As explained above, the development of second-entry topical products should follow a stepwise approach. Ideally, for ethical and economic reasons, equivalence should be demonstrated with in vitro data in the first step. The main limiting factor that complicates not only the first step (in vitro data for a biowaiver), but also the second step (kinetic permeation as a surrogate of local availability) is the clinical relevance of excipients. In the oral systemically acting products, it is generally assumed that excipients do not affect the bioavailability, efficacy and safety of the medicinal product; this is not always correct [8,9]. For locally acting products, the excipients are known to be relevant, especially in topical products, where, e.g., the occlusive function of some excipients, moisturizers and emollients is known to contribute to the therapeutic effect [10,11] or enhance absorption when evaporating; this increases the drug concentration and its thermodynamic activity by disrupting the skin barrier. Consequently, second-entry products that intend to demonstrate therapeutic equivalence to the reference medicinal product with in vitro data only in step 1, or with kinetic studies in step 2, i.e., without conducting therapeutic equivalence clinical trials, need to employ the same active substance in the same form, e.g., salt, which is not required for generics in the EU, and the same excipients, including grade, in the same or very similar amounts.

There are exceptions to this qualitative (Q1) and quantitative (Q2) sameness of excipient. First, some excipients may differ qualitatively. In the EMA draft guideline [5], excipients whose primary function is not related to product performance, i.e., antioxidants, antimicrobial preservatives, and colors, and do not have any other functions or effect that influences the active substance solubility, thermodynamic activity or bioavailability, and product performance, can differ qualitatively. This draft also accepts different paraffin homologues whose function relates to the vehicle or emolliency, and who do not influence the active substance solubility, thermodynamic activity or bioavailability and product performance. However, these differences might be detected through the physicochemical properties of the semisolid, which would preclude the waiver. Second, small quantitative differences are permitted. The difference in the nominal quantitative composition of the excipients should not be greater than ±5%. For excipients whose function only relates to the vehicle properties or emolliency, differences of 10% are acceptable. For excipients whose function is not related to product performance, i.e., antioxidants, antimicrobial preservatives, and colors, 10% differences are also acceptable, but this limit is not logical since qualitative differences are acceptable. In addition, the obligation to use the same salt form is also questionable if equivalence is demonstrated with kinetic studies.

Furthermore, Q1 and Q2 sameness of excipients, as described above, is also required in case of kinetic similarity (step 2). It has been claimed that these kinetic studies are single-dose studies and differences in absorption after repeated administration might occur if the excipients are different, since the excipients that are absorbed in subsequent applications may dissolve the crystalized drug located in the epidermis and promote the absorption of that drug partly absorbed in a previous application [12,13]. However, it is questionable that these different excipients will promote the absorption of subsequent applications differently without promoting the absorption in the first application differently. Therefore, Q1 and Q2 sameness is necessary because of the direct contribution of excipients to the efficacy, e.g., occlusive or hydrating effect. Consequently, if excipients had no contribution to efficacy, e.g., because the efficacy depends on drug concentrations in the synovial liquid and the product is not indicated for a skin disease or it is known that the vehicle has no contribution to efficacy, the Q1 and Q2 sameness could be waived in step 2, where kinetic data is used to support therapeutic equivalence.

For the US FDA, quantitatively the same implies ≥95%, but ≤105% of the reference concentration or amount [14]. Any qualitative or quantitative deviations from the reference should be accompanied by an appropriate in vivo bioequivalence (BE) study(ies). However, changes are allowed for some excipients, i.e., color agents, preservatives, buffers, and antioxidants, which are exceptionally considered inactive ingredients. Applicants should identify and characterize the differences, and submit information demonstrating that the differences do not affect the efficacy and safety of the product.

The nomenclature for describing the dosage form of topical products, e.g., solutions, suspensions, lotions, gels, creams, ointments, sprays, shampoos, pastes, etc., does not correspond to their structural and physicochemical properties. Therefore, the classification based on the dosage form does not ensure products with similar properties. As indicated by the US FDA: “a product designated as a cream may be comprised of a classic oil-in-water emulsion microstructure, or it may be an aqueous dispersion of different components. An ointment may be comprised of different types of components with different types of Q3 attributes; as examples, an ointment may have an oleaginous hydrocarbon base as a single phase with particles of suspended active ingredient(s), or it may be a water-in-oil emulsion, or it may be comprised of a polyethylene glycol base. In addition, although lotions are typically considered to be more fluid than creams, this may not always be true, and some creams may contain a substantially greater percent composition of water and volatiles than some lotions. Also, although creams and lotions are typically considered to be emulsions, structural features like globules or droplets may not always be evident, and conversely, some gels may be emulsion dosage forms” [6].

