VISUAL loss after nonocular surgery is a rare but devastating perioperative complication. 1It has been documented in a wide variety of procedures, the most common of which are cardiopulmonary bypass, head and neck operations, and prone spine operations. 2Ischemic optic neuropathy is the most common diagnosis in these cases. Factors believed to be associated with its occurrence are large intraoperative blood loss, long duration in the prone position, and intraoperative hypotension and anemia. 3,4We present a case of unilateral posterior ischemic optic neuropathy occurring after an uneventful prone lumbar spine operation in a relatively healthy man who did not have decreased blood pressure and hematocrit perioperatively.
A 58-yr-old, 80-kg man presented for an L2–L3 posterior spinal instrumentation and fusion. The patient had a 22-pack-year history of tobacco, but had stopped smoking at the age of 36 yr. He had undergone radiation therapy for Graves disease 10 yr previously but was not noted to be proptotic at the time of presentation. Preoperative hematocrit was 50% (high end of normal at our institution), and baseline blood pressure was 134/92 mmHg (mean arterial pressure [MAP], 106 mmHg). A 12-lead electrocardiogram showed normal sinus rhythm with left ventricular hypertrophy and left atrial enlargement. Fentanyl (3 μg/kg), propofol (2.5 mg/kg), and lidocaine (0.5 mg/kg) were used for induction of anesthesia. Rocuronium (1 mg/kg) was administered for muscle relaxation before tracheal intubation. Isoflurane (0.8–0.9% end-tidal) in a 50:50 air:oxygen mixture and a sufentanil infusion (0.2–0.3 μg · kg−1· h−1) were used for anesthetic maintenance. The patient was turned prone onto a Wilson frame with his head supported solely by a soft foam cushion with a cutout for the eyes, nose, and mouth. The head was slightly dependent with an approximate 15° tilt from the longitudinal axis of the body. After turning prone, a low-dose phenylephrine infusion (0.01%) was administered for the majority of the 6.5-h case and was titrated at a rate of 1.67 μg · kg−1· min−1or less to maintain an MAP of predominately 80–90 mmHg (range, 70–95 mmHg), despite adequate volume replacement. At the time of initiation of the phenylephrine infusion, 2,000 ml crystalloid had been administered to the patient (calculated fluid deficit, 1,440 ml), and urine output was 600 ml. The requirement for phenylephrine was thought to result from loss of vascular tone during general anesthesia. The eyes were checked approximately every 30 min throughout the procedure. The patient was turned supine after approximately 320 min in the prone position and was noted to have an extremely edematous face. He underwent extubation uneventfully after confirmation of a leak around his endotracheal tube with the cuff deflated and was transported to the recovery room. Estimated blood loss for the case was 800 ml, and urine output was 700 ml. Eight thousand milliliters crystalloid was administered intraoperatively. No blood products were administered. The hematocrit at the end of the procedure was 40% in the operating room and 39% in the recovery room.
That evening, the patient was noted to be comfortable, with intact lower extremity motor and sensory function. Early the next morning, the patient reported decreased vision in his left lateral field to the orthopedic service for the first time. He had noted blurry vision in the recovery room the day before, but did not verbalize any complaints because he thought it was caused by residual anesthesia. An ophthalmology consult was obtained. There were no signs of facial bruising or trauma. Fundoscopic examination yielded normal results bilaterally. An afferent pupillary defect was present in the left eye, and visual acuity was decreased to 20/800 compared with 20/20 on the right. Intraocular pressures were 18 mmHg bilaterally. Laboratory workup revealed a normal erythrocyte sedimentation rate and negative syphilis serologies. A magnetic resonance image of the head, performed approximately 24 h after surgery, was read by the radiologist as a probable lesion in the left posterior optic nerve on the T2-weighted images, but it could not be confirmed because of lack of enhancement. Fine sections through the optic nerve were not obtained because magnetic resonance imaging was performed on a “stroke protocol.” Re-review of the image with a neuroradiologist did not reveal any definitive lesions. Visual evoked potentials and electroretinography were not performed. The final diagnosis was posterior ischemic optic neuropathy. No treatment was recommended by the ophthalmologic consultants.
At the time of discharge, the visual acuity in his left eye had improved to 20/400. At 3 weeks’ follow-up, the patient’s ophthalmologic examination revealed a pale optic disc and a persistent relative afferent pupillary defect in the left eye. Color vision was significantly decreased in the left eye, and the visual field was still restricted with an inferotemporal quadrant defect, but central visual acuity had improved to 20/60. Six months after the injury, the left eye had improved central visual acuity to 20/25 and improved color vision but remained with a persistent temporal field deficit.
