The cause of peri-operative ischemic optic neuropathy (ION) remains elusive. It is most frequent in spine fusion or cardiac surgery,1 and presents as anterior ION (AION) or posterior ION (PION). Many case reports described hypotension and anaemia/haemodilution, although relative fluid excess, head-down positioning, lengthy surgery and vasopressors, among others, have been suggested as intra-operative risk factors. In a retrospective case–control study of 83 ION cases, the risk in spinal fusion was increased by male sex, positioning on a Wilson frame, lengthy surgery, large blood loss, obesity and decreased colloid to crystalloid ratio for fluid resuscitation.2 In a retrospective case–control study of more than 2.5 million operations in the US National Inpatient Sample, significantly associated with ION in spinal fusion surgery were age, transfusion and obesity.3
These studies assessed association but not causation. The tendency of the Wilson frame to keep the head lower than the body, combined with decreased colloid/crystalloid fluid ratio, suggest that orbital or optic nerve oedema may be involved; however, these hypotheses have to date, not been possible to study in humans.2 Prospective clinical studies are limited by confounding factors and ethical problems in manipulating suspected risk factors. Accordingly, mimicking ION in a rodent under conditions resembling those developing peri-operatively is a novel means to examine peri-operative ION. In pigs, decreasing haematocrit by 37%, along with significant hypotension, and occluding the jugular venous drainage, significantly decreased optic nerve blood flow. However, visual neurophysiology, and anatomical or biochemical/molecular changes in the optic nerve were not examined, and the porcine eye circulation differs from human.4
Although rodents differ from primates in having minimal lamina cribrosa in the optic nerve,5 their optic nerve structure and circulation otherwise resemble humans.6 In this study, we examined the impact of haemodilution, procedure duration and head-down positioning. We tested the hypothesis that prolonged head down with and without haemodilution damages the optic nerve.
Detailed Methodology is available in the data supplement, http://links.lww.com/EJA/A157.
Procedures conformed to the Association for Research in Vision and Ophthalmology Resolution on the Use of Animals in Research and were approved by the Animal Care Committee of the University of Chicago. Animal Care Protocol number 70935, initial approval date 6/1/09, study start date 8/10/09 and chairman of the committee Catherine Reardon-Alulis, Ph.D. Male Sprague-Dawley rats (mean weight 220 g, range 200 to 250 g, mean 7 weeks of age, from Taconic, Hudson, New York, USA; with an indwelling jugular vein cannula, surgical model #JVC-R), were housed at ambient light less than 325 lx, and automated light cycling 12 h on/12 h off to prevent light damage. Ketamine (35 mg kg−1) and xylazine (5 mg kg−1) were injected via an intraperitoneal catheter about every 20 min. This anaesthetic regimen was used as it was well tolerated and efficacious in our previous experiments and was deemed the best compromise between efficacy and safety and potential impact upon optic nerve and retinal ganglion survival.7
To simulate head down positioning, rats were secured on their dorsal side on a covered wooden platform and inclined at 70° head down. The rationale is that there is no prone position (as in spine surgery) for rodents, but profound head down tilt increases intra-ocular and intracranial pressure and alters MRI-based parameters of cerebrospinal fluid dynamics in the optic nerve sheath.8–11 Groups were 5 h supine (SUP), 5 h tilted head-down only (TILT), 5 h tilted head-down with haemodilution (TILT-HD) and 5 h SUP with haemodilution (SUP-HD). For isovolumic haemodilution, 2-ml blood was withdrawn from the venous catheter every 30 min over a 90-min period; during this time the animals remained positioned according to group assignment. Blood was replaced with 6% Hetastarch (Hospira, Lake Forest, Illinois, USA)12 following each 2-ml blood withdrawal, until target haematocrit (approximately 40% reduction) was achieved.
Temperature was maintained at 36 to 37 °C with a heating blanket (Harvard Apparatus, Natick, Massachusetts, USA). O2 saturation was measured with a pulse oximeter (Ohmeda, Louisville, Colorado, USA) on the tail, as was noninvasive blood pressure (BP) (IITC Life Science, Woodland Hills, California, USA). Supplemental O2, when necessary to maintain saturation more than 94%, was administered by a plastic cannula in front of the nares and mouth.