The US FDA has detailed the following physicochemical properties to characterize topical products [6]:

(a)

Appearance and texture, including the look, feel and smell of the dispensed product;

The EMA draft guideline [5] has indicated only appearance but the scope is the same.

(b)

Phase states, including high-resolution micrographs to show the absence or presence of undissolved particles to identify single-phase and multiple-phase products;

(c)

Structural organization of the matter, including particle size distribution and crystal habit, and/or emulsion globule size distribution and identification of the type of emulsion, e.g., oil in water or water in oil;

(d)

Polymorphic form of the active substance within the product for products with a suspended active substance.

The EMA draft guideline [5] indicates the particle size distribution and polymorphism for suspensions and the globule size distribution for emulsions, but crystal habit is not mentioned; however, it is mentioned as a factor that may affect bioavailability and should be included in stability studies;

(e)

Rheological behavior of liquid and semisolid dosage forms, including

• Complete flow curves plotting shear stress vs. shear rate, and viscosity vs. shear rate until low or high-shear plateaus are achieved. Apparent viscosity should be reported at least at low, middle, and high shear rates.
  The EMA draft guideline specifies that these shear stress flow sweep experiments should comprise multiple data points across the range of increasing and decreasing shear rates so that any linear portions of the up-curves or down-curves are clearly identified. In contrast to the US FDA, for the EMA, the resulting curves should be characterized by fitting to (modified) power law equations so that numerical data can be produced [15]; however, the minimum of the three share rates to estimate apparent viscosities is not defined, only the need to measure the apparent viscosity at specified shear rates across the rheograms, e.g., η100, and the quantification of the thixotropic relative area (S_R) if appropriate.
• Yield stress, if the products exhibit plastic flow behavior.
• Linear viscoelastic response (storage and loss moduli vs. frequency).
  The EMA draft guideline specifies that viscoelastic storage and loss moduli (G’ and G”), and loss tangent (tan δ) should be determined in these oscillatory strain sweep (shear strain oscillatory amplitude sweep) experiments [15] to obtain parameters that can be compared objectively. The EMA draft guideline also mentions the creep test.

(f)

Water activity and/or drying rate, relevant for products with volatile excipients including water;

(g)

pH and buffering for products with an aqueous component;

(h)

Oleaginous components, which should be characterized according to the tests listed in the US Pharmacopeia;

(i)

Specific gravity (density);

(j)

Metamorphosis of the product when dispensed from the containers.

The EMA draft guideline defines the pH, buffering capacity, viscosity, density, surface tension and osmolality for solutions and suspensions, and the pH density and rheological behavior for semisolids, although this separation is not so straightforward. The characterization of the oleaginous components, metamorphosis, water activity and drying rate are not included.

Both the US FDA and EMA require that batches of different ages or storage periods should be characterized.

The in vitro tests required for the characterization and comparison of the physicochemical properties of these semisolid dosage forms depend on the characteristics of the specific reference product; they should be defined case-by-case, since not all of them are applicable. In this regard, a “Quality Attributes Data Comparison Protocol” should be employed in the EU, as indicated in the reflection paper on statistical methodology for the comparative assessment of quality attributes in drug development [16], in order to pre-define those quality attributes or physicochemical properties that define the semisolid product. This task is facilitated by the product-specific guidances (PSGs) for generic drug development, elaborated by the US FDA [17], since the EMA has not developed any product-specific guideline for topical products at present [18]. The Draft Guideline on the Quality and Equivalence of Topical Products defines only the usual physicochemical properties that characterize a semisolid [5], but additional characteristics may be necessary depending on the composition of the formulation, e.g., characterization of the oleaginous components, metamorphosis, water activity and drying rate. For example, it is necessary to characterize the water activity of those formulations where the solvent activity affects the performance of the formulation. Water activity depends on excipient composition and manufacturing variables and, in addition to the viscosity and thermodynamic activity of the drug, controls the drug diffusion rate within the vehicle and affects drug output from the formulation. Vehicles with low water activity alter the hydrodynamics of skin and cause structural changes in the stratum corneum. Small molecule humectants, such as propylene glycol, retain skin hydrodynamics [19].

