Chapter 1: Ophthalmic Formulation

Poor sight constitutes a major and expensive burden to society, with risk of accidental self harm, the need for specialised provision and often the extensive involvement of family members in care. Sight loss arises from trauma, disease, infection and old age, with thinning of the retinal nerve fibre layers at a rate of 3 µm per decade or 0.27% per year over an age range of 18 to 80 years.¹  Some diseases of old age such as cataracts are now routinely treated but as we live longer, the principal eye diseases in the Western world impact the elderly population. These ocular conditions include age-related macular degeneration, diabetic retinopathy and glaucoma.

In ophthalmic medicine, the scope for intervention is uniquely difficult as the structures defend the inner target – the retina, by the blood–retina barriers preventing penetration of drugs circulating in the systemic circulation and the outer tissues preserve ocular acuity by attempting to dilute and wipe away any medicine applied topically by blinking. The eyes are vulnerable anatomically, and recently it has been hypothesised that COVID might be spread by droplet to eye surface contact. This suggests that the wearing of glasses might be a simple protective measure against aerosol-borne infection.

The internal drug delivery targets are shown in Figure 1.1. Following administration of a topical formulation, the drop mixes with the tear film and is drained through the nasolacrimal apparatus (Figure 1.2). The drainage from the eye moves drug through the facial viscera with venous return ultimately to the heart. Modern delivery systems for ocular drug treatment must therefore serve the following tasks: sustained efficacy, residence without blurring, increased transport without hyperaemia and minimisation of side effects especially cardiovascular insult. For the eye, packaging systems are also important as those for daily use must remain clean, stable and appropriate for use.

The usual technique of dosing is to pull down the lid to make a pocket to accommodate the drop and therefore most formulations encounter the mucous layer of the conjunctivae rather than the cornea. Patients should be trained to achieve safe and effective dosing but often will apply multiple drops. The excess is lost through the nasolacrimal duct or spillage on the cheek. On relaxing the pinch, the lid closes, expelling most of the dose. The reflex tearing mechanism, activated by the change in surface temperature and a few seconds later by the drop composition, further dilutes the formulation.

The globe protrudes from the skull and the eye is very vulnerable. Additionally, the wet surface is prone to desiccation and accumulates particulate matter which must be cleared to maintain visual acuity. This is because the lens and wet cornea form a compound lens structure, with the cornea providing about 43 dioptres and the lens 16 dioptres. The dioptre is the unit of refractive power equal to the reciprocal of the focal length of a given lens. The addition of a drop of liquid to the cornea therefore causes a large change in refraction and triggers blinking and a reflex upward movement of the eye. The pull back on the blink generates negative pressure in the lacrimal sac, assisting the flow of tears from the gland. The blink compresses the tear film and the thickness of the film increases, and the eye no longer focuses an image.

Tears are secreted through the lacrimal gland and supplemented by lipids provided by the Meibomian glands. The film is adherent and very thin over the cornea (thickness 2–5.5 µm), as illustrated in Figure 1.3. Note that the lower marginal strip acts as a reservoir for the applied drop. Mucin is secreted by the lacrimal gland and to a lesser extent the goblet cells of the conjunctivae. The mucin gel is composed of heavily O-glycosylated protein domains making up 50–80% of the weight of the mucin. This carbohydrate/protein structure increases adherence and without adequate mucin the tear film is unstable. The tear film is then unable to wet the surface, promoting the formation of dry spots which can become a difficult problem to treat. The mucins, being thixotropic, allow stability at rest and shear thinning during blinking. Overall, this provides protective and lubricative functions but tears are variable in composition. Basal flow provides the resting coat but reflex tears are produced by contact stimulation or blowing the nose. Emotional tears are also produce by central arc stimulation and have higher protein content than reflex tears²  as do closed-eye tears collected after sleeping. This is due to the serum content produced by leakage of the conjunctival blood vessels. The rate of tear secretion, estimated at between 0.5 and 2.2 µL min⁻¹, is affected by environmental conditions, age, disease and psychological state.³  Osmolarity of the tear film is estimated as 290–310 mOsm kg⁻¹ in healthy adults, rising to 320–325 mOsm kg⁻¹ in dry eye disease.

The pH of freshly secreted normal tears is measured at 7.2–7.4 but old studies of contact lens wearers who wore rigid lenses suggest that the tear film can be more acidic (pH 6.6) due to the build up of carbon dioxide under the lens.⁴  Similarly, pH is lowest on awakening as a result of acid by-products associated with relatively anaerobic conditions in prolonged lid closure and pH increases because of loss of carbon dioxide as the eyes open. The same authors found disease in non-contact wearers can make the tear film more alkaline. The chemical composition of tears was reviewed by Tiffany, who with Prof. Bron, made great advances in understanding the nature of the tear film.⁵  Tiffany measured tear viscosity and described the film as weakly elastic due to the lipid layer. He measured values of 4.4–8.3 mPa seconds for normal tears and 28–31 mPa seconds in patients with dry eye. An excellent description of the solids which constitute 1.8% of the tear film was given in a recent e-book.⁶ 

Meibomian glands are situated at the base of the lashes and secrete a lipid mix of phospholipids, neutral oils, sterol esters and other waxy lipids with a melting point range of 19–32 °C.^7,8  Branched and unsaturated lipids are responsible for a lowering of the melting point.⁹  The tear film lipid layer can be observed by interferometry and is approximately 40 nm thick and extends upwards over the film to the gland like a curtain, travelling upwards during the upswing of a blink. The instillation of a solution containing drugs or adjuvants that lower the surface tension may disrupt the outermost lipid of the tear film into oily droplets, which become solubilized. The protective effect of the oily film against evaporation of the tear film aqueous layer will then disappear. Approximately 3 µL of a solution can be incorporated in the precorneal film without causing it to destabilise. Alteration in the resident tear film can be noted by a change in the tear break-up time when the tears are stained with fluorescein. Alteration in the composition of tears, usually a change in mucin or lipid composition, results in decreased wettability and faster tear evaporation, leading to painful dry spots on the cornea or “dry eye disease”.¹⁰  Wetting agents are often added to dry eye treatment drops to improve contact on a surface that with drying becomes more hydrophobic.

The eye is not a portal for systemic delivery although topical delivery to the eye can result in profound systemic effects; for example, the administration of a beta blocker to exercising students decreased heart rate¹¹  and the effects in other adult groups was markedly affected by formulation.¹²  Generally the purpose of ocular delivery is to treat that organ, albeit that dosing can be ineffective. Missing the target and missing the opportunity to treat are common problems in ophthalmic medical practice.¹³  The issue of comfort and lack of fuss in the dosing becomes important in silently advancing diseases such as glaucoma because, as far as the patient is concerned, there is no signal of immediate benefit.

Diseases affecting the front of the eye can often be treated topically which, as ophthalmic solutions are easy to prepare is a good low cost option. The optimum volume for a drop delivered to the eye is estimated to be between 5 and 20 µL;^14,15  however, the typical volumes delivered by commercial eyedroppers are in the range of 25 to 56 µL¹⁶  and smaller volumes delivered by sprays are equally effective.¹⁷  The general presentation is a sterile preparation in an eye-drop bottle. Multi-dose plastic eye-drop containers have replaced the traditional glass bottle and bulb pipette. The wall has to be sufficiently flexible to allow compression, yet rigid enough to prevent flooding of the ocular surface. In addition, wall thickness affects potential loss of water on long-term storage. The handleability of the bottle is also important and small very rigid single drop dispensers such as those used in clinics may be difficult for the elderly person to use. Squeezing the sides of the bottle allows a dose to form at the dropper tip with an approximate volume of 30–40 µL depending on the angle of the tip and the surface tension of the formulation. Preservative agents such as benzalkonium chloride alter the size of the drop and viscous solutions may take a long time for the drop to form on the tip so pseudo-plasticity of the formulation may be an asset.

Plastic containers do have some significant problems including adsorption of antimicrobial agents such as benzalkonium chloride from the solution into the polyethylene container wall and allowing leaching of inks through labels and plasticisers from the wall into the bottle contents.¹⁸  Polyethylene will not withstand autoclaving and alternatives such as ethylene oxide or irradiation must be employed. Polypropylene will withstand autoclaving but the plastic is much stiffer, which may cause problems for the patient with reduced grip strength.

Contamination of the tip is a serious problem. The patient must form the dose near to the eye, increasing the chance of lash-drop contamination and as thumb-finger pinch on the bottle is relaxed, the contaminated remainder of the drop returns to the bulk solution. The sequence is illustrated in Figure 1.4.

