Pars plana vitrectomy (PPV) with gas tamponade is commonly performed in the management of rhegmatogenous retinal detachment.1 Gas was first utilized by Ohm in 1911, and since then, our knowledge of gas usage has mainly been based on experimental animal model data.2–5 The type, concentration, and volume of gas injected are all important when considering efficacy, duration of tamponade, and the potential side effect profile.
Different gases are available (sulphahexafluoride [SF6], hexafluoroethane [C2F6], or perfluoropropane [C3F8]), variable on location, and they are usually injected at a “nonexpansile” concentration, to reduce the likelihood of a potentially sight-threatening intraocular pressure (IOP) rise. Postoperatively the gas enters a brief expansile phase where equalization of the partial pressures of nitrogen, oxygen, and carbon dioxide with the blood gas partial pressures occurs. After reaching maximal volume, gas absorption then occurs over 2.5 to 9 weeks into the blood by partial pressure forces, depending on the type of gas used.6,7
In vivo modelling of gas dynamics is challenging due to problems with accuracy of volume measurement and, therefore, mathematical models have been developed to further our understanding. Hutter et al. produced a model assessing pressure and fluid flow dynamics of intraocular gases in pneumatic retinopexy which was further modified for use in vitrectomized eyes.8,9 Recently, we have developed a mathematical model to determine the concentrations of different gas tamponades in air to achieve 100% fill of the vitreous cavity postoperatively.10 This model demonstrated that variation in axial length and the size of the eye does not appear to significantly alter the gas concentration required.10
Another influencing factor on gas behavior and total fill is the aqueous outflow ability of the eye. This is important because gas requires aqueous to leave the eye to allow it to fully expand and form a complete tamponade. It has been shown that eyes with ocular hypertension have a reduced trabecular meshwork outflow and uveoscleral outflow rate, and therefore it is possible that such patients could have a reduced maximal gas bubble size. They therefore run the risk of being under-filled at the time of surgery with the inherent risk of reduced tamponade time.11 This has the potential to adversely affect surgical outcome in certain situations. Furthermore, drugs such as apraclonidine (Iopidine; Novartis, Basel, Switzerland), which alter aqueous dynamics, theoretically may influence gas dynamics. We have further improved our mathematical model to control for aqueous fluid dynamics and quantify its effect on required gas concentrations to achieve 100% fill. In particular, we were interested in assessing the effect of ocular hypertension and apraclonidine treatment on gas dynamics.
Materials and Methods
The initial gas mathematical model was based on that of Hall et al., composed of five equations and previously described thoroughly.9,10 In that paper, various configurations of the eye were tested, including those involving a vitreous cavity with a concave anterior indent from the crystalline lens. The mathematical model of a spherical eye was found to be the most accurate, with no relevant difference if lens shape was included. Therefore, this was used for all subsequent work, partially filled with a liquid, air, and gas mixture. The first four equations describe the mass transfer of the tamponade gas, oxygen, nitrogen, and carbon dioxide across the retina. The final equation describes fluid balance, and we have extended this to include aqueous outflow from the uveoscleral pathway in an attempt to achieve a more accurate assessment of the ocular fluid dynamics. From our previous investigations, an 80% to 85% fill has been calculated to represent the best fit to the animal model.9 This percentage takes into account anterior chamber volume, lens volume, residual vitreous, and residual cavity fluid volume. Alberti et al. have previously found the median percentage fill during gas tamponade to be 78%.12
The model allows variation in different inputs, such as percentage operative gas fill, percentage gas concentration, aqueous inflow, aqueous outflow via the trabecular meshwork, and uveoscleral outflow. Variation in these parameters allows estimation of duration and maximum expansion of the gas bubble, as well as IOP over time. The following aqueous flow, outflow facility, and uveoscleral outflow measurements were used, based on the best available evidence in the literature11,13 (Table 1).
Rates of Aqueous Production and Outflow in Different Clinical Scenarios
In order to validate our model, we used data from a human study which measured change of gas volume (SF6 and C3F8) with time in vitrectomized eyes undergoing gas tamponade.14 This study in 1988 was small and measured gas bubble volume at various postoperative timepoints using A-scan ultrasound. The concentration and volume of gas injected varied depending on the authors opinion of what would be required clinically.
After validation, we used the model to calculate the percentages of gas (SF6, C2F6, and C3F8) in air required to achieve a 100% fill of the vitreous cavity post-operatively for an eye of 7.2 mL volume in different states of aqueous dynamics. We calculated the change in gas volume with time and estimated the intraocular pressure in normal eyes, OHT eyes and OHT eyes receiving apraclonidine. We also calculated the fill ratio of the vitreous cavity achieved when injecting gases at their standard isovolumetric concentrations, to give an assessment of true clinical practice with our updated model.
