Visual loss after surgery is a rare but catastrophic complication. Its incidence has been estimated as 0.01%–1%, depending on the type of surgery (1–3). In recent years, visual loss after spine surgery in the prone position received increased attention (4–6), and postoperative blindness has been ranked as a major patient-safety issue among anesthesiologists (7). Of note, direct pressure to the eye can be excluded as a factor contributing to visual loss in the majority of these cases. Most of these patients did not recover their vision after surgery. Postoperative visual loss is most often caused by ischemic optic neuropathy (3). In anterior ischemic optic neuropathy, ischemia occurs in a watershed zone of the anterior optic nerve supplied by the branches of the posterior ciliary artery (Fig. 1) (8). Perfusion pressure of the anterior optic nerve is the difference between pressures in the posterior ciliary arteries and the venous drainage of the eye (Fig. 1), which are approximated by mean arterial blood pressure (MAP) and intraocular pressure (IOP), respectively (9). Increased IOP may substantially decrease the perfusion pressure of the anterior optic nerve (10). Posterior ischemic optic neuropathy (PION) affects the intraorbital portion of the optic nerve from the optic foramen to the entry of the central retinal artery, which derives its arterial supply from terminal centripetal pial vessels (3). Whether increased IOP is a useful, albeit indirect, marker for the risk of PION is not known (11,12).
Body position affects IOP. The degree of inclination of the body has been shown to affect IOP in subjects lying supine (13). In supine subjects, head-up tilt decreased, whereas head-down tilt (HDT) increased, IOP. The increase of IOP during HDT follows a sinusoidal function of tilt angle (14), with a threefold increase in complete inversion and slight IOP increases for angles <10°. Most previous studies (1–4) during HDT (<3 days) reported an increase of IOP from 2 to 5 mm Hg between the sitting and −10° HDT positions (14–17). Prone positioning has been shown to increase IOP in both awake and anesthetized subjects (11,18). In these studies, increases in IOP were typically detected within minutes of assuming the prone or head-down recumbent position (11,13,18). This observation suggests that the absolute pressure in veins draining the eye (Fig. 1) and the redistribution of venous blood are important factors that influence IOP. In fact, because of the absence of venous valves, changes in central venous pressure (CVP) translate into concomitant changes of ocular venous pressures and therefore affect IOP (19,20).
Because CVP correlates with IOP, head-up inclination of the operating room (OR) table might be a practical way to decrease both ocular pressure and CVP and thereby decrease IOP (19,20). Similarly, an OR table that is best designed to relieve pressure from the abdomen during prone positioning might have a beneficial effect on IOP. Two positioning devices often used for surgery in the prone position are the Jackson table (Fig. 2A) and the Wilson frame (Fig. 2B). They differ in the degree to which they free the abdomen from compression. We hypothesized that the reverse Trendelenburg position ameliorates the increase in IOP caused by the prone position. To test this hypothesis, we determined the IOP in awake volunteers during prone positioning at various degrees of OR table inclination in a randomized crossover study comparing the Wilson frame and the Jackson table.
After IRB approval, written informed consent was obtained from 10 volunteers, all of whom were ASA physical status I or II. Exclusion criteria were a body mass index ≥30 kg/m2 or medical therapy with β-adrenergic blockers, cholinesterase inhibitors, or muscarinic agonists. Individuals allergic to proparacaine or latex were not recruited. None of the subjects had any preexisting eye disorder except for refractive errors within ±4 diopters. With the exception of 0.5% proparacaine hydrochloride for topical anesthesia, no ophthalmic medication was used before or during the study. Two drops of topical anesthetic were applied to each eye before the first IOP measurement and were repeated before subjects assumed the prone position.
IOP measurements were performed by applanation tonometry with a handheld device (Tono-Pen™ XL; Medtronic Ophthalmic, Jacksonville, FL). This device has been validated in humans (21,22) and in comparison with invasive IOP measurements in animal studies during HDT (23). Its tip has a central plunger to measure the force applied to the center of the cornea. The mean of four successful readings was displayed on the screen of the device, along with the sd of the measurements expressed as a percentage of the mean. Any measurement series with an sd more than 5% was excluded and repeated.
