Population-based Incidence and Outcomes of Compressive Optic Neuropathy

Abstract

Purpose:

Prior studies on compressive optic neuropathy (CON) have come from large tertiary centers, which may contain referral bias toward more severe or atypical disease. To our knowledge, there are no studies to determine the population-based etiologies and clinical outcomes of compressive optic neuropathy (CON). This study aims to bridge that gap using the Rochester Epidemiology Project database.

Design:

Retrospective, population-based cohort

Methods:

Medical records of all residents living in Olmsted County, Minnesota from January 1, 2000 through December 31, 2018 were screened for CON. Demographic and clinical information were collected before and after treatment.

Results:

23 patients had a confirmed diagnosis of CON during our study period, which provided an overall incidence of 1.14 per 100,000 per year. Average age at onset of CON was 51 years (SD 24), and 39% were male. The most common etiologies were pituitary adenoma and meningioma. There was significant improvement in visual fields (p < 0.003) but not in visual acuity (p = 0.08) after patients underwent treatment for CON. There was also a significant relationship between the time until treatment and the degree of visual field improvement at follow-up (Pearson correlation rho = −0.58, p < 0.047).

Conclusion:

To our knowledge, this study provides the first population-based incidence of CON. The finding that earlier treatment leads to better visual outcomes stresses the importance of having CON on the differential diagnosis of patients with optic neuropathy.

Table of Contents Statements

The Rochester Epidemiology Project was used to conduct a retrospective chart review for all residents living in Olmsted County, Minnesota from 2000 through 2018 to determine the incidence, population-based etiologies, and clinical outcomes of compressive optic neuropathy (CON). 23 patients had a confirmed diagnosis of CON, which provided an incidence of 1.14 per 100,000 per year. There was a significant relationship between the time until treatment and the degree of visual field improvement at follow-up.

Introduction:

Compression anywhere along the optic nerve can lead to compressive optic neuropathy (CON), which can cause symptoms such as visual field loss, loss of color perception, and decreased visual acuity. Common neoplastic etiologies of CON include meningiomas, intracranial aneurysms, pituitary adenomas, craniopharyngiomas, and gliomas of the anterior visual pathway.¹ Non-neoplastic etiologies, such as thyroid eye disease, can also cause optic nerve compression.² Progression of vision loss in CON is typically gradual and progressive, but also can be rapid and dramatic, such as in cases of an aneurysm. Thus, it is essential to correctly identify the cause of CON, as the differential is broad, and clinical course as well as treatment varies accordingly.

Studies on outcomes and/or etiologies of CON have mostly come from large tertiary centers, which could potentially be biased toward less common or more severe cases.^3-8 To our knowledge, there are no population-based incidence studies of CON to determine the etiologies and outcomes. This study aims to bridge that gap by using the Rochester Epidemiology Project database to evaluate the incidence, etiologies and clinical outcomes of CON.

Methods:

Medical records of all patients residing in Olmsted County, Minnesota were screened for CON from January 1, 2000 through December 31, 2018 for this retrospective chart review. Patients of interest were identified and retrospectively reviewed using the Rochester Epidemiology Project, which is a is a medical records linkage system that links and indexes diagnostic and procedure information from Mayo Clinic in Rochester, Minnesota and other facilities that provide health care to residents in this community.^9-12 The total number of patients screened was based on the initial search for any optic neuropathy diagnosis codes that could potentially be compressive optic neuropathy (optic neuritis, optic neuropathy, optic atrophy, compressive optic neuropathy, papilledema, and papillitis). A total of 1,070 patients were initially identified in this manner. Then, these patients were subsequently screened for a confirmed CON diagnosis and to ensure that they were residents of Olmsted county at the time of diagnosis in our study criteria timeline.

This cohort study was approved by the Institutional Review Boards of the Mayo Clinic and Olmsted Medical Center. This study conforms to the requirements of the U.S. Health Insurance Portability and Accountability Act and the Declaration of Helsinki.

Electronic and paper medical records were used to confirm each diagnosis of CON for this retrospective chart review. CON was diagnosed using MRI, CT, or biopsy findings, and patients were only included if their first episode of CON was diagnosed during our study period and were a resident of Olmsted County at the time of diagnosis. Patients were excluded if the original diagnosis was made outside of this study window or if they were a non-Olmsted county resident. For each patient, demographic information, including age, sex, and race was collected. Clinical markers, including duration between symptom onset and diagnosis, duration between diagnosis and treatment, visual acuity, visual fields, disc appearance, presence of eye pain or headache, and treatments were also collected. Clinical outcomes were measured using the same markers as clinical presentation, but during a follow-up appointment that took place closest to 1 year after treatment. This time period allowed for post-operative visual stability while reducing confounding variables such as the development of cataracts, particularly in older patients.