Compliance aids may reduce the incidence of the tip touching the lashes. The pinch force required to dispense a single drop varies greatly across bottle types, ranging between 0.5 to 5.34 kgf across 17 different single dose bottles with the bottle in the vertical position.¹⁹  These levels approach mean pinch strength and for the elderly patient, exerting and sustaining such forces can result in bad tremor. Missing the eye completely is therefore a common occurrence across populations. Self-administration of ocular formulations often encounters difficulties. Connor and Severn found that at least 50% of patients admit to difficulty in instilling their own eyedrops, coordinating aim and squeeze.²⁰  A study in Northern India conducted in patients with glaucoma reported that 5/165 patients missed the eye completely, whereas 49/165 touched the eye whilst dispensing the drop.²¹  In the study reported by Tatham and colleagues, 12% of patients missed the eye and 42% touched the surface of the eye or the eyelids and lashes.²²  Tatham comments that his patients had not been instructed on how to use drops properly and a similar issue was reported in the Dehli study. Education regarding this very basic practice is therefore needed for all.

The physical difficulty of dosing is compounded by fear: the natural protective response of a patient is to turn the face away from objects approaching the eye, being instinctively cautious that they risk damaging their sight with the sharp point of the tip. Poor eyesight, limited grip strength and inability to sustain muscle tension compound the problem. The net result is erratic delivery and potential failure of treatment. Approaches to tackling the problem are ingenious but sometimes extremely simple.²³  A study in New Zealand examined the effect of sticking black tape to the tip of the bottle so it could be easily discerned.²⁴  Devices to improve compliance have variable rates of uptake by patients. In a study of relatively inexperienced patients, Gomes and colleagues used the Xal-Ease^® (SHL Medical) dispenser.²⁵  The device fits over the existing eye dropper bottle and helps with the removal of the bottle cap and the positioning of the drug over the eye; it then dispenses only a single drop of the formulation into the eye. In their small study, 61% preferred to use the device whilst 39% decided to stay with non-assisted drop delivery. A colleague in Germany noticed some of his patients using eyelash curling tongs to squeeze the bottle. This seems to be an ad hoc version of the Autosqueeze™ Eye Drop Dispenser (Owen Mumford), a tongs-like device to assist with bottle squeezing. Higher levels of sophistication are used in the Mystic blister-based platforms, now existing for the delivery of powders (VRx2) or liquids (Versidoser), delivering a precise volume into the eye (ranging from 10 to 50 μL) in virtually any hand/head orientation. Finally, bottles with flexible areas, or a pump action, have been used to facilitate the action of dispensing a drop accurately.

Periocular disease and treatment of the anterior segment in diseases such as glaucoma are treated with classical topical sterile formulation variants: solutions, suspensions and ointments. Suspensions are useful when a material is unstable in dilute solution or when a sustained effect is required. Topical administration is performed by squeezing a tube or dropper bottle to achieve a variable-sized dose on the surface of the eye. Blurring of vision, sensation of cold, irritation and sometimes discomfort are common unwanted effects and the formulation can be sensed in the nose and mouth due to nasolacrimal drainage. Once opened, the bottle and contents must be discarded after a suitable time limit (typically 28 days) and kept clean to avoid bacterial contamination. Similarly, ophthalmic products must be manufactured in a sterile, pyrogen-free environment and packaged to ensure good shelf life.

The objective of topical dosing is therefore to deliver sufficient active pharmaceutical ingredients to be absorbed at the eye surface but not too much to cause overspill. Tonicity or solubility considerations usually limit the concentration of the active ingredient to around 2% w/v, equating to a 500–600 µg dose in a single drop. More often, solution strengths of 0.5% to 1% are routinely employed but for prostaglandins such as travoprost, concentrations of 40 µg mL are encountered. Nevertheless, even at this low concentration there is a need for a non-ionic oil-in-water solubiliser such as macrogolglycerol hydroxystearate 40 (Kolliphor RH40). All the appropriate components have to be incorporated into the drop, to produce a robust dosage form. These are summarised in Figure 1.5 below and the inclusion of each varies according to single and multi-dose presentation.

The physical state of the API might generate additional variables including presentation as gels, suspensions, ointment-based formulations, wafers and devices. Usually an appropriate salt is chosen to ensure good solubility or a solubilising excipient is used. Other chemical strategies such as use of prodrugs are also encountered.

Lacrimal fluid is a glandular secretion isotonic with blood, from which it is formed as an ultra-filtrate. It contains a mixture of electrolytes, weak organic acids and protein and is able to neutralise unbuffered solutions by a combination of contributions of protein and the body's bicarbonate-carbon dioxide system. Addition of tonicity agents allows formulations to be formulated to have tonicities equivalent to the range 0.7 to 1.5% w/w sodium chloride.

Solubility enhancement is a common interest in formulation and the formulator must work with a delicate and unforgiving tissue. Even 10 minutes exposure to a 100 mM citrate buffer can cause a decrease in conjunctival cell viability.²⁶  However, most solutions are quickly cleared from the eye as shown in Figure 1.6 with a half-life of around 20 s for a simple aqueous solution as measured by lacrimal scintigraphy.^27,28  The later phases of retention of label, as measured in the central corneal region of interest (ROI), reflect the manner in which polymers such as hyaluronic acid can assist in stabilising the tear film. The conjunctival membrane acts as a secondary reservoir for the drug and sustains the drug effect for a few minutes. Where multiple, individual drops are given, it is therefore important to wait 5 minutes between administering the two doses. This also provides a rationale for dose combinations to reduce nursing time and allow convenience for the patient.

Most ophthalmic drugs are weak bases or weak acids because to cross the epithelium and the stroma requires a mix of ionisable and lipophilic character (to partition into the epithelium). For acidic drugs such as NSAIDS, the sodium and tromethamine salt have been used.²⁹  Indomethacin is unstable in alkaline solution and poorly soluble below pH 6 and therefore may be preferentially presented as a suspension in polyvinyl alcohol (PVA) or HPMC or solubilised in an oil-in-water suspension. Flurbiprofen base is practically insoluble in water and presented as the sodium salt at a pH 6–7 0.03% concentration, and solutions at a concentration of greater than 0.2% are reported to be quite irritant. Generally, drug solutions contain a high concentration to try and drive transcellular absorption through the cornea. Permeation of flurbiprofen is reported to increase in the presence of the preservative benzalkonium chloride because of the formation of ion pairs and the solution becomes opalescent.³⁰ 

The chemical stability of alkaloids is usually increased in acid solution and the use of a weak buffer allows the lacrimal fluid to pull up the pH so that more drug is in an unionised form and available for absorption.

Although the tear film has been described in several reviews as having good buffering characteristics, that is somewhat misleading as although tears can deal with unbuffered solutions, buffering agents may be added to increase drug stability and to promote partition into tissues. Tear film pH recovery after instillation of a drop is largely due to tear turnover³¹  and buffer strength directly impacts patient sensation of eye comfort. For this reason it is desirable that the buffers used in ophthalmic formulations are weak. Commonly encountered anions include maleate, citrates, acetates, borates and phosphates. Phosphates at high concentration could cause problems as the solubility limit of calcium phosphate is low and corneal calcification can occur.³²  The EMA has recommended that this is a risk in a compromised cornea that should be mentioned in the product information.³³ 

Bitartrates are strong buffers. In the past, before the introduction of selective adrenoreceptor agonists, adrenaline bitartrate was formulated at pH 4 and the strong acidity caused issues. Presentation as the hydrochloride at a higher pH or borate reduced the stinging. Boric acid is a Lewis acid with a pK_a of 8.9–9.2, reacting reversibly with alcohols such as polyvinyl alcohol and is used in ophthalmic formulations to increase antimicrobial effectiveness.³⁴  These materials are widely used in contact lens preservatives, reducing the contamination potential of Aspergillus niger.³⁵ 

Another consideration is autoclaving, stability and the effect of pH. The degradation of solutions at elevated temperatures is significant for labile compounds. Atropine for example is stable at 25 °C with a half-life of 2 years at pH 6.8 and 130 years at pH 5. When heated to 121 °C, the degradation is much faster: 1 hour at pH 6.8 and 60 hours at pH 5.