Following dark adaptation, rats were injected intraperitoneally with ketamine (35 mg kg−1) and xylazine (5 mg kg−1) about every 20 min.13 After corneal analgesia with 0.5% proparacaine, pupils were dilated with 0.5% tropicamide and cyclomydril (Alcon, Fort Worth, Texas, USA). To assess functional outcomes, visual function was evaluated at baseline (prior to experiments), 14 and 28 days after experiments, using a UTAS-E4000 system and a Ganzfeld stimulator (LKC Technologies, Gaithersburg, Maryland, USA).
We measured the electroretinogram (ERG) a-wave, b-wave, oscillatory potentials and P2 (Supplemental Fig 1, http://links.lww.com/EJA/A157), the scotopic threshold response (STR) and visual evoked potentials (VEP). The ERG measures the response to bright light flashes, yielding the downward (negative) direction a-wave, generated by electrical activity of retinal rod cells.14 The upward (positive) b-wave is generated by depolarising retinal bipolar cell and Müller glial cell currents.15 Damage to the inner retina is represented by decreased amplitude of the b, and other waveforms (refer to Supplemental Fig 1, http://links.lww.com/EJA/A157), whereas damage to the outer retina, by changes in the a-wave.16 The STR, consisting of negative (nSTR), and positive (pSTR), is elicited by ultra-dim light stimuli and specifically reflects retinal ganglion cell (RGC) function in rodents (Supplemental Fig 1, http://links.lww.com/EJA/A157). VEP reflects postretinal function to the primary visual cortex, thus it includes the optic nerve and visual pathway projections.17
To evaluate the optic nerve noninvasively, we imaged the rat eye fundus. Photographs were assessed by a neuro-ophthalmologist (NJN) blinded to the groups.
Intra-ocular pressure measurement
To examine the impact of positioning and haemodilution on intra-ocular pressure, intra-ocular pressure (IOP) was measured with a Tono-Pen XL (Mentor, Norwell, Massachusetts, USA) on the topically anaesthetised cornea to achieve four consecutive values with variation less than 0.5%. At least two such readings were obtained and results were averaged.18
Eyes and optic nerves from euthanised rats at 28 days were evaluated using immunohistochemistry. To stain optic nerve axons, cholera toxin subunit B (Alexa Fluor 594 conjugate; Invitrogen-Life Technologies, Grand Island, New York, USA) had been intravitreally injected (2 μl, 10 μg μl−1) 48 h before euthanasia.19 Primary antibody (1 : 50) for glial acidic fibrillary protein (GFAP) (Sigma, St. Louis, Missouri, USA, mouse monoclonal) was used to stain glial cells. Standardisation, validation and quantitation of immunostained images have been described by us previously.13
Data handling and statistical analysis
Amplitude, time course and intensity (log cd s m−2) of ERG a-wave, b-wave, oscillatory potentials, P2-wave and STR were analysed in Matlab 2011a (MathWorks, Natick, Massachusetts, USA).20 For STR, the response to three flash intensities generated a stimulus-intensity plot,21 with amplitude on the y-axis, and flash intensity on the x-axis. pSTR was calculated as maximum positive amplitude, whereas nSTR was calculated as maximum negative amplitude (Supplemental Fig 1, http://links.lww.com/EJA/A157).22 Results of three flash intensities were compared at baseline vs. 14 and 28 days in each rat; in the “Results” section, only the response to the highest intensity flashes are shown. The Data Supplement contains the complete data set, http://links.lww.com/EJA/A157. Animal numbers were estimated by power analysis based upon previous experiments, with visual function considered the primary outcome.21,23 Assuming 20% difference between means, SD of one group 5 to 15%, and the other 20% of mean, it would require 9 to 13 animals (β 0.8, α 0.05).
We used Stata v10.0; StataCorp, College Station, Texas, USA. Analysis of variance was followed by paired t tests comparing between animals in the same group studied over time (e.g. electroretinography, BP, haematocrit and IOP), and an unpaired t test to compare results between matched groups (e.g. immunostaining). Bonferroni correction was applied for multiple comparisons. Data are expressed as mean ± SD, and in the tables, include 95% confidence intervals.
