Ocular Deformations in Spaceflight-Associated Neuro-Ocular Syndrome and Idiopathic Intracranial Hypertension

Abstract

Purpose: Spaceflight-associated neuro-ocular syndrome (SANS) shares several clinical features with idiopathic intracranial-hypertension (IIH), namely disc edema, globe-flattening, hyperopia, and choroidal folds. Globe-flattening is caused by increased intracranial pressure (ICP) in IIH, but the cause in SANS is uncertain. If increased ICP alone causes SANS, then the ocular deformations should be similar to IIH; if not, alternative mechanisms would be implicated.

Methods: Using optical coherence tomography (OCT) axial images of the optic nerve head, we compared “pre to post” ocular deformations in 22 patients with IIH to 25 crewmembers with SANS. We used two metrics to assess ocular deformations: displacements of Bruch's membrane opening (BMO-displacements) and Geometric Morphometrics to analyze peripapillary shape changes of Bruch's membrane layer (BML-shape).

Results: We found a large disparity in the mean retinal nerve-fiber layer thickness between SANS (108 um; 95% confidence interval [CI] = 105-111 um) and IIH (300 um; 95% CI = 251-350.1 um). The pattern of BML-shape and BMO-displacements in SANS were significantly different from IIH (P < 0.0001). Deformations in IIH were large and preponderantly anterior, whereas the deformations in SANS were small and bidirectional. The degree of disc edema did not explain the differences in ocular deformations.

Conclusions: This study showed substantial differences in the degree of disc edema and the pattern of ocular deformations between IIH and SANS. The precise cause for these differences is unknown but suggests that there may be fundamental differences in the underlying biomechanics of each consistent with the prevailing hypothesis that SANS is consequent to multiple factors beyond ICP alone. We propose a hypothetical model to explain the differences between IIH and SANS based on the pattern of indentation loads.

In 2011, Mader et al.¹ described the spaceflight-associated neuro-ocular syndrome (SANS) in crewmembers who completed long duration space flights to the International Space Station (ISS). The principal signs include optic disc edema, acquired hyperopia, flattening of the globe, and choroidal folds. The placement of an optical coherence tomography (OCT) device on the ISS has made it possible to detect subclinical ocular changes based on an increase in the peripapillary retinal nerve fiber layer thickness (RNFLT), peripapillary total retinal thickness (TRT), and choroidal thickness. It has also enhanced our detection of choroidal and retinal folds as well as deformations of the optic nerve head.^2–6 That several crewmembers had borderline or slightly elevated intracranial pressures (ICPS) post-flight, led to the notion that headward fluid shifts in space might elevate the ICP and result in a disorder similar to idiopathic intracranial hypertension (IIH). However, certain features appear to be inconsistent with elevation of the ICP alone, namely the absence of headaches, other symptoms of intracranial hypertension (ICH), choroidal congestion, and the high frequency of choroidal folds, among others. Although the precise mechanism of SANS is uncertain, the prevailing view is that it is a multifactorial consequence of headward fluid shifts.^1,6,7 

Globe flattening, in IIH, is consequent to an elevated ICP; the cause in SANS is less certain. Recent studies have used a variety of quantitative methods to assess globe flattening using OCT and magnetic resonance imaging (MRI).^8–14 A morphometric comparison of the deformations in SANS and IIH might provide insights into the underlying biomechanics and the role of ICP in SANS. If SANS was simply caused by an elevation in ICP alone, then the ocular deformations should be similar to those that occur in IIH, and, if not, then alternative mechanisms would be implicated. This study compares changes in peripapillary shape of Bruch's membrane layer (BML) and displacements of Bruch's membrane opening (BMO) in crewmembers on long duration space flights to patients with papilledema caused by IIH. 

Methods

Subjects

Two data sets were used for this retrospective comparison: (1) crewmembers who completed long duration space flights on the ISS comparing data gathered preflight and inflight; and (2) patients with IIH with papilledema before and after successful treatment. 

We reviewed the ophthalmic examinations and OCT images of 36 crewmembers collected for medical monitoring and approved for use by the NASA Johnson Space-Center Lifetime Surveillance of Astronaut Health Advisory Board, Institutional Review Board, and Human Research Multilateral Review Board. Crewmembers provided informed consent consistent with the Declaration of Helsinki. The diagnosis of SANS was defined by at least one of the following criteria in either eye: (a) ophthalmoscopic evidence of optic disc edema, (b) folds (e.g. peripapillary wrinkles, outer or inner retinal folds, or choroidal folds),^15,16 or (c) an increase in peripapillary TRT, within an elliptical annular region 250 µm from the BMO,² that statistically exceeded day to day variability of the TRT by ≥20 µm.^3,4 Twenty-five of 36 (69%) patients met the inclusion criteria for SANS. There were 21 men and 4 women with a mean age of 46.6 ± 6.3 years and an average body mass index (BMI) of 25.5 ± 2.6 kg/m² (see the Table). 

