Intracranial Effects of Microgravity: A Prospective Longitudinal MRI Study

Introduction

Approximately 60% of the crew members of the International Space Station have reported altered visual acuity after long-duration exposure to microgravity (1). Postflight evaluation has shown variable degrees and combinations of optic disc edema, retinal nerve fiber layer thickening, retinal hemorrhage, cotton wool spots, posterior globe flattening, and choroidal folds (1). Lacking a terrestrially based clinical correlate, this medical condition is generically referred to by the National Aeronautics and Space Administration (NASA) as the spaceflight-associated neuro-ocular syndrome. The discovery of spaceflight-associated neuro-ocular syndrome has inevitably raised concerns for the long-term health of astronauts on extended-duration interplanetary travel.

MRI findings in postflight astronauts, in whom similarities to idiopathic intracranial hypertension such as posterior globe flattening are found, have implicated elevated intracranial pressure (ICP) as a hypothesized mechanism of spaceflight-associated neuro-ocular syndrome (2,3). However, contrary to expectations, ICP measured in volunteers during brief microgravity exposures in aerial parabolic flight was not elevated (ICP, 13 mm Hg), but instead remained between the supine (ICP, 15 mm Hg) and 90° upright (ICP, 4 mm Hg) baseline values (4). It has been suggested that the inherent absence of postural ICP variability in microgravity could expose astronauts to increased mean diurnal ICP because of the inability to intermittently lower ICP by upright positioning that depends on a gravitational environment (4). This is potentially problematic because chronic supine-like ICP, without intermittent decompression, simulated by 30 days of strict head-down-tilt bedrest, is associated with optic nerve edema in otherwise healthy adults (5).

Although an unremitting mild increase in the mean diurnal ICP may play a role in the development of spaceflight-associated neuro-ocular syndrome, a potential cofactor was implicated in a structural MRI study (6) of crew members, which showed a mild increase in ventricular volume with long-duration microgravity exposure. Ventricular enlargement also occurred in a microgravity analog study (7) associated with increased aqueductal cerebrospinal fluid (CSF) peak velocity, inferring reduced intracranial compliance and enhanced ICP pulsatility. We hypothesize that ventricular enlargement acquired during spaceflight could result in similar physiologic consequences.

Increased postflight aqueductal peak CSF velocity was reported in a retrospective study (8) of 14 astronauts; however, there was no quantification of ventricular volume or long-term follow-up to identify residual effects. Additionally, in the absence of invasive measures of ICP in microgravity, temporal changes in pituitary morphologic structure could help infer exposure to elevated ICP (9), but initial reports of pituitary gland abnormalities in astronauts were without preflight comparisons (2). Therefore, the purpose of our study was to create a more cohesive model of microgravity by combining measures of intracranial volumetry, pituitary morphologic structure, and CSF hydrodynamics into one prospective study, and by establishing a comprehensive model of recovery following return to Earth.

Materials and Methods

Study Participants

The Human Research Multilateral Review Boards of NASA and the International Space Station approved this Health Insurance Portability and Accountability Act−compliant prospective study, which required written informed consent to participate. Consecutive groups of astronauts who had a planned long-duration mission on the International Space Station were given an option to volunteer for the study after attending a briefing at the Johnson Space Center on available research protocols. Exclusion criteria consisted of contraindication to MRI.

There are multiple intracranial volumetric studies related to spaceflight (6,10–14). Excepting the latter two prospective studies (13,14) composed entirely of cosmonauts, the former studies (6,10−12) consisted of astronauts obtained from clinical archives at NASA and were retrospective by design. According to NASA, two of our 11 participants may have been included in the data set of the retrospective studies. Although a portion of these studies shares similar measures of intracranial volumetry, our prospective study used two different segmentation techniques for independent analysis of soft tissue and fluid spaces. Compared with the previous studies, we combined these volumetric measures with CSF hydrodynamic and pituitary morphologic analysis, both relevant to the ICP hypothesis. Furthermore, we used five postflight MRI examinations performed up to 1 year after landing, exceeding the number and duration found in the previous studies, to depict and characterize residual effects of spaceflight with greater confidence.