In use, a multi-dose preparation will quickly lose sterility and represent a risk to the patient. Facial organisms, especially Staphylococcus aureus, will colonise non-preserved formulations with ease.³⁶ 

Preservatives are added to increase stability and attempt to prevent bacterial growth during the in-use cycle for a multi-dose container. Patients with severe ocular disease may instil drops frequently; Noeker reported that patients with keratoconjunctivitis sicca may be instilling drops as frequently as every 20 minutes.³⁷  The most commonly used modern preservatives are perborates, polyquads and oxy-hypochlorites (Purite). Borate has mild antimicrobial activity and 1.2% w/v borate at pH 7 has inhibitory activity against the growth of Staphylococci and Pseudomonas sp.³⁸  Benzalkonium chloride (BKC, BAK) is probably the most commonly used of the polyquad preservatives and is an effective bactericide and fungicide. It is a mixture of alkylbenzyl dimethylammonium chlorides of different alkyl chain lengths, from C8 to C18, but mostly C12 and C14. As a mix of cationic surfactants, the preservative lyses the pathogen's outer membrane but it also affects the host cells, leading to loss of corneal cells and increase in drug permeability. At 0.01% w/v BAK is useful over the range of pH encountered in ophthalmic formulation and is stable to autoclaving. Unsurprisingly, in daily use, continued exposure to preservatives poses a risk and is associated with a worsening of periocular symptoms including superficial punctate keratitis in dry eye disease. It can cause allergic symptoms including eczema and blepharitis.³⁹  Polyquad, a derivative of BAK, is said to be kinder for the eye but does reduce goblet cell count which will affect tear film lipids. Purite has been investigated in our laboratory and was found to have a lower toxicity to cultured epithelial cells than BAK or hydrogen peroxide.^40,41 

Older preservatives which are rarely used include chlorhexidine, sorbic acid or sorbates, benzchlorobutanol and the mercury-based preservatives such as thimerosal, which increased the effectiveness of a preparation containing chloramphenicol plus neomycin sulfate.⁴²  The very old mercurials phenylmercuric acetate and phenylmercuric nitrate have almost disappeared. Parabens are still found in some preparations but their use is decreasing. The similarity of nasal and ophthalmic mucosa dictates that similar formulations are used for both routes. An excellent series of reviews on antimicrobial preservatives was published by Elder and Crowley.⁴³ 

An alternative is to encourage the use of preservative-free presentation, and the switch has shown a decrease in adverse symptoms.^44,45  Device approaches include the incorporation of a 0.2 µm filter to prevent reflux of organisms back from the contaminated tip into the bulk of solution (ABAK system) and an airless pump sealed system with non-return (Comod).⁴⁶ 

Ethylenediaminetetracetic acid (EDTA) increases the effectiveness of preservatives since it chelates Ca²⁺ and Mg²⁺ needed for bacterial or fungal metabolism. In various combinations, it has been shown to have a low toxicity to cultured corneal cells;⁴⁷  it causes similar (although milder) effects to BAK including a disruption of the plasma membrane, generally increasing cellular permeability. In combination with a high molecular weight hyaluronic acid, these deleterious effects seem to be reduced.⁴⁸ 

Originally described by Neuberg in 1916, a hydrotrope increases the solubility of a compound through a non-micellar mechanism. Usually relatively large amounts of the compound are employed relative to the API and solubilisation occurs by alteration of water structure and the formation of solute-hydrotrope complexes which can include complete shielding around the drug;⁴⁹  however, this is essentially nanomicellisation.

Increasing the viscosity of the formulation has a direct effect on efficacy since the rapid clearance is avoided. The bioavailability of a preparation containing either prednisolone acetate or prednisolone phosphate as 1% suspensions in carbopol 940 gel showed a 4 to 5 times increase in aqueous humour samples taken over a 12 hour interval;⁵⁰  however, a separate study of the anti-inflammatory effects of prednisolone acetate delivered in gel or a suspension showed no difference in magnitude of effect but a prolongation of duration.⁵¹  The negative attribute is that filtration of viscous vehicles may generate problems in sterilisation, since filtration pressures are markedly increased. The use of polymers in ophthalmic preparations has been reviewed by Calles et al.⁵²  and by Wagh.⁵³  The basic considerations for selections of polymers are that they should be easy to filter, stable for heat sterilisation if possible and fully compatible with other formulation components.

Cellulosic polymers, polyvinyl alcohol and gum derivatives are commonly employed as viscosifying excipients. The principal polymers used in ophthalmic formulations are shown in Table 1.1. Hydrophilic compounds such as the cellulosics hold water by weak hydrogen bonding and wet surfaces, resisting surface drying. The excipients interpenetrate tear and surface mucins: chain length, polymer flexibility and chain segment mobility are key properties of nonionic polymers.

They provide a wide range of viscosities (400–15 000 cps) and are compatible with many topically applied drugs. Polyvinyl alcohol is also a widely used drug delivery vehicle and a component of artificial tear preparations. This polymer can reduce interfacial tension at the oil/water interface, enhancing re-wetting of the surface. Polymeric solutions in their own right are therefore functional ingredients in the supplementation of mucin-deficient tears in postmenopausal dry eye.

Studies with a gamma camera conducted in our laboratories demonstrated the effects of the polymers on precorneal residence in humans and animals. In this technique, a small amount of a sterile, non-absorbed gamma-emitting radio pharmaceutical such as [^99mTc]-diethylene triaminepentacetic acid is mixed with the formulation and the clearance measured without contact to the eye. The technique is known as lacrimal scintigraphy and is particularly useful for the assessment of ophthalmic products because it is a non-contact technique which avoids provoking lacrimation and dilution of marker.²⁴ 

Although the effect of viscosity on ocular retention is obvious, a further advantage is that cardiac exposure to the unabsorbed drug is slowed. There is a point of maximum advantage; for example, for hydroxyethylcellulose at a concentration of more than 0.3% HEC, there is a significant increase in contact time which reaches a usable maximum at 0.5% w/v (Figure 1.7).

Increasing the viscosity to make semi-solid implants can be used to dramatic effect. Pilocarpine, an old drug used in ophthalmology, is an example of a drug with difficult characteristics. It decomposes rapidly at pH 6 to 7 and is normally formulated at a low pH (4.5–5.5). As a weak base with a pK_a of 6.61 it is extensively ionised at this pH. By formulating the base into an anhydrous PVA film at pH 7, the neutral pilocarpine species could penetrate the epithelium more readily, resulting in an eightfold increase in bioavailability compared to a standard eye drop. Unfortunately, manufacturing difficulties prevented the exploitation of these findings. Polyvinyl alcohol is however a useful polymer commonly used in ophthalmic solutions at concentrations of 0.25% to 3% w/w dependent on molecular weight. The pharmaceutical-quality grades that are available include low-viscosity (20 000 g mol⁻¹), medium-viscosity (130 000 g mol⁻¹) and high-viscosity grades (200 000 g mol⁻¹). PVA as used in the pilocarpine-containing film forms gel-like structures and as such may cause transient blurring that is noticeable by the patient.

When extensive hydrogen bonding occurs between a polymer and surface or a solute macromolecule, the polymer may be classified as a bioadhesive. Bioadhesion is an interfacial phenomenon in which a synthetic or natural polymer becomes attached to a biological substrate by means of interfacial forces; however, if the viscosity is too high, the formulation will offer resistance to blinking and will feel uncomfortable.

A common goal in ophthalmic formulation is to seek polymers that attach to mucin or mucous-covered membrane such as the conjunctivae or cornea, remaining in contact with the precorneal tissues until the surface mucin is turned over.⁵⁴  In this context, the narrower term “mucoadhesion” is employed. The polymer tails must be long enough and mobile to facilitate molecular entanglement. The threshold has been defined as around 100 000 Da in flexible chain motif polymers. When anionic polymers are used, the maximum interaction occurs at an acid pH, suggesting that the polymer must be in its protonated form for viscoelastic synergy with surface mucins.