Table.

 

Summary of Patient Demographics, Time Intervals, Clinical Signs, and OCT Metrics in Crewmembers With SANS and Patients With IIH

 

The records from a clinical neuro-ophthalmology service database were retrospectively reviewed for patients with a complete set of OCT images who met the updated modified Dandy criteria for IIH¹⁷ with papilledema between 2017 and 2019. All were successfully treated with acetazolamide or topiramate or both. We identified 22 patients (21 women and 1 man) who met these criteria. Mean age was 28.0 ± 9.3 years, average BMI was 35.7 ± 7.1 kg/m², and mean opening pressure was 38.1 ± 8.4 cm H₂O (opening pressure in one patient was recorded by the neurologist as “very high”; see the Table). This part of the study was approved by the Stony Brook Institutional Review Board, conformed with the US Health Insurance Portability and Accountability Act and the Declaration of Helsinki, and qualified for waiver of consent (“per 45CFR46.116.d; CORIHS 2015-1233-R2”). 

To determine if the degree of optic disc edema might influence the ocular deformations, we decreased the disparity in swelling by comparing two subgroups: (1) patients with IIH with “mild papilledema” defined by a Frisén Grade ≤2 and a mean RNFLT of ≤200 um (9 subjects, 16 eyes), and (2) crewmembers with “advanced SANS” that included those with choroidal folds, funduscopic disc edema, or a change (Δ) in TRT ≥40 um (16 subjects, 25 eyes). We also compared BML-shape deformations between those crewmembers with anterior BMO-displacements (>+18.5 um; 12 subjects, 18 eyes) to those with posterior BMO-displacements (<−18.5 um; 15 subjects 24 eyes). 

To maintain comparative terminology, the “pre” condition (without optic disc edema) refers to the baseline preflight optic nerve head (ONH) in SANS and resolved papilledema in IIH (as an approximation of the ONH before the development of papilledema). The swollen ONH will be considered the “post” condition (inflight in SANS, at presentation in IIH). 

Optical Coherence Tomography

For shape and displacement analysis, pre and post 20 degrees transverse axial images from radial or volume scans centered on the optic nerve head from both eyes were acquired with Spectralis OCT1 or OCT2 systems (Heidelberg Engineering, Heidelberg, Germany) 8 to 16 automatic real-time tracking (ART) levels, enhanced depth imaging (EDI), with the TruTrak-AutoRescan feature enabled to ensure follow-up registration and alignment for both cohorts. To eliminate vertical stretching of the image for shape analysis, the Z axis (scaled at 3.87 µm/px) was adjusted to the individual's X axis scale (1:1 um). 

The details on the methodology used to assess the RNFLT, TRT, and choroidal thickness using the Spectralis OCT among crewmembers with SANS have been previously described.^2–4 Briefly, TRT was quantified using a radial scan pattern centered on the optic nerve head. RNFLT was quantified using approximately 3.5 mm diameter circular scan pattern centered on the ONH. Automated segmentations of the internal limiting membrane (ILM), retinal nerve fiber layer (RNFL), and Bruch membrane (BM) were manually corrected and verified. Peripapillary TRT was calculated from a radial scan pattern centered on the ONH as the average distance between the ILM and BM layers within an elliptical annular region extending from the margin of the BMO to 250 µm, whereas RNFLT was calculated as the average difference between the ILM and RNFL layers along an approximately 3.5 mm diameter circular scan centered on the ONH. Optic disc edema originates at the prelaminar ONH and spreads radially as the edema progresses; therefore, the TRT was measured adjacent to the ONH to facilitate detection of mild edema that has not extended out to the circular path used to quantify RNFLT. The TRT is more reliably segmented than the RNFLT in this region and thus more sensitive in detecting subclinical disc edema. 

Peripapillary RNFLT for patients with IIH were obtained on the Cirrus OCT (Zeiss-Meditech, Dublin, CA, USA; version 5.0.0) with their proprietary RNFLT analysis mode, which uses a 200 × 200 optic disc cube that extracts 256 A scan samples along a calculation circle (3.46 mm diameter) centered on the optic disc. Scan quality of >7/10 was required. Measurements of TRT on patients with IIH were not available and thus not included in the analysis. The follow-up OCT for crewmembers was obtained between 76 and 181 days in flight (mean 154 ± 24 days, or 5.1 ± 0.8 months). Follow-up OCT for patients with IIH to resolution averaged 430 ± 255 days (14.3 ± 8.5 months). 