MRI Parameters

All participants were imaged on a dedicated 3-T MRI system (Verio 3T, vB19; Siemens Healthineers, Erlangen, Germany) by using a 32-channel head coil. Magnet hardware and software remained unchanged during the study. All imaging was performed between 9 am and 4 pm CST to minimize diurnal variations in CSF production rate. We requested that participants avoid caffeine, alcohol, tobacco, and prescription medication such as antihypertensives, antihistamines, hypnotics, anti-inflammatories, stimulants, proton pump inhibitors, and decongestants at least 8 hours before imaging to avoid any potential effect on cardiac output, blood pressure, cerebral perfusion, or ICP. Each longitudinal data set included one preflight and five postflight MRI examinations performed after spaceflight (return day is indicated by R+): R+1, 2.6 days ± 1.6; R+30, 30 days ± 10; R+90, 90 days ± 10; R+180, 180 days ± 30; and R+360, 360 days ± 30. The time between preflight MRI and launch was 530 days ± 193. Figure 1 summarizes study design. Detailed pulse sequence parameters are available in Table E1 (online).

Figure 1:

Download as PowerPointOpen in Image Viewer

Quantitative Volumetric Analysis

Quantitative volumetric analysis was performed independently on a three-dimensional T1-weighted magnetization-prepared rapid gradient-echo sequence of the whole brain by a dedicated MRI physicist (K.M.H., with 20 years of quantitative MRI experience) who was blinded to flight status, clinical data, and other measurements. White matter and gray matter volumes were derived by using a segmentation tool (SPM12, r6906; https://www.fil.ion.ucl.ac.uk/spm/software/spm12/) (15). Regional CSF volumes were derived by using an online software package (MRI Cloud; BrainGPS, https://braingps.mricloud.org/) (16). Both segmentation techniques were semiautomated and did not require user input. Dural tissue and intracranial vascular spaces such as dural sinuses, cortical veins, and cerebral arteries were not quantified.

Quantitative CSF Flow Analysis

CSF flow characteristics were derived from a pulse-synchronized phase-contrast cine sequence, retrospectively gated, acquired perpendicular to the midcerebral aqueduct, as previously described (8). Phase-contrast data were analyzed with a freely available software (Segment v1.9R3656; http://segment.heiberg.se) (17) to determine the following: (a) CSF peak-to-peak velocity magnitude, representing the total amplitude from the peak (maximum) to the trough (minimum) of the CSF velocity waveform within the cerebral aqueduct (Fig 2a); (b) aqueductal stroke volume, representing the total volumetric CSF flow passing bidirectionally through the cerebral aqueduct per cardiac cycle (Fig 2b); and (c) aqueductal cross-sectional diameter. CSF flow measurements were performed independently by a radiologist (L.A.K., with 30 years of MRI experience) blinded to flight status, clinical data, and other measurements. A pixel-by-pixel technique was used to accurately determine aqueductal margins (8).

Figure 2a:

Download as PowerPointOpen in Image Viewer

Figure 2b:

Download as PowerPointOpen in Image Viewer

Pituitary Gland Evaluation

Pre- to postflight (day 1) three-dimensional T1- or T2-weighted sagittal images of the pituitary gland were corrected for head position variation by using a three-dimensional multiplanar reconstruction tool (Osirix MD v9.5.1; Pixmeo, Bernex, Switzerland, https://www.osirix-viewer.com/osirix/osirix-md/). Electronic slides composed of pre- to postflight images of each participant were created but they were not labeled regarding flight status or participant number. This approach allowed for direct comparative analysis of matched studies but avoided any bias of knowing flight status or participant identification. Two radiologists (L.A.K. and A.S., with 35 years of imaging experience) assigned points by consensus interpretation according to increasing degrees of deformity of the pituitary dome by using the following grading scale modified after Yuh et al (9): upward convexity of the dome of the pituitary gland (normal appearance, 0 points); flattened dome (1 point); mild concavity (2 points); moderate concavity (3 points); and severe concavity or empty sella (4 points). Disagreements were reviewed until consensus was achieved. Preflight points were deducted from postflight points to determine the net change or pituitary deformity score. The pituitary stalk was evaluated for posterior deviation. Performed more than 14 days after the qualitative analysis, pre- to postflight (day 1) midgland height was measured independently on separate workstations by the same expert raters (L.A.K. and A.S.), as shown on Figure 3. Head position correction was independently adjusted per rater to prevent any bias introduced by the other rater. Raters were blinded to flight status, clinical data, and each other’s measurements.