Thixotropic solutions show shear thinning, yielding behaviour that is an advantage in ocular formulation. Essentially, this characteristic imitates the properties of mucin. Sodium hyaluronate is a high-molecular-weight polymer with a molecular weight of 1–3 MDa extracted by a patented process from sources including animal sources such as rooster comb, bacterial fermentation and most recently directly using hyaluronate synthetase acting on UDP-sugar monomers. It consists of a linear, unbranched, nonsulfated, polyanionic glycosaminoglycan, composed of one repeating disaccharide unit of d-sodium glucuronate and N-acetyl-d-glucosamine. It forms a flexible open coil configuration, with the random orientation of the molecules providing high resistance to shear at low shear rates. The polymer is mucoadhesive: the carboxyl groups of hyaluronate form hydrogen bonds with sugar hydroxyl groups of mucin, producing an intimate contact with the cornea. The pseudoplastic behaviour (where viscosity is higher at the resting phase) provides a thickened tear film, slows drainage and ensures an improved distribution on the cornea during blinking. A closer tear mimetic is produced when sodium hyaluronate is mixed with xanthan gum, the objective being to produce a tear mucin mimetic with similar viscoelasticity and rewetting properties. This formulation provides approximately 8 times longer residence times in humans compared to simple saline solution.⁵⁵ 

In situ gelling systems provide extended residence in the eye through sol–gel transition. This can be triggered by a change in pH, temperature or ionic strength of the formulation,⁵⁶  upon instillation in the eye. Cellulose acetate phthalate solutions have a low viscosity at pH 5, but coacervation occurs in contact with the tear fluid at pH 7.4, forming a gel in a few seconds. Unfortunately, the low pH of the gel is uncomfortable for the patient.

A poloxamer F127 solution shows thermosetting behaviour in the eye. It is a solution at room temperature, but when it is instilled onto the eye surface (34 °C), the solution becomes a gel, thereby prolonging ocular contact. As with cellulose acetate phthalate (CAP), the system requires a high polymer concentration (25% poloxamer) and the surfactant properties of this excipient may be detrimental to ocular tolerability.

Gellan gum is an anionic polysaccharide formulated in aqueous solution, which forms clear gels under the influence of an increase in ionic strength. The gelation increases proportionally to the amount of either monovalent or divalent cations. The concentration of sodium in human tears (∼2.6 μg μL⁻¹) is particularly suitable to induce gelation of gellan gum following topical instillation into the conjunctival sac. The reflex tearing further enhances the viscosity of the gellan gum by increasing the tear volume and thus increasing the cation concentration. Scintigraphic studies showed that Gelrite (0.6% w/v) significantly prolongs ocular retention in humans by forming a gelled depot on the scleral margin.⁵⁴  The product, Timoptic XE^®, has been marketed for the beta-blocker timolol maleate with less frequent dosing required compared to conventional ophthalmic formulations of the same drug.

Carbomers comprise poly(acrylic acid) polymers that undergo both temperature- and pH-dependent changes in structure. They are acidic, low-viscosity, aqueous dispersions that transform into stiff gels when instilled into the conjunctival sac upon instillation. When anionic polymers interact with mucin (which is also anionic), the maximum interactive adhesive force occurs at an acid pH, suggesting that the mucoadhesive in its protonated form is responsible for the mucoadhesion. The observed precorneal residence of the carbomer formulation in a scintigraphic study was attributed to this type of interaction.⁵⁷  Carbomers offer several advantages for ophthalmic delivery, including high viscosities at low concentrations, strong adhesion to mucosa, thickening properties, compatibility with many active ingredients and low-toxicity profiles. The polymer is more fluid at pH 5 and therefore easier to dispense at this pH. DuraSite^® (InSite Vision, Inc.) is a synthetic polymer of cross-linked poly(acrylic acid) that stabilises drug molecules in an aqueous matrix, maintaining therapeutic doses of a drug on the eye surface for up to 6 hours. The technology is used in Azasite^® (azithromycin ophthalmic solution) for the ocular delivery of the antibiotic azithromycin in the treatment of bacterial conjunctivitis: with the DuraSite DDS, only once a day dosing is required.⁵⁸  It has been modified with an additional chitosan moiety to reduce clearance rates as DuraSite-2.⁵⁹ 

Combinations of different phase-transition polymers are also being investigated in order to improve the gelling properties while also reducing the total polymer payload in the system, thereby improving tolerability and reducing discomfort. Simple mixing sometimes produces synergistic effects on thickening but may result in reduced mucoadhesion.

Lipid-based emulsions are presented as three types: oil-in-water emulsions (most common), water-in-oil emulsions and bicontinuous lipid emulsions.⁶⁰  The emulsions are stabilised with a suitable surfactant. The bioavailability of flurobiprofen axetil is significantly improved in castor oil/Tween 80 with a more than sixfold increase in AUC_(0–10 h) in rabbits.⁶¹ 

Pharmaceutical emulsions such as Restasis, an anionic lipid emulsion containing 0.05% cyclosporin, established an important toe-hold in treatment of chronic dry eye and subsequently a number of cationic lipid emulsions, including Cationorm^® and Novasorb^®, were launched for treatment of dry eye disease. Many of the newer oil-based systems are classified as lipid-based nanocarriers and include nano-emulsions and liposomes.⁶²  A positive charge on a liposome formulation markedly decreases clearance in rabbits compared to negatively charged or neutrally charged liposomes.⁶³  Other candidates for ophthalmic emulsions include ester forms of drugs such as flurbiprofen.

Eye ointments are useful for nighttime application as the white petrolatum/liquid petrolatum base causes persistent blurring of vision. The long retention means that they are useful for treating periocular eye infections and agents such as chloramphenicol, bacitracin, ciprofloxacin and tobramycin have been included in ointment compositions. Ointments are also used as ocular lubricants in severe dry eye. They are more difficult than solutions to manufacture as sterile preparations and techniques such as hot filtration or dry heat have been used but practically the market has embraced non-greasy alternatives including polymer-based gels. An advantage of ointment vehicles is that they can be used in an anhydrous form which would be useful for water-labile drugs.

There is interest in novel non-aqueous, non-lipophilic vehicles based on the perfluoroalkanes which were first used as blood substitutes as the liquid dissolves oxygen (Green Cross). The use in tamponade in retinal surgery was well established 30 years ago although the physical compression causes damage to the tissues and has to be removed once the condition has been controlled.⁶⁴  When injected for tamponade of the retina, perfluoroctane shows retinal toxicity and has to be purified in order to be used for ocular surgery.⁶⁵  Perfluorodecalin can be used externally as a novel way of delivering a small drop suspension that shows flash dispersion on the eye.⁶⁶  Because the drop size is small, surface tension is extremely low and the suspended particles are dry, the system flash disperses to the edge of the eye and doses the Meibomian gland margin. This suggests that it could be used to target the mucus-secreting cells at the edges of the eye lashes. The vehicle is being investigated in several new systems for the delivery of cyclosporin.

The preferred dosing vehicle is usually a solution but stability in dilute solutions may limit shelf life. In addition, important anti-inflammatory drugs such as dexamethasone and prednisolone have the highest partitions into the corneal epithelium and may achieve more flux, even from suspensions, compared to water-soluble salts. Particle size is important because particles 15 µm or larger are irritants. Accordingly, typical specifications are 95% average particle size below 10 µm. Milling to produce smaller particles will obviously produce an opportunity for a sustained delivery based on control of the rate of dissolution. Larger particles theoretically provide prolongation of effect due to the increased size of the depot; however, there is a risk of irritation and suspended particles being washed out before dissolution, leading to lower bioavailability.⁶⁷  It was demonstrated that for radiolabelled dexamethasone, administered as suspensions at 5.75, 11.5 and 22 µm, dissolution at the greater particle size proceeded so slowly that the particulates were ejected from the eye before dissolution was complete. A further complication is agglomeration. The mucins in the tear film will coat particles with glycoprotein, causing the concretion of a mass which is subsequently ejected. This may be reduced by incorporation of polymers such as polyvinyl alcohol or polyvinylpyrrolidone as crystallisation inhibitors, and viscosifiers which also maintain dispersal. Suspensions are kinetically stable but thermally unstable as a system and problems occur when left due to cyclical exposure to heat and cold which may encourage crystal growth and sedimentation on the bottom of the dropper bottle. It is therefore important that the particle is in a deflocculated state.⁶⁸ 

Nanotechnology-assisted delivery employing amphiphilic molecules has been exploited for anterior segment delivery and the technologies include nanomicelles, liposomes, dendrimers, nanospheres and nanocapsules. These systems occupy the border between intra-cellular entry mechanisms and membrane fusion. Endocytotic mechanisms have been proposed to explain increased permeation.⁶⁹  Nanoparticles placed externally will be cleared by lymphatic drainage, and in live animals that barrier function is maintained. The nanotechnology constructs are generally at early stages, with experimental platforms described for isolated cells and laboratory animals. There are however some interesting data from preclinical and early clinical studies.

Liposomal formulations which could supply phospholipids and other tear-stabilising factors have been widely investigated as carriers since they can encapsulate both hydrophobic and hydrophilic drugs. It is the carrier used for verteporfin. The corneal pI is 3.2 and is therefore negatively charged with regard to most carriers. This can be exploited in this technology as positively charged liposomes show more attraction to the cornea than neutral or negatively charged liposomes.⁶³  Liposomal formulations are usually prepared aseptically as they are easily degraded by heat.