BMO-Displacement

Using transparent overlays (Adobe Photoshop, Adobe Systems Inc., San Jose, CA, USA), pre and post images were registered by vertically aligning the BMO and superimposing the nasal and temporal reference landmarks on BML, each located 2 mm from the BMO (Fig. 1). To ensure accurate manual placement of reference points at the outer border of Bruch's membrane layer, the images were vertically stretched three times. BMO-displacements were measured at the midsection of BMO. Following the placement of the reference points and BMO-displacements, the images were returned to their unstretched state. Left eye images were flipped horizontally to align the nasal and temporal sides with the right eye. To determine the reproducibility of BMO-displacements, we examined 2 transverse axial OCTs obtained on normal eyes of 10 subjects (8 women and 2 men, mean age 34 ± 10 years) at different times (2–12 months apart). We repeated the assessment independent of the original measurements for a total of 20 comparisons. The difference in the position of the BMO with repeated measurements ranged between 0 and 19 µm with an absolute mean of 8.1 ± 5.2 µm (99% confidence interval [CI] = 4.8–11.5 µm). We considered any displacement between plus two standard deviations of the mean (that is between −18.5 µm and +18.5 µm) to be insignificant and presumably related to measurement variability or physiological fluctuations. 

Figure 1.

 

Twenty degree transverse axial optical coherence tomography (OCT) showing the placement of 20 equidistant “semi-landmarks” (landmarks that outline a curved structure) along Bruch's membrane layer (BML) in a patient with resolved papilledema (top, yellow points, “pre”-like) and papilledema (middle, red points, “post”). Both images (bottom inset) were superimposed and registered by aligning Bruch's membrane opening (BMO) along the vertical white lines and superimposing the reference points (a and b) located 2 mm from the BMO. Positional displacement of the BMO is measured at the midsection (bidirectional white arrow) of the yellow and red lines.

Geometric Morphometric Shape Analysis

Geometric morphometrics (GMs) is a method of analyzing morphological shape and its covariation with other variables. GM defines shape as the property that remains after normalizing for differences in position, scale, and rotation (Procrustes superimposition).^18–20 Conventional permutation or multivariate statistical methods can be used to assess differences in shape variables. We used the TPS software suite¹⁹ to analyze our data. The application of GMs to peripapillary shape deformations of the ONH has been detailed in several earlier studies.^13,14,21,22 Only a brief description specific to this study follows. 

After image registration (described above), using imaging software (Adobe Photoshop), we applied a semitransparent 2 mm square grid overlay to position 10 equidistant “semi-landmarks” (defined as points positioned on a curved structure) spanning a distance of 2 mm on both sides of the BMO along the posterior border of the BML (see Fig. 1). To ensure accurate manual placement of semi-landmarks we vertically stretched the image (3:1 um) after which the image was returned to its unstretched state (1:1 um) for shape analysis. The statistical analysis of shape variations compares the sums of squared Procrustes differences between and within the samples expressed as an F-ratio. The evaluation compares the observed F-value to an empirical distribution based on 10,000 random permutations of the group memberships of the individuals for the groups being compared. The proportion of Goodall's F statistics from these permutations (within blocks) equal to or larger than the observed Goodall's statistic is interpreted as the “P value” for the test (i.e. the probability that a random assignment of the individuals to the groups could show group differences as large or larger than those observed using the known groupings to avoid correlation bias between eyes within an individual, all statistical analyses on shape were performed on the right eye only. Principal component analysis (PCA) was used to analyze the change in shape (Procrustes distances) between pre and post images for each cohort. The application of PCA to analyze peripapillary shape changes has been described elsewhere.^13,18,22 

OCT parameters (TRT, BMO-displacement, RNFLT, Procrustes distance, and choroidal thickness) were analyzed with Generalized Estimating Equation (GEE) models to address the repeated eyes within subjects and robust standard errors to allow for non-homogenous variance. Group (IIH or SANS) was included as categorical fixed effect. Changes corresponding to each group were estimated, tested, and compared through the marginal means of these models. Relationships between OCT change measures were conducted by modeling one in terms of the other as a covariate. For these models, the group categorical variable was modeled as both a main and interaction effect, allowing for modification of both slope and intercept within the models. Statistical association was assessed through the modification of the slope coefficient (interaction term). The frequency distribution of categorical variables was compared using Fisher's exact probability test when N ≤300, otherwise using the chi-square test for independence. 