For the purposes of our study, we defined an IOP of more than 23 mm Hg as grossly abnormal. This value exceeds the mean IOP value of the nonglaucomatous population, i.e., 15.5 ± 2.6 mm Hg, by 3 sd (24). As in our study population, normal IOP ranges from 10 to 20 mm Hg. However, these values apply to an examination in the sitting or supine position. To exclude more than 99% of values that could potentially be considered normal, we chose a more stringent definition of 23 mm Hg as an abnormal IOP.
The study was designed as 2 separate 25-min sessions, at least a week apart, for each subject. The sessions differed only in the setup used for prone positioning (Fig. 2). IOP was checked in both eyes in six different positions that were assumed in the following sequence: sitting upright (sitting), supine on a horizontal OR table (supine), prone on a horizontal OR table (prone-horizontal 1), prone with a 10° reverse Trendelenburg position (prone-head up), prone with a 10° Trendelenburg position (prone-head down), and, finally, prone with the table back to horizontal (prone-horizontal 2). A handheld goniometer was used to measure the degree of table inclination. After subjects assumed a position, 5 min was allowed to pass before the IOPs were measured. The rationale for the 5-min interval was threefold: 1) it maximized subject comfort, 2) it avoided the need to reapply topical anesthetic in the prone position, 3) position-related changes in IOP reportedly occur within minutes (11,13,18). Heart rate and MAP were measured by plethysmography and oscillometry, respectively, before each IOP measurement.
One setup consisted of the Jackson spinal table (Orthopedic Systems, Inc., Union City, CA). It was assembled with its prone positioning frame and supports for the thorax, hips, and legs (Fig. 2A). To allow access to the eyes, we modified the headrest of the table by enlarging its central opening to conform to the T-shaped cutout of a prone foam face cushion. This modification neither deformed the cushion nor affected the stability of the headrest. In the prone position, the subjects’ faces rested on the foam cushion with the eyes, nose, and mouth centered in the cutout of the foam pillow. A neutral position of the neck was maintained by choosing an appropriate cushion height. The frame was adjusted to support the body at the upper thorax and bilateral iliac crests. Arms were supported with the designated armrests of the Jackson table as shown in Figure 2A.
The other setup consisted of an OR table (Amsco 3080; Amsco Inc., Erie, PA), a Wilson frame (Orthopedic Systems, Inc.), and a horseshoe headrest (Fig. 2B). The lateral margins of the gel-padded rails of the Wilson frame were adjusted to approximate the midclavicular lines for each subject (Fig. 2B). Subjects positioned themselves prone with their shoulders at the upper end of the Wilson frame. They were asked to rest their faces on the horseshoe headrest so that the weight of the head was supported at the forehead and cheeks. The curved gel pads of the headrest were adjusted to avoid any pressure on the eye. The height of the headrest was adjusted to keep the neck in neutral position. Arms were abducted, with the forearms flexed and positioned on arm boards at each side of the head, ventral to the axis of the body.
Sample size calculations (StatMate Version 1.01; GraphPad Software, San Diego, CA) were centered around our primary hypothesis that the reverse Trendelenburg position ameliorates the increase in IOP caused by the prone position. To this end, we estimated the difference between IOP in the prone horizontal position and the supine position to be 5 mm Hg. Similarly, we estimated the sd to be approximately 4 mm Hg. Using alpha and beta values of 0.05 and 0.8, respectively, we estimated that seven to nine subjects would be needed. Normality of distribution was determined with the Kolmogorov-Smirnov test with the Lilliefors correction (SigmaStat 2.03; SPSS, Inc., Chicago, IL). Demographic and hemodynamic data are presented as mean ± sd, whereas IOP data are presented as median (25th–75th percentile). Data were analyzed with two-way repeated-measures analysis of variance, followed by Tukey post hoc testing as appropriate. A difference was considered statistically significant if P < 0.05.
