Circulation 84,122 • Volume 23, No. 4 • Winter 2008   Issue PDF

Reader is Poised to Rethink Ischemic Optic Neuropathy in Light of “POISE” Study

Paul M Kempen, MD, PhD; Lorri A. Lee, MD

To the Editor

Several authors presented significant pertinent information in the Spring 2008 ASPF Newsletter, which deserve consideration together as very significant interrelated themes regarding patient safety. Dr. Lehner described the impact and course of his personal battle with posterior ischemic optic neuropathy (PION).1 Dr. Lee summarized contemporary recommendations for management in avoiding PION, while pointing to the only 2 identified associated factors: long and sanguine surgery.2 I was particularly surprised that Dr. Lee failed to reference or discuss her recently published findings from May 2008, incriminating hypovolemic deliberate hypotension using beta blockade (labetalol) and anemia as specific liabilities to optic nerve metabolism in a porcine model.3 While deliberate hypotension and hemodilution are anesthetic procedures per se, they are typically also performed at surgical request/direction and thus may be more rationally events requiring surgical rather than anesthetic consent. The recently published PeriOperative ISchemic Evaluation (POISE) study results deserve mention here and did clearly confirm the cardioprotective effects of perioperative beta blockade (PBB)—a message not to be ignored. However, the cost was that of increased morbidity and mortality from stroke in the metoprolol treated group. Clearly, the doses of metoprolol used in POISE were fixed and quite aggressive (100 mg oral bid or 15 mg q 6 hr iv), with metoprolol held only when systolic blood pressures or heart rate remained <100 mmhg or 50 bpm, respectively.4 An atherosclerotic patient presenting preoperatively with untreated hypertension of 170/95 (mean of 120 mmHg), heart rate of 95 bpm, and a known LV ejection fraction of 30% might well be expected to have an adverse outcome, when allowed to remain at 100/35 (mean 56 mmHg) at 45 bpm, and with a, now, unknown acute drug-induced depression of ejection fraction. Under this prescribed significant dosing regimen, it may be surprising that POISE could be completed, and the increased morbidity documented may well be that of drug overdose. Thus, Drs. Kleinman and Corey’s observations regarding perioperative hemodynamics, goals, and consequence of drug administration are especially pertinent and timely, raising the serious question of just what goals we should accept in light of these very recent studies.5,6 While contemporary patients have become increasingly polymorbid, and while the anesthetic ability to manipulate hemodynamic parameters has been increasing to extreme degrees, the continued use of “historic hemodynamic guidelines” (when patients were healthier and surgeries shorter) may be a significant liability. The custom of accepting mean pressures of 60 (or worse 50) mmHg under modern anesthesia appears questionable for short periods in the polymorbid, and apparently can become disastrous even in “otherwise typically normal” individuals. With normally 120/80 and mean pressures of 90-100 mmHg, the historic acceptance of 60 mmHg mean values under anesthesia (derived from healthy animal studies demonstrating autoregulation and minimal urine production) also becomes suspect, especially in the comorbid and geriatric patient.7 A 20% deviation from the “usual or normal” baseline (i.e., to a mean of 70-80 mmHg?) may prove a much more rational, safe, and physiologic intraoperative goal. Introducing acute beta blockade will reduce cardiac output as well as blood pressure, while modern PBB recommendations promote judicious introduction over weeks to adequately monitor for side effects and facilitate remodeling of the cardiovascular system. In spite of the availability of multiple modern methods to quantify cardiac output, they are typically not utilized in spine surgery, with this important parameter remaining “unknown.” Support of blood pressure with alpha agonists is also commonplace to prevent tachycardia and may further compromise overall substrate delivery to important, marginal tissues, by increasing systemic vascular resistance or reducing systemic venous capacitance/compliance at any measured CVP.8 Should striving for “normal” intraoperative hemodynamics perhaps receive renewed interest as a factor to eliminate PION? Anesthetists must now decide if participation in induced hypotension or anemia, and to what degree, is really in the patient’s best interests.