Results

The differences between crewmembers with SANS and patients with IIH with respect to gender, age, and BMI were all statistically significant (P < 0.0001; see the Table). The difference in the degree of papilledema based on mean RNFLT between pre and post images in patients with IIH was much greater than crewmembers with SANS (P < 0.0001). Only six (24%) of the SANS crewmembers had mild ophthalmoscopically evident optic disc edema. An additional 19 of 25 (76%) of SANS cohorts had subclinical disc edema (based on ΔTRT ≥20 µm) in at least one eye. Six of the 25 (24%) crewmembers with SANS had choroidal folds, whereas 3 of the 22 (14%) patients with IIH had choroidal folds (Fisher's exact test, P = 0.47) that met the criteria for this analysis.^2,6 

BMO-Displacements

The magnitude and direction of BMO-displacements for each group are shown in Figure 2. In patients with IIH with papilledema, the BMO was preponderantly displaced anteriorly (positive, toward the vitreous) in 80% of the eyes, 11% were posteriorly displaced with a mean displacement of +104.8 µm (95% CI = 62.5–147.1 µm); and 9% of eyes did not exceed the ± 18.5 µm threshold for meaningful displacement (“nil”). In contrast, the BMO-displacements among eyes of SANS crewmembers were bidirectional (36% displaced anteriorly, 48% posteriorly, and 16% “nil”), with a mean of -2.8 µm (95% CI = −25.6 to 19.9 µm). There was a difference in the ratio of anterior to posterior displacements between IIH and SANS (Fisher's exact test, P < 0.001). BMO-displacement between eyes was directionally concordant in 21 of 25 (84%) of the crewmembers with SANS and 21 of 22 (95%) patients with IIH (Fisher's exact test, P = 0.19). 

Figure 2.

 

Displacements of Bruch's membrane opening (BMO, from pre to post condition) for each eye sorted in order of magnitude and direction. A positive shift along the ordinate indicates that the “post” BMO-position relative to the “pre” was displaced toward the vitreous, negative shifts are displaced away from the vitreous. (A1) Includes the eyes from the entire idiopathic intracranial hypertension (IIH) cohort; (A2) is a subgroup of A1 with “mild papilledema” (i.e. Grade 1–2, mean RNFL ≤200 µm); (B3) includes the eyes from all of the crewmembers with spaceflight-associated neuro-ocular syndrome (SANS); (B4) is a subgroup of B3 with more “advanced SANS” (i.e. optic disc edema, choroidal folds, or TRT ≥40 um). Grey bar along the abscissa marks a “nil” zone with BMO-displacements between 18.5 to +18.5 µm considered to be insignificant. The inflection between anterior and posterior displacements is marked by a small grey vertical line on the abscissa. The associated table summarizes the probabilities, frequencies, and per cent ratios of anterior to posterior BMO-displacements for each group. BMO, Bruch's membrane opening; IIH, idiopathic intracranial hypertension; SANS, spaceflight-associated neuro-ocular syndrome.

We minimized the disparity in swelling by comparing BMO-displacements in two subgroups: (1) the subgroup of IIH with “mild papilledema” had a mean RNFLT of 163.4 µm (95% CI = 142.1–184.6 µm), and (2) the subgroup with “advanced SANS” crewmembers had a mean RNFLT of 111.3 µm (95% CI = 106.9–115.6 µm). Although the magnitude of the BMO-displacements decreased (see Fig. 2A) to 61 µm (95% CI =  10.8 – 109.6 µm; P = 0.02) in those with “mild papilledema,” the ratio of anterior to posterior displacements was essentially the same. For SANS (see Fig. 2B), there was no difference between the entire cohort and the “advanced” subgroup with respect to both the mean BMO-displacement (P = 0.87) or the ratio of anterior to posterior displacement within the group (P = 0.59). 

BML-Shape Deformations

Figure 3 summarizes the consensus (or mean) BML-shape deformations in patients with IIH (see Figs. 3a, 3b) and SANS (Figs. 3c–f). Those patients with papilledema (“post condition,” solid red line; see Fig. 3a) were symmetrically deformed anteriorly (toward the vitreous) relative to the consensus shape after the papilledema resolved (“pre-condition”; black dotted line; [permutation statistic, P < 0.001]). Among the subgroup with “mild papilledema” (see Fig. 3b), there was a decrease in the magnitude of the BML-shape deformation (permutation statistic, P = 0.04), but the pattern remained the same. 

Figure 3.

 

Geometric morphometric consensus (mean) shape deformations of Bruch's membrane layer (magnified 5 times) among patients with idiopathic intracranial hypertension (IIH) and crewmembers with spaceflight-associated neuro-ocular syndrome (SANS). (a) Consensus shapes of patients with IIH and papilledema (red solid line, “post”) and resolved papilledema (black dashed line, “pre”). (b) Consensus shapes of a subgroup with “mild” (Grade 1–2, mean RNFL <200 µm) papilledema (solid red line) and resolved mild papilledema (dashed black line). (c) Consensus shape comparison of crewmembers with SANS preflight (dashed black line) and inflight (solid red line). (d) Preflight-inflight consensus shapes from a subgroup of crewmembers with “advanced SANS” (only those with choroidal folds, disc edema, or Δ TRT >40 µm). (e) Preflight-inflight consensus shapes from those crewmembers’ eyes with anterior BMO-displacements (>18.5 µm). (f) Preflight-inflight consensus shapes from those crewmembers’ eyes with posterior BMO-displacements of <−18.5 µm.