A liposomal formulation loaded with triamcinolone acetonide was administered as a topical liposomal formulation to a small group of patients with cystoid macular oedema in a commercially sponsored trial.⁷⁰  The 12 patients were refractive to treatment following cataract operations and received one active drop every 2 hours for 12 weeks. No placebos or alternative treatment was used in the study. Foveal central field thickness was decreased from baseline, indicating exposure of the retina from a topical dose and resolution of the condition. Enhancement of retinal electrophysiological function was noted by another group after treatment of patients with open angle glaucoma (OAG) with a liposomal formulation of cytidine 5′ diphosphocholine, citicoline, a drug proposed for the treatment of brain injury, Parkinson's disease and dementia. The vehicle also contained sodium hyaluronate. It is an endogenous compound and an intermediate in the production of phosphatidylcholine. Only one eye was treated and the patients showed improvement in the retinal bioelectric responses against the baseline, attributed to an effect on the retinal basal ganglion.⁷¹ 

Since ophthalmic products are sterile products, the processes and procedures used in production are essentially similar to those for parenteral administration. Manufacture must be carried out under clean conditions with scrupulous attention to maintenance of quality attributes and absence of contamination. The design of airflow, fittings and reducing contact with particulates in small-volume ophthalmic medicine production is addressed by specialist companies as these facilities and the knowledge base are highly specialised. Unlike high-volume parenteral production, the buildings do not generally require mezzanine structures to support equipment in an overall shell, but the design from production to product must be carefully organised. All elements of the process are subject to rigorous tests. The packaging must ensure that pharmacopeial standards are maintained during shelf life and, according to use, the product may be subjected to freeze–thaw cycles to check API integrity. The most common source of contamination is the human operator, so systems are highly robot controlled, using process analytical control where possible.

Manufacture follows three main routes. Generally, regulators prefer terminal post-production processes such as moist heat sterilisation as the primary pathway. If the material is thermolabile, the sterilisation is performed by filtration through 0.22 µm filters. These must be sterilised prior to use to avoid post filter transfer. Preservatives such as benzalkonium chloride may bind to plastics used in filter construction and this must be accounted for in production.

All issued drug products for ophthalmic use conform to regulatory standards and release testing attributes. Gibson provides a useful list for a multi-dose ophthalmic product as follows (Table 1.2).

It can be quickly appreciated that many new drug compounds are extremely fragile and cannot be sterilised during manufacture. Most larger biologic constructs for example will not survive heat, ethylene oxide, or E-beam irradiation and must be prepared aseptically.

Overall there have been relatively few successful implant devices used in ophthalmology. The devices are divided into solid platforms that are removed on exhaustion and include contact lenses and bioerodible constructs that can be left to disperse. Soaking a contact lens in a drug solution allows slow release from the pre-lens film (in contact with air) and is absorbed through the conjunctiva, whereas the material trapped under the lens is absorbed through the cornea or radially diffuses to the conjunctiva, aided by small movements of the lens on the eye.⁷²  As an alternative, drugs could be incorporated during polymerisation or added by coating.⁷³  Copolymerisation with cyclodextrin to create adsorption possibilities for hydrophobic drugs, loading using super-critical fluids and the use of nanocarriers such as liposomes, nanoparticulates and micro-emulsions have also been attempted. Scleral lenses, dimensions 18–24 mm diameter have been used post refractive surgery and in glaucoma treatment as stated in an excellent review by Dubald and colleagues, which focuses on ocular antibiotic therapy.⁷⁴ 

Ocusert was a breakthrough innovation for delivering pilocarpine from Alza, then the world's leader in advanced zero-order drug delivery concepts. The device had to be placed in the inferior fornix by the patient where it released drug into the tear film below the lower eyelid. Unfortunately, it was a commercial failure. Ocusert lasted only 7 days and a new unit needed to be re-inserted. Patient compliance was poor, the unit was occasionally lost from the eye and the product is no longer marketed.⁷⁵ 

So far we have considered solubilisation of drug substances and aid to provide sustained delivery at the eye surface. Treatment of diseases beyond the cornea such as glaucoma is not particularly efficient as typically 3–5% of the dose reaches the anterior chamber. Robust control of drug delivery beyond the lens generally requires invasive procedures as transfer to the posterior eye from the cornea is low. We must now consider the eye structures to understand the nature of the barriers in ocular drug delivery.

Following topical dosing, the formulation ingredients mix with the tear film and movements of the upper lid draw the mixed film upwards. This is opposed by fluid flow downward from the lacrimal gland. With the increase in viscosity, materials mix poorly and tend to remain in the lower cul de sac as illustrated previously in Figure 1.3. This lower reservoir is important because it favours distribution through the conjunctival lining, rather than the cornea.

The cornea is a significant barrier because it is composed of sequential layers, with the cells of the outer epithelium forming a pavement of hexagonal cells, tightly fused together. The sclera and the cornea form the tough outer substantial support, resisting an internal pressure of 13–19 mm, or greater in disease. In the adult the cornea measures 11–12 mm horizontally and 9–11 mm vertically and is only 1/6 of the globe's surface, the remainder being sclera. In the front of the eye, the sclera is covered with the conjunctival membrane and although initially the cornea was thought to be the principal barrier for drug absorption, the smaller surface area of the cornea compared to the sclera/conjunctiva probably means that for most compounds, the corneal absorption does not play the dominant role. The molecular size limit for corneal absorption is around 500 Da, whereas conjunctival tissue shows no clear relationship with regard to flux with molecular radius and log distribution coefficient and is much more permeable.⁷⁶  Nevertheless, it is the tissue that researchers study because it is the easiest to harvest. Corneal thickness varies, being thinnest centrally (550 μm) and gradually increasing to the periphery (700 μm). It has an avascular five-layered structure comprising (anteriorly to posteriorly) the epithelium which is 5–6 cell layers thick and hydrophobic, Bowman's membrane, stroma which forms the main hydrophilic tissue in the cornea, Descemet's membrane and endothelium, as shown in Figure 1.8. The endothelium controls the transparency and thickness of the cornea when alive by pumping out sodium ions, withdrawing water and preventing swelling.

The isoelectric point of the cornea is approximately 3.2 and as a result carries a negative charge, favouring the absorption of cations. Cationic surfactants therefore readily associate with the membrane. Disruption of the outer corneal architecture with surfactants leads to loss of barrier function. This has been described by some inventors as the action of absorption enhancement, whereas the action of such substances on the outer cornea is best described as transient damage. The thin hydrophobic epithelium contains approximately 100 times more lipid than the stroma and with the conjunctival lining provides a temporary tissue reservoir for fat-soluble drugs. To cross the cornea, the drug should also have sufficient hydrophilicity to traverse the stroma. Intermediate solubility in both epithelium and stroma is needed. This provides the rationale for ester prodrugs such as prednisolone acetate, which must partition into the epithelium and then hydrolyse to a more hydrophilic parent product. Because of the nature of this double barrier, partition through the cornea is seen best with a log D of around 2–2.5 at neutral pH. Mild acidification is tolerated but alkaline solutions form soaps and irritate.

The drug flux pathways are summarised in Figure 1.9. Drug applied to the tear film distributes and reaches the anterior chamber by two routes: through the cornea and through the sclera and conjunctivae. Overall, the bioavailability is low: typically 3%.

The sclera is composed of an outer episcleral layer, the stroma and the interior lamina fusca and is fairly devoid of blood vessels and human beings are distinct from other primates by having a white sclera. The blood vessels are largely near the surface and originate from small arteries of the anterior ciliary vessels. The matrix (stroma) has bundles of collagen fibres which form a sheet structure rather like the structure of a tennis ball. In sclera excised from an animal, lateral permeability is similar to forward permeability, but when the eye is perfused, lateral movement is visible with solvent drag occurring in the mid-sclera, moving the drug faster laterally than it penetrates.⁷⁷  The thickness of the sclera varies across the globe which might indicate the most favourable site of application for drug delivery. From an examination of tissue in eye banks, the region just beyond the corneoscleral limbus, the equatorial area, is thinner than the corneoscleral limbus,⁷⁸  which has been confirmed by MRI.⁷⁹  Scleral permeability is similar to corneal stroma and has a permeability to moderately polar compounds up to 70 kDa. This is an important distinction since the cornea is more selective with regard to relatively hydrophobic drugs.

