand at the site of injection. A gel's capacity to swell is dependent on factors such as the concentration of polymer, the degree of crosslinking and the process by which the gel was hydrated. A strong correlation between swelling factor and cohesion has also been reported: the further away the product is from equilibrium swelling (the point beyond which the product phase separates between a polymer-rich and a water-rich phase), the more cohesive the product is. 32 Limited swelling is expected in tighter gel networks (higher extent of crosslinking, hence higher G 0 ), 13,23,32 leading to a lower propensity for fluid uptake. However, while this relationship may be true within a given filler series, it may break-down across series due to different crosslinking technologies. 33 F I G U R E 3 Schematic representation of how the elastic (G 0 ) and viscous (G 00 ) moduli reflect the capacity of the gel to deform and flow and impact its shape when exposed to shearing forces (adapted from 25  In a typical swelling test, an aqueous solution is added to a precisely weighted quantity of dermal filler and left for a given period of time. The resulting mixture is then centrifuged, the supernatant removed, and the resulting hydrated gel weighted. The swelling ratio can be calculated using the following equation: where, W s is the mass of the swollen gel and W d is the mass of the dry gel. Kablik et al. 13  A strong correlation between "dilution durability" and gel-to-fluid ratio was observed: the more fluid (i.e., soluble HA) in the product, the higher extent to which the product can swell before phase separation occurs.
The swelling factor (SwF) is defined by V/V 0 where V 0 is the initial volume of the gel and V is the fully swollen volume. 14 Table 3 shows data obtained on the Restylane's OBT filler range. The reveal some correlation between SwF and G 0 : the lower the G 0 , the higher the SwF. However, it was not possible to establish whether the degree of crosslinking (CrD) correlates with swelling; as explained above, this correlation often breaks down across different brands due to different technologies used.

| Enzymatic degradation
HA is a substrate for the enzyme hyaluronidase: the enzyme cleaves HA molecular strands into smaller oligosaccharides, making them susceptible to metabolism and clearance from the body. 30 Increasing the degree of crosslinking is expected to reduce the propensity to enzymatic degradation and achieve longer residence. 34 In addition to the degree of crosslinking, HA concentration also influences the rate of degradation. Many authors have reported this trend through qualitative in vitro studies, 35 quantitative in vitro studies 36,37 animal models, 38 and clinical testing of human subjects. 39 To measure enzymatic degradation, typically, a solution of hyaluronidase is mixed with the gel. The mixture is centrifuged, the fluid phase filtered out and the remaining gel weighted. This process can be performed over several time points, until the gel is completely degraded. 30 This measure is important to understand how a given filler will respond to the injection of hyaluronidase, routinely used in case of occlusions.

| Cohesion tests
Cohesion is a parameter often mentioned in the literature of dermal fillers, which, while intuitively quite accessible, remains poorly characterized. According to Micheels et al., 33  A qualitative assessment of cohesivity has also been reported by La Gatta et al., 29 using a protocol originally proposed by Sundaram et al., 40 where the gel is stained by toluidine blue. A small amount (ca. 1 g) is then extruded from a syringe into a 1 L beaker of water, with continuous stirring. The droplets are filmed, and images evaluated at specific times, and a value of cohesivity assigned. Both the viscosity of the gel and its cohesivity (Figure 4) are likely to determine how the gel distributes within the tissue: in order to obtain good tissue integration, both of these parameters need to be carefully considered. If the tissue is firm (such as cheek bones or the jaw line), 41 a high viscosity gel that is highly cohesive can be molded by massaging the area after injection, allowing more precise placement without fragmentation. Instead, a HA gel with low viscosity may be more suitable for superficial indications (such as fine lines and wrinkles) as the product will flow and spread more, giving a profile more akin to the area that is being treated. 25 4.5 | Particle size and particles per ml HA dermal fillers are not "bulk" gels but are constituted by crosslinked HA particles, resulting from fragmentation. 26 The process of sizing down the gel mass is performed by passing the gel through a series of sieves and screens. 16 Depending on the sieving method, the various products will have a distinct average gel particle size and shape, which will impact the final product's performance.
For dermal fillers, there is a maximum particle size beyond which the gel cannot be extruded easily and may clog the needle during injection. 13 On the other hand, larger HA particles have a limited total surface area for enzymatic breakdown, while it is smaller for HA particles, which therefore degrade faster.
A good trade-off can be achieved by breaking down the mass by a homogenization process. This process results in a broader distribution of gel particle sizes than obtained by sieving, and "softer" gels with lower G 0 values. 36 Particle size may be linked to the residency of the HA filler in situ: it is generally thought that the larger the particle size, the longer they reside in the tissue. 42 However, the link between particle size and important parameters such as the elastic and viscous moduli, or longterm performance, is not obvious. 26

| Extrusion force
Extrusion force is another important parameter of high clinical relevance as it relates to the force that the physician must apply to allow the HA filler to flow through the needle. 41 There are no quantitative data reported on the extrusion force profile. While this aspect is linked to practice, it would be useful to produce quantitative data to better understand the optimum force profile required to extrude a given filler from the syringe in an even and smooth manner.