Symptomatic visual defects have been reported to occur as infrequently as 1 in 60,965 anesthetic procedures for nonocular surgery, 1but the incidence in selected operations, such as cardiac surgery, may be as high as 4.5%. 5Asymptomatic retinal microemboli have been documented in up to 100% of patients undergoing cardiopulmonary bypass procedures. 6Other procedures that account for a large proportion of the cases of perioperative visual deficits are head and neck surgery (e.g. , sinus surgery and radical neck dissections) and spine surgery in the prone position. 2The etiology of the injury to the eye is unknown in most cases but is probably multifactorial, and a number of potential risk factors have been implicated. 2,4Although prolonged direct compression of the globe can cause blindness, it is not the usual cause. Blindness has occurred after operations in the supine position and after prone positioning with the head in Mayfield pins. 7Visual injury due to direct compression of the globe is caused by retinal vascular occlusion, 2which can also occur as a result of emboli and possibly increased intraocular pressure. 2,4The most common diagnosis in patients in whom perioperative visual deficits develop after nonocular surgery is ischemic optic neuropathy. 2Ischemic optic neuropathy is categorized as either anterior or posterior, depending on the location of the lesion in the optic nerve. Retinal vascular occlusion and cortical blindness are less common diagnoses. Procedure-related factors that have been implicated in this complication are emboli, extreme hypotension of prolonged duration, direct pressure on the globe, large blood loss, massive transfusion of blood and fluid, and anemia. 2–5However, some cases of perioperative visual injury, such as the current case, do not have any of these speculated associated factors, except perhaps for a large amount of intravenous fluids.
A number of case reports and case series of perioperative visual complications occurring after spine surgery in the prone position have been published. 2,3,7,8Myers et al. 3compared a group of 28 patients in whom visual deficits developed after prone spine surgery to a group of 28 controls matched for age, type of surgery, approach, number of spinal levels being fixed, instrumentation, and primary versus revision surgery. Factors that were significantly different between groups in the study of Myers et al. 3were operative time (mean 430 min in the blindness group vs. mean 250 min in the control group) and estimated blood loss (mean 3,600 ml in the blindness group vs. mean 880 ml in the control group). Despite these significant differences in mean operative time and estimated blood loss between groups, most patients who undergo spine surgery in the prone position for a prolonged duration with a large estimated blood loss do not wake up with visual deficits. Therefore, long operative time and large estimated blood loss may place patients at higher risk for development of visual deficits, but they are not absolute determinants for this complication. Clearly, other factors, intrinsic either to the patient (e.g. , ocular vascular anatomy, coexisting illnesses) or to the operation (e.g. , positioning on frames, type and amount of fluid and blood replacement), may have important contributory roles in the development of postoperative visual deficits.
The operation described in this case report had a prone duration intermediate between the times for the blindness group and the control group in the study of Meyers et al. , 3and the estimated blood loss in this case report was approximately the same as in the control group of Myers et al. 3The blood pressure was maintained within 25% of baseline, and the eyes were checked frequently to ensure there was no direct pressure on the globes. This patient may have had mild untreated hypertension based on his preoperative diastolic blood pressure of 92 mmHg and his electrocardiogram, which showed left ventricular hypertrophy and left atrial enlargement. Graves disease is known to cause proptosis, which theoretically may place a patient at higher risk for visual loss, but his thyroid had been irradiated 10 yr before this operation, and he was not proptotic. Although his preoperative hematocrit of 50% was in the high normal range, he was hemodiluted with crystalloid to a hematocrit of 40% by the end of the procedure. This makes the possibility of high viscosity causing decreased blood flow an unlikely explanation.
To what extent the phenylephrine infusion, the 8,000 ml crystalloid, or the positioning on the Wilson frame may have contributed to this complication is not clear. The effects of phenylephrine on the vascular supply to the eye have not been studied, and the available literature is contradictory as to whether vasoconstrictors are beneficial or harmful. 9,10It has been speculated that posterior ischemic optic neuropathy may be caused by increased central venous pressures in the prone position and venous congestion in the orbit of patients who are in the prone position for a long duration while receiving large amounts of fluid. 2,8Many of the orthopedic frames (e.g. , Relton-Hall and Wilson) are designed so that the head is in a dependent position to the body, which decreases venous outflow from the head. The prone position is also known to cause an increase in intraocular pressures (IOP), 11which may decrease the ocular perfusion pressure (PPop) by the equation PPop= MAP − IOP.
The current case is unusual because it lacked excessive intraoperative blood loss, hypotension, or anemia. Although the patient may have had mild hypertension, his degree of vasoocclusive disease does not fit the usual description of ischemic optic neuropathy that occurs spontaneously in the community. 12If a similar patient were to present to the hospital today for the same procedure, he or she would not be considered at high risk for developing this complication. This suggests that, despite the apparent increase in the incidence of perioperative visual deficits over the past 5–10 yr, we do not know the precise etiologic factors, nor do we know how to prevent it. Other factors that may never be discernible preoperatively, such as unique vascular anatomy or ocular hemodynamics, may be the most important risk factors for developing perioperative visual deficits.