The BML-shape deformations in crewmembers with SANS are shown in Figures 3c to 3f. Compared to the baseline preflight consensus shape (black dashed line), the inflight deformation was small and asymmetrically tilted temporally (see Fig. 3c; posteriorly on the temporal side, anteriorly on the nasal side, [permutation statistic, P = 0.02]). The subgroup with “advanced” SANS (see Fig. 3d) exaggerated the temporal tilt (permutation statistic, P = 0.01) but again did not change the basic pattern of deformation. Subgroup comparison of crewmembers with anterior BMO-displacements (see Fig. 3e) had a BML-shape deformation that was nasally skewed anteriorly (permutation statistic, P = 0.01), and those with posterior BMO-displacements (see Fig. 3f) had a shape that was slightly asymmetrically deformed posteriorly (permutation statistic, P = 0.001). The consensus shape (see Fig. 3c) of the entire cohort is a summation of subgroups 3e and 3f. PCA of the shape changes are summarized and discussed in Supplementary Figure S1. It is noteworthy that the pattern reflected in PC2 closely mirrors the BMO-displacements described above. 

Relationships Between Ocular Deformations and Other OCT Parameters

In patients with IIH, eyes with greater degrees of papilledema, based on ΔRNFLT, had larger anterior displacements of the BMO (see Supplementary Fig. S2a; P < 0.02). There was a strong correlation between BMO-displacements and changes in shape (expressed as a ΔProcrustes distance; see Supplementary Fig. S2b; P = 0.004) but no relationship between the change in shape and ΔRNFLT (see Supplementary Fig. S2c; P = 0.27). Among crewmembers with SANS there was an association between ΔRNFLT and shape change (see Supplementary Fig. S2f; P = 0.0002) but not between BMO-displacements and ΔRNFLT (see Supplementary Fig. S2d; P = 0.37) or between BMO-displacements and shape change (see Supplementary Fig. S2e; P = 0.40). There was a statistically significant positive relationship between change in ΔTRT and several other OCT parameters, including the ΔRNFLT (see Supplementary Fig. S2g; P < 0.0001), BMO-displacement (see Supplementary Fig. S2i; P = 0.01), Δ shape (see Supplementary Fig. S2h; Procrustes distance; P = 0.0003; see Supplementary Figs. S2a–S2i). 

Relationships between Δ choroid thickness and ocular deformations were also investigated for crewmembers with SANS. There was a significant relationship between ΔTRT and Δ choroid thickness (Δ choroid thickness = 21.08 + 0.1928 ΔTRT, R² = 0.11, P = 0.01). The Δ choroid thickness was inversely related to BMO-displacement (ΔBMO = 20.2805 – 0.8010 Δ choroid thickness, R² = 0.09, P = 0.01) but not ΔProcrustes distance (P = 0.72). 

Case Examples

Supplementary Figure S3 showing infrared images of the optic disc (top row) and corresponding 20 degrees transverse axial OCT images (bottom row) through the center of the optic disc illustrating the differences in ocular deformations between IIH and SANS. Yellow (arrows, lines, and text) indicates the baseline “pre” condition (preflight in SANS and resolved papilledema in IIH); red indicates the “post” condition (inflight in SANS, swollen in IIH). Supplementary Figure S3a shows a patient with IIH, high-grade papilledema, and a large anterior BMO-displacement and BML-shape deformation. Supplementary Figure S3b shows a crew member with SANS and subclinical optic disc edema with similar anterior deformations but smaller compared to IIH. Supplementary Figure S3c shows the infrared image of a crew member with inflight mild optic disc edema, supero-nasal choroidal folds, and radially oriented inner retinal folds temporally. The image also shows a subtle widespread anterior indentation of the peripapillary retina. The flashing vertical arrows and horizontal lines in the transverse axial OCT image below shows inflight thickening of the prelaminar ONH, expansion of the choroid, deformation of BML and posterior displacement of the BMO. 

Discussion

The most pronounced differences between crewmembers with SANS and patients with IIH was the disparity in the degree of optic disc edema and the dissimilarities in ocular deformations (i.e. BMO-displacements and peripapillary BML-shape deformations). Although the precise cause is unknown, these differences suggest that SANS is not entirely explained by an elevation in ICP alone and that the underlying biomechanics causing optic disc edema and globe flattening in each disorder are different. 