We were able to collect a complete set of data from all 10 subjects. Demographic data on the subjects are summarized in Table 1. No subject sustained a corneal abrasion as a result of the IOP measurements. Similarly, topical anesthesia, as described previously, proved adequate for the duration of the study.
Both body position and OR table inclination profoundly (P < 0.001) influenced the IOP in awake volunteers (Fig. 3A). The prone position caused a sustained increase in IOP over values measured in the sitting or supine positions. IOP in the sitting position was 15.0 mm Hg (12.8–16.3 mm Hg) (median [25th–75th percentile]), whereas that in the supine position was 16.8 mm Hg (14.0–18.3 mm Hg). Median IOPs measured in the prone horizontal and Trendelenburg positions were 46% and 58% higher, respectively, than those measured in sitting position. The reverse Trendelenburg position ameliorated the increase in IOP caused by prone positioning but did not completely normalize it. Values for IOP in the prone horizontal, Trendelenburg, and reverse Trendelenburg positions were 20.3 mm Hg (16.3–22.5 mm Hg), 22.5 mm Hg (19.8–25.3 mm Hg),* and 23.8 mm Hg (21.5–26.3 mm Hg),*† respectively (*P < 0.001 versus reverse Trendelenburg; †P < 0.001 versus horizontal).
Many of the IOP values in the prone position were increased beyond the normal range of 10–20 mm Hg and exceeded the mean IOP of the general population by more than 3 sd. IOP values exceeding 23 mm Hg occurred in 16 of 40, 8 of 40, and 32 of 40 individual eyes in the horizontal, reverse Trendelenburg, and Trendelenburg positions, respectively (Fig. 3B). Therefore, the reverse Trendelenburg position decreased the number of grossly abnormal IOP values by 50% compared with the prone horizontal position and by 75% compared with the Trendelenburg position.
Although the IOP was strongly affected by body position and OR table inclination, the choice of the OR table and frame for prone positioning played no role in the IOP increase caused by the prone position (Fig. 4). The IOP of the subjects measured either on the Jackson table or Wilson frame was not significantly different at identical body positions. Likewise, the number of abnormal IOP values was similar with both setups. The effect on heart rate of the setup used for prone positioning was statistically, but not clinically, significant (Table 2). Heart rate decreased by up to 18% in all prone positions with the Wilson frame (P < 0.001). In contrast, heart rate did not change significantly during the sessions with the Jackson table. Neither setup nor position caused a clinically significant change in MAP (Table 2).
Our study shows that a 10° reverse Trendelenburg position ameliorates the increase in IOP caused by the prone position in awake volunteers. Although the IOP is strongly affected by body position and OR table inclination, the setup used for prone positioning played no role in the IOP increase caused by the prone position. We believe that the alleviation of IOP with table inclination is clinically significant because the number and degree of grossly abnormal IOP values were significantly lower in the reverse Trendelenburg position than in either the horizontal or the Trendelenburg positions.
Our results have both similarities with and differences from the results of other studies on the effects of body position and inclination on IOP. Similar to other studies (18,25), IOP increased with the prone position in our study. Likewise, inclination of the OR table in our study changed the IOP in the same direction as in a previous study in supine subjects (13). In contrast, the magnitude of the increase in IOP with the prone position in our study was less than that reported during surgery in anesthetized subjects (11). Changing the body position from supine to prone more than doubled the IOP in subjects under general anesthesia, whereas the IOP of the subjects in our study increased by only 32% upon being positioned prone. This difference can be due to a number of factors, including the effect of anesthesia, time spent in the prone position, or the amount of IV fluid administered. General anesthesia decreases the IOP in supine subjects (26), but its effects on IOP in the prone position have not been extensively studied. Thus, general anesthesia might increase the intraocular blood volume in the prone position by impairing the autoregulatory mechanism of the choroidal circulation (25). Another factor is the time spent in the prone position, which may affect the circulation of aqueous humor and, therefore, IOP. The administration of IV fluids may further exacerbate increases in IOP in the prone position by increasing ocular venous pressures (27).