Figure 3:

Download as PowerPointOpen in Image Viewer

Statistical Analysis

Statistical analyses were conducted by using software (Stata IC v15.2; Statacorp, College Station, Tex) with an emphasis on characterizing the observed effects in addition to reporting statistical significance and was performed independently by a statistician (R.J.P., with 20 years of statistical experience). A two-tailed critical α was set to reject the null hypothesis at .05. Our experimental design was a mixed factorial, with repeated observations over time. All outcomes were continuously scaled and assumed to follow a normal distribution; however, all statistical assumptions were tested before interpreting results.

Reliability of individual and average pituitary height observations by the two pituitary raters (L.A.K. and A.S.) were assessed by intraclass correlation coefficients treating raters as random. Pre- and postflight intraclass correlation coefficients were evaluated separately.

We evaluated the effects of long-duration spaceflight on outcomes in separate mixed-effects models per dependent variable. Each model included a random y-intercept to accommodate the within-participants and longitudinal experimental design by using full information maximum likelihood estimation assuming compound symmetry variance-covariance structure of the random effects to accommodate the repeated measures within the participant over time. We included fixed-effects parameters by comparing the preflight observations to each postflight observation, enabling our a priori hypothesized comparisons of the differences in outcomes relative to preflight baseline. We also adjusted for the variability in previous flight experience of the astronaut (ie, number of days exposed to microgravity before our experiment) by incorporating it as a continuously scaled covariate in our modeling.

Each statistical test underwent a rigorous examination of the statistical assumptions and distributions of the model residuals before hypothesis testing. There were some interruptions to the standardized research protocol that resulted in observations that were collected outside the time of planned analyses; these observations were not evaluated in the current report. It was also necessary to eliminate occasional observations that were overly influential on the model, with standardized residuals exceeding ± 2.0. These adjustments were few and resulted in satisfactorily meeting the required statistical assumptions. Two participants were missing volumetric data and five participants were missing CSF hydrodynamic data; however, this mixed-model approach can accommodate for missing participant data. This report characterizes the effect of spaceflight on the mean of each outcome ± 95% confidence intervals of the mean unless otherwise indicated.

Results

Participant Characteristics

Of approximately 52 eligible active international crew members, 11 astronauts (mean age, 45 years ± 5; 10 men, one woman; mean body mass index, 24 kg/m² ± 2) volunteered for the study, and the remainder elected to participate in another research study. None of the 11 volunteers had contraindications to MRI. Five participants had previous spaceflight experience and six participants had no previous spaceflight experience. Mean exposure to microgravity for this study was 171 days ± 71. Demographics are summarized in Table 1.

Volumetry

Volumetric results are summarized in Table 2 and Figure 4. Compared with before spaceflight, summated mean brain (white and gray matter) and mean CSF (extraventricular and ventricular) volumes were increased at R+1 (33 mL of 1615 mL; 2.0%; P < .001) and remained above baseline at all subsequent postflight observation periods: R+30, 1631 mL (P = .02); R+90, 1635 mL (P < .01); R+180, 1634 mL (P < .01); and R+360, 1643 mL (P < .001).

Figure 4:

Download as PowerPointOpen in Image Viewer

We observed a similar pattern of increased mean brain (27 mL of 1285 mL; 2.1%; P < .001), lateral ventricular (2.2 mL of 19.6 mL; 11.2%; P < .001), total ventricular (2.8 mL of 26.1 mL; 10.7%; P < .001), and white matter (26 mL of 444 mL; 5.9%; P < .001) volumes at R+1. This remained above baseline values for all subsequent observation periods (Table 2) as follows: brain, lateral ventricular, total ventricular, and white matter volume, respectively, R+360: 1311 mL (P < .001), 20.6 mL (P < .001), 27.2 mL (P < .001), and 466 mL (P < .001).