The sclera is overlaid by a thin membranous covering of conjunctival tissue which is composed of squamous non-keratinised cells. The tissue, visible as a pink flat mass covering the anterior sclera, is the bulbar conjunctiva and is loosely attached; that lining the inside surface of the eyelids (palpebral conjunctiva) is more firmly anchored.⁸⁰  The conjunctiva has a large surface area—approximately 18 cm² in an adult. The relative area of the conjunctival lining is about 17 × that of the cornea.⁸¹  This large ratio is approximately double that found in rabbits and might account for PK differences in comparison between the two species.

The conjunctival membrane covers the frontal part of the eye, apart from the cornea, and is connected to the epidermis of the eyelids. It is important for producing the aqueous component of tears, aiding film stability and maintaining anti-inflammatory status. Bulbar and palpebral conjunctiva had comparable permeability for polyethylene glycol oligomers of molecular weights below 1000 g mol⁻¹, confirming that neutral and hydrophilic therapeutic compounds should have equal degrees of absorption through conjunctival tissue.⁸²  It has been shown that more than 20–50% of an instilled low-molecular-weight drug was absorbed through palpebral conjunctiva into systemic circulation.⁸³  The conjunctivae serve as a secondary reservoir for drugs and instillation of one preparation within 5 minutes of the first usually results in a reduction of the concentration.⁸⁴  It is not surprising therefore that absorption of drug compounds into the conjunctiva is favoured over the cornea. Larger solutes and particulate systems were unable to cross the conjunctiva epithelia barrier, suggesting that other routes such as endocytic transport are important. It has been shown that nanoparticle uptake could be detected an hour earlier in the conjunctiva compared to the cornea.⁸⁵ 

The role of the lymphatic pathway in the conjunctivae and removal of nano-constructs is of particular interest in the delivery of complex pharmaceuticals. The conjunctival membrane covers the lids and most of the frontal sclera as a thin membrane. In terms of surface area, the relative area is 12 × that of the cornea. In the living eye, the sclera effectively clears particles towards lymph nodes and uptake can be seen 6 hours after administration.⁸⁶  As shown in this study of 20 µm particles injected subconjunctivally in rats, the particles were cleared rapidly to the spleen and liver and did not transport across the sclera to the retina (Figure 1.10).

The eye is a relatively isolated structure with a vasculature largely confined to the periphery of the globe supporting nutrition of the retina, as shown in Figure 1.11. It is composed of three layers which from the outer surface are (a) the sclera and cornea which are connective tissue elements, (b) the uveal tract at the front of the eye which consists of the choroid, iris and the ciliary body and finally (c) the innermost retina, which has an outer pigmented layer and the inner neuroretina with photoreceptors, bipolar cells and ganglions. The large areas devoid of cellular material are made up of a connective tissue matrix and serve to support the lens and cornea so that images can be focused on the retina. The lens and iris, situated towards the front of the eye, divide the globe into two chambers: the fluid-filled anterior chamber and the posterior chamber extending from the lens to the retina. This chamber contains a gel—vitreous humour—that acts as a shock absorber. The system is under hydraulic pressure generated by fluid flow from the ciliary body as aqueous humour, a thin watery fluid containing nutrients and antioxidants. The pressure shapes the cornea, preserving the curvature and thus the dioptric power of the front of the eye. The volume of the chamber is higher in the young than in the elderly: 247 ± 39 μL in 20–30 years adults versus 160 ± 39 μL in the 60 + years group.⁸⁷ 

The tissue behind the lens forms the posterior chamber and is filled with a structured hydrogel collagen type II fibres, coated with type IX fibres bridged by glycosaminoglycans giving the gel its non-uniform, viscoelastic property. In the young adult, the volume is approximately 4 mL.⁸⁸  In laboratory animals, the anatomy profoundly alters the distances between injection towards the posterior pole, so for example the pathway topical surface to retina in mouse and rat is short. In the rabbit, the lens occupies a much bigger volume than in primates.⁸⁹  The gel is transparent and gradually collapses with age, as shown by the elegant work of Seebag.⁹⁰ 

As has been mentioned, solutions placed on the cornea will be cleared quickly by tear drainage and venous and lymphatic outflow, and loosely attached external drug-eluting implant structures may be dislodged unless the patient becomes deliberately tolerant. Any temporary degradation of the image provokes tearing, forced blinking, movement of facial muscles and the urge to rub the eyes. Flow will dilute and distribute drug released from devices and only that fraction that reaches the retina allows treatment. Around the eye, broaching the sclera will allow closer access of a drug from an exterior reservoir but deeper penetration will be strongly influenced by the choroidal circulation.

Dry eye conditions associated with changes in hormone balance, contact lens wearing and staring at computer screens are often treated with artificial tear solutions or with topical solutions of pharmacological agents which reduce surface inflammation including cyclosporin, lifitegrast and rebamipide⁹¹  which may be presented as solutions or emulsions. Topical and systemic fatty-acid-containing formulations have also been used for treatment.⁹²  Small tablets and other dissolvable delivery devices have been used as an alternative to drops to provide sustained treatment and minimise nursing care. These include soluble inserts based on PVA (e.g., NODS⁹³ ), gelatine foams impregnated with insulin for diabetic patients or mydriatic drugs to widen the pupil and other devices anchored under the lower eyelid.⁹⁴  Many of these concepts were discontinued due to problems in use or manufacturing difficulties.

Devices placed under the lower eyelid, an invention from early ALZA days that was marked as Ocusert^® which delivered pilocarpine in a zero order rate are still investigated, for example the delivery of bimatoprost in a flexible silicone ring. The original issue that these systems would also fall out of place or would be uncomfortable is dealt with by supply of multiple sizes for a best anatomical fit. Another issue that emerged was that constant delivery of prostaglandin analogues produces a compensation and a lowering of the magnitude of the effect over a period of time.⁹⁵  Other minor issues noted are mechanical stimulation leading to an increase in mucus discharge and the device being partly visible at the edge of the eye.

The addition of polymers to increase lubrication and decrease evaporation is partially successful, but higher viscosities lead to shear effects, causing discomfort. In part, dry eye may result from iatrogenic interventions including cosmetic surgery, corneal refractive surgery (as in laser-assisted in situ keratomileusis, LASIK) and from some of the ingredients of glaucoma medication, especially preservatives such as benzalkonium chloride.⁹⁶ 

Post-LASIK surgery, the use of silicone- or collagen-based punctal occlusion plugs has been one of the alternatives to surgical interference with the duct; however, many patients lose their plugs with a consequent reduction in effect. The use of a plug to increase the efficacy of anti-glaucoma treatment is an old and successful idea,⁹⁷  which again surfaced in treatment with prostaglandins.⁹⁸  The systems are of two types, dissolvable drug-eluting plugs and occlusion devices, as illustrated in Figure 1.12.

The use of a travoprost-eluting punctal plug based on polyethylene glycol to treat glaucoma was described in a review by Young et al.⁹⁹  The rod-shaped system consists of encapsulated travoprost spheres in the dissolvable matrix. The authors draw attention to understanding how the drug-loaded reservoir is able to enter the tear film and that patients may not be aware when implants dislodge or fall out, necessitating retreatment.

In order to increase flux of drug into the eye, iontophoresis, electroporation and sonophoresis have been utilised in relatively early clinical practice or as experimental tools.¹⁰⁰  The approach is to provide a drug reservoir on the outer eye surface and drive the drug in by physical means. It is required that minimum manipulation occurs and desirable that the treatment time is kept short. In iontophoresis, the drug is soaked into a pad that is applied to the sclera under the lower eyelid and connected to the controller. The amount of drug as a salt delivered is proportional to the current applied, the length of time or treatment and the surface area of the drug-loaded contact.¹⁰¹  An appropriate polarity is selected for the drug: an anode for a positively charged salt and a cathode for negatively charged drug. The principle is thus repulsion: the drug is pushed through the tissue at low voltages at currents of a few milliamperes so heating effects are small.

Gene therapy is hampered by the need to deliver charged molecules, and Behar-Cohen and others have pioneered the use of electroporation to load ciliary muscle with large constructs including plasmids.¹⁰²  In electroporation, the cell membrane is temporarily opened up with a short pulse sequence allowing the introduction of material before significant leakage occurs.