Increased IOP in the prone position may be relatively more important in the intraoperative setting secondary to the changes in MAP and hematocrit. MAP may decrease substantially secondary to blood loss or may be deliberately maintained at a low level to decrease surgical blood loss. In either case, the perfusion pressure of the anterior optic nerve (MAP − IOP) is decreased. However, a decrease in hematocrit would increase the minimum blood flow needed for the viability of the anterior optic nerve. As a result, when both MAP and hematocrit are low, IOP becomes an even more critical factor that influences the blood supply to the anterior optic nerve (4,6,28).
Although we did not measure the episcleral venous pressure (EVP) in this study, we believe EVP to be an important factor in explaining the changes we have observed in IOP. The episcleral veins are connected to the central circulation by a valveless system such that increases in CVP and cephalad shifts in venous blood caused by changes in body position will increase EVP (29,30). Indeed, EVP has been shown to increase in supine or prone recumbence, as well as during head-down vertical inversion (13,18,29). Although the effect of body inclination on EVP during prone positioning has not been studied, it is likely that the reverse Trendelenburg position ameliorates the increase in EVP caused by prone positioning.
EVP is positively correlated with IOP. In fact, there are disease states such as carotid cavernous fistula that can increase EVP and cause a secondary glaucoma (31–33). The relationship between EVP and IOP is expressed by the Goldmann equation, which states (34)
More than 80% of the aqueous humor outflow occurs through Schlemm’s canal into the episcleral veins (10,35). Outflow is passive and therefore depends on a gradient between the IOP and the EVP (34). In contrast to the outflow, the rate of production of aqueous humor is affected by IOP to a much lesser degree and only at very high IOPs (10,36). Therefore, a change in the EVP affects the IOP by changing the total aqueous humor volume. Turnover of aqueous humor is rapid and requires approximately 2 hours for one complete exchange. Because the eye is relatively nondistensible, even small changes in the volume of aqueous humor may significantly change IOP.
Intraocular blood volume is a second factor that may have been responsible for the changes in IOP in our study. The major portion of the total intraocular blood volume is within the choroid (Fig. 1), a vascular structure with very high blood flow (37). Most of the choroidal blood volume resides in its venules, and the venular filling of the choroid is believed to depend on the pressure in the orbital veins (19). Body position is an important factor in determining the pressure within the orbital venous system (19,29,38). Specifically, prone and/or Trendelenburg positions increase, whereas the reverse Trendelenburg position decreases, orbital venous pressure. An increased orbital venous pressure may cause congestion of the choroid and thereby increase the IOP by augmenting the intraocular blood volume (39).
Increased IOP is a risk factor for anterior ischemic optic neuropathy and a proposed mechanism for postoperative visual loss. Perfusion pressure of the anterior optic nerve is defined by the difference between MAP and IOP (9). Blood flow to the anterior optic nerve is autoregulated by endocrine and metabolic factors. Similar to the cerebral circulation, autoregulation of the anterior optic nerve is also effective over a critical range of perfusion pressures, but this range is not clearly known for humans (9). Disease states that affect the circulatory system, such as hypertension, diabetes mellitus, and atherosclerosis, derange this autoregulation and therefore increase the critical perfusion pressure below which the autoregulation of the anterior optic nerve fails (9). Moreover, autoregulation might be defective or absent at some locations of the anterior optic nerve even in certain healthy individuals (9,40). In short, a safe lower limit for the perfusion pressure of the anterior optic nerve is difficult to establish.
