Comparing the Efficiency of Fluid Infusion Systems for Pars Plana Vitrectomy

Abstract

Purpose: To compare the efficiency of infusion sets including distal tubing, infusion, and trocar cannulas (23G, 25G, 27G) of leading brands (Alcon, Bausch & Lomb, BVI, DORC, Optikon) in minimizing the intraocular pressure drop on aspiration. The study reports the pressure drop along infusion as a function of the flow rate and correlates performance to design.

Methods: We measured the pressure drop of the distal silicone tubing, infusion cannula and trocar cannula, connected to a model eye and the corresponding flow rate. The piezometric height of the balanced salt solution (BSS) reservoir open to atmospheric pressure ranged from 10 to 120 cm above outlet. All infusion set components were measured in length and inner lumen diameter.

Results: Pressure drop as a function of flow rate varied significantly: BVI and DORC proved more efficient at all gauges. Pressure drop at 10 mL/min flowrate varied between 9–16 mm Hg for 23G, 11–25 mm Hg for 25G, and 16–50 mm Hg for 27G. The tubing friction and kinetic energy loss, respectively responsible for the linear and quadratic component of the head-loss to flow rate parabolic function, also differed significantly.

Conclusions: Pressure drop secondary to flow rate varied by a factor of two among manufacturers. Excessive pressure drop during aspiration may lead to dangerous hypotony or force the surgeon to set anomalously high pressures at rest to avoid it. The present study gives useful insights to help improve infusion system performance.

Translational Relevance: A thorough understanding of head loss mechanisms along infusion sets allows the design of more efficient and safer infusion sets.

Introduction

Pars plana vitrectomy (PPV) is an effective surgical procedure for several vitreoretinal diseases, including retinal detachment, proliferative diabetic retinopathy, epiretinal membranes, macular holes, and vitreous hemorrhage. 

Since its introduction by Robert Machemer¹ in the late 1970s, technical improvements such as dual-action cutting,² high cut rates,³ valved trocar cannulas,⁴ and the introduction of smaller gauges,5 have widened indications and improved outcomes. 

Instrument miniaturization from 18G of the original VISC⁶ to 25G and 27G⁷ that represent today's standard have enhanced accuracy and reduced surgical invasiveness while increasing technical complexity, and limiting aspiration and infusion fluidics. 

Aspiration flowrate more than doubled at the higher cut rates with the introduction of dual action cutting, which allows invariant duty cycle, thus mitigating the negative effect of gauge reduction.⁸ On the infusion side, infusion cannulas and valved trocar cannulas have not always improved comparably to keep up with increasingly efficient cutter probes and minimize intraocular drop of pressure (i.e., head loss) as flowrate increases⁹ during vitreous base shaving or fluid exchanges. 

The purpose of this article is to compare the efficiency of different infusion systems, measuring the drop of pressure as a function of the flow rate and correlating the fluidic performance to their respective design. 

Material and Methods

Infusion Cannula and Trocar Cannula Design

The infusion sets, complete with infusion cannula, trocar cannulas and tubing of leading manufacturers (Alcon [Geneva, Switzerland], Bausch & Lomb [Rochester, NY, USA], BVI Medical [Waltham, MA, USA], DORC International [Zuidland, The Netherlands], and Optikon [Rome, Italy]) have been evaluated. For each manufacturer three gauges were tested: 23G, 25G, and 27g. 

For the purpose of present study, we defined as “infusion terminal” the part of the infusion line starting from the Luer lock connector at the proximal upstream end of the silicone rubber tubing, possibly equipped with a hollow metal cylinder (“the infusion cannula”) at its distal end to be inserted into the “trocar cannula” (i.e., the hollow metal cylinder to be inserted into the sclera with a proprietary blade [“the trocar”] that is then removed). It should be noted that BVI (Fig. 1c) and DORC (Fig. 1d) have developed a different type of infusion that does not have a metal cannula and connects directly the tubing and trocar cannula in that DORC has a proprietary docking system and BVI simply left open the silicone tubing terminal to be placed around the trocar cannula, relying on it elasticity to avoid inner lumen reduction (Fig. 1c). We also define as the “infusion set” the entire infusion terminal including tubing, infusion cannula (if present), and connection to the trocar cannula. 

