Adaptation of the Proximal Femur to Skeletal Reloading After Long-Duration Spaceflight


  • The authors state that they have no conflicts of interest.


We studied the effect of re-exposure to Earth's gravity on the proximal femoral BMD and structure of astronauts 1 year after missions lasting 4–6 months. We observed that the readaptation of the proximal femur to Earth's gravity entailed an increase in bone size and an incomplete recovery of volumetric BMD.

Introduction: Bone loss is a well-known result of skeletal unloading in long-duration spaceflight, with the most severe losses occurring in the proximal femur. However, there is little information about the recovery of bone loss after mission completion and no information about effect of reloading on the structure of load-bearing bone. To address these questions, we carried out a study of the effect of re-exposure to Earth's gravity on the BMD and structure of the proximal femur 1 year after missions lasting 4–6 months.

Materials and Methods: In 16 crew members of the International Space Station (ISS) making flights of 4.5–6 months, we used QCT imaging to measure the total, trabecular, and cortical volumetric BMD (vBMD) of the proximal femur. In addition to vBMD, we also quantified BMC, bone volume, femoral neck cross-sectional area (CSA), and femoral neck indices of compressive and bending strength at three time-points: preflight, postflight, and 1 year after mission.

Results: Proximal femoral bone mass was substantially recovered in the year after spaceflight, but measures of vBMD and estimated bone strength showed only partial recovery. The recovery of BMC, in the absence of a comparable increase in vBMD, was explained by increases in bone volume and CSA during the year after spaceflight.

Conclusions: Adaptation of the proximal femur to reloading entailed an increase in bone size and an incomplete recovery of vBMD. The data indicate that recovery of skeletal density after long-duration space missions may exceed 1 year and supports the evidence in the aging literature for periosteal apposition as a compensatory response for bone loss. The extent to which this compensatory effect protects against fracture remains to be seen.


The long-term success of human space exploration depends on finding countermeasures to overcome the deleterious effects of the space environment on human health. Bone loss is one of the most serious medical problems associated with prolonged weightlessness. Skeletal fragility incurred over the course of a long-duration space voyage may result in increased risk of a debilitating bone fracture on return to Earth or during an excursion onto another planet.

Bone atrophy is a well-known consequence of extended skeletal disuse and was predicted as a consequence of extended space travel well before the initial U.S. Skylab and Soviet Salyut missions in the early 1970s. In the Skylab missions, bone loss was measured in the heel using early X-ray absorptiometry techniques, and biochemical methods were used to observe negative calcium balance and increased blood levels of molecular markers of bone resorption.(1,2) The most recent observational studies of the Russian MIR and International Space Station (ISS) crews have used modern bone densitometry methods to characterize the extent of bone loss in different skeletal regions.(3–5) These studies showed trends toward minimal changes at non–load-bearing bones such as the distal radius and pronounced bone loss at the load-bearing regions of the lower skeleton, with the highest rates of bone loss found at the proximal femur. Based on data from the MIR and ISS, crew of long-duration space mission typically lose on the order of 1.2–1.5% of bone mass from their proximal femurs per month of spaceflight. This compares to age-related changes of <1%/year in elderly white females, the population group most predisposed to osteoporosis.(6)

Whereas there are extensive data on bone loss in prolonged spaceflight, there are few data on bone recovery after return to Earth. Animal and human studies of skeletal reloading after prolonged disuse have indicated incomplete recovery of bone mass and architectural parameters, even after recovery periods longer than the period of unloading.(5,7–9) To address this gap in our knowledge, we have extended a previous study of bone loss in prolonged spaceflight(4) to examine, in 16 crew members of the ISS who made flights lasting 4–6 months, the effect of resumed load bearing on measures of proximal femoral structure. We focused on the proximal femur because it is a load-bearing bone subject to extensive loss in spaceflight and is the most clinically important osteoporotic fracture site. Scanning preflight, postflight, and 1 year after mission completion, we used volumetric X-ray QCT (vQCT) to characterize the recovery of the cortical and trabecular compartments of the proximal femur, as well as changes in indices of proximal femoral size and strength. Because of the growing body of evidence supporting an increase in bone size as a biomechanical response to age-related bone loss, we hypothesized that a similar response would occur in skeletal reloading after long-duration spaceflight.


Human subjects

Sixteen crew members of the second through the eighth ISS missions visited the imaging clinic (Methodist Hospital, Houston, TX, USA) before their missions, 1–3 weeks after landing, and 1 year after return. Each subject provided informed consent before enrolling in the study, which was approved by the Institutional Review Boards of the NASA Johnson Space Center, the Baylor College of Medicine and the University of California, San Francisco.