Preflight mean gray matter volume (841 mL), mean extraventricular CSF volume (305 mL), and mean CSF volume (331 mL) did not undergo a postflight change: for gray matter, R+1 was 842 mL (P = .70) and R+360 was 844 mL (P = .39); for extraventricular CSF, R+1 was 307 mL (P = .71) and R+360 was 305 mL (P = .99); and for CSF, R+1 was 336 mL (P = .34) and R+360 was 332 mL (P = .75).

An example of lateral ventricular enlargement is shown in Figure 5 and Movie (online).

Figure 5a:

Download as PowerPointOpen in Image Viewer

Figure 5b:

Download as PowerPointOpen in Image Viewer

CSF Hydrodynamics

CSF hydrodynamic results are summarized in Table 2 and Figure 4. Pre- to postflight CSF flow waveform examples are shown in Figure E1 (online). Compared with before spaceflight, aqueductal stroke volume and CSF peak-to-peak velocity magnitude were increased at R+1 (11.4 μL of 95.9 μL [11.9%], P = .045; and 2.0 cm/sec of 15.1 cm/sec [13.2%], P = .01, respectively), with subsequent postflight observations returning to baseline levels (eg, R+360, 93.9 μL [P = .88] and 93.9 cm/sec [P = .30], respectively). Mean preflight aqueductal cross-sectional area was 3.7 mm² and remained unchanged in all postflight measures (eg, R+360, 3.6 mm² [P = .76]; Table 2).

Pituitary Gland Deformity

The mean preflight height of the pituitary gland was 5.9 mm compared with 5.3 mm at R+1 (P < .01). Excellent individual and average rater agreement and intraclass correlations (with 90% confidence intervals) were observed before flight (individual and average rater agreement, 0.92 [90% confidence interval: 0.79, 0.97] and 0.97 [90% confidence interval: 0.88, 0.99], respectively) and after flight (individual and average rater agreement, 0.95 [90% confidence interval: 0.86, 0.98] and 0.96 [90% confidence interval: 0.92, 0.99], respectively). Qualitative analysis showed that after flight, six of 11 (55%) astronauts acquired pituitary gland deformation and one astronaut also developed posterior deviation of the pituitary stalk. All six of the astronauts with no previous microgravity exposure had a normal convex pituitary dome before spaceflight; however, after spaceflight, three astronauts developed flattening and one astronaut developed mild concavity. Only one of the astronauts with previous microgravity exposure had a normal convex pituitary dome that became flattened after spaceflight. Four of the remaining astronauts with previous microgravity exposure had a concave dome at the preflight examination and only one of these pituitary domes acquired additional concavity. The highest pituitary deformity score was 2 (n = 1), followed by a pituitary deformity score of 1 (n = 5). Examples of pituitary deformation are shown in Figure 6 and Movie (online).

Figure 6a:

Download as PowerPointOpen in Image Viewer

Figure 6b:

Download as PowerPointOpen in Image Viewer

Figure 6c:

Download as PowerPointOpen in Image Viewer

Discussion

Spaceflight-associated neuro-ocular syndrome is an unresolved medical phenomenon of spaceflight with symptoms that can persist for years after returning to Earth. In a cohesive search for potential contributory factors, our prospective study identified augmented cerebrospinal fluid (CSF) hydrodynamics (aqueductal stroke volume, 11.4 μL [P = .045]; CSF peak-to-peak velocity magnitude, 2.0 cm/sec [P = .01]), increased pituitary deformation (midgland height loss of 0.6 mm; P < .01; subjective dome depression in six of 11 participants), and expansion of summated brain and CSF volumes (33 mL; P < .001) after long-duration microgravity exposure. The latter was predominantly composed of expanded white matter (26 mL; P < .001) and, to a lesser degree, lateral ventricular (2.2 mL; P < .001) volumes. Summated brain and CSF volumes remained increased 1 year into recovery (28 mL; P < .001).