Figure 1.

 

Tested infusion sets: (a) Alcon, (b) Bausch & Lomb, (c) BVI, (d) DORC, and (e) Optikon. For each pane, from left to right are shown the trocar cannula, the infusion terminal, and the trocar blade. Figure 1a also reports the terminology used throughout the article: (1) trocar cannula, (2) infusion cannula, and (3) trocar blade.

Experimental Set-Up

The setting used during the measurements is sketched in Figure 2. All infusion systems were connected and fed by the same irrigating line consisting of a 1.96 m long tubing (6.0 mm in diameter) connected to a reservoir (1) with its free surface kept at a known, constant height, and exposed directly to the atmospheric pressure. A three-way stopcock (3) at the end of the irrigating line allowed parallel connection of the tested infusion sets (2) and one side of a differential pressure probe (4) (PD-33X; Keller Pressure, Winterthur, Switzerland). The trocar cannula was inserted into the rubber membrane of a model vitreous chamber using the same procedure as during surgery. 

Figure 2.

 

Sketch of the setup used during the experiments: (1) open bottle; (2) tubing (internal diameter 6 mm; external diameter of 8 mm); (3) three-way stopcock; (4) differential manometer; (5) infusion terminal; (6) model vitreous chamber; (7) tubing (inner diameter of 8 mm; outer diameter of 10 mm); (8) flowmeter; (9) constant head reservoir.

The model vitreous chamber (6) consisted of a hemisphere cavity, carved in a Plexiglas cylindrical block, closed at the top by a 1 mm-thick rubber membrane allowing the trocar insertion and a stainless-steel ring with screws securing water-tightness. The internal volume of the model vitreous chamber was 4ml and the distance between the rubber membrane and the opposite wall was 18 mm, thus mimicking the human vitreous chamber. Special care was taken to keep the trocar orthogonal to the membrane to reproduce the surgical setting. 

The second side of the differential pressure probe (4) was connected to the model vitreous chamber (6) to measure the overall pressure loss along the infusion set (in mm Hg) from the Luer lock connector (3) to the chamber (i.e., along the whole infusion set). A further tubing (7) connected to the chamber allowed the fluid outlet. Along that tubing an ultrasonic flow meter (8) (Sonoflow IL. 52 V2.0 , full-scale of 100 mL/min; Sonotec, Halle, Germany) measured continuously the infusion rate, Q, in milliliters per minute. The outlet tubing was conveyed within a water-filled vessel (9) large enough to ensure a negligible change in fluid level during the test run. The outlet (7) was oriented upward to avoid trapping air bubbles. 

During the experiments, the flowrate variation was obtained by changing the height of the feeding reservoir level between 0.10 m and 1.50 m above the water level of the discharge vessel (indicated with z in Fig. 1) in 18 identical steps, and the infusion was carried out by gravity. Pressure drop and flowrate were acquired synchronously by a National Instrument Acquisition Board (NI USB 6211). 

Experimental Data Acquisition and Statistical Analysis

Two different infusion sets were tested for each manufacturer, and the procedure was repeated twice for each of them. During each experiment, the volumetric flow rate, Q, and the pressure drop through the infusion set were recorded for 15 seconds at a 100 Hz sampling rate, thus collecting 1500 samples. Mean and standard deviation were calculated. Before each experiment was run, the pressure drop at zero flow was measured with the same procedure, and its average value was subtracted from the successive measures. Thus the pressure drops, Δp, presented hereafter are derived from the hydrostatic gradient and reflect only the energy dissipation from the beginning of the infusion set to the vitreous chamber. The Δp(Q) represents the intraocular pressure (IOP) drop that a surgeon experiences when aspiration starts at a given flow rate, Q, with a vitreous cutter. 