Scanning and image analysis

At each visit, crew members were imaged in their proximal femora using a standard whole body X-ray CT system using previously described protocols to acquire, calibrate (to equivalent concentration of calcium hydroxyapatite), and analyze the scans.(4,10,11) The analytical software isolated the left proximal femur and defined measurement regions encompassing the entire proximal femur and the femoral neck. Within each region, the program characcterized the volumetric BMD (vBMD), volume, and mass of the total tissue envelope and the outer cortical bone layer, as well as the vBMD of the trabecular bone in the medullary cavity. Within the femoral neck, the program extracted the cross-section of minimal area (MNCS) and computed its cross-sectional area as well as indices of bending/torsional and compressive strength. The tissue volumes and femoral neck cross section quantified by our analysis program are displayed in Fig. 1. According to methods recently described in this journal,(4) we computed two indices of femoral neck strength at the MNCS. A femoral neck bending/torsional strength index (NBSI) was calculated as an elastic modulus-weighted effective polar moment of inertia (Ix + Iy) of the MNCS cross-section divided by the calculated bone width W.

equation image

(x̄,ȳ) are the elastic modulus weighted centroid of the cross-section. The ei are the elastic modulae, which were determined parametrically from the BMD of each voxel in the entire MNCS cross-section using relationships developed by Keyak et al.,(12) and eb is the elastic modulus of cortical bone. A femoral neck compressive strength index (NCSI) was computed as the square of the integral femoral neck vBMD (iBMD) multiplied by the MNCS area.

equation image
Figure Figure 1.

(A) 3D reconstruction of CT scan of proximal femur showing coronal midplane (COR), femoral neck axis (F), and transverse femoral neck plane (AX) used to calculate the minimal cross-sectional area and strength indices of the neck. (B) Volumes of analyzed tissue are overlaid in white on coronal midplane through proximal femur. b1 and b2 are femoral neck cortical and trabecular subregions, respectively, and b3 and b4 are, respectively, the cortical and trabecular subregions of the overall proximal femur region. b5 is the image corresponding to the transverse plane in AX.

Statistical methods

Statistical analyses were performed using the SAS statistical analysis program (SAS Institute, Cary, NC, USA). We computed percentage changes in our measurements between pre- and postflight time-points and between postflight and 1-year time-points. Because of small sample size and nonnormally distributed data for these percentage changes, we used the Wilcoxon sign-rank paired-differences test to determine statistically significant levels of change. To assess completeness of recovery, we computed the ratio of 1-year to preflight measurements. The Wilcoxon sign-rank test was used to determine if the 1-year values were significantly different from the preflight values. We used the Spearman correlation coefficient to determine associations between bone loss and changes in bone size during the year after mission completion. A 5% significance level was used throughout the paper for statistical significance.


Skeletal unloading/inflight changes

Consistent with our previously published study,(4) comparison of pre- and postflight image measurements in Tables 1 and 2 show that the crew members, on average, lost roughly 11% of the total bone mass from their proximal femora (p < 0.01). On average, the trabecular bone mass and density declined by 14.4–16.5% over the course of the mission (p < 0.001). There was no statistically significant change over the course of the flight in any of the bone size parameters, indicating that the substantial loss of cortical tissue volume and mass occurred by thinning of the cortex from its inner margin. The loss of bone mass without the compensating effect of increased bone size was manifested in the sharp decreases observed for the bending (15.8% p < 0.001) and compressive strength indices (16.8% p < 0.001) in Table 3.

Table Table 1.. Femoral Neck vBMD, Mass, and Volume Measured Preflight (PrFL), Postflight (PoFL), and 1 Year After Mission, Followed by the Percentage Changes Measured During the Flight and Recovery Periods and the Ratio of1 Year to PrFL Values
original image
Table Table 2.. Total Femur vBMD, Mass, and Volume Measured Preflight (PrFL), Postflight (PoFL), and 1 Year After Mission, Followed by the Percentage Changes Measured During the Flight and Recovery Periods and the Ratio of1 Year to PrFL Values
original image
Table Table 3.. MNCS, FBSI, and NCSI Measured Preflight (PrFL), Postflight (PoFL), and 1 Year, Followed by the Percentage Changes Measured During the Flight and Recovery Periods and the Ratio of 1 Year to PrFL Values
original image

Resumption of load-bearing/postflight changes

To study the response of the proximal femur to resumed load-bearing after a disuse period conferring an average 11% decline in total proximal femoral bone mass, we compared postflight and 1-year measurements. The changes observed 1 year after spaceflight are summarized in Tables 1–3. Compared with the losses of roughly 11% incurred during the spaceflight, total bone mass increased by 8.1% (p < 0.05) and 12.2% (p < 0.001) in the femoral neck and the overall proximal femoral regions, respectively, during the year after mission, resulting in 1-year values within statistical error of the preflight values (p > 0.1 for both regions). However, the substantial recovery of bone mass was not matched by a recovery in total vBMD, which increased by only 0.9% (not significant) in the femoral neck and by 4.4% in the overall proximal femoral region (p < 0.05), resulting in 1-year values that were 91% and 93% of preflight values, respectively (p < 0.001). As shown in Tables 1 and 2, the discrepancy between the recovery of bone mass compared with the vBMD was explained by the substantial increase in total bone tissue volume of 7.2% observed for both the femoral neck (p < 0.05) and overall proximal femoral regions (p < 0.01). The increase in total bone tissue volume in these regions was supported by a 2.4% increase in the femoral neck cross-sectional area (p < 0.05). At 1 year, the femoral neck cross-sectional area was larger than at preflight (1-year/preflight ratio = 1.03, p < 0.001), with a trend for the total bone volumes at the femoral neck and total proximal femur to be 5–6% larger at 1 year than at preflight (ratio = 1.05–1.06, statistically insignificant at 0.07 < p < 0.18).