Baseline values differed between the groups, but Table 1 shows that amplitudes were significantly different over time after experiments within groups. Specifically, at 28 days, there was a significant decrease in a-wave amplitude in TILT-HD, and an increase in TILT. Oscillatory potentials significantly decreased at 28 days in SUP-HD. P2 and the b-wave significantly increased in TILT.
Scotopic threshold response responses
In TILT-HD, the nSTR amplitude significantly decreased (Table 2) at 14 and 28 days. In SUP-HD, there was a significant difference in nSTR at 28 days. In all groups except TILT-HD, the pSTR increased at 28 days (Table 2). More extensive data on the STRs may be viewed in the Supplemental Data file, http://links.lww.com/EJA/A157.
Visual evoked potentials
There were significant changes only in N2-P3 amplitude, which decreased in TILT-HD vs. SUP (P = 0.01, Fig. 1). Changes were present also in the implicit times (time from flash to peaks, Table 3) in several VEP components.
Immunostaining and histology
In optic nerve longitudinal sections, GFAP, an astrocytic activation marker (Supplemental Fig 2, http://links.lww.com/EJA/A157),24 in comparison with SUP (1161 +/− 398 U), was decreased in SUP-HD (886 +/− 318 U, P < 0.017), increased in TILT (1565 +/− 512 U, P < 0.004) and increased in TILT-HD (1684 +/− 689 U, P < 0.01). No terminal deoxynucleotidyl transferase deoxyuridine triphosphate nick end labeling (TUNEL), reflecting apoptosis, was in the retinas at 1, 14 or 28 days. H&E-stained 4-μm paraffin-embedded cross sections of retina at 28 days showed no thinning, or cell loss as would be expected in retinal ischaemia.25
Systemic blood pressure, haematocrit and tonometry during experiments
At the end of experiments, haematocrit significantly decreased by about 40% in TILT-HD and SUP-HD, restored to baseline by the 28th day (Table 4). The mean arterial BP significantly decreased in SUP-HD vs. SUP at end measurements, and in SUP-HD, TILT and TILT-HD at midpoint measurements. As anticipated, IOP was increased during experiments in the two head-down tilted groups, TILT-HD and TILT.
In general, at baseline, rat fundi had very red and hyperaemic optic discs with surrounding red edges (i.e. normally appearing mildly swollen), and the retina surrounding the optic disc was relatively white with attenuated arteries (Fig. 2). Fundi from both groups of haemodiluted animals appeared diffusely pale immediately after experiments. TILT-HD and, to a lesser degree, TILT showed increased optic nerve swelling beyond baseline and dilated veins at 28 days.
We found decreased nSTR amplitudes, decreased amplitudes and latencies of VEP, optic nerve swelling and glial activation, showing that haemodilution and head down tilt injured the optic nerve. Given the lack of b-wave and P2 change, and no histological damage, or TUNEL, it seems reasonable to conclude there was no retinal injury.
No RGC loss or TUNEL in retina of TILT-HD and SUP-HD might also be explained by not allowing enough time for injury. However, although in human ION, it has been assumed that RGCs die over months, recently, with ocular coherence tomography, thinning of the retinal nerve fibre layer is observed within weeks of an acute event, suggesting that neuronal injury proceeds at a faster pace than previously realised.26 In future studies, it may be advantageous to examine longer time periods beyond 28 days for a more complete picture. In this study, there was disc oedema in TILT-HD at 28 days. This could be reflective of damage to the optic nerve head, but could also be extension of retro-bulbar oedema forward to become visible at the optic nerve head, combined with retro-bulbar optic nerve oedema compressing venous outflow via the central retinal vein. Our study methods do not enable defining this model as AION or PION.