The study by Jacobs et al. measured change in 20% SF6 volume over time in seven eyes, and C3F8 (varying concentrations) in five eyes.14 Our model showed good concordance for both gases, at a range of concentrations and injected gas volumes (Figure 1).
Model predictions after injection of 20% SF6 or C3F8 following vitrectomy. Change in gas volume with time after injection of 20% SF6 or C3F8 (at different concentrations) following vitrectomy. For each of the eyes considered, measurements reported by Jacobs et al. are shown as a symbol, whereas the prediction made using our model is shown as a continuous line of matching grayscale.14 An average eye volume of 7.2 mL was assumed, unless the maximum volume reported in the original paper was larger, in which case it was used instead. The initial volume being unknown, it was set to 90% of the first measurement to allow for gas expansion.
A table of different gas concentration was generated using our model to provide values for achieving maximal fill of the vitreous cavity in scenarios of different aqueous outflow. In order to simulate true clinical fill volume, in which 80% to 85% fill of the cavity with gas is achievable, we found the concentration of SF6 required for a postoperative full gas fill after the expansile phase was 17% to 23% for a normal eye, increasing to 22% to 28% in OHT, and returning to 19% to 25% in an OHT eye receiving apraclonidine (Table 2).
Percentage Fill of the Vitreous Cavity at Time of Surgery to Achieve Maximal Fill in Different Aqueous Outflow States
We calculated the gas concentration required of C2F6 and C3F8 to give a full postoperative gas fill in normal, OHT and apraclonidine treated eyes. For C2F6, 11% to 14% was required in a normal eye, 14% to 17% in OHT, and 12% to 15% in an OHT eye with apraclonidine. The results for C3F8 were 9% to 12% in a normal eye, 11% to 14% in OHT, and 10%to 13% in OHT with apraclonidine. The required gas volumes for all three gases in different aqueous flow states are shown in Figure 2.
The effects of ocular hypertension (OHT) and apraclonidine treatment on concentration of gas required to achieve 100% fill postoperatively. The gas concentration required varies depending on aqueous dynamics. Eyes with normal aqueous outflow require lower gas concentrations than those with reduced outflow states, such as OHT, to achieve maximal fill.
Based on the assumption that 22.6% SF6 provides 80% vitreous cavity fill at operation and a postexpansile full gas fill, we calculated the change in volume and IOP over time at this concentration. As expected, an initial rise in volume occurred followed by a gradual reduction during the absorption phase. Eyes with OHT and those with OHT receiving apraclonidine had an increased rate of gas volume loss compared to eyes with normal aqueous dynamics. In the immediate post-op period, the IOP reduced rapidly before becoming reaching a relatively constant plateau level in all eyes. Apraclonidine was an important adjunctive treatment to rapidly reduce the IOP postoperatively and maintained it at a safer lower level in the OHT model over the gas absorption period (Figure 3).
The effect of ocular hypertension (OHT) and apraclonidine treatment on postoperative volume reduction and intraocular pressure (IOP). The top graphs demonstrate the increased gas volume during the expansile phase, followed by a gradual reduction over time for 22.6% SF6, 13.9% C2F6, and 11.6% C3F8. Eyes with OHT do not achieve the same maximal fill as normal eyes, although the difference is reduced with apraclonidine treatment. The bottom graphs display change in IOP with time for all three gases in each aqueous outflow scenario.
As expected, 13.9% C2F6 and 11.6% C3F8 resulted in longer duration of tamponade than 22.6% SF6 (Figure 3). We found a similar trend in earlier gas absorption in OHT patients for both gases, and the IOP remained stable.
On the assumption that an 80% initial fill is achievable, we used our model to analyse gas dynamics at concentrations commonly used in clinical practice (20% SF6, 16% C2F6, and 12% C3F8). We found a maximal fill of 97% with SF6, compared to 100% with C2F6 and C3F8 in normal eyes. OHT resulted in a reduction in maximal fill achievable, more noticeable with SF6 and C3F8 (93% and 95%, respectively), than C2F6 (98%).
The model revealed the optimal concentration of gas was 22.6% for SF6, 13.9% for C2F6, and 11.6% for C3F8 to completely fill a normal eye with an 80% initial fill of the vitreous cavity. Despite this, with the aqueous fluid dynamics in OHT, the percentage fill achieved was 95%, 95%, and 94% for SF6, C2F6, and C3F8, respectively. When the model included apraclonidine therapy in OHT, the percentage fill achievable increased to 98%, 97%, and 97% for SF6, C2F6, and C3F8, respectively at the above concentrations.