Conducting this study on awake volunteers rather than on patients undergoing surgery has both advantages and limitations. The main advantage of having awake volunteers is that pressure on the eye either during prone positioning or during the actual measurement can be reliably avoided. Furthermore, the effects of two different prone positioning setups can be easily compared for each volunteer. However, general anesthesia and surgery may affect IOP and are not accounted for in our study. Although general anesthesia decreases IOP, this decrease fails to offset the increase in IOP caused by prone positioning. In fact, Cheng et al. (11) found a more pronounced increase in IOP than we did, after changing anesthetized subjects from supine to prone positions. The duration of prone recumbence has been shown to positively correlate with the IOP and is likely to be substantially longer during a surgical procedure than our study sessions. Another factor that affects the IOP is CVP, which may change secondary to blood loss, fluid shifts, and IV fluid replacement during surgery. Our present study cannot predict how and to what degree these factors would interact with the effect of body inclination on IOP during prone positioning. Although none of our subjects complained of decreased vision, the outcome of postoperative visual loss was not assessed in our study.
In summary, we have shown that a 10° reverse Trendelenburg position can ameliorate the increase in IOP caused by prone positioning. Although this study does not include all the factors that may affect the IOP in an intraoperative setting, it suggests a strategy to beneficially influence the IOP by changing the operating table inclination for patients in the prone position. A similar study on patients undergoing surgery in the prone position will be needed to determine the full clinical applicability of our findings.
The Tono-Pen XL™ used in this study was a generous gift from Medtronic Solan, Jacksonville, FL.
1. Roth S, Thisted RA, Erickson JP, et al. Eye injuries after nonocular surgery: a study of 60,965 anesthetics from 1988 to 1992. Anesthesiology 1996;85:1020–7.
2. Warner ME, Warner MA, Garrity JA, et al. The frequency of perioperative vision loss. Anesth Analg 2001;93:1417–21.
3. Williams EL, Hart WM Jr, Tempelhoff R. Postoperative ischemic optic neuropathy. Anesth Analg 1995;80:1018–29.
4. Cheng MA, Sigurdson W, Tempelhoff R, Lauryssen C. Visual loss after spine surgery: a survey. Neurosurgery 2000;46:625–30.
5. Lee LA. ASA postoperative visual loss registry: preliminary analysis of factors associated with spine operations. ASA Newslett 2003;67:7–8.
6. Myers MA, Hamilton SR, Bogosian AJ, et al. Visual loss as a complication of spine surgery: a review of 37 cases. Spine 1997;22:1325–9.
7. Stoelting RK. APSF survey results identify safety issue priorities. APSF Newslett 1999;14:6–7.
8. Hayreh SS. Anterior ischemic optic neuropathy. Clin Neurosci 1997;4:251–63.
9. Hayreh SS. Blood flow in the optic nerve head and factors that may influence it. Prog Retin Eye Res 2001;20:595–624.
10. Bill A. Blood circulation and fluid dynamics in the eye. Physiol Rev 1975;55:383–417.
11. Cheng MA, Todorov A, Tempelhoff R, et al. The effect of prone positioning on intraocular pressure in anesthetized patients. Anesthesiology 2001;95:1351–5.
12. Lee LA, Vavilala MS, Sires BS, et al. Intraocular pressures during prone spine surgery do not predict visual deficits [abstract]. Anesthesiology 2001;95:A298.
13. Krieglstein GK, Waller WK, Leydhecker W. The vascular basis of the positional influence of the intraocular pressure. Albrecht Von Graefes Arch Klin Exp Ophthalmol 1978;206:99–106.
14. Linder BJ, Trick GL, Wolf ML. Altering body position affects intraocular pressure and visual function. Invest Ophthalmol Vis Sci 1988;29:1492–7.
15. Carlson KH, McLaren JW, Topper JE, Brubaker RF. Effect of body position on intraocular pressure and aqueous flow. Invest Ophthalmol Vis Sci 1987;28:1346–52.
16. Frey MA, Mader TH, Bagian JP, et al. Cerebral blood velocity and other cardiovascular responses to 2 days of head-down tilt. J Appl Physiol 1993;74:319–25.
17. Mader TH, Taylor GR, Hunter N, et al. Intraocular pressure, retinal vascular, and visual acuity changes during 48 hours of 10 degrees head-down tilt. Aviat Space Environ Med 1990;61:810–3.
18. Lam AK, Douthwaite WA. Does the change of anterior chamber depth or/and episcleral venous pressure cause intraocular pressure change in postural variation? Optom Vis Sci 1997;74:664–7.