In the EU, the physicochemical properties to be compared for solutions are pH, buffering capacity, viscosity, density (or specific gravity), surface tension and osmolality [5]. In the USA, in some old PSGs, these parameters are not identified, e.g., [20,21,22,23,24,25,26,27,28,29], but the same can be observed in the most recently revised PSG, e.g., [30,31,32,33]. In the Ciclopirox topical solution PSG, the physicochemical properties refer to the polymeric resin (molecular weight distribution, number of butyl groups/g of resin) [34], because the resin defines the properties of the formulation and the nail coat. In the PSG for Efinaconazole topical solution [35] and Tavaborole topical solution [36], their specific properties are as follows: appearance, specific gravity, viscosity, evaporation (drying) rate and surface tension; pH is added to this list for hydrogen peroxide [37]. Therefore, it can be concluded that the physicochemical properties defined in the EMA draft guideline should be adapted to each specific product. For example, the pH and buffering capacity are not applicable for non-aqueous solutions and the drying rate may be necessary for those products containing volatile solvents.

For suspensions, the EMA draft guideline [5] identifies the same physicochemical parameters as noted above for solutions. Obviously, for drug particles in suspension, additional characterization, in terms of active substance particle size distribution and polymorphic form, including photomicrographs, is required. In addition, an in vitro release test (IVRT) should demonstrate a similar release rate and the total amount released at the end of the study, since the concept of extended pharmaceutical equivalence coined in this draft guideline includes equivalent performance. In the US FDA PSG for betamethasone dipropionate and calcipotriene topical suspension [38], the physicochemical and structural characterization details are as follows: (a) visual appearance and texture; (b) phase states and structural organization of the matter by means of (i) microscopic examination and (ii) particle size distribution, crystal habit, and polymorphic form of the drug substance(s) in the drug product; (c) rheological behavior, which includes (i) characterization of shear stress vs. shear rate and viscosity vs. shear rate to obtain numerical viscosity data at three shear rates (low, medium, and high), (ii) a complete flow curve across the range of attainable shear rates, until low or high shear plateaus are identified and (iii) yield stress values that should be reported if the material tested exhibits plastic flow behavior, but the linear viscoelastic response does not have to be reported; (d) specific gravity; and (e) equivalent rate of betamethasone dipropionate and calcipotriene release that must be shown in an IVRT according to the draft guidance in vitro Release Test Studies for Topical Drug Products Submitted in ANDAs [39]. In the PSG for Spinosad topical suspension [40], pH and water activity are added to the previous list, but the analyses of particle size distribution, crystal habit and polymorphic form are not required. The above rheological parameters are required in the EMA draft guideline for semisolids, but not for suspensions. However, as stated above, the nomenclature used to describe the dosage form of topical products, e.g., solutions, suspensions, gels, lotions, creams, ointments, sprays, shampoos, pastes, etc., does not correspond to the compositional, physicochemical, or structural attributes of the drug product. Therefore, some suspensions may need the characterization defined for semisolids in the EMA draft guideline.

In contrast, for ciclopirox topical suspension [41] and Ketoconazole shampoo (suspension) [42], a waiver is not possible in the US FDA. In this regard, the waiver of the EMA draft guideline is always applicable unless

As indicated in the last point, the difficulty of developing second-entry topical products increases when considering products applied to the mucous membrane or damaged skin, where the skin models described below (IVPT, PK BE and TS) may not be representative. In those cases, e.g., in vaginal semisolids, the same principles could be followed, but using skin models as surrogates for mucous membranes may not be possible. Therefore, if the formulation is considered “simple”, the same Q1, Q2 and Q3, including IVRT, may be enough to demonstrate equivalence. On the contrary, if the formulation is complex, the most convenient methodology to demonstrate therapeutic equivalence is to conduct a PK BE study if the drug is absorbed into the systemic circulation from the site of action, if that absorption is not saturated, or at least if its less-than-proportional increase for the dose–AUC relationship is closer to proportionality than the dose–therapeutic response curve. In the cases where the semisolid is applied both in the skin and mucous membranes, the skin model could be used for the cutaneous indications and the extrapolation to the mucous membrane indications or site of application may need to be justified.

For gels, the physicochemical and structural characterization always includes visual appearance and texture, phase state and structural organization of the matter by microscopic examination, particle size distribution and crystal habit. Considered also are the polymorphic form of the drug substance(s) if it is in suspension in the drug product, or the globule size distribution if it contains an emulsion.

The rheological behavior always includes the characterization of shear stress vs. shear rate and viscosity vs. shear rate, in order to obtain numerical viscosity data at three shear rates (low, medium, and high) and the yield stress if the material tested exhibits plastic flow behavior. However, the complete flow curve across the range of attainable shear rates is not always required and the linear viscoelastic response is only required in a few cases. Specific gravity is always required, but other physicochemical parameters, such as pH and drying rate, are recommended case by case. For some gels, therapeutic equivalence trials are still required, e.g., [64,65], or a PD blanching assay [66,67].