Both components of the STR in rodents are believed to reflect function of the RGCs, with a significant contribution from amacrine cells attributed to the nSTR, whereas pSTR is thought to originate primarily from RGCs.27 Elevated IOP in rodent glaucoma models decreased nSTR.28 However, studies are conflicting on the magnitude as well as the direction of change in pSTRs in rodent models of optic nerve injury.29–31 To our knowledge, there are no previous studies concerning STR, head down positioning and/or haemodilution. In our study, TILT-HD was notably the only group in which nSTR decreased without increase in pSTR. We cannot determine the mechanism of increased pSTR in the three other groups. But decreased nSTR suggests functional injury to the RGCs due to retrograde injury originating in the optic nerve. Decreased amplitude and increased latency of VEP reflect optic nerve injury. This has been reported in other eye injury models including anterior ION,32 optic neuritis33 and glaucoma.34
There are limitations. Although visual function referenced specifically to the optic nerve and its retinal connections were affected, as shown by STR and VEP, we cannot state definitively there was visual impairment. Techniques utilising animal behaviour to measure vision in rodents are limited, and mostly performed in the mouse.35 In peri-operative ION, onset has typically been upon awakening, but also as late as 24 to 48 h, mostly in sedated patients.36 We do not know the onset time of injury in this model. Conclusions should be applied cautiously to peri-operative ION in spine surgery as there is no rat prone position. Our results may not be applicable to peri-operative ION in another high-risk patient group, cardiac surgery, in which other systemic alterations including inflammation may be a component of the injury.1
In pigs,4 decreasing haematocrit to 15%, significant hypotension and occluding the jugular veins significantly decreased blood flow in the optic nerve, but damage to the nerve was not assessed. In our study, haematocrit was decreased about 40%, to approximately 25%, but we did not directly occlude the venous outflow. The influence of hypotension in this study should not be neglected and could have influenced the results, as MAP was below baseline at some point during experiments in all of the groups except for SUP, and was below baseline at the end of experiments only in TILT-HD.
We did not include surgical manipulation, specifically, surgery on the spine and anaesthesia was maintained with ketamine. Although the latter causes the overall procedure to differ from human spine surgery anaesthesia (in which a combination of intravenous and inhaled agents are typical), ketamine is known not to alter physiology significantly. The role of surgery itself (i.e. degree of invasiveness and stimulation etc) in the development of peri-operative ION has not been directly studied. It may be significant as surgically induced systemic inflammation exacerbated neuronal dysfunction in rodents with ischaemic injury.37
In summary, our study described a new means of inducing optic nerve injury that resembles peri-operative ION. Haemodilution (SUP-HD) or head-down position (TILT) were associated with a few changes in nSTR or VEP but to a lesser extent than in the TILT-HD group. This suggests that mild injury to the optic nerve under conditions of haemodilution or head-down tilt may become significant when prolonged for more than 5 h. One human study found transient alteration in visual field in seven patients with head-down positioning for robotic surgery of the prostate, showing that our study in rats may be relevant to assess the effect of prolonged mild retinal damage.38 Our study also raises concern that significant haemodilution, when maintained for lengthy periods, may not be tolerated by the optic nerve, which is also a novel finding.
It is important to interpret our results in the context that we studied healthy animals. The impact of pre-existing diseases such as diabetic retinopathy and other conditions such as glaucoma in which the optic nerve may be already compromised, as shown in recent retrospective studies,39 remains to be studied. As there is currently no medical strategy for regenerating the optic nerve,40 it is essential to develop reliable means to prevent peri-operative ION. This preparation may enable us to gain further insights into the mechanisms and preventive strategies for this peri-operative complication.
Acknowledgements relating to this article
Assistance with study: David Salek, B.S., Brian Savoie, B.S., Michael Alexander, B.S., Kelsey Y. Tupper, B.S. and Venkat Boddapati, B.S.
Financial support and sponsorship: SR: North American Neuro-ophthalmological Society (St Paul, Minnesota, USA), the Brain Research Foundation (Chicago, Illinois, USA), The Glaucoma Foundation (New York, New York, USA), the Michael Reese Foundation (Chicago, Illinois, USA) and National Institutes of Health (NIH, Bethesda, Maryland, USA) Grants EY10343, EY10343-15S1 and EY027447. No name: University of Chicago Institute for Translational Medicine (NIH RR024999), and NIH Visual Sciences Core Grant EY001792 to the University of Illinois. NJN: Research to Prevent Blindness, Inc. (New York, New York, USA), and NIH EY06360 to Emory University.
