We would like to thank Drs. Kempen, Raw, and Larson for their interest in our study on determining risk factors for perioperative ischemic optic neuropathy (ION) after spinal fusion surgery in the prone position.1Dr. Kempen's suggestion to perform spine surgery in the lateral position, instead of the prone position, is intriguing. We have also considered this possibility in the past and queried our surgical colleagues. In special situations, such as second or third trimester pregnancy, when postponement of surgery is not feasible, spine surgery has been performed in the lateral position. However, the “up-down” manipulations required in the lateral position are technically more difficult than the more symmetric “right-left” manipulations in the prone position. Achieving ideal spinal alignment is much more challenging technically in the lateral position. Many surgeons rely on the lordosis imparted by some of the spinal frames in the prone position to provide optimal “anatomic” alignment for fusion.
We agree with Dr. Kempen that further study should be performed to examine the relative risks and benefits of staging very prolonged spine surgery with expected high blood loss, as we noted on pages 22 and 23 of our article.1The supposition that this injury may reflect the coincidental occurrence of spontaneously occurring ION in a general nonsurgical population is not supported by national incidence data. Data from the Nationwide Inpatient Sample published by Shen et al. demonstrates a greatly increased odds ratio of 6.96 for developing ION in spinal fusion surgery compared with the referent abdominal surgery.2Its occurrence in children and relatively healthy adults after spinal fusion surgery is not consistent with the high incidence of atherosclerotic risk factors found in the subpopulation of nonsurgical patients who develop nonarteric anterior ION.3In addition, the high percentage of cases with bilateral profound loss of vision in perioperative ION is not consistent with the clinical course of nonarteritic anterior ION, which typically presents with unilateral disease with less severe loss of vision.
Lastly, we also considered the possibility, like Dr. Kempen, that vasopressors may be a contributory factor in the development of ION. We did not find a significant association with vasopressor use in the univariate analysis, as noted in table 1 on page 18 of our article.1
Dr. Raw offers the interesting hypothesis that perioperative inflammation may contribute to the development of ION after spinal fusion surgery. Given the low incidence of perioperative ION, it may prove difficult to examine this hypothesis with prospective studies. Dr. Raw also notes that increased plasma homocysteine levels and lower vitamin B6 levels are independently associated with the occurrence of nonarteric anterior ION in the nonsurgical population. For this reason, nitrous oxide could play a contributory role. In our 2006 study, less than one-quarter of the 83 patients with ION after spinal fusion received nitrous oxide, making it unlikely that nitrous oxide administration is an important factor in perioperative ION.4But a systemic inflammatory syndrome may result from prolonged, complex surgery, and the notion that inflammation is a pathogenic factor in axonal injury or brain injury is supported by experimental studies in animals.5,6
We appreciate Dr. Larson's continued interest in perioperative ION and his efforts to provide precise limitations on the amount of crystalloid (40 ml/kg) administered. We remain curious about the 40 ml/kg crystalloid limit. It is not clear if he is encouraging a practice of intentional hypovolemia, or if he is advocating the use of colloid along with crystalloid to maintain euvolemia, as recommended in the American Society of Anesthesiologists' practice advisory.7The former practice of intentional hypovolemia in these cases with large blood loss and prolonged duration would subject patients to a high potential for end organ ischemia, or ultimately, cardiovascular collapse. The latter practice of using colloids along with crystalloids may reduce the incidence of ION, as suggested by the results of our multicenter case-control study. However, our studies and understanding of the current literature do not suggest that a specific limit to crystalloid administration, such as 40 ml/kg, will prevent perioperative ION. Of note, the mean crystalloid infusion for the control patients in our study was 4.6 l ± 2.3 l, well above Dr. Larson's limit of 40 ml/kg for most patients. The highest amount of crystalloid infused in a control patient was more than 18 l. Conversely, crystalloid limitation did not protect all patients from ION, as the lowest amount of crystalloid infused in an ION case was 2.2 l. Based on our observations, we do not believe that the 40 ml/kg crystalloid limit prevents ION, nor does it help predict those who might develop ION. Dr. Larson's supposition that the increased risk of ION seen in men was because men received more crystalloid than women was not supported by our data. There were no significant differences in the amount of crystalloid received between men and women, either in cases or controls.
Although we agree that increased venous pressure is likely to increase blood loss and fluid resuscitation, there are many types of surgery where arterial bleeding results in much greater blood loss and fluid resuscitation, but without an associated risk for ION. Therefore, we believe that the increased venous pressure – in the head – is one of the most important risk factors placing prone spinal fusion surgery patients at increased risk for developing ION. This same feature of increased venous pressure in the head is also present in other surgical procedures that carry a high risk for ION groups: bilateral radical neck dissections with ligation of bilateral external and internal jugular veins8,9and laparoscopic/robotic prostatectomies with the head placed in steep Trendelenburg for prolonged duration.10It is interesting that Dr. Larson has “specialized” fluid management plans for these types of procedures with increased venous pressure in the head and high risk for ION, yet dismisses venous congestion as a significant contributory factor for ION.
We are impressed by Dr. Larson's efforts to prevent ION in robotic prostatectomy patients who require steep Trendelenburg position for 4–6 h. It is not clear to us if cases of that duration for this procedure are at risk for ION. The duration of surgery for the five reported cases of ION after laparoscopic prostatectomy ranged from 6.5 to 9.9 h, with four of these cases lasting 7.9 h or more.9,10One additional case of ION occurring after a laparoscopic proctocolectomy also lasted greater than 6 h.11We are not aware of cases of ION after 4–6 h of robotic prostate surgery that were associated with Trendelenberg position and 5–10 l of crystalloid administration.
We applaud Dr. Kempen's, Dr. Raw's, and Dr. Larson's interest in this topic and their efforts to minimize the occurrence of perioperative ION. This is a devastating perioperative complication that deserves continued reflection and sound, methodical investigation.