Since one of the clinical hallmarks of CON is reduced visual fields and/or visual acuity, those markers were our key indicators of disease progression and response to treatment. Visual fields were primarily measured using automated static perimetry tests, where mean deviations were recorded for quantitative analysis. Kinetic visual field testing that provided visual field patterns were used for qualitative, pattern analysis. Best corrected Snellen visual acuities were converted to logMAR for statistical calculations. To compare overall amount of visual field loss, mean deviations were measured in decibels (dB) using standard automated perimetry and compared before and after treatment.

Incidence was calculated using the population denominators from the Olmsted population during the study period for this study. Population figures were used from the census populations, and individual years between census periods were linearly interpolated to provide the yearly populations to calculate the incidence across the study period. The Olmsted population in 2010 was 86% white; thus, the incidence rates were adjusted for age and sex to the 2010 U.S. population of whites, as has been done for prior Rochester Epidemiology Project incidence studies. ¹³ Two-sided P values of < 0.05 were considered statistically significant. Analysis was completed using MATLAB 2020b software (MathWorks, Natick, MA).

Results:

A total of 23 Olmsted residents had a confirmed diagnosis of CON during our study period from January 1, 2000 to December 31, 2018 (Table 1). The overall age and sex adjusted incidence was 1.14 per 100,000 per year (95% confidence interval [CI], 0.67-1.62). The median age of CON diagnosis was 55 years (interquartile range [IQR], [37 65] – Figure 1), with 61% of patients being female. Seventy eight percent of cases were diagnosed using magnetic resonance imaging (MRI), while 22% were diagnosed using computed tomography (CT).

Table 1 -.

Clinical Characteristics of Compressive Optic Neuropathy Based on Etiology.

Etiology		Age of diagnosis (mean / median if different)	Gender	Race	Eyes affected	Presenting VF pattern	Diagnosis Modality	Treatment	Average change in visual acuity (logMAR)	Average change in visual fields (mean deviation , dB)	Median time from symptom onset to results (days)	# of Patients with headache	# of Patients with eye pain	# of Patients Recurrence(s)
Tumor	Adenoma	62/67	Male - 6 female - 2	White - 5 Black – 2 Asian – 1	OU - 8	bitemporal hemianopsia - 6 junctional scotoma - 1 N/A - 1	MRI - 7 CT - 1	ETSS - 7 None - 1	−0.24	11.02	125	3	2	2
	Craniopharyngioma	42	Male - 1 female - 1	White - 1 Asian - 1	OU - 2	bitemporal - 2	MRI - 2	Craniotomy - 2	−0.136	−4.98	312	1	0	0
	Germinoma	14	Male - 0 female - 2	White - 1 Asian - 1	OU - 2	bitemporal hemianopia - 2	MRI - 1 CT - 1	Chemotherapy, radiation - 2	−0.699	15.805	438	2	0	0
	Glioma	6	Male - 0 female - 1	White - 1	OU - 1	N/A - 1	MRI - 1	Craniotomy, chemotherapy, and radiation - 1	−0.125	N/A - 1	N/A - 1	0	0	1
	Hemangioma	53	Male -1 female - 0	White - 1	OD - 1	cecocentral scotoma - 1	MRI - 1	Radiation - 1	0	3.35	1352	0	0	0
	Meningioma	45/45	Male - 0 female - 4	White - 4	OU - 2 OD - 2	left homonymous - 1 paracentral scotoma - 1 advanced – 1 N/A - 1	MRI - 4	Craniotomy + radiation - 1 Craniotomy + radiosurgery - 1 Radiation - 2	−0.03	5.90	609	2	0	2
NonTumor	Aneurysm	60/58	Male - 0 female - 3	White - 3	OU - 1 OS - 2	arcuate - 1 mild generalized depression - 1 N/A - 1	MRI - 2 CT - 1	Clip - 1 Coil embolization Surpass embolization - 1	−0.063	0.49	532	3	1	0
	Grave's	60	Male - 0 female - 1	Asian - 1	OS - 1	advanced - 1	CT - 1	Steroids - 1	0.097	13.5	49	0	0	0
	IOI	95	Male - 1 female - 0	White - 1	OD - 1	N/A - 1	CT - 1	Steroids - 1	−0.602	N/A - 1	33	0	0	0

ETSS = Endoscopic Transsphenoidal Surgery. IOI = Idiopathic orbital inflammation. N/A = missing parameter in patient chart.

Figure 1 -.