Conflicts of interest: SR has provided expert witness evaluation and testimony in cases of peri-operative visual loss on behalf of physicians, hospitals and injured patients.
Prior presentation: in part, at the Annual Meetings of the American Society of Anesthesiologists, and the North American Neuro-ophthalmological Society. Information regarding depositing article into PubMed Central: Research supported by National Institutes of Health grants RO1 EY10343, UL1 RR024999, P30 EY06360, P30 EY001792 and EY027447.
1. Shen Y, Drum M, Roth S. The prevalence of perioperative visual Loss in the United States: a 10-year study from 1996 to 2005 of spinal, orthopedic, cardiac, and general surgery. Anesth Analg
2. Lee LA, Roth S, Todd MM, et al. The Postoperative Visual Loss Study Group: Risk factors associated with ischemic optic neuropathy after spinal fusion surgery. Anesthesiology
3. Rubin DS, Parakati I, Lee LA, et al. Perioperative visual loss in spine fusion surgery: ischemic optic neuropathy in the United States from 1998 to 2012 in the nationwide inpatient sample. Anesthesiology
4. Lee LA, Deem S, Glenny RW, et al. Effects of anemia and hypotension on porcine optic nerve blood flow and oxygen delivery. Anesthesiology
5. Ribak CE, Shapiro LA, Perez ZD, et al. Microglia-associated granule cell death in the normal adult dentate gyrus. Brain Struct Funct
6. Morrison JC, Johnson EC, Cepurna WO, et al. Microvasculature of the rat optic nerve head. Invest Ophthalmol Vis Sci
7. Mathew B, Poston JN, Dreixler JC, et al. Bone-marrow mesenchymal stem-cell administration significantly improves outcome after retinal ischemia in rats. Graefe Arch Clin Exp Ophthalmol
8. Porciatti V, Feuer WJ, Monsalve P, et al. Head-down posture in glaucoma suspects induces changes in IOP, systemic pressure, and PERG that predict future loss of optic nerve tissue. J Glaucoma
9. Marshall-Goebel K, Terlevic R, Gerlach DA, et al. Lower body negative pressure reduces optic nerve sheath diameter during head-down tilt. J Appl Physiol
10. Gerlach DA, Marshall-Goebel K, Hasan KM, et al. MRI-derived diffusion parameters in the human optic nerve and its surrounding sheath during head-down tilt. NPJ Microgravity
11. Blecha S, Harth M, Schlachetzki F, et al. Changes in intraocular pressure and optic nerve sheath diameter in patients undergoing robotic-assisted laparoscopic prostatectomy in steep 45 degrees Trendelenburg position. BMC Anesthesiol
12. Schimmer RC, Urner M, Voigtsberger S, et al. Inflammatory kidney and liver tissue response to different hydroxyethylstarch (HES) preparations in a rat model of early sepsis. PLoS One
13. Dreixler JC, Barone FC, Shaikh AR, et al. Mitogen-activated protein kinase p38alpha and retinal ischemic preconditioning. Exp Eye Res
14. Hood DC, Birch DG. Phototransduction in human cones measured using the alpha-wave of the ERG. Vision Res
15. Pugh EN, Falsini B, Lyubarsky AL. Williams TP, Thistle AB. The origin of the major rod- and cone-driven components of the rodent electroretinogram and the effect of age and light-rearing history on the magnitude of these components. Photostasis and related phenomena
. Boston, MA, USA: Springer; 1998. 93–128.