To physically interpret the results, data were approximated by a second-order polynomial defined by the equation:  

\begin{eqnarray} \Delta p\left( Q \right) = a\,Q + b\,{{Q}^2}.\quad \end{eqnarray}

(1)

The best fit for the measured dataset was obtained using the least squares method. The values of the coefficients, R² and root mean square error (RMSE) are reported in the Table. As we previously reported,¹⁰ in the above equation, the linear contribution (Δp₁ = a Q) should be ascribed mainly to the friction of the laminar flow along the tubing of the infusion system, according to the Hagen-Poiseuille law, whereas the quadratic term (Δp₂ = b Q²) should be related mainly to the energy dissipation occurring at the sudden geometrical changes of the lumen and at the distal end of the infusion set because, after balanced salt solution (BSS) exits from the infusion set, it moves erratically within the vitreous chamber until it dissipates all its kinetic energy, thus determining an outlet pressure loss proportional to the squared flowrate. 

Table.

 

Best-Fitting Analysis: a and b Represent the Coefficients of the Generic Δp(Q) = aQ + bQ²

 

The Analysis of Variance (ANOVA), with t-test and Bonferroni correction when applicable, were applied to numerical variables. The Shapiro-Wilk test was used for the Gaussian distribution of data. In all cases P values <0.05 were considered statistically significant. 

Results

Best fit curves resulting from the experimental data, showing the loss of pressure, Δp, as a function of the flowrate, Q, for all the tested infusion sets, are reported in panel a of Figures 3–5 for 23, 25, and 27G systems, respectively. All curves showed R² values >0.993 and RMSE < 3 mm Hg, demonstrating excellent adherence to experimental data (Table). The linear (Δp₁) and quadratic (Δp₂) contributions are reported in panels b) and c) of the same figures for each caliper. 

Figure 3.

 

(a) Best fits (Equation 1) obtained from pressure drops, Δp, generated by 23G infusion sets as a function of flowrate. (b) Linear contribution, Δp₁, representing friction along the tubing and (c) quadratic contribution, Δp₂, representing the dissipation of kinetic energy at the outlet of the infusion set in the vitreous chamber. The red, dashed line at 10 mL/min flow rate indicates the typical flowrate of the surgical setting.

Figure 4.

 

(a) Best fits (Equation 1) obtained from pressure drops generated by 25G infusion sets as a function of flowrate. (b) Linear contribution, Δp₁, representing friction along the tubing and (c) quadratic contribution, Δp₂, representing the dissipation of kinetic energy at the outlet of the infusion set in the vitreous chamber. The red, dashed line at 10 mL/min flow rate indicates the typical flowrate of the surgical setting.

Figure 5.

 

(a) Best fits (Equation 1) obtained from pressure drops generated by 27G infusion sets as a function of flowrate. (b) Linear contribution, Δp₁, representing friction along the tubing and (c) quadratic contribution, Δp₂, representing the dissipation of kinetic energy at the outlet of the infusion set in the vitreous chamber. The red, dashed line at 10 mL/min flow rate indicates the typical flowrate of the surgical setting.

The DORC infusion system yielded the lowest pressure drop, Δp, regardless of the gauge and flowrate. The BVI infusion system was a close second at 23G (Fig. 3a) and completely overlapped the DORC response at 25G (Fig. 4a) and 27G (Fig. 5a), while other sets exhibited a much higher pressure drop with increasing differences as caliper is reduced (Figs. 3a, 4a, and 5a). 