Imperative questions arise:

  1. Who has the obligation for discussions regarding perioperative visual loss?
  2. Is the planned use of hypotension or hemodilution desirable, and to what degree? Does this also require specific informed consent?
  3. Can the use of CVP vs. pulmonary catheter monitoring to avoid hypovolemia become an effective standard measure to prevent hypovolemia and blindness, particularly given perturbations in intrathoracic pressures? How does morbid obesity or beta blockade affect these variables?
  4. Will vasopressor support, the choice of vasopressor, and the target blood pressure be protective or merely shift morbidity to other organ systems?
  5. How does liability relate to the allowed degree of anemia and what is the minimal sampling interval to determine intraoperative hematocrit?
  6. Just how long should a prone procedure be allowed to persist and will “staging” of surgical procedures only increase cost, infection, morbidity, and mortality by methods of repeated insult?
  7. Have hypotension and clinical anemia remained unidentified as important factors because of a widespread acceptance of marginally low values as a false definition?
  8. Given the frequent discordance of oscillotonometric vs. transduced arterial blood pressures in many patients, does the arterial trace represent a consistently reliable measure or must interval correlation to non-invasive blood pressure (NIBP) measured values become an important standard to accurately assess blood pressure over prolonged periods?
    Modern banked blood and conservation/cell salvage techniques are quite safe and effective. Perhaps it is time to define and stress “normal” or “optimal” over “acceptable” and “desired” induced hemodynamic parameters intraoperatively. Are the surgical requests to suppress patient hemodynamic parameters to extreme margins, creating significant stresses with only questionable benefit in reducing blood loss, really acceptable? Does this approach only serve to promote the prolonged, more extensive, and sanguine surgical trespass, in which PION is promoted? Should such prolonged/extensive surgeries be reserved for senior and efficient surgeons, rather than relegated to trainees under limited supervision? While hemodilution can be safely maintained when adequate filling pressures are insured, hypovolemia is often the nature of such surgical bleeding and can compromise systemic oxygen delivery. Perhaps the historical optimum hemoglobin level of 10 gm may be more appropriate in sanguine situations with sudden, episodic loss.
  9. Would maintenance of (more) normal hemodynamic parameters eliminate PION? We as professionals stress a preoperative evaluation to insure patients are optimized for surgery, recognizing the cardiovascular depressant effects of anesthetic agents. Should we then actively strive for marginal hemodynamic states intraoperatively, and if so, why, to what degree, and with what goals? Are iatrogenic blindness and stroke acceptable alternatives to transfusion risks and myocardial infarction, respectively? Are we asking the right questions? Should patients be asking more? Research may not readily yield an answer on PION, with the rarity of occurrence.

Paul M Kempen, MD, PhD Wexford, PA


  1. Lehner AD. If my spine surgery went fine, why can’t I see? Postoperative vision loss and informed consent. APSF Newsletter 2008;23:1-3.
  2. Lee LL. Solutions to POVL mystery requires research. APSF Newsletter 2008;23:3. ChloraPrep™ labeling (Cardinal Health). Accessed on August 28, 2008.
  3. Lee LA, Deem S, Glenny R, Townsend I, Moulding J, An D, Treggiari MM, Lam A. The effects of anemia and hypotension on porcine optic nerve blood flow and oxygen delivery. Anesthesiology 2008;108:864-72.
  4. POISE Study Group. Effects of extended-release metoprolol succinate in patients undergoing non-cardiac surgery (POISE trial): a randomised controlled trial. The Lancet 2008; 371:1839-1847.
  5. Kleinman B. Reader calls attention to change from baseline pressure. APSF Newsletter 2008;23:10.
  6. Corey PW. Labetolol affects hemodynamics. APSF Newsletter 2008;23:10.
  7. Guyton AC, Hall JE. Chapter 19, Textbook of Medical Physiology. 9th ed. Philadelphia: WB Saunders; 1996.
  8. Gelman S. Venous function and central venous pressure: a physiologic story. Anesthesiology. 2008;108(4):735-48.