The installation of an OCT on the ISS has provided a unique opportunity to observe the earliest subclinical stages of optic disc edema by comparing normal preflight images to inflight images of the ONH in otherwise healthy crewmembers.^3,4,23 An increase in the TRT, minimum rim width, and decrease in the cup volume are some of the earliest measurable signs of a “subclinical” optic disc edema.^2–5 Because the optic disc edema that develops in SANS is so mild, it oftentimes may not reach the eccentricity of the circle scan used to measure mean RNFLT. Whereas the RNFLT is an effective metric for diagnosing and managing papilledema in IIH,^24,25 the peripapillary ΔTRT appears to be more sensitive in identifying and monitoring subclinical or mild optic disc edema and its covariation with other OCT parameters in SANS.^2–4,6 Corroborating earlier studies,^3,7,26 17% (6/36) of crewmembers developed clinical signs of SANS that increased to 69% (25/36) using OCT criteria (ΔTRT) over a mean inflight time of 5.3 months (159 ± 37 days). 

We found a strong correlation between ΔTRT and ΔRNFLT in SANS, but for each µm increase in the mean Δ RNFLT there was an 8.33 µm increase in the mean Δ TRT. The ΔTRT correlated with several other structural parameters including BMO-displacement and change in shape (ΔProcrustes distance). 

This study also showed that the magnitude and pattern of ocular deformations (both BML-deformations and BMO-displacements) between crewmembers with SANS and patients with IIH are different (see Figs. 2, 3). BMO-displacements in IIH were preponderantly anterior (toward the vitreous) whereas crewmembers with SANS were bidirectionally displaced from anterior to mostly posterior displacements (see Fig. 2). Similarly, Figure 3 shows that the IIH cohort displayed a relatively large symmetrical anterior shape deformation, whereas the consensus (mean) shape in SANS was characterized inflight by a small asymmetric temporal tilting (temporally posterior and nasally anterior). The consensus shape for crewmembers with SANS appears to be a composite of two shape patterns. Subjects with anterior BMO-displacements (≥18.5 um) had an asymmetric anteriorly deformed peripapillary BML-shape skewed nasally toward the vitreous whereas those with posterior BMO-displacements (≤−18.5 um) had a slightly asymmetrical posterior BML-shape deformation. 

Reducing the disparity in swelling by comparing a subgroup of IIH with “mild papilledema” to crewmembers with large increases in TRT (“advanced SANS”) did not alter the basic patterns of ocular deformations in either cohort (see Figs. 2, 3). Thus, it appears that differences in deformations between IIH and SANS may not be entirely explained by the inequality in the degree of optic disc edema alone. 

In general, deformations of the ONH and peripapillary tissues are consequent to a combination of loading forces and conditions, structural anatomy, and tissue material properties.^27,28 The principal loads include the translaminar difference between the intraocular pressure (IOP) and perioptic CSF pressure, the retrolaminar and prelaminar ONH tissue pressure, and the orbital tissue pressure. The mechanical response will also depend on the stiffness (compliance) of the load bearing lamina cribrosa, sclera, and dura.^29–34 Experimental studies on mechanical indentation of pressurized spherical shells have shown that both the properties of the spherical shell itself and the characteristics of the indentation load (e.g. magnitude, indentation depth, indenter shape, and indenter contact zone) are also critical factors in determining the size and pattern of deformations.^35–39 

Papilledema is known to be a consequence of an elevated ICP extending into the perioptic subarachnoid compartment that increases pressure within the retrolaminar tissue and scleral flange.^{29,30,32,40,41} In addition to obstructing axoplasmic flow that distends the prelaminar axons,⁴⁰ it also imposes a large concentrated retrolaminar indentation load that anteriorly displaces the lamina cribrosa and deforms the peripapillary tissues, flattening the globe.^{13,14,29,30,42,43} 

The precise cause of the anterior deformations among a subgroup of crewmembers is unknown; however, the pattern is similar to patients with IIH albeit smaller in magnitude. ICP has not been measured in space; however, post-flight lumbar punctures in some cases have shown borderline or mildly elevated ICP.¹ The rise in ICP among crewmembers was thought to be caused by headward fluid shifts that increased jugular venous congestion, obstructing cranial venous outflow.^{1,7,23,44–46} Post mission MRI studies have also shown indirect signs of elevated ICP.^47–49 Parabolic flight⁵⁰ and head down tilt and bedrest studies^51,52 have shown a mild increase in ICP compared to the upright position at 1G. Moreover, mild disc edema and folds can develop in normal individuals subjected to 30 to 60 days of strict head-down tilt bedrest.^53,54 Lawley⁵⁰ has suggested that the ICP in space is not pathologically elevated to levels commonly seen in IIH, but instead crewmembers develop a mild but sustained rise in the daily average ICP because the normal daytime orthostatic drop in ICP (at 1G) is lost in space. 

Some have suggested additional factors caused by a unidirectional orbital inflow of cerebrospinal fluid (CSF) that is chronically sequestered within the optic nerve sheath compartment^1,55,56 because of headward fluid redistribution and impaired lymphatic resorption.^56–58 This may explain the persistence of disc edema with normal ICP for 6 to 12 months postflight,^3,55,59 as well as the anterior deformations in selected crew members; however, it would not, itself, explain the posterior deformations discussed below. The low-grade disc edema and the anterior deformation among some crewmembers described in this study are consistent with both a mild elevation in the daily average CSF pressure or perioptic CSF sequestration or both and thus may, in part, contribute to the development of SANS. 