The results of our study agree with previous work demonstrating enlargement of the ventricles after spaceflight (6,11–13) and confirm that the changes are small in magnitude and remain within the range for healthy adults of similar age (18). Normative data showing that expected ventricular growth of 0.65 mL per year (19) was exceeded within our pre- to postflight time range suggest that ventricular expansion is not exclusively age related.

There was excellent agreement with the absolute value and pre- to postflight stability of our mean CSF volume with a prospective study of 10 cosmonauts (14). However, increased summated mean brain and CSF volumes found throughout our postflight period contradicted other spaceflight studies that showed no significant change in these parameters after spaceflight (14,20). This distinction is important because increased summated brain and CSF volumes can potentially augment ICP according to the cranial pressure-volume relationship (21).

White matter fractional volume has a quadratic relationship to age that is independent of sex (22). It progressively increases from childhood, peaks between the 3rd to 4th decade, and then steadily declines (22). The mean age of our cohort approximated the region of steady white matter decline, which suggested that pre- to postflight white matter expansion is not likely age related. The increase in white matter volume we reported here is consistent with three other spaceflight studies in which smaller nonsignificant 1.5–4.0-mL increases were identified (11,14,20).

Acquired pituitary gland deformity is a surrogate marker of chronic intracranial hypertension (9). Our longitudinal study confirmed a causal relationship to spaceflight in the majority of astronauts. In idiopathic intracranial hypertension, elevation in ICP is hypothesized to cause the development of a CSF-filled arachnoid diverticulum that protrudes through an acquired or congenital defect of diaphragma sellae (9). Superimposed chronic CSF pulsations gradually expand the diverticulum, causing flattening or concavity of the pituitary gland and posterior deviation of the pituitary stalk (9).

In a retrospective study (23) of 500 participants without endocrine abnormalities, the mean pituitary height in men aged 41–50 years was 6.1 mm, decreasing to 6.0 mm in men older than 50 years. In women, pituitary height increased 0.3 mm between the same age groups, attributed to an increase in gonadotropin-releasing hormone in the older age group (23). These normative values suggest that a postflight pituitary height loss of 0.6 mm is unlikely to be age or sex related. In a study of 208 healthy women, chronic intake of oral contraceptives was associated with a minimal reduction in mean pituitary height (0.03 mm; P > .05) (24) and thus also unlikely to account for the 0.6-mm loss of mean pituitary height in our study. Chronic pituitary apoplexy is another potential cause of pituitary height loss (25) but has not been emphasized in the literature for astronauts. Similarly, idiopathic intracranial hypertension causes pituitary height loss but is rare in the western world (incidence, 0.9 of 100 000 persons) (26).

Our combined observations could support an ICP model of spaceflight-associated neuro-ocular syndrome pathogenesis. Critical to this model is the Monro-Kellie doctrine, which states that in the presence of a noncompliant skull, the summation of the brain, blood, and CSF volumes should remain constant and that an increase in one should be offset by a decrease in another or ICP increases (27,28). The development of pituitary gland deformation in the majority of astronauts indicates that summated expansion of brain and CSF volumes is incompletely offset by a decrease in blood volume, resulting in elevated ICP. Furthermore, increased CSF peak-to-peak velocity magnitude infers reduced intracranial compliance (29), which could accentuate ICP pulsatility related to cerebral perfusion (30).

Although summated brain and CSF volume expansion persisted up to R+360, augmented CSF peak-to-peak velocity magnitude returned to baseline at R+30. This apparent mismatch could be related to the exponential relationship of the pressure-volume curve in the region of reduced intracranial compliance (poor compensatory reserve) (21). If applicable at R+30, the relatively small decrease in summated brain and CSF volumes observed could cause a proportionately larger increase in intracranial compliance, reflected by a rapid decline in CSF peak-to-peak velocity magnitude.