For lipophilic ointments, it is necessary to characterize the oleaginous components in a few cases, in addition to the previously described parameters, such as particle size distribution, crystal habit, and polymorphic form when particles are in suspension, and globule size distribution for emulsions in the ointment. The biowaiver for the combination of betamethasone dipropionate and calcipotriene is only applicable for the calcipotriene component because a PD vasoconstriction study is recommended for betamethasone dipropionate [68].

For hydrophilic ointments, the characterization is slightly simpler, e.g., mupirocin ointment containing polyethylene glycol 400 and polyethylene glycol 3350 [69].

For some old drug products, grandfathered drugs, gentamicin sulfate [70], nystatin [71], triamcinolone acetonide [72,73] and the combination of the last two [74], the recommendations for these ointments do not include the Q1 and Q2 sameness. Only the physicochemical characterization is needed, but the parameters to compare them are not defined in the PSG. The same can be seen in some creams.

All the above properties need to be addressed for creams depending on the reference product characteristics.

For example, in acyclovir cream, it is necessary to characterize the particle size distribution, crystal habit, and polymorphic form of acyclovir in the drug product, but not the globule size distribution [82]. The contrary is required for ammonium lactate [84], and both are needed for others such as the combination of acyclovir and hydrocortisone [83], calcipotriene [88] and docosanol [89]. A complete rheological characterization, pH and specific gravity are generally required for creams. On the contrary, water activity and drying rate should be characterized in a few cases. Although included in Table 5, the biowaiver for the combination of betamethasone dipropionate and calcipotriene is only applicable for the calcipotriene component because a PD vasoconstriction study is recommended for betamethasone dipropionate [85]. The waiver is not possible for a fluorouracil product containing microspheres [111].

Similar requirements can be found for lotions as for emulsions, e.g., appearance and texture, complete rheological characterization, specific gravity, and particle size distribution and polymorphism [112], globule size distribution [113,114], or both [115]. pH needs to be compared in all cases, except in the combination of miconazole nitrate, white petrolatum, and zinc oxide due to its fatty nature, for which the oleaginous components need to be characterized [116]. The drying rate only needs to be compared for ammonium lactate [112]. These recommendations do not apply to triamcinolone acetonide, which, as a grandfathered drug, only needs a comparison of undefined physicochemical properties without Q1 and Q2 sameness [117]. In addition, some lotions are recommended based only on Q1 and Q2 sameness in solutions [118,119].

To illustrate the diversity of dosage forms and tests to characterize topical products, it can be highlighted that the tests mentioned above differ from those required for topical aerosol-foam [124], which include the following:

(a)

Microscopic birefringence analysis of the dispensed foam after complete collapse, in order to determine whether any crystals of undissolved active substance form during dispensing.

(b)

Time to break (from dispensing to complete foam collapse) analysis, conducted at 30 °C, 33 °C, 35 °C, and 40 °C under controlled relative humidity conditions.

(c)

Weight per volume of uncollapsed foam.

Nystatin powders [125] are waived based on comparative physicochemical and structural (Q3) characterization, without defining the necessary in vitro tests. Ciclopirox [126] and clobetasol propionate [127] shampoos can be waived based on Q1 and Q2 sameness, like solutions. The same criterion is applied for sprays of clobetasol propionate [128] and desoximetasone [129], as well as clindamycin [130] and erythromycin [131] swabs. For glycopyrronium tosylate cloth [132], the same dimensions and content are recommended, and it is necessary to compare pH and absorbency.

Importantly, the US FDA PSG does not waive all products with the same systematic approach due to the complexity of these products. Therapeutic equivalence trials with clinical endpoints or PD endpoints are required without the option of a biowaiver for some suspensions [41], gels [133] (even where other gels of the same drug can be waived), ointments [134], lotions [120] (even where other lotions of the same drug can be waived), creams [111] (even where other creams of the same drug can be waived), and aerosol-foam [135] (even where the other topical dosage forms such as cream and ointment of the same drug can be waived). In contrast, the EMA draft guideline intends to apply the general principles to all products, and it is the responsibility of the applicant to develop the necessary methods for the waiver. The fact that a well-established PD model, such as skin blanching, or that a clinical trial design and endpoint are available should not be a reason to preclude the biowaiver. Only the lack of adequate methodology for the comparison, e.g., IVRT for all semisolid simple and complex formulations, IVPT, TS, or PK BE studies for complex formulations, should preclude the biowaiver; this is because the means by which the active substance reaches the local site of action should be known, and the approved products should be fully characterized with respect to quality attributes.