In Reply:

Dr. Kempen provides some very thoughtful and astute comments regarding perioperative management of complex surgical spine procedures. He asks the numerous salient questions in the minds of most anesthesiologists and anesthetists who care for these patients with respect to ischemic optic neuropathy (ION). Numerous specialized anesthetic management techniques such as deliberate hypotension, hemodilution, and fluid restriction with resultant hypovolemia have converged in spine surgery. As spine operations have become much longer in duration and with larger blood loss, anesthesiologists must be clinically astute to assess whether or not their patient is an appropriate candidate for any of these techniques. Combinations of these techniques, or use of a single technique for a prolonged duration in the prone position may exhaust compensatory mechanisms designed to maintain adequate perfusion to end organs.

However, the lack of any randomized controlled trial or other clinical study demonstrating an association between hypotension, anemia, and perioperative ION limits the ability of any guideline or advisory-setting group to rationally devise any evidence-based parameters for hemodynamic management and transfusion threshold. Seventeen percent of patients with ION in the ASA POVL Registry had their nadir hematocrit>30%.1 Choosing a transfusion threshold>30% would require speculation with significant potential for increased risk and without any guarantee of benefit. Using non-primate animal studies as the sole supporting evidence when differences in vascular anatomy exist is less than ideal. The effects of anemia and hypotension on porcine optic nerve oxygen delivery may only apply to a small subset of patients with similar blood supply to their optic nerve, and perhaps only under these very severe physiologic stresses.2 Though the benefit of using deliberate hypotension is dubious, and is not a technique that this author advocates, condemnation of the practice in association with perioperative ION is also not supported by the literature.3 There are many anesthesiologists and anesthetists who have utilized deliberate hypotension in prone spine surgery for decades, reportedly without significant complications—albeit without adequate power to detect an influence on ischemic optic neuropathy. Making changes in clinical care without evidence-based medicine can cause harm in 2 ways. First, harm can occur because of the intervention itself—e.g., maintaining a higher mean arterial pressure with phenylephrine could mask hypovolemia and result in other end organ ischemia to the kidneys or heart as Dr. Kempen notes in his letter. The immunosuppressive and infectious risks of blood transfusion are well known. Second, it may divert attention away from the actual cause of the problem. Recall that most anesthesia professionals, surgeons, and ophthalmologists were undeniably certain that pressure on the globe was responsible for all POVL after spine surgery without cortical strokes. It was perfectly logical—made perfect sense. However, it took a very long time to collect enough data to refute this misperception, and then continue the search for the actual cause of ischemic optic neuropathy (ION). In the meantime, many anesthesia professionals were wrongly accused of being negligent in protecting the eyes in the prone position.

Dr. Kempen is correct that randomized clinical trials for perioperative ION are unlikely to be accomplished soon, because they would require an enrollment of an extremely large number of subjects for sufficient power. The ongoing multicenter case control study matching cases from the ASA POVL Registry to controls who underwent similar procedures, but did not develop ION, may identify specific risk factors. One hopes that results from this study will be able to determine whether maintenance of a specific blood pressure or hematocrit decreases the risk of developing ischemic optic neuropathy. Dr. Kempen and many other anesthesia professionals may well be proven correct.

Lorri A. Lee, MD Seattle, WA


  1. 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 2006;105:652-659.
  2. Lee LA, Deem S, Glenny R, et al. The effects of anemia and hypotension on porcine optic nerve blood flow and oxygen delivery. Anesthesiology 2008;108: 864-872.
  3. Myers MA, Hamilton SR, Bogosian AJ, et al. Visual loss as a complication of spine surgery: a review of 37 cases. Spine 1997;22:1325-1329.