To summarize the behavior of the different calipers and sets, we present in Figure 6 the total pressure drop, Δp (Fig. 6a), and its linear (Δp₁, Fig. 6b), and quadratic (Δp₂, Fig. 6c) components at the surgically relevant flow rate of 10 mL/min. As expected, Δp increases significantly as the gauge reduces: all 23G systems showed Δp lower than 40 mm Hg, whereas three of the 27G exceeded 60 mm Hg (Fig. 6). At the same time, the difference between the worst and best-in-class rises. When comparing 27G to 23G at the 10 mL/min flowrate, Alcon exhibits the largest pressure drop ratio, with an approximately quintupled Δp, whereas the lowest ratio is shown by BVI and Optikon with a nearly tripled Δp. Decreasing the caliper from 23G to 25G, the pressure drop increases approximately by one-and-a-half (BVI, Optikon) to two times (Alcon, DORC). 

Figure 6.

 

Pressure drops at 10ml/min flow rate for all tested gauges. (a) Total pressure drop, Δp; (b) Linear contribution, Δp₁; (c) Quadratic contribution Δp₂.

The contribution of linear and quadratic components to the overall pressure drop is further compared in Figure 7, showing the ratio of the quadratic to linear pressure drop, Δp₂/Δp₁ at the reference flow rate of 10 mL/min for all the tested infusion systems. In all cases the quadratic component was lower than the linear component (Δp₂/Δp₁<1) except for the Alcon system at 25G and 27G. 

Figure 7.

 

Quadratic to linear contribution ratio Δp₂/Δp₁ at 10 mL/min flow rate. Note that the quadratic contribution (Δp₂) prevails on the linear contribution (Δp₁) for both Alcon 25G and 27G.

All the infusion systems we compared include silicone tubing, possibly equipped with an infusion cannula, and a trocar cannula. Although the silicone rubber tubing varies only slightly in diameter and length, the combination of trocar and infusion cannulas radically changes in concept between brands as seen in Figure 1. It is therefore of interest to analyze the role of the different infusion cannula/trocar cannula combination designs on the infusion set performance. 

However, measuring the pressure drop along that part of an infusion set is practically unfeasible. Therefore, to elucidate how its design affects the performances, we estimated the theoretical pressure drops resulting by the friction along the tubing walls and the pressure drop at the outlet. 

The former was evaluated, by the well-known Poiseuille-Hagen law, assuming a fluid with viscosity µ = 1.0 Pa ∙ s flowing in a laminar direction within a cylinder of diameter D and length L, at a Q = 10 ml/min rate:  

\begin{eqnarray}{\rm{\Delta }}p = 128\frac{{\mu L}}{{\pi {{D}^4}}}Q\# \end{eqnarray}

(2)

and the latter was estimated by assuming that the fluid dissipates all its kinetic energy after entering the vitreous chamber by erratically moving within the chamber until rest. Consequently, the final pressure drop equals the kinetic energy of the flow at the outlet, which reads:  

\begin{eqnarray}{\rm{\Delta }}p = \frac{{8\rho }}{{{{\pi }^2}{{D}^4}}}{{Q}^2}\# \end{eqnarray}

(3)

where ρ = 998 Kg/m³ is the BSS fluid density. Although theoretically consistent, the above estimation suffers from possible dimensional measuring errors and does not include fundamental effects such as the energy dissipation at the sudden lumen variations along the infusion set, which substantially contribute to the pressure drop attained by the real sets. Figure 8 displays the calculated pressure drop in mm Hg for the same cases as above. As expected, the pressure loss increases significantly as the caliper decreases. At 23G the differences are quite small but increase significantly at 27G, with the highest value given by Bausch & Lomb. Irrespective of the caliper, the two settings without cannula give the best results. 

Figure 8.

 

Calculated contribution of the cannula, trocar and outlet at 10 mL/min flow rate. (a) 23G, (b) 25G, (c) 27G. Note the advantage of BVI and DORC at all calipers and the increasing role of the cannula on the overall pressure drop as the caliper reduces.