Intraocular gas tamponade is commonly used for rhegmatogenous retinal detachment, but our understanding of gas dynamics is relatively primitive, based on animal model experiments from many years ago. Human data are limited to a small study by Jacobs et al., the results of which are unlikely to be widely known by the vitreoretinal surgical community.14 Instead, anecdotal data and personal experience is the likely basis for individual decisions on which type, concentration and volume of gas is to be used. In an era of increased medicolegal scrutiny, a solid understanding of gas dynamics is imperative, and therefore research on this subject is important to support clinical decision making. Incorrect gas usage can result in visually damaging IOP rises, problems at altitude and interactions with anesthetic gases (NO2).15–17
The importance of furthering our knowledge of gas dynamics has driven our development and modification of this mathematical model. It allows estimation of tamponade duration, IOP, and maximal fill of different intraocular gases, and now allows the investigator to alter decision making on the basis of a pre-existing OHT diagnosis. In clinical practice, many clinicians give a short course of apraclonidine in the early post-operative period mainly to reduce the likelihood of a spike in IOP. Our model shows that apraclonidine is of significant importance in OHT patients to ensure that a near full postoperative fill can be achieved with good IOP control, which has resultant implications on the likelihood of successful primary retinal reattachment.
For all three gases, an eye with OHT, if given the same gas concentration as a normal eye, does not achieve 100% gas fill of the vitreous cavity postoperatively. Cases most at risk with underfill would be those in which an early full fill is important, such as retinal detachment secondary to inferior retinal breaks. It would be expected that in these clinical scenarios the gas of choice would be C2F6 or C3F8. Although the reduction in maximal fill for these gases with coexisting OHT is less pronounced than SF6, it is still significant, and therefore adjustment of gas concentration (eg, from 12% to 14% for C3F8) could be considered to improve surgical outcomes. An alternative would be to treat these cases with apraclonidine, and then only a smaller, if any, adjustment in gas concentration is required. We recognize that many clinicians would try to overcome these volumetric issues by posturing the patient in the early postoperative period. Although this is an important adjunct, patient compliance with such instructions is often highly variable, and therefore we recommend this is performed in conjunction with gas concentration adjustment to give the greatest chance for surgical success.
It is important to note that we used 80% to 85% fill at time of surgery to calculate gas concentrations. We recognize that this figure may vary slightly on a case by case basis, but it is important to note that 100% is not achievable during PPV due to residual anterior vitreous and a small liquid layer. In addition, the mathematical model is based on a sphere, and therefore the anterior chamber and lens volume needs to be included. Based on these assumptions, in normal eyes, we found the optimal concentrations to be 22.6% for SF6, 13.9% for C2F6, and 11.6% for C3F8. These concentrations are in line with those often used in clinical practice.
A limitation of our study and modified mathematical model is that the aqueous humour dynamics flow rates are based on old studies, in which the accuracy of measurement may be suboptimal. However, we were unable to find more up to date information and therefore used these in our calculations. If further research altered these figures, they could be easily adjusted in our calculations.
In conclusion, our model further improves our understanding of gas dynamics and now includes the ability to input variables for aqueous inflow, traditional outflow and uveoscleral outflow. This should provide clinicians with adjunctive information when making decisions regarding gas tamponade procedures.
- Gupta B, Neffendorf JE, Williamson TH. Trends and emerging patterns of practice in vitreoretinal surgery. Acta Ophthalmol. 2018;96(7):e889–e890. doi:10.1111/aos.13102 [CrossRef] PMID:27213838
- Neffendorf JE, Gupta B, Williamson TH. The role of intraocular gas tamponade in rhegmatogenous retinal detachment: A Synthesis of the Literature. Retina. 2018;38Suppl 1:S65–S72. doi:10.1097/iae.0000000000002015 [CrossRef] PMID:29280936
- Abrams GW, Edelhauser HF, Aaberg TM, Hamilton LH. Dynamics of intravitreal sulfur hexafluoride gas. Invest Ophthalmol. 1974;13(11):863–868. PMID:4431486
- Lincoff H, Maisel JM, Lincoff A. Intravitreal disappearance rates of four perfluorocarbon gases. Arch Ophthalmol. 1984;102(6):928–929. doi:10.1001/archopht.1984.01040030748037 [CrossRef] PMID:6329150
- Peters MA, Abrams GW, Hamilton LH, Burke JM, Schrieber TM. The nonexpansile, equilibrated concentration of perfluoropropane gas in the eye. Am J Ophthalmol. 1985;100(6):831–839. doi:10.1016/S0002-9394(14)73376-8 [CrossRef] PMID:3000186
- Kontos A, Tee J, Stuart A, Shalchi Z, Williamson TH. Duration of intraocular gases following vitreoretinal surgery. Graefes Arch Clin Exp Ophthalmol. 2017;255(2):231–236. doi:10.1007/s00417-016-3438-3 [CrossRef] PMID:27460279
- Intraocular Gases SC. Retina. St. Louis, MO: CV Mosby Co; 1989.