Age distribution of compressive optic neuropathy diagnosis.

Among the 23 patients who had CON, eight (35%) had pituitary adenomas, four (17%) meningiomas, three (13%) aneurysms (right middle cerebral artery and internal carotid artery, left internal carotid artery, and left cavernous carotid aneurysm), two (9%) germinomas, two (9%) craniopharyngiomas, and one each from thyroid eye disease, idiopathic orbital inflammation (IOI), hemangioma, and glioma (4% each) (Figure 2). Eighteen (78%) of the 23 cases of CON were due to intracranial tumors, while five (22%) cases were not tumor related. The individual incidence rates for each etiology can be found in Table 2.

Figure 2 –

Source of compressive optic neuropathy among 23 diagnosed patients in Olmsted County from January 1, 2000 to December 31, 2018. IOI = idiopathic orbital inflammation.

Table 2 –

Incidence of Compressive Optic Neuropathy Based on Etiology.

Etiology	N	Overall age and sex adjusted incidence per 100,000 per year (95% CI)
Total	23	1.14 (0.67, 1.62)
Tumor	18	0.87 (0.46, 1.28)
Adenoma	8	0.43 (0.13, 0.74)
Craniopharyngioma	2	0.10 (0.00, 0.25)
Germinoma	2	0.07 (0.00, 0.18)
Glioma	1	0.04 (0.00, 0.11)
Hemangioma	1	0.05 (0.00, 0.15)
Meningioma	4	0.17 (0.00, 0.35)
Non-tumor	5	0.27 (0.03, 0.51)
Aneurysm	3	0.16 (0.00, 0.33)
Grave’s	1	0.06 (0.00, 0.17)
IOI	1	0.06 (0.00, 0.17)

IOI = Idiopathic orbital inflammation.

Visual field patterns were analyzed and compared before and after treatment. Six (75%) of patients with pituitary adenomas presented with a bitemporal hemianopsia, while one (12.5%) presented with a junctional scotoma (the remaining pituitary adenoma patient did not have visual fields taken upon presentation). All patients with craniopharyngiomas presented with a bitemporal hemianopsia. The remaining CON lesions produced a variety of visual defect patterns, including cecocentral scotomas and generalized depression.

The mean presenting logMAR visual acuity was 0.51 ± 0.40 (Snellen equivalent of 20/65), and the final logMAR visual acuity was 0.28 ± 0.37 (Snellen equivalent of 20/40). The change in visual acuity (−0.23 logMAR) was not significantly different (p = 0.08). However, there was a significant improvement in mean deviation in visual fields after treatment (+7.95 dB, p = 0.003). The average presenting mean deviation in visual fields was −15.76 ± 7.14 dB and the average final mean deviation after treatment was −7.94 ± 5.53.

Among the different etiologies, patients with germinomas had the most improvement in both visual acuity and visual fields after treatment with a median of approximately 3 lines of improvement on the Snellen chart (ΔlogMAR −0.699) and a 15.81 dB improvement in mean deviation on visual fields. In the other etiologies, visual acuity for adenoma patients improved by approximately 1 line (ΔlogMAR −0.240); craniopharyngioma and glioma by approximately < 1 line (ΔlogMAR −0.136 and −0.125 respectively); IOI by approximately 2-3 lines Δ(logMAR −0.602); and hemangioma, meningioma, aneurysm, and thyroid eye disease by less than 0.1 ΔlogMAR (ΔlogMAR 0, −0.03, and −0.063 respectively). Average visual fields improved by 10.96 dB for adenoma patients, 3.35 dB in hemangiomas, 2.87 dB in meningiomas, 0.49 dB in aneurysm, 13.5 dB in thyroid eye disease, and decreased by −4.98 in craniopharyngiomas. Pre and post-operative visual fields were not measured in the IOI or the glioma patients.

Patients with pituitary adenomas were all treated with endoscopic transsphenoidal resection except for one elderly patient who expired prior to surgery. One pituitary adenoma patient had tumor recurrence after 10 years, causing recurrent CON and requiring multiple repeat surgeries and radiation therapy. Similarly, one meningioma patient experienced tumor recurrence three years after craniotomy + radiation with repeat CON and received antineoplastic treatment prior to passing away. The one glioma patient also experienced tumor recurrence without significant visual field change 8 years after her initial craniotomy. After treating the recurrence with radiation, the tumor size stabilized with no repeat CON. The remaining tumors were treated with a combination of craniotomy, chemotherapy, and/or radiation and did not have recurrent CON.

Eye pain and/or headache was a variable associated symptom among the etiologies of CON. All patients who had CON secondary to an aneurysm experienced headache (3 out of 3), and one experienced additional eye pain (1 out of 3).