Cell membrane permeation can also be achieved by ultrasound probes by generating a micro-bubble containing the payload—for example, the plasmid—at the tip and then employing the pulse train to cavitate the micro-bubble.¹⁰³  SonuVue^® micro-bubbles are phospholipid vesicles containing sulfur hexafluoride and are used by many workers, and Luo et al. describe the application of ultrasound to target siRNA into cells.¹⁰⁴ 

Injection into the space between the cornea and the lens is known as intracameral injection. Injection into the anterior chamber will not achieve significant drug concentrations at the back of the eye, but the injection can be used to deliver antibiotics when the risk of endophthalmitis is high or steroids such as triamcinolone acetonide to reduce inflammation after cataract surgery. However, if the lens is removed (pseudophakia) during cataract surgery, intracameral implants treat the posterior tissues more effectively.¹⁰⁵  More recently, interest in intracameral implants has increased, especially for the treatment of glaucoma. This disease is ‘silent’ and treatment is less effective because of poor patient compliance, even when patients know that they are being monitored electronically.¹⁰⁶  Open-angle glaucoma is a chronic, progressive disease resulting in loss of retinal ganglion cells with atrophy of the optic nerve and consequently irreversible vision loss. Treatment with longer-acting prostaglandins such as latanoprost and travoprost was established in topical therapy, and industry players with newer prostamides such as bimatoprost, which appeared later in development, elected to go down the implant route. The slow-release implants are designed to release drug for 4–6 months and prolong effectiveness. The dose is extremely small; in the case of bimatoprost SR based on Novudur^® technology, the administered dose is 15 µg. After administration to beagle dogs, the implant could be seen in the inferior angle of the dog eye.¹⁰⁷  These systems are more selective and effective than much larger doses given topically.

The current treatment for cataractous lens is the removal of the severely clouded lens and replacement with a synthetic intraocular lens. The inflammatory response is considered to be normal and is routinely managed by topical NSAIDs or corticosteroids as eyedrops for 2 weeks post surgery.

If untreated, a persistent inflammatory response can lead to complications including uveitis, iritis, glaucoma and an increase of intraocular pressure (IOP). The FDA recently approved a sustained-release suspension of dexamethasone (5 µL) injected into the intracameral space forming a depot alongside the IOL and behind the iris.¹⁰⁸  An alternative would be to incorporate the drug in the lens, and a recent publication reviewed the scope of the applications.¹⁰⁹  The strategies used are illustrated in Figure 1.13, which shows a range of techniques that can be used to load polymer matrices.

Other applications include treatment of posterior capsule opacification. The implant is placed into the capsular bag from which the lens was phaco-emulsified and a complication is that a scar tissue layer can form months to 5 years later. This is sometimes known as secondary cataract. The condition is treated by laser capsulotomy but the feasibility of the incorporation of an anti-metabolite such as methotrexate has been investigated.¹¹⁰ 

Removal of the vitreous affects clearance throughout all of the interior tissues including the anterior chamber. Triamcinolone acetonide has a mean half-life of 18.6 days in non-vitrectomized eyes and 3.2 days in post-vitrectomized eyes.¹¹¹ 

The vitreous humour (VH) is a large hydrogel compartment filling the space between the lens and the retina. Its appearance is almost featureless and transparent, which allows 85–95% of light with wavelengths between 300–1400 nm to be transmitted across to the retina. The gel consists of 98–99% water and is structured from collagen type II fibres, coated with type IX fibres. Collagen bridged by glucosaminoglycans gives the gel a viscoelastic nature, which is non-uniform. In the young adult, the volume is approximately 4 mL.

The VH acts as a hydraulic damping mechanism that protects the internal structure of the eye from friction and vibration caused by sudden eye movements and strenuous physical activity. It provides form to the eye and although the solid content of the vitreous humour is relatively small (∼1%), the alignment of collagen and glucosaminoglycans as shown is sufficient to maintain the rigidity and stability of the gel and more critically to act as a molecular barrier to diffusion. It also assists as a sink for oxygen, with the ascorbate protecting the lens from elevated oxygen levels.¹¹² 

The arrangement of the collagen and hyaluronate forms a mesh that resists diffusive movement. Foulds found that when treated with hyaluronidase, movement of water from mid vitreous to choroid is faster 20 + 7 minutes in treated animals versus 36 + 13 minutes in controls.¹¹³  The clearance of compounds in humans is slower than in laboratory animals, as shown in Table 1.3, although there are considerable differences in values obtained in human studies.¹¹⁴ 

In addition, globular-shaped molecules can cross the vitreous network more easily compared to linear molecules since 50K radio-iodinated albumin has been documented to be cleared faster than 50K labelled dextran. This again suggests a sieve-like vitreous microstructure that selectively limits the diffusion of injected molecules. The distance from the site of injection to the macular tissue is short enough to allow simple diffusive movement from the site of application to the target in mice, so results from that species will mislead us. The size of the lens in rabbits is also important as the distance from an IV injection to the posterior pole is shorter than in humans, leading to a higher concentration. Thus, data obtained from small animals is unlikely to translate to clinical outcomes.

Overall, drug substances administered intravitreally distribute from the point of administration by diffusion, as illustrated in Figure 1.14, with a small amount refluxing, dependent on the volume of the injection. The pool of drug spreads back to the retina and is cleared through the uveal tract or by advective flows around the lens as shown in Figure 1.14B, but there are some important observations indicating that the flows are complex. Surgeons observe a transient increase in IOP following intravitreal injections. The volume of the injected drugs varies from 0.05 (e.g., ranibizumab) to 0.2 ml (e.g., high-dose triamcinolone). This is of concern as it could lead to a short-term occlusion of the central retinal artery and has been observed after administration of a number of agents.¹²⁴ 

Currently, intravitreal drug administration appears to be the surest option in achieving therapeutic drug concentrations in the posterior eye, although issues of maintaining effective concentrations at the target remain.¹²⁵  Although the vitreous humour is avascular in nature and is a static gel, the circulation systems within its vicinity, including suprachoroidal and episcleral vascular flows, allow adequate drainage of injected materials or removal of metabolic waste from the vitreous.

All tissue spaces that have directional flow components associated with them result in temporal and physical gradients of exposure, which is most evident when the device is delivered by bolus or is physically much smaller than the space that it is put into. Knowledge of advection and convective forces becomes important in a closed space, because a flow away from a target determines an optimum physical spacing or, failing that, maintenance of a flux sufficient to ensure sufficient exposure; for example, an implant in the vitreous space to control angiogenesis must generate sufficient anti-VEGF activity to control vessel leakage. In general, large molecules appear to be cleared forwards and therefore a large proportion of the dose is lost. This situation worsens when the vitreous becomes more fluid, as in old age.

On ageing, the vitreous humour undergoes progressive structural and biochemical changes with loss of gel content. Neither the vitreous humour nor the inner limiting membrane undergo renewal in later life and do not regenerate if the vitreous gel is removed surgically (vitrectomy). It was found that the vitreous of young human adults of around 20 years of age was 80% gel phase, which decreased to almost 50% beyond 60 years.¹²⁶  These processes are usually associated with vitreous synaeresis (contraction) and synchisis (liquefaction), with the rate of occurrence increasing with age. Vitreous diffusivity and convective forces appear to be enhanced in the partially liquefied vitreous, leading to a faster rate of drug clearance.

The principal advantage of delivery to the vitreous cavity is that it provides access to a reservoir, which is able to supply the retinal tissues by crossing the inner limiting membrane into the neuroretina that lines the globe rather like the inner surface of a ball, as discussed later. The retina receives a fast blood flow per mg tissue but the fraction of the circulation it receives (approximately 1/37) will always mean that systemic delivery is associated with large bystander effects.

There are certain procedures in which the drug is delivered into the circulation by intravenous injection—for example, liposomally entrapped verteporfin—and release triggered in the eye by thermal effects from a laser beam. This is known as photodynamic therapy (vPDT). Although originally a first line intervention, the advent of potent monoclonal antibodies (mAbs) began to supplant this approach, which took more nursing time than simpler office procedures. However, the EVEREST II trial in which vPDT was combined with ranibizumab intravitreal therapy suggests this approach is superior to monotherapy with the mAbs alone.¹²⁷ 

Direct injection of a monoclonal antibody into the vitreous quickly arrests blood vessel leakage. The formation of new vessels under the retinal layer causes oedema that lifts the retina away from base layer causing loss of function and eventually atrophy. As the cells are not replaced, there is a gradual loss of sight.

