Experimental investigations of the fluid dynamics of vitreoretinal surgery

Vitrectomy is a surgical procedure by which the vitreous humour is removed from the vitreous chamber of the eye and is replaced with a tamponade fluid. Vitrectomy is performed by means of a system consisting of several components, including the vitreous cutter, which aspirates the vitreous body, a viscoelastic fluid that occupies the vitreous chamber of the eye. The vitreous cutter port geometry may have great influence in vitrectomy surgery from the fluidics and safety standpoints. For this reason, there has been an increase interest from ophthalmic companies in optimising the fluidics within the vitreous chamber during surgical maneuvers, in order to avoid retinal tractions and intraocular pressure (IOP) variations. Once removed, the vitreous is often replaced with Silicone Oils (SOs), which, however, inevitably tend to emulsify in the aqueous solution present in the eye, limiting the chances of using SOs as long term vitreous substitutes. Many months are indeed often required for complete healing of retinal damages. Moreover, formation of emulsions triggers local inflammatory processes, which can cause further complications. This thesis is an experimental work related to the physics of ocular surgical procedures and it consists of two main parts. The first part concerns the fluid mechanics of the vitrectomy procedure. In particular, experiments with different approaches have been performed on the fluidics of various vitrectomy systems, with the objective of identifying ideal working conditions that minimise the risks of retinal damages, while keeping the surgical procedure efficient. The second part is related to the characterisation of the interfacial rheology between SOs and aqueous solutions containing endogenous proteins released by the ocular tissue as a response to an inflammatory state. Aim of this part is to investigate to what extent biomolecules present in the aqueous solution can modify the rheological properties of the interface between the two fluids and, in turn, facilitate the formation of an emulsion. This work is carried out in collaboration with Dr. Libero Liggieri at CNR-ICMATE. Each part consists of several chapters, which address various projects within these two topics and are briefly summarised below. PART I. PERFORMANCE ASSESSMENT OF VITRECTOMY DEVICES An important issue associated with vitrectomy is the risk of generation of retinal tractions by the vitreous cutter. It has been demonstrated that an intermittent flow generates fluid accelerations, with consequent pressure variations within the vitreous chamber, which may result in pulsatile tractions on the retina, with significant risk of iatrogenic retinal damage. Rossi et al. (2014b) proposed a conceptual diagram according to which, for a safe and efficient vitrectomy, optimisation of fluidics can be achieved by maximising the flow rate and minimising the acceleration around the cutter port. Another important aspect of the surgical procedure is prevention of large IOP changes within the eye. In particular, ocular hypertension or ocular hypotony can have very important clinical consequences. In order to maintain a normal IOP value, the balance between the aspiration and the irrigation pressures during vitrectomy plays a very important role. In chapter 1 we reviewed the literature concerning fluid dynamics of vitreoretinal surgery. In chapter 2 we evaluated the performance of vitreous cutters in vitro in terms of flow rate, using vitreous phantoms with different rheological properties (artificial vitreous, AV). On each fluid solution we performed rheological tests in order to obtain a good control of the properties of the working fluid and thus guarantee reproducibility of the experiments. This allowed us to compare the efficiency of the instruments in a highly controlled environment and to assess how their behavior is affected by fluid properties. Our results showed that all cutters are higly sensitive to fluid rheology and this is particularly true for double blade cutters, which, however, always produce larger flow rates compared to single blade ones. These findings imply that large flow rate fluctuations should be expected when a vitreous cutter is operated in a medium with heterogeneous mechanical properties (Nepita et al., 2021). In chapters 3 and 4 we show two experimental studies aimed at assessing the performance of two vitrectomy systems, characterised by different technologies (guillotine and ultrasound probes), tested both with balanced salt solution (BSS) and AVs. The following types of experiments have been performed in a cubic testing chamber. - Flow rate measurements: mean flow rate has been measured using a digital image tracking process for all cutters, changing the operational parameters within a wide range (cutting frequency, aspiration pressure, etc.). Results for guillotine cutters show that for viscoelastic fluids the averaged flow rate only slightly decreases as the cutting frequency increases. The ultrasound cutters produce larger flows regardless the stroke (the amplitude of the ultrasound vibrations of the needle). - Flow fields measurements: two dimensional PIV experiments have been performed with the goal of characterising the flow fields generated by all cutters. For each set of parameters, a sequence of images has been acquired on two different planes: “frontal view” and “lateral view”. Moreover, for a specific solution, two areas of interest have been used for the acquisition: a “large field” of view, aimed at describing the entire circulation induced by the cutters at the scale of the experimental box, and a “small field” of view, i.e. a zoom around the cutter tip, to describe more accurately flow structures around the cutter port. The flow induced by ultrasound cutters is strongly three dimensional and time dependent. Flow structures are generated at different spatial scales, from large scales similar to the experimental domain to smaller scales close to the cutter port (Stocchino et al., 2020). - Temperature measurements: a specific task was dedicated to measure temperature variations induced by ultrasound cutters as a function of the stroke, maintaining the vacuum pressure equal to zero. The two dimensional variation of the temperature field has been monitored with an infrared thermal camera. The acquired thermal images were used to obtain the temperature variations in time at a particular spot, or to generate two dimensional temperature maps. The temperature slightly grows but this effect is likely to be clinically irrelevant, since infusion of a different fluid will mitigate it. Moreover, the cutter head reaches maximum values close to body temperature, meaning that it will not transmit much heat to the needle and eventually to the vitreous during surgery (Stocchino et al., 2020). - Pressure measurements: for this specific type of experiments a model of the human vitreous chamber at real scale and with a realistic geometry has been employed. The eye model has been designed considering both infusion and aspirations lines, thus mimicking the real surgical procedure. The main goal of this part of the research activity is to perform pressure measurements at two different positions of the vitreous