16. Aleman TS, LaVail MM, Montemayor R, et al. Augmented rod bipolar cell function in partial receptor loss: an ERG study in P23H rhodopsin transgenic and aging normal rats. Vision Res
17. Waterhouse BD, Azizi SA, Burne RA, et al. Modulation of rat cortical area 17 neuronal responses to moving visual stimuli during norepinephrine and serotonin microiontophoresis. Brain Res
18. Mittag TW, Danias J, Pohorenec G, et al. Retinal damage after 3 to 4 months of elevated intraocular pressure in a rat glaucoma model. Invest Ophthalmol Vis Sci
19. Aizawa H, Yanagihara S, Kobayashi M, et al. The synchronous activity of lateral habenular neurons is essential for regulating hippocampal theta oscillation. J Neurosci
20. Dreixler JC, Shaikh AR, Shenoy SK, et al. Protein kinase C subtypes and retinal ischemic preconditioning. Exp Eye Res
21. Dreixler JC, Poston JN, Balyasnikova I, et al. Delayed administration of bone marrow mesenchymal stem cell conditioned medium significantly improves outcome after retinal ischemia in rats. Invest Ophthalmol Vis Sci
22. Bui BV, Edmunds B, Cioffi GA, et al. The gradient of retinal functional changes during acute intraocular pressure elevation. Invest Ophthalmol Vis Sci
23. Dreixler JC, Poston JN, Shaikh AR, et al. Delayed postischemic conditioning significantly improves the outcome after retinal ischemia. Exp Eye Res
24. Stevenson L, Matesanz N, Colhoun L, et al. Reduced nitro-oxidative stress and neural cell death suggests a protective role for microglial cells in TNFalpha−/−
mice in ischemic retinopathy. Invest Ophthalmol Vis Sci
25. Roth S, Li B, Rosenbaum PS, et al. Preconditioning provides complete protection against retinal ischemic injury in rats. Invest Ophthalmol Vis Sci
26. Aggarwal D, Tan O, Huang D, et al. Patterns of ganglion cell complex and nerve fiber layer loss in nonarteritic ischemic optic neuropathy by Fourier-domain optical coherence tomography. Invest Ophthalmol Vis Sci
27. Wang J, Mojumder DK, Yan J, et al. In vivo electroretinographic studies of the role of GABAC receptors in retinal signal processing. Exp Eye Res
28. He Z, Bui BV, Vingrys AJ. The rate of functional recovery from acute IOP elevation. Invest Ophthalmol Vis Sci
29. Yukita M, Machida S, Nishiguchi KM, et al. Molecular, anatomical and functional changes in the retinal ganglion cells after optic nerve crush in mice. Doc Ophthalmol
30. Perez de Lara MJ, Santano C, Guzman-Aranguez A, et al. Assessment of inner retina dysfunction and progressive ganglion cell loss in a mouse model of glaucoma. Exp Eye Res
31. Choh V, Gurdita A, Tan B, et al. Short-term moderately elevated intraocular pressure is associated with elevated scotopic electroretinogram responses. Invest Ophthalmol Vis Sci
32. Bernstein SL, Guo Y, Kelman SE, et al. Functional and cellular responses in a novel rodent model of anterior ischemic optic neuropathy. Invest Ophthalmol Vis Sci
33. You Y, Klistorner A, Thie J, et al. Latency delay of visual evoked potential is a real measurement of demyelination in a rat model of optic neuritis. Invest Ophthalmol Vis Sci
34. Heiduschka P, Julien S, Schuettauf F, et al. Loss of retinal function in aged DBA/2J mice – new insights into retinal neurodegeneration. Exp Eye Res
35. Grillo SL, Koulen P. Psychophysical testing in rodent models of glaucomatous optic neuropathy. Exp Eye Res
36. Lee LA, Roth S, Posner KL, et al. The American Society of Anesthesiologists Postoperative Visual Loss Registry: analysis of 93 spine surgery cases with postoperative visual loss. Anesthesiology
37. Degos V, Maze M, Vacas S, et al. Bone fracture exacerbates murine ischemic cerebral injury. Anesthesiology
38. Taketani Y, Mayama C, Suzuki N, et al. Transient but significant visual field defects after robot-assisted laparoscopic radical prostatectomy in deep tRendelenburg position. PLoS One
39. Rubin DS, Matsumoto MM, Moss HE, et al. Ischemic optic neuropathy in cardiac surgery: incidence and risk factors in the united states from the national inpatient sample 1998 to 2013. Anesthesiology
40. Goldberg JL, Guido W. AGIWP: Report on the National Eye Institute Audacious Goals Initiative: regenerating the optic nerve. Invest Ophthalmol Vis Sci
Supplemental Digital Content
© 2018 European Society of Anaesthesiology