Figure 9 shows the percentage contribution of trocar cannula (green bar), infusion cannula (cyan bar, when present), and outlet (red bar) to the pressure drop calculated by the above equations at 10 mL/min flowrate for the three considered calipers. When present, the infusion cannula plays a predominant role because of the high sensitivity (the fourth power) to the diameter of this section of the tubing, which is the smallest of the entire infusion set. This predominance increases at increasing gauges. The contributions of the trocar cannula and the outlet are mostly of the same order but for Alcon, where the outlet prevails on the trocar because of the small length affected by the flow (see Fig. 1). 

Figure 9.

 

Calculated relative contribution of the infusion cannula, trocar cannula, and outlet at 10 mL/min flow rate. (a) 23G, (b) 25G, (c) 27G. Note the predominant role of the infusion cannula as the gauge reduces.

Discussion

Ensuring optimum fluid infusion during PPV^11,12 is of high importance because pressure and volume stability are key to maintaining ocular homeostasis and preventing potentially blinding complications including optic nerve damage,¹³ choroidal hemorrhage,¹⁴ and ciliary effusion syndrome.¹⁵ 

The introduction of double action blades, allowing invariant flow irrespective of the duty cycle,¹⁶ improved the efficiency of vitrectomy probes, especially at higher cut-rates because of the increased aspiration flow rates achievable. This makes it more difficult for infusion systems to match aspiration flow rates and maintain IOP stability. Higher flow rates, in fact, increase pressure loss along the infusion line, which, in turn, translates into an equal IOP decrease when the cutter aspiration starts,¹⁰ if not otherwise compensated. 

The infusion sets we tested, showed significant variability with more than twice the difference in pressure drop, at a given flow rate (Figs. 3a, 4a, 5a). Although we only tested the distal end of the infusion system for each manufacturer (the infusion set), and different consoles may compensate for pressure drop with different strategies, the terminal parts have the highest impact on overall performance. 

The reduction in IOP because of flow (i.e., the pressure drop, Δp, we measured) could be theoretically compensated for by simply increasing the infusion pressure, but this would instantaneously translate into intraocular pressure as aspiration stops, and flow rate drops to zero. Since we measured that 10 mL/min flow may cause a pressure drop exceeding 20 and 30 mm Hg at 25g and 27g, respectively (i.e., the most widely used gauges), an effective IOP compensation would require a potentially dangerously high intraocular pressure “rebound” when aspiration stops. 

The surgical relevance of the pressure drop difference found across brands is further shown in Figure 10, displaying the potential effect of adjusting the working flowrate as a function of the probe gauge. The flowrate obtained during surgery greatly varies as a function of vitreous gel hydraulic resistance,¹⁷ which is related to age, degree of syneresis, and many other factors¹⁸ including the amount of infused BSS¹⁹ as surgery progresses.²⁰ Especially during vitreous base shaving phase, when most core vitreous gel has been removed and replaced by BSS, flowrates increase, reaching the values we took (e.g., 15, 10, and 5 mL/min, respectively) for 23G, 25G, and 27G probes. 

Figure 10.

 

Pressure drops at varying “surgically relevant” flow rates: 15 mL/min @ 23G, 10 mL/min @ 25G, and 5 mL/min @ 27G.

In such cases, the corresponding pressure drop (Δp), varies from 12 to 55 mm Hg, depending on manufacturer and gauge (Fig. 10). Then, if the surgeon sets 25 mm Hg as a “safe” working IOP, he/she will need to add the flow-induced reduction in IOP (Δp; i.e., the head loss) to that value and therefore set the infusion pressure between 37 and 80 mm Hg (25 + 12 = 37 or 25 + 55 = 80 mm Hg), accepting the unavoidable drawback that, when aspiration stops, IOP will instantaneously build up to potentially harmful values. 