Finally, our results suggested a potential intriguing relationship to normal pressure hydrocephalus, which is similarly characterized by a positive correlation between ventricular enlargement and increased aqueductal stroke volume associated with impaired intracranial compliance and augmented ICP pulsatility (31,32). Glymphatic pathway dysfunction has been implicated in normal pressure hydrocephalus (31) and, if applicable in microgravity, it could help explain increased white matter free water in postflight astronauts (10). Continued research into the pathophysiologic causes of ventricular enlargement in astronauts with attention to the glymphatic pathway could potentially improve our understanding of both normal pressure hydrocephalus and spaceflight-associated neuro-ocular syndrome.

Our study had limitations. First, there was no age- or sex-matched longitudinal healthy control group to ascertain the stability of the measurements in normal gravity, and the time delay between the preflight baseline imaging and the launch was longer than ideal. These limitations required reliance on literature-based normative data to help discriminate between the effects of aging versus spaceflight. Second, this study was limited to a subset of astronauts and therefore these results may not capture the variability of the entire long-duration astronaut population. Third, compression of the more compliant dural sinuses (33) (ie, blood volume) could have completely offset summated expansion of brain and CSF volumes; however, the extent of offset could not be determined given the limitations of the segmentation technique. Fourth, the intracranial effects of spaceflight may not have a direct application to terrestrially based medicine because of the uniqueness of the microgravity environment. Future microgravity-related research could, however, indirectly improve our understanding of CSF hydrodynamics and brain structural changes in diseases such as normal pressure hydrocephalus. Fifth, there was no direct preflight, inflight, or postflight measurement of intracranial ICP or compliance, which would help confirm the nature of altered CSF hydrodynamics or pituitary morphologic structure after spaceflight. Because of this limitation, it is important to express caution regarding whether our reported physiologic and anatomic changes accurately represent inflight altered ICP or compliance, or if they are simply a readaptation response to the postflight gravitational environment.

Long-duration spaceflight was associated with increased pituitary deformation, augmented aqueductal cerebrospinal fluid (CSF) hydrodynamics, and expansion of summated brain and CSF volumes. We speculate that prolonged spaceflight is associated with progressive elevations in mean and pulsatile intracranial pressure (ICP) induced by expansion of summated brain and CSF volumes within a noncompliant skull predicted by the Monro-Kellie doctrine (28). Future work will be needed in a larger cohort of astronauts to understand if the augmented mean and ICP pulsatility hypothesized in our study are transmitted to ocular structures and have a role in the development of spaceflight associated neuro-ocular syndrome.

Disclosures of Conflicts of Interest: L.A.K. Activities related to the present article: disclosed money to author’s institution for grant from NASA Human Research Program; disclosed money to author’s institution for provision of writing assistance, medicines, equipment, or administrative support from Kellogg Brown and Root. Activities not related to the present article: disclosed no relevant relationships. Other relationships: disclosed no relevant relationships. K.M.H. disclosed no relevant relationships. M.B.S. disclosed no relevant relationships. A.S. disclosed no relevant relationships. S.S.L. disclosed no relevant relationships. C.O. disclosed no relevant relationships. R.J.P. Activities related to the present article: disclosed money to author’s institution for grant from NASA Human Research Program; disclosed money to author’s institution for travel support from NASA. Activities not related to the present article: disclosed money to author’s institution for grant/grants pending from NASA. Other relationships: disclosed no relevant relationships. K.M. disclosed no relevant relationships. R.F.R. disclosed no relevant relationships. B.R.M. Activities related to the present article: disclosed money to author’s institution for grant from NASA Human Research Program. Activities not related to the present article: disclosed money paid to author for royalties from World Scientific. Other relationships: disclosed no relevant relationships.

Author Contributions

Author contributions: Guarantors of integrity of entire study, L.A.K., R.J.P., B.R.M.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, L.A.K., K.M.H., S.S.L., B.R.M.; clinical studies, L.A.K., K.M.H., A.S.; experimental studies, K.M.H., M.B.S., A.S., C.O., R.J.P., R.F.R., B.R.M.; statistical analysis, K.M.H., R.J.P., R.F.R., B.R.M.; and manuscript editing, all authors