Among the patients who had their optic discs examined before and after treatment, 10 subjects (53%) had optic disc pallor before treatment and that number increased to 11 (58%) after treatment.

The median time elapsed between symptom onset and diagnosis date for all subjects was 90 days. There was a significant correlation between the time elapsed from symptom onset to treatment and the degree of visual field improvement; the earlier the treatment, the greater the visual field improvement (Pearson correlation rho = −0.58, p <0.047 – Figure 3). The correlation between time elapsed and improvement in visual acuity, however, was not significant (Pearson correlation rho = 0.31, p = 0.24 – Figure 4). A meningioma patient and a craniopharyngioma patient were excluded in the Pearson correlation analyses due to lack of information about the time elapsed between symptom onset to diagnosis date.

Figure 3 -.

Relationship between time until compressive optic neuropathy treatment and corresponding change in visual field defects, measured using mean deviation (dB). Pearson correlation coefficient −0.58 (p < 0.05). IOI = idiopathic orbital inflammation.

Figure 4 -.

Relationship between time until compressive optic neuropathy treatment and corresponding change in visual acuity. A negative change in logMAR indicates improvement in visual acuity. IOI = idiopathic orbital inflammation.

Discussion:

Our population-based study evaluated the etiologies and clinical outcomes of patients with CON, providing an incidence of 1.14 per 100,000 per year. Seventy-eight percent of the cases were neoplastic, with the most common etiologies being benign tumors (pituitary adenomas and meningiomas), which supports findings discussed in previous literature from non-population based studies.¹⁴

We found a significant improvement in visual fields (p < 0.003) but not in visual acuity (p = 0.08) after patients underwent treatment for CON. Given that visual acuity can be normal or impaired depending on whether the central visual field is affected,¹ the larger improvement in visual fields we observed likely reflects the relatively high percentage of patients presenting with chiasmal compression where visual fields are more affected than visual acuity at presentation. An earlier study also found rapid recovery of central visual fields within 15 days post-operation in mixed-pathology CON.¹⁵ In contrast, a prior study on surgical outcomes for CON found only 24% of eyes had improvement in visual fields after optic nerve decompression, which likely represents a higher percentage of sphenoid wing meningiomas that are harder to fully resect.¹⁶ There is conflicting evidence on the change in visual acuity after CON treatment; one study found that 53% of eyes with CON (36 patients with surgical decompression of 36 optic nerves) resulted in the same or worse visual acuity at the last follow-up compared to preoperative values.¹⁶ Meanwhile, an earlier study found that over 90% of eyes with CON (15 eyes of 11 patients) that was treated with craniotomies improved in both visual fields and visual acuity immediately after surgery.³

We found a significant relationship (rho = −0.58, p < 0.05) between the duration of symptoms prior to treatment and improvement in visual fields after treatment. However, there was no significant relationship between duration of symptoms prior to treatment and change in visual acuity. This is in agreement with separate studies that also found no significant relationship between duration of preoperative symptoms and the change in visual acuity achieved after surgery.^16,17 Unfortunately neither study tested for a significant relationship between change in visual fields and duration of preoperative symptoms. The significant relationship between change in visual fields and duration of symptoms but not between change in visual acuity and treatment supports the notion that visual fields is a more reliable marker of improvement for CON and that early identification and treatment of CON may lead to better visual field outcomes.

Patients with germinomas had the most improvement in visual acuity and visual fields after treatment (approximately 3 lines of improvement on the Snellen chart and a +15.81 dB improvement in visual fields). This may have been due to their relatively younger age of diagnosis (median age 14 years) compared to the median age of 58 years for the other etiologies of CON in our study. Prior studies suggest that younger patients have better outcomes for a variety of optic neuropathies, including optic neuritis, Leber hereditary optic neuropathy, and others.^18,19

Limitations of our study include the homogeneity of the Olmsted County cohort, which was 86% white in 2010, as well as the small sample size, with 23 subjects identified by the Rochester Epidemiology Project over the 18-year period. However, it is important to note that we chose not to include all patients with CON seen at Mayo Clinic during this time. Though this would increase our sample size, it would also have risked introducing referral bias to our study and thus skewed the disease severity, presence of complications, and other factors.²⁰ Because of this, we believe that our estimates of CON are more representative of the general population. This study provides clinicians the epidemiological data on CON, thus providing the frequency of underlying etiologies, and provides the severity and outcomes from a population-based cohort of CON. This can help guide the workup, with expedited identification and treatment of CON associated with better outcomes, and provides a framework to discuss the expected prognosis of CON.