| LINKING EXPERIMENTAL PARAMETERS TO PERFORMANCE
The clinical performance of a HA based crosslinked dermal filler is dictated by their physico-chemical properties. Depending on the specific clinical indication, many parameters need to be considered to optimize the formulation or to select the most appropriate product.
Within the range of parameters reviewed in this article, the elastic modulus, G 0 , is the most widely reported and perhaps most relevant parameter; strong correlations have been established between G 0 and other parameters such as swelling, the degree of crosslinking and cohesion. 23,25,26 F I G U R E 4 Schematic representation of the relationship between cohesivity and viscosity of a HA gel to show its capacity to remain or spread at the injection site (adapted from 25 ). HA, hyaluronic acid F I G U R E 5 Extrusion curve of HA dermal filler with yield point, where extrusion force, F is plotted as a function of displacement, D. HA, hyaluronic acid "Stiffer" gels possess higher G 0 values; these gels swell less, tend to be classified as more "cohesive," and, as a result, are also more resistant to enzymatic degradation. These types of gels, such as Revolax Sub-Q (G 0 = 281 Pa) (unpublished data), Teoxane RHA4 (G 0 = 296 Pa), 23 Juvéderm Volux (G 0 = 307 Pa), 23 and Restylane Defyne (G 0 = 260 Pa) 23 (all values taken at 0.1 Hz), are more suitable for areas where bony structures need to be replicated; these stiffer gels are able to resist the high shear forces found under the muscles). 41 However, high cohesion is not always a desirable property, as facial movements could result in a "bunching up" effect where the gels aggregate into clumps. The high cohesion could prevent the gel from spreading, thus producing an uneven contour.
On the other hand, "softer" gels, with lower G 0 values, such as Revolax Fine (G 0 = 93 Pa) (unpublished data), Teoxane Redensity II (G 0 = 37 Pa) (unpublished data), Juvéderm volbella (G 0 = 159 Pa), 23 and Restylane Fynesse (G 0 = 10 Pa) 23 (all at 0.1 Hz). These gels tend to be formulated with lower HA concentrations and/or lower degrees of crosslinking. 30 Due to the lower extent of crosslinking, possibly higher amount of free HA, resulting in a looser gel network, they are more susceptible to enzymatic degradation, and will be more readily eliminated by the body. These fillers are more suited to finer corrections and can provide a more natural feel when injected, such as for less dynamic wrinkles (tear troughs, soft tissue found in lips and the periorbital region). However, these gels tend to have a shorter residence time in the tissues. 34 The drop weight test reported in various studies considers that the lighter the average drop weight, the less cohesive the gel is. 23,32 This characteristic also correlates with gels with lower G 0 values. Less cohesive gels have looser polymer networks, thus once extruded from the needle will drop quicker, therefore producing smaller and lighter drops. The less cohesive nature of these gels is more likely to produce smoother contours once injected. However, this could also pose a problem of gel migration from the injection site.
A property that also correlates with a low G 0 value is the swelling factor: increased gel fluid uptake is usually observed for gels with lower G 0 values (unpublished data). This could pose an issue for clinical applications because if gels with low G 0 are recommended for superficial applications, such as in the tear trough, the swelling of the gel could cause a convexity, which is an extremely undesirable outcome.
Clinicians must take into account a gel's propensity to swell and warn patients that swelling will subside or inject a smaller amount of filler, because swelling will contribute to the volume injected.
Beyond dynamic rheological measurements, the lack of standardized measurements is a challenge when attempting to compare physicochemical properties from different fillers. Differences in crosslinking technologies, HA concentration, particle size and degree of crosslinking across brands make the establishment of universal correlations challenging. More importantly, there is yet no reported physico-chemical parameters that take into consideration clinical and anatomical effects on the overall long-term performance of the gel once injected into the body.
When choosing a suitable product, it is important to consider the site of injection, including which anatomical facial layer. Different areas and layers of the face are subject to different magnitudes, frequencies and type of forces, which need to be considered to achieve the optimal outcome. There are two types of forces that govern the mechanical stresses within the face: intrinsic forces (dynamic tensions that occur between tissues within the face, such as bone, muscles, fat and skin) and extrinsic forces (environmental forces that result from daily activities: sleep, nutrition, exercise, etc.). 41 Steady-state shear rheology (measuring viscosity as a function of shear force) can provide a measure of how gels will react to these forces, however more work needs to go into modeling the shear, compression and torsion forces each facial zone may experience and measure the properties of fillers under conditions that better mimic them.
As mentioned above, the phase angle obtained from dynamic rheology is another parameter often quoted to advertise fillers which claim to have a "low percentage to migrate," however no conclusive evidence has been found across the literature that supports this claim.
Better methods must be put forward to measure the migration of a product.

| CONCLUSION
A growing body of literature reports experimental data on HA dermal fillers, however, correlation within clinical application and long-term performance is lagging behind.
Overall, the methods described in this review and commonly reported on fillers do not predict the long-term performance of the gels. However, a large set of physico-chemical parameters can act as an indicator of performance, which, alongside clinical experience, is extremely helpful in guiding clinicians in choosing the optimal product for a specific application.
Future work should focus on establishing standardized experimental protocols and link them to clinical data.

CONFLICT OF INTEREST
We declare no conflict of interest and no funding sources.

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study.