Ischemic Optic Neuropathy Following Spine Surgery

Case Control Analysis and Systematic Review of the Literature

Anshit Goyal, MBBS; Mohamed Elminawy, MBBCh; Mohammed Ali Alvi, MBBS; Timothy R. Long, MD; John J. Chen, MD, PhD; Elizabeth Bradley, MD; Brett A. Freedman, MD; Mohamad Bydon, MD


Spine. 2019;44(15):1087-1096. 

In This Article


Case Series

We found 12 patients eligible for inclusion from our institutional records with a mean age of 66.9 ± 11.2 years and 75% (n = 9) males with a mean body mass index of 29.3 ± 5.24. We stratified patients using the risk stratification criteria used by Sadda et al.[5] On an average, each patient had more than two vascular risk factors. Table 1 summarizes baseline characteristics and comorbidities of our cohort.

The majority of our cases (66.6%, n = 8) underwent a fusion. A few cases also underwent vertebrectomy (n = 3, 25%) and posterior decompression (n = 1, 8.3%). Almost all of our cases underwent multi-level surgery (≥4 levels in 58.3% and ≥3 levels in 92%). The median blood loss was 1.5 L (inter-quartile range: 0.1—10 L) while mean operative time was 561 ± 219 minutes (>9 h). Mean change in hemoglobin and hematocrit was 3.3 ± 1.5 g/dL and 9.9 ± 4.2%. The majority of our cases were operated at the thoracolumbar or lumbar levels (75%) while one patient received cervical and another received sacral surgery. While the majority of cases (10/12, 83%) were positioned prone, we also recorded a case with supine and another with circumferential (prone followed by supine) positioning. Postoperatively, facial/periorbital edema was recorded in 50% (6/12) of our cases. Table 2 summarizes perioperative parameters for our case series.

PION was the most common diagnosis (10/12, 83%) while the rest were AION. Symptom onset was within 24 hours postoperatively in 75% of cases (9/12). Afferent Pupillary Defect was noted in five cases (41%). Five cases (41%) were blind at initial presentation as per WHO grading of visual acuity. All cases had a recorded visual field defect (VFD) while color vision deficit was observed in 10 cases (83%). Bilateral symptoms were observed in the majority of the cases (75%, n = 9). Table 3 summarizes the postoperative ophthalmic examination.

Follow-up ophthalmologic examination was available in 10 (83%) patients. All patients were followed at 3 months from diagnosis. Disc pallor was observed in all cases at follow-up. Visual field defect had improved in three (30%) and unchanged in five (50%) patients. Visual acuity as defined by WHO grading remained unchanged in four (40%) patients and improved in three (30%) of patients. Color vision deficit remained unchanged in 70% (n = 7), improved in 20% (n = 2), and worsened in 10% patients (n = 1). Overall outcome as defined by change in either visual acuity, field defect or color vision deficit was improved in three (30%) patients, unchanged in four (40%), and worsened in three (30%) of cases. Table 4 highlights individual patient data on perioperative parameters and ophthalmologic examination at presentation and follow-up.

Case–control Comparison

We compared each of our cases to four randomly selected controls matched by age and year of surgery (Table 5). Using conditional logistic regression, we found that cases with ION were more likely to have undergone a fusion (odds ratio [OR]: 23.8, confidence interval [CI]: 2.94–192.40, P = 0.003), with more number of vertebral levels operated (unit OR: 2.99, CI: 1.42–6.32, P = 0.004), sustained higher blood loss (OR: 1.003, CI: 1–1.005, P = 0.02), and decrease in perioperative hemoglobin (OR: 2.28, CI: 1.20–4.34, P = 0.01), and hematocrit (OR: 1.32, CI: 1.07–1.65, P = 0.01). While the majority of cases (n = 10, 83.3%) underwent a fusion procedure, only 18.75% (n = 9) of controls received a fusion. Similarly, 92% (n = 11) of cases with ION underwent multiple level surgery with more than or equal to three operated vertebral levels compared with only 12.5% (n = 6) of controls. There were no significant differences observed in terms of sex, comorbidities, number of vascular risk factors, and type of positioning (prone vs. non-prone).

Literature Search

A total of 182 patients were identified from 42 studies, of which 62 were from individual case reports, 37 from a case series by Myers et al[12] and 83 from the ASA Postoperative Visual Loss Registry.[2] Data obtained from individual studies were heterogeneous due to variability in reporting. Supplemental Tables 1–3, summarize the patient level data from these reports including baseline characteristics and comorbidities, perioperative parameters, final diagnosis, and outcome.

Mean age at the time of surgery was 48.3 ± 14.5 years with 100/182 (55%) of patients being males. The majority of cases reported in the literature underwent a fusion (94.6%, 160/182). Very few patients received single level surgery (n = 23, 19%). Prone positioning was the most common (156/165, 95%). The median ranged from 2000 to 3500 mL while the mean operative time was 497 ± 180 minutes. Compared with our series, mean change in hematocrit and hemoglobin was larger (25% and 5 g/dL, respectively). Very few cases reported on presence or absence of facial edema and it was found to be positive in 62% (18/29) of cases. Most cases (105/176, 58.5%) were blind (WHO grade IV) and had bilateral deficit (51%, 93/182) at initial presentation, 92% (22/24) of patients also had a color vision deficit. The onset of visual disturbance was within 24 hours after surgery in most cases (72%, 32/50).

No follow-up data were available in 15 patients. Assessment of final visual field and acuity was recorded in for 52 patients while follow-up color vision was available for 26 patients. Optic disc pallor was noticeable in the majority of cases (36/43, 84%) on follow-up assessment, consistent with the pathophysiology. While visual field and color vision was found to be unchanged in 62% (32/52) and 58% (15/26) of cases respectively, 48% (25/52) of cases demonstrated improvement in visual acuity. Overall, the outcome remained unchanged from baseline in 55% (101/182) of cases.

Factors Associated With Severity of Presentation

We pooled results from our own case series and the literature review to identify factors associated with severity of visual disturbance at initial presentation and improvement at follow-up. Age, sex, presence of vascular risk factor, blood loss, and operative time were not found to be associated with blindness at initial presentation or odds of improvement in overall vision outcome. PION patients were found to be more likely to be blind at initial presentation (OR: 6.45, CI: 1.04–54.3, P = 0.04). No difference in odds of improvement was however, observed between patients presenting with PION versus AION.