In this scenario, the IOP would keep bouncing between dangerously high and low values as the surgeon depresses or releases the cutter foot pedal, because aspiration generated flow, and the magnitude of such variability would match exactly head loss values displayed in Figure 10. It is therefore apparent that, regardless of the compensation strategy, there are surgical safety benefits from the lowest possible pressure drops. 

The two best-in-class infusion sets (DORC and BVI) increase their advantage as gauge reduces (compare Fig. 5a with Fig. 3a), and fluid friction along the infusion cannula and trocar cannula (the linear contribution) is more important than energy losses at the sudden lumen variations and distal outlet (the quadratic contribution), although the proportion changes as gauge reduces (compare panels b to c of Figs. 3–5). This observation has surgical significance and explains why design efficiency is important and becomes even more so at small gauges. 

Infusion set design and specifically the infusion cannula to trocar cannula docking, in fact, are the main explanations for the pressure drop differences, since BVI and DORC have developed infusion cannula - trocar cannula joints that avoid inner lumen reduction by removing the restriction of the infusion cannula inserted into the trocar cannula. All other manufacturers have retained the usual solution of an infusion cannula within the trocar lumen (tube in the tube) as the simplest way to lock them, thus introducing a lumen decrease to the detriment of infusion flow. The difference between the two brands excluding the cannula (BVI and DORC) and the other competitors (Alcon, B&L and Optikon) increases at 25G and 27G compared to 23G (see Figs. 4 and 5 compared to 3). 

We further investigated the role of the design of the distal part of the infusion set, consisting of the infusion cannula and trocar cannula, calculating the ideal contribution of the infusion cannula (when present) and trocar cannula to the pressure drop (Δp) by Equation 2, and the sudden variations in flow occurring along the tubing and at the outlet by Equation 3. Results are shown and compared in Figures 8 and 9. 

The infusion cannula is responsible for most of the pressure drop, as gauge reduces, and the advantage of DORC and BVI at 25G and 27G can be almost completely ascribed to the lumen increment because of its removal (compare the blue blocks in Figures 8 and 9). Because the outlet pressure drop is related to final lumen of the infusion set, the infusion cannula, which is the narrowest lumen of the entire set, is also involved in the outlet pressure drop if its distal end is very close to the distal end of the trocar cannula. This seems to be the case with the Optikon and, to a lesser extent, Alcon sets. 

It should be noted that other factors may have instructed the difference in design among brands, such as the decision to use the same infusion cannula for viscous fluid injection at higher pressure (i.e., silicone oil) and the age of the respective consoles, because some infusion sets have been released or updated more recently, also according to marketing strategies. 

In summary, we found significant differences across different manufacturers: Our results provided the experimental evidence that the infusion set and particularly its distal portion are responsible for significant IOP fluctuations occurring every time the surgeon starts and stops aspiration. Most vitrectomy consoles adopt forced infusion systems to limit pressure fluctuations, for example by controlling the gas pressure in the BSS bottle or using mechanical actuators. However, solving the problem at the source is preferable. 

From the fluidics point of view, the metal infusion cannula inserted into the trocar cannula is counterproductive, especially for small gauges and large flows (Figs. 4–6) and, when present, is responsible for the main contribution to infusion flow loss (Fig. 8). Maximization of the inner lumen diameter at the outlet is also a major factor. Therefore the infusion cannula, if present, should be designed to terminate upstream of the distal end of the trocar. In this way the outlet lumen is the maximum allowed by the chosen gauge. Surgeons should be aware of the pressure drop imposed by their instruments, because the infusion system may not be able to compensate completely, especially during transitions between aspiration and non-aspiration phases of surgery. 

Acknowledgments

The authors thank the “Fondazione Roma” and the Italian Ministry of Health for financial support. The authors also thank Donatella Petti and Manlio Valentini for technical assistance. 

Disclosure: T. Rossi, None; G. Querzoli, None; G.B. Angelini, None; V. Santoro, None; C. Pellizzaro, None; D.H. Steel, None; M. Parravano, None; M.R. Romano, None