chamber (anterior and posterior): for this purpose two high precision pressure transducers (Viatran model 422) have been used. Results show that dynamic variations of the pressure due to the motion of the cutter blade are always small compared to the average pressure in the eye. Pressure fluctuations related to fluid motion are felt more intensely by the pressure sensor placed on the back of the vitreous chamber, opposite with respect to the irrigation line. This is probably due to the generation of a jet that impinges the chamber wall opposite to its position. When the aspiration phase begins the pressure in the eye model drops, which is related to hydraulic head (pressure) losses along the pipelines of the system. This pressure drop is typically reduced (sometimes very efficiently) when the vitrectomy system can actively compensate pressure variations. Results suggest that maximum irrigation pressure that can be used for compensating pressure losses during aspiration should be selected depending on several parameters such as the minimum irrigation pressure and the aspiration pressure, in order to optimise the desired compensation. PART II. STUDY OF SILICONE OIL EMULSIFICATION IN RELATION TO VITRECTOMY The tendency of SOs to emulsify in water is influenced by the chemico-physical properties of the fluids and interface and by external factors, such as temperature and mechanical energy provided to the system. Exposition of the SO to endogenous molecules (proteins, lipids, etc.) in the vitreous chamber, the presence of which is favored by the post-surgical inflammatory state of ocular tissues, is also considered to be important for the generation of an emulsion. The mechanisms at the basis of the effects of these molecules on the emulsification process depend on their capability to adsorb at the water-oil interface, modifying the mechanical interfacial properties, such as the interfacial tension (IT) and the dilational viscoelasticity (DV), i.e interfacial rheological response to surface variations. In chapters 5 and 6 we give an overview of the possible mechanisms for SO emulsion formation and stabilisation, with a particular focus on biomolecules adsorption processes at the SO-aqueous interface. In chapter 7 we propose an experimental study aimed at achieving a chemico-physical characterisation of the interfacial properties between SOs and aqueous solutions and at investigating the effect of surfactant biomolecules, such as albumin and γ-globulins, possibly responsible for emulsion formation. In particular, measurements of IT on pendant drops of the protein aqueous solution in the SO have been performed. These measurements have been carried out by a Drop Shape Tensiometer (PAT1-Sinterface, Berlin). This technique allows one to quantify the DV by measuring the IT response to small sinusoidal perturbations of frequency imposed on the interfacial area. Results show that adsorption of these proteins at the interface reduces the IT to values compatible with the generation of an emulsion and provides values of the DV compatible with a good stability of the resulting emulsions (Nepita et al., 2020). This study allowed us to understand the effect of individual, key blood proteins on the mechanical properties of the interface. In order to evaluate the overall effect of the presence of these proteins, we conducted a second series of experiments adopting whole human serum blood (WHSB) samples obtained from donors. The results obtained with the serum dissolved in the solution showed that, already in concentrations of 1-2 orders of magnitude smaller than that of single proteins, the IT significantly decreases, up to values compatible with the formation of an emulsion, indicating therefore possible synergistic effects with other components of the serum, in addition to the proteins previously studied. Moreover, the DV modulus increases up to rather high values, compatible with the stabilisation of the emulsions against the coalescence. Drugs used as intraoperative tools in vitrectomy can also act as surfactants. An example is the intraocular injection of an anti-inflammatory steroid, composed of micro crystal particles. To assess whether the drug has a surfactant role at the interface, we conducted a first series of experiments to quantify the IT and the DV properties. We performed the measurements on the aqueous phase of the dispersion (supernatant), which separates after the sedimentation of crystals. The experiments with the supernatant neither show a significant decrease of IT that can be associated with the formation of an emulsion, nor lead to significant DV values. However, it is known that dispersions of partially hydrophobic solid particles can give rise to particularly stable emulsions, known as Pickering emulsions (Binks, 2002). In order to verify the capability of biomolecules to form emulsions stable over the timescale of months, comparative emulsification tests have been performed and shown in chapter 8. Emulsions were produced using to the Double Syringe method (Gaillard et al., 2017). Although this emulsification method may be very different from what occurs in practice, it is significant for investigating the stability of the emulsion formed and is, therefore, used as a simple screening test. For all samples containing blood proteins or the steroid drug, the presence of an emulsion layer was detected, the thickness of which decreases with increasing dilution, and which was stable over long times. On the contrary, the sample containing the Dulbecco buffer alone was already separated a few minutes after the emulsion formation. Thus, we can conclude that blood proteins are able to form and stabilise emulsions. Moreover, the steroid drug is able to stabilise Pickering-type emulsions. Eye movements are the main source of mechanical energy applied to the interface between SO and aqueous solution inside the eye. A set of measurements in a realistic domain have been performed. A qualitative and a quantitative analysis of the risults are reported in chapter 9 and 10, respectively. In particular we have used a Plexiglass model of a vitreous chamber with realistic geometry. The experimental setup is equipped with an optical diagnostics for monitoring the emulsion, a mechanical system controlled by a computer to simulate eye saccadic movements and thermalised at 35±C. Once filled with SO and aqueous solution, the model is mounted on a support connected to a computer-controlled motor. In this study, harmonic rotations were considered. The temporal evolution of the emulsion has been studied through acquisition of images with a digital camera. In the case of emulsions characterised by very small droplets, a microscope (Leica Microsystems, Switzerland) has been used to determine their distribution and size. The main result obtained from these experiments consists in the formation and persistence of emulsions even when a low concentration of albumin has been dissolved in the aqueous solution (1% of the physiological concentration in blood). Moreover, the majority of the drops had a very small diameter: in the range of 0 – 10 μm. Droplets with these characteristic dimensions are considered more dangerous from the clinical point of view, since they are able to favor the onset of ocular complications.