19. Hvidberg A, Kessing SV, Fernandes A. Effect of changes in PCO2
and body positions on intraocular pressure during general anaesthesia. Acta Ophthalmol (Copenh) 1981;59:465–75.
20. Lee LA, Vavilala MS, Sires BS, et al. Intraocular pressure is partially dependent on central venous pressure during prone spine surgery [abstract]. Anesthesiology 2003;99:A289.
21. Hessemer V, Rossler R, Jacobi KW. Comparison of intraocular pressure measurements with the Oculab Tono-Pen vs manometry in humans shortly after death. Am J Ophthalmol 1988;105:678–82.
22. Kao SF, Lichter PR, Bergstrom TJ, et al. Clinical comparison of the Oculab Tono-Pen to the Goldmann applanation tonometer. Ophthalmology 1987;94:1541–4.
23. Setogawa A, Kawai Y. Measurement of intraocular pressure by both invasive and noninvasive techniques in rabbits exposed to head-down tilt. Jpn J Physiol 1998;48:25–31.
24. Schottenstein EM. Intraocular pressure and tonometry. In: Ritch R, Shields MB, Krupin T, eds. The glaucomas. St. Louis: Mosby, 1996:407–28.
25. Hyams SW, Friedman Z, Neumann E. Elevated intraocular pressure in the prone position: a new provocative test for angle-closure glaucoma. Am J Ophthalmol 1968;66:661–72.
26. Murphy DF. Anesthesia and intraocular pressure. Anesth Analg 1985;64:520–30.
27. Brucculeri M, Hammel T, Harris A, et al. Regulation of intraocular pressure after water drinking. J Glaucoma 1999;8:111–6.
28. Shahian DM, Speert PK. Symptomatic visual deficits after open heart operations. Ann Thorac Surg 1989;48:275–9.
29. Friberg TR, Sanborn G, Weinreb RN. Intraocular and episcleral venous pressure increase during inverted posture. Am J Ophthalmol 1987;103:523–6.
30. Smith TJ, Lewis J. Effect of inverted body position intraocular pressure. Am J Ophthalmol 1985;99:617–8.
31. Greenfield DS. Glaucoma associated with elevated episcleral venous pressure. J Glaucoma 2000;9:190–4.
32. Jorgensen JS, Guthoff R. Sturge-Weber syndrome: glaucoma with elevated episcleral venous pressure. Klin Monatsbl Augenheilkd 1987;191:275–8.
33. Jorgensen JS, Guthoff R. The role of episcleral venous pressure in the development of secondary glaucomas. Klin Monatsbl Augenheilkd 1988;193:471–5.
34. Goldmann H. Die Kammerwasservenen und das Poiseuille’sche Gesetz. Ophthalmologica 1949;118:496–511.
35. Maepea O, Bill A. The pressures in the episcleral veins, Schlemm’s canal and the trabecular meshwork in monkeys: effects of changes in intraocular pressure. Exp Eye Res 1989;49:645–63.
36. Moses RA, Grodzki WJ Jr, Carras PL. Pseudofacility: where did it go? Arch Ophthalmol 1985;103:1653–5.
37. Weiter JJ, Shiyoung R. Anatomy and physiology: retinal and choroidal circulation. In: Yanoff M, ed. Ophthalmology. St. Louis: Mosby, 1999:1–4.
38. Kaskel D, Muller-Breitenkamp R, Wilmans I, et al. The influence of changes in body position on intraocular pressure, episcleral venous pressure, and blood pressure. Albrecht Von Graefes Arch Klin Exp Ophthalmol 1978;208:217–28.
39. Kiel JW. Choroidal myogenic autoregulation and intraocular pressure. Exp Eye Res 1994;58:529–43.
© 2004 International Anesthesia Research Society
40. Pillunat LE, Anderson DR, Knighton RW, et al. Autoregulation of human optic nerve head circulation in response to increased intraocular pressure. Exp Eye Res 1997;64:737–44.