The magnitude and frequency of posterior BMO-displacements and BML-shape deformations was the feature that most distinguished crewmembers with SANS from patients with IIH. Similar posterior BMO-displacements, during and after spaceflight, have been noted in two earlier studies using slightly different reference points.^2,5 The cause in SANS is unknown and ostensibly inconsistent with clinical and MRI findings^8,60 showing that the globe is anteriorly indented toward the vitreous. The occurrence of posterior ocular deformations suggests that the indentation load in this subgroup may differ from crew members with anterior deformations and patients with IIH. There are several possible explanations. 

Acute elevations of the IOP can cause small posterior displacements and deformations of the prelaminar and peripapillary tissues as well as the lamina cribrosa.^31,61–64 Although IOP may transiently increase initially inflight, measurements of IOP after an initial rise during spaceflight, head down tilt, and bed rest studies have been mixed.^53,65–68 A decrease in the ICP could also cause posterior deformations,² but there is no evidence that ICP is pathologically decreased during weightlessness or experimentally. In contrast to the axonal distension of early papilledema,^69,70 some have suggested that headward fluid shifts compromise the microvasculature of the ONH and cause an interstitial optic disc edema.^4,71 Additionally, based on numerical modeling, choroidal thickening that occurs in SANS^3,4,55,67 can impose strains on the prelaminar ONH comparable to elevating the IOP to 30 mm Hg.⁷² 

It is also possible that the headward fluid shifts in microgravity alter the distribution and pattern of indentation loads on the globe. Orbital-ocular fluid congestion expands the choroid volume,^1,3,4 that in theory can flatten the globe and form choroidal folds. It is noteworthy that the mean preflight-inflight change in choroidal thickness was greater in the posterior subgroup (38.64 µm, 95% CI = 29.35 to 47.94) than the anterior subgroup (22.06 µm, 95% CI = 13.04 to 31.08; P = 0.01). Headward fluid redistribution, which causes the puffy face syndrome with nasal congestion and facial edema,^44,73,74 may also increase orbital tissue pressure. In contrast to the concentrated load at the BMO in IIH, indentation loads caused by orbital and ocular fluid shifts would presumably be more broadly distributed across the posterior pole and spare the BMO. Thus, if the BML reference points are anteriorly displaced more than the BMO then the BMO would appear to shift posteriorly (relative to the reference points), when in fact both were unequally indented toward the vitreous. This may explain the correlation we found between the BML-shape deformations and BMO-displacement in IIH (see Supplementary Fig. S2b) but not in SANS (see Supplementary Fig. S2e). In sum, a peripapillary indentation load (caused by choroidal expansion or orbital tissue pressure), interstitial prelaminar optic disc edema, changes in the compliance of the load bearing structures, and the unequal displacement of the reference points and the BMO acting in part or together could potentially explain the “posterior” ocular deformations.⁷² However, the relative importance of each of these factors in SANS is unknown. 

This study alone does not entirely rule out an elevated ICP as the sole cause of SANS. Posterior displacements occurred in our patients with IIH but were less frequent (11% of the eyes), proportionately smaller and more apt to occur with higher grades of papilledema (mean RNFLT 356 ± 102 µm). When swelling is severe, distended bulging prelaminar axons can form a “peripapillary hyper-reflective ovoid mass-like structure (PHOMS)”^75,76 that may posteriorly displace the BMO. Numerical models on the effects of translaminar pressure differences have shown that in some scenarios,^33,43,77 depending on dural stiffness and scleral flexibility, the peripheral lamina can be displaced posteriorly in response to an elevated ICP.⁴³ The difficulty in attributing the deformations we observed to an elevated ICP alone is that it implicitly minimizes the potential impact of headward fluid shifts and microgravity on the globe and orbit in space and assumes a large increase in ICP (which appears unlikely in SANS^50,78). 

Figure 4 summarizes the differences in ocular deformations between IIH and SANS and speculates on how retrolaminar and peripapillary indentation loads might potentially interact in SANS compared to IIH. 

Figure 4.