- Hutter J, Luu H, Schroeder L. A biological model of tamponade gases following pneumatic retinopexy. Curr Eye Res. 2002;25(4):197–206. doi:10.1076/ceyr.220.127.116.1187 [CrossRef] PMID:12658552
- Hall SK, Williamson TH, Guillemaut JY, Goddard T, Baumann AP, Hutter JC. Modeling the dynamics of tamponade multicomponent gases during retina reattachment surgery. American Institute of Chemical Engineers AIChE Journal. 2017;63(9):3651–3662. doi:10.1002/aic.15739 [CrossRef]
- Williamson TH, Guillemaut JY, Hall SK, Hutter JC, Goddard T. Theoretical gas concentrations achieving 100% fill of the vitreous cavity in the postoperative period: A Gas Eye Model Study. Retina. 2018;38Suppl 1:S60–S64. doi:10.1097/iae.0000000000001963 [CrossRef] PMID:29232331
- Toris CB, Koepsell SA, Yablonski ME, Camras CB. Aqueous humor dynamics in ocular hypertensive patients. J Glaucoma. 2002;11(3):253–258. doi:10.1097/00061198-200206000-00015 [CrossRef] PMID:12140404
- Alberti M, la Cour M. Nonsupine positioning in macular hole surgery: A Noninferiority Randomized Clinical Trial. Retina. 2016;36(11):2072–2079. doi:10.1097/IAE.0000000000001041 [CrossRef] PMID:27046458
- Toris CB, Tafoya ME, Camras CB, Yablonski ME. Effects of apraclonidine on aqueous humor dynamics in human eyes. Ophthalmology. 1995;102(3):456–461. doi:10.1016/S0161-6420(95)31000-7 [CrossRef] PMID:7891985
- Jacobs PM, Twomey JM, Leaver PK. Behaviour of intraocular gases. Eye (Lond). 1988;2(Pt 6):660–663. doi:10.1038/eye.1988.121 [CrossRef] PMID:3256505
- Costarides AP, Alabata P, Bergstrom C. Elevated intraocular pressure following vitreoretinal surgery. Ophthalmol Clin North Am. 2004;17(4):507–512, v. v. doi:10.1016/j.ohc.2004.06.007 [CrossRef] PMID:15533743
- Lincoff H, Weinberger D, Stergiu P. Air travel with intraocular gas. II. Clinical considerations. Arch Ophthalmol. 1989;107(6):907–910. doi:10.1001/archopht.1989.01070010929043 [CrossRef] PMID:2730410
- Fu AD, McDonald HR, Eliott D, et al. Complications of general anesthesia using nitrous oxide in eyes with preexisting gas bubbles. Retina. 2002;22(5):569–574. doi:10.1097/00006982-200210000-00006 [CrossRef] PMID:12441721
Rates of Aqueous Production and Outflow in Different Clinical Scenarios
|Aqueous Flow Qin(uL/min)||Outflow Facility Kp(uL/min/mmHg)||Uveoscleral Outflow Qus(uL/min)|
|Normal||2.58 ± 0.10 (SEM, n = 55)||0.27 ± 0.02 (SEM, n = 55)||1.09 ± 0.11 (SEM, n = 55)|
|Ocular Hypertension||2.62 ± 0.11 (SEM, n = 55)||0.17 ± 0.01 (SEM, n = 55)||0.66 ± 0.11 (SEM, n = 55)|
|Apraclonidine vs. Ocular Hypertension||Decreased by 11.5%||Increased by 52.9%||Decreased by 82.3%|
Percentage Fill of the Vitreous Cavity at Time of Surgery to Achieve Maximal Fill in Different Aqueous Outflow States
|Fill % of the Vitreous Cavity by the Gas at Surgery||Normal||OHT||OHT + Apraclonidine|
|5||100 (13)||100 (12)||100 (12)|
|10||100 (25)||100 (23)||100 (24)|
|15||100 (37)||100 (35)||100 (36)|
|20||100 (49)||100 (46)||100 (48)|
|25||100 (61)||100 (57)||100 (59)|
|30||100 (73)||100 (68)||100 (71)|
|35||100 (84)||100 (78)||100 (82)|
|40||100 (96)||100 (89)||100 (93)|