The driving force is the release of vascular endothelial growth factor, which are a family of materials which have essential roles in vascular development. Of these, VEGF-A is identified in causing the leakage of the microvasculature and early attempts at amelioration used laser therapy to seal vessels and corticosteroids to quieten dampen inflammation. Modern anti-VEGF agents were developed which became the mainstay of the present day approach with biologics. These are bevacizumab, ranibizumab and aflibercept and future anti-VEGF strategies are focused on extending the interval between treatments.¹²⁸ 

Controlled delivery systems have been recognised as an avenue for exploration and include suspensions, gels, nanotechnology-based approaches and devices, as summarised in Figure 1.15. Cao et al., Gote et al. and Kim and Woo have reviewed the approaches in sustained ocular drug delivery.^129–131  Of the various innovative research strategies, non-biodegradable implants have proven themselves to be the closest to reaching clinical approval.¹³² 

Because the eye is an enclosed space into which the agent is directly introduced, local inflammatory reactions related to the substance itself, the formulation components or the repeated intraocular injections can arise. Non-specificity of the pharmacological actions of the agent is a concern. For example, the broad blockade of all VEGF actions can cause capillary regression, changes in pericyte numbers and basement membrane in the adult normal vasculature, such as seen with tyrosine kinase inhibitors in non-ocular tissues.¹³³  The shape of an object provokes a foreign body reaction. Studies in cynomolgus monkeys showed retinal degeneration and epiretinal membrane formation around poly(lactic-co-glycolic acid) (PLGA) rods at the temporal pars plana.¹³⁴  In a study of a copolymer microparticle, polyethylene glycol–polybutylphthalate (PEG–PBT hydrogel) containing anti-VEGF therapy and administered to cynomolgus monkeys, microparticles were observed to migrate from the posterior to the anterior chamber and pool in the anterior chamber; however, the same phenomenon was not found in rabbits. In addition, the slow degradation of the carrier particles appeared to provoke ocular inflammation.¹³⁵ 

The first intravitreal devices to reach useful commercial exploitation were ganciclovir implants used in the treatment of cytomegalovirus infections, which were a prominent cause of blindness in AIDS patients. The implant is based on a silicone-based elastomer encasing a drug core with a release port at one end covered with a PVA-based membrane that controls drug release. To introduce the implant, a small incision is made into the pars plana. Retisert is the modern iteration that releases 3–4 µg fluocinolone per day over 30 months in the treatment of non-infectious posterior uveitis.¹³⁶  The unit is secured behind the lens as shown in Figure 1.16.

A higher rate of cataract progression associated with the Retisert^® device has generally encouraged surgeons to examine other alternatives such as the Ozurdex^® implant containing 0.7 mg of dexamethasone in a PLGA matrix.

A lower-dose non-degradable implant, Illuvien^®, which contains 250 µg fluocinolone acetonide is administered through a 25G needle into the vitreous and releases 0.2 or 0.5 µg of drug per day. A common complication, that all the patients so treated developed cataract, is a well established sequelae of IV steroid treatment and is expected.¹³⁷ 

The gradual removal of the matrix after the drug has been released is desirable, although there is a risk that the fragile device will fragment and material will float to the front of the eye. Ozurdex, developed by Allergan, releases steroid as the rod-shaped implant slowly degrades to lactic and glycolic acid, with peak daily release in the first 2 months after implantation using a purpose-made injector.^138 ,139  The use of steroids in the eye to control inflammation has been described by physicians since the 1950s and technology has enabled a significant reduction in systemic side effects by control of rate and position of release.¹⁴⁰  The key requirements as noted by the retinal physicians are a very sharp needle and to keep the needle size as narrow as possible. Studies are underway to assess materials which are more benign to protein than PLGA, which on hydrolysis generates acidic pockets in the implant which can denature a protein.

There are many polymer-based platforms in development for ocular delivery but the drug particles when presented in this manner will migrate, causing potential problems as discussed earlier. Much clinical experience is being gained by purposing these systems for use in osteoarthritis or from peri-ocular injections (e.g., sub-Tenon), Triamcinolone acetonide is administered as a suspension and by virtue of slow dissolution has sustained release properties.

Particles can be introduced into the suprachoroidal space. This is often described as a virtual anatomical space, since it is visible on injection of a fluid with a short needle through the sclera or the eye is perfused from a sclerally mounted microneedle array.¹⁴¹ 

The SurModics I-vation implant is a triamcinolone-coated helical coil made from titanium coated with polymer that is injected with a 25G needle into the sclera. It delivers drug into the eye for a period of 2 years. The small device can be recovered as necessary. The system raised a lot of excitement in the treatment of diabetic macular oedema when the concept was supported by Merck but photocoagulation was judged to be more successful and interest has waned over the years.¹⁴² 

A small refillable reservoir-based device for continuous delivery of ranibizumab was recently introduced by Genentech and the general features are illustrated in Figure 1.17. The reader is referred to the manufacturer for specifications and the figure is provided for general information only. This device, The Port Delivery system, is surgically implanted through a small incision at the pars plana. A self-sealing septum in the centre of the scleral flange allows refills to be completed at the clinic.¹⁴³ 

The ultimate target of drug delivery is usually the preservation of the function of the neuroretina. The retina is well protected from the outside from chemical agents by the retinal pigment epithelium containing melanin and by efflux transporters. From the inside, medium-sized proteins cross the retina but for larger constructs such as dextrans or mAbs the influence of synchisis becomes important.^144,145  On ageing, the vitreous liquefies as the hyaluronans separate from the collagen fibres that make up the vitreous gel and collapse toward the back surface of the lens, allowing greater convective clearance forward. In this population, sustained delivery devices may fail to be effective.

The first barrier for drug in the vitreous to reach the inner retina is the inner limiting membrane. This is an extracellular matrix of type IV collagen, laminin and fibronectin and forms the basement membrane of the Mueller cells.¹⁴⁶  It is not of uniform thickness and in primates is considerably thicker than in laboratory animals – mouse 100 nm versus human 4 µm and thickens with age.¹⁴⁷  In studies with labelled mAbs such as bevacizumab, the large protein is shown to accumulate at the inner limiting membrane (ILM) before showing signs of crossing the retina. Although not a quantitative assessment, the authors suggest that Mueller cells may in part be responsible for transport.¹⁴⁸  This seems reasonable in view of the association of ILM and the Mueller cell layer. The presence of laminins is thought to play a role in accumulating viral vectors at the ILM, aiding transport through the retina.¹⁴⁷ 

The neural retina is a relatively dense structure consisting of the layers of outer ganglion cells, plexiform layers and photoreceptors and probably freely transports proteins up to 76 kDa.¹⁴⁹  Because larger proteins can penetrate, it is probable that active transport and translocation processes are partially responsible for distribution throughout the retina. This detailed information is outside the scope of this article but for a review of anatomical factors affecting the efficacy of drug delivery within the eye, please refer to ref. 150.

Three factors were identified in the WHO global action plan (2014–2019) with the goal of reducing the impact of vision impairment. To provide a remedy appropriate to need, the magnitude of visual impairment was assessed in a systematic analysis. The estimated growth in the number of blind increased from an estimated 30.6 million in 1990 to 36 million in 2015. The factors driving the change were identified as population growth and population ageing, countered by a reduction in age-specific prevalence.¹⁵¹  In addition, mild to moderate visual impairment climbed from 160 million to 217 million over the same period. It was confirmed that women bear the majority of sight impairment as revealed in population-based studies. Female gender is particularly an issue in South East Asia regarding cataract surgery and if addressed, cataract blindness could be reduced by 6.28%.¹⁵²  In addition to cataract, women have a higher incidence of diabetic retinopathy than men.¹⁵³  Most importantly, the Bourne report that formed an important pillar of the WHO action drew attention to a study showing that health care intervention provided one of the largest returns on investment (ROI). For example, the ROI calculated for ranibizumab therapy in the treatment of neovascular age-related macular degeneration (ARMD) was 450%, that for cataract surgery was 4500% and that for medical OAG therapy was 4000%.¹⁵⁴ 

The health burden is significant globally but the patterns of blindness in the 50 years + age group vary considerably across the world with a high prevalence of ARMD in the high-income regions. This promotes investment by pharmaceutical companies to provide new approaches in the management of visual diseases and span activity across drugs and devices (Figure 1.18). It can be seen that future approaches are multifaceted, drawing from inputs in chemistry, biology and engineering.

The ocular drug delivery space is a lucrative commercial market. It is supported by a growing proportion of the population who are older and better informed at seeking new and more effective treatments, the appetite for preventive medicine to reduce burden of eye diseases and the increasing health expenditure that we are promised will improve quality of life.