 

Summarizes the differences in ocular deformations between IIH (A) and SANS (B). We speculate on how indentation loads may hypothetically interact. (A)The mechanism in IIH is well understood and consequent to a concentrated retrolaminar indentation load caused by a large rise in ICP and retrolaminar tissue pressure, much greater in most cases to any increase in the prelaminar tissue pressure caused by swelling of the optic nerve head. (B) The mechanism in SANS remains uncertain; however, it is likely that ICP alone does not explain the anterior and posterior pattern of ocular deformations. We speculate that there may be a dynamic interaction between two types of loads. (B1) A mild retrolaminar indentation load akin to a IIH (caused by intracranial fluid shifts with a low grade elevation in the ICP or perioptic CSF sequestration or both) and (B2) a peripapillary indentation load that spares the BMO possibly caused by the combined effects of orbital-ocular fluid shifts that expands the choroid and possibly increases orbital tissue pressure. The direction of the ocular deformation, if any, is determined by the relative magnitude of each load and individual factors that may depend on gender,⁵⁰ the material properties of the load bearing structures (e.g. sclera, lamina cribrosa, or optic nerve sheath) and structural anatomy (e.g. anatomic communication between the cranial-perioptic CSF compartments and the choroidal geometry).^33,72 BML, Bruch's membrane layer; BMO, Bruch's membrane opening; CSF, cerebrospinal fluid; ICP, intracranial pressure; IIH, idiopathic intracranial hypertension; IOP, intraocular pressure; PLTP, prelaminar tissue pressure; RIL, retrolaminar tissue pressure; PIL, peripapillary indentation load; SANS, spaceflight-associated neuro-ocular syndrome.

The findings and conclusions in this study must be tempered by a number of limitations. The study was retrospective, and the sample size was small. We unavoidably compared two entirely different cohorts: morbidly obese young women to healthy mostly middle-aged men. The inflight average duration of 5.3 months may be insufficient to bring about the full syndrome in many crew members. 

An important limitation of shape studies of the posterior pole using OCT is an optical discrepancy between the OCT image (which can appear artificially flatter) and an MRI image. The OCT artifact is affected by axial length, the distance between the imaged eye and OCT system, and is greater the farther away the scan is from the image center. Kuo et al.^79,80 has developed an optically based algorithm to correct this discrepancy on OCT. 

We were unable to apply these corrections because the necessary parameters were not recorded at the time of the scan. However, this imaging artifact was largely mitigated by several factors. First and most importantly, our analysis was based on the paired “pre” versus “post” changes in BML-shape and BMO-displacements within an individual. The active eye tracking and follow-up mode in the Heidelberg OCT (TruTrak/AutoRescan features), used in this study, uses a robust registration system that allows precise comparisons between baseline and follow-up scans. To the degree that the artifact played a role, it did so in both the “pre” and “post” paired images. Second, the artifact can be affected by axial length changes in space but the changes in this SANS cohort were small, with a mean decrease of −0.14 mm ± 0.15. This small change in axial length (7% of the 2 mm image depth of the OCT image) coupled with the active eye tracking and follow-up mode limited differences in the eye to OCT system distance within the same individual. Third, limiting our analysis to the peripapillary retina and optic nerve head on a 20 degrees transverse axial OCT minimized the effect of the image distortion because the artifact becomes more prominent farther away from the scan center. 

We acknowledge that this is a limitation that future studies will need to consider, particularly if absolute morphology is the end point. Although the OCT artifact may have had an incremental effect on the magnitude of the deformations (which we may have underestimated by not increasing the curvature of the globe or using wide angle transverse axials), it is unlikely to account for the large quantitative and qualitative differences between IIH and SANS. 

Measurements of RNFLT were derived from two different OCT devices: Spectralis OCT (for crewmembers with SANS) and Cirrus OCT (for patients with IIH). We do not believe that this had any meaningful impact on our results or conclusions because the differences in RNFLT measurements between these devices is far smaller than the extraordinarily large disparities in swelling (confirmed by clinical Frisen grading) between IIH and SANS. Moreover, all pre and post scans were paired within the same individual's eye that was tested using the same device and follow-up registration modes. 

The model proposed in Figure 4 emphasizes the differences in retrolaminar and peripapillary indentation loads in IIH and SANS; however, little is known about the effects of microgravity on the tissue material properties (e.g. choroid, BML, orbital tissue, neural tissue and the load bearing lamina, and dura sclera) or structure (e.g. scleral or laminar thickness, and intracranial CSF communication with the perioptic nerve compartment) of the globe and orbit and how they interact.^{32,33,43,72,81,82} Further studies are needed to consider how these factor into the pathogenesis of SANS. 

Notwithstanding these limitations, this study showed substantial differences in the degree of ONH swelling and the magnitude and pattern of ocular deformations between IIH and SANS. Although we cannot identify the precise cause, it seems likely that the indentation loads, the loading conditions, and possibly the material properties of the globe itself are different. The observations in this study support the prevailing hypothesis that SANS is not entirely explained by an elevation in ICP alone but is likely to be a multifactorial consequence of headward fluid redistribution in space. 

Acknowledgments

Supported by the National Aeronautics and Space Administration (NASA) under the Human Research Program (directed research). 

Disclosure: P.A. Sibony, None; S.S. Laurie, None; C.R. Ferguson, None; L.P. Pardon, None; M. Young, None; F.J. Rohlf, None; B.R. Macias, None