When we examined the cortical and trabecular components of the proximal femur separately, we observed that they exhibited different patterns of change after skeletal reloading. The vBMD of trabecular bone increased by 6.7% and 6.8% in the femoral neck and overall proximal femoral regions (p < 0.01 and p < 0.001), respectively, whereas the vBMD of cortical bone changed by <1% in both regions. At 1 year, both measures were still below preflight values (ratio = 0.89–0.97, 0.0001 < p < 0.03). The larger percentage increase of trabecular vBMD was consistent with previous reports(5) and was expected given the higher metabolic activity of trabecular compared with cortical bone. Although the vBMD of the cortical bone remained constant, there were large increases of 8% (p < 0.05) and 11.5% (p < 0.001) in the volume of new cortical bone in the femoral neck and overall proximal femoral regions, respectively, resulting in recovery to preflight levels by 1 year after the mission (0.99 < ratio < 1.01, p > 0.1 between preflight and 1-year values).

Changes in indices of bone strength

To relate the changes in BMD, bone mass, and bone size to the structural integrity of the bone tissue, we compared the femoral neck bending/torsion and compressive strength indices between the postflight and 1-year time-points. As shown in Table 3, these indices tended to increase between the postflight and 1-year measurements, but neither change was statistically significant (0.14 < p < 0.19). At 1 year after spaceflight, the bending and compressive strength indices were 15% and 20% below their preflight values, respectively.


Our study provides important new data that document the effect on a key load-bearing skeletal site of restored weight bearing after a period of disuse atrophy. These effects can be considered in light of the linkage between mechanical use and skeletal structure elaborated by Frost,(13,14) who proposed a strain-based model by which bone tissue is locally remodeled, with addition or resorption of new bone, as an adaptation to changing mechanical loads. This concept is supported by findings of bone loss and deterioration of bone architecture in studies of mechanical unloading both in human bedrest models(9,15) and rodent hindlimb unloading models.(16) On the other hand, studies of high impact exercise have shown that increased mechanical loading entails new bone formation and alterations of bone geometry in adaptation to increased strains.(17,18) The dynamic relationship between increased mechanical loading and skeletal structure have been elegantly shown in four-point bending studies, in which dynamic bending moments were applied to the rat forelimb and in which the magnitude and frequency of applied loads were found to correlate to changes in bone size and shape.(19,20)

Before reloading, our study subjects had sustained a 4- to 6-month period of partial skeletal disuse in which they, on average, sustained a striking loss of proximal femoral bone mass, cortical thinning, and decline of estimated bone strength indices. The mean loss of proximal femoral BMC per month exceeded the annual decline observed for elderly females.(6) The amount of bone loss in our study, which occurred over a period of 4–6 months, corresponded to roughly one third of the age-related bone loss incurred in men over a lifetime.(21) Unlike the aging process, in which the skeleton is thought to alter its geometry to adapt to continued mechanical loading in the presence of bone loss,(21,22) no such compensatory response was observed in our data. There were no changes in any of the overall tissue volume and cross-sectional area parameters over the course of the flight, indicating that the substantial loss of cortical tissue volume and mass occurred by thinning of the cortex from its inner margin. In keeping with the loss of cortical bone mass without the compensatory effect of periosteal apposition, there was a decline of estimated bending and compressive strength with a magnitude exceeding that of overall bone mass loss.

This study is the first to examine the resumption of weight bearing on proximal femoral structure after a prolonged period of skeletal disuse. A previous study had described postlanding changes in cortical and trabecular tibial vBMD(5) in crew of the MIR spacecraft, but this study did not report changes in bone geometry. The observed increases in bone cross-sectional area and cortical bone volume measures are consistent with an adaptive response in which high rates of bone formation occur in the cortical femoral neck region, which is subject to elevated bending strains with resumed loading of the proximal femur in normal ambulation. Although our imaging method did not have sufficient spatial resolution to directly measure bone formation in the periosteum, periosteal apposition is consistent with the observed increase in the year after flight in CSA and total bone volume measures. Periosteal apposition at the femoral neck and long bones is widely accepted as a continuous skeletal response to protect skeletal strength in the context of age-related bone loss.(23–27) Riggs et al.(21) used QCT to cross-sectionally examine measures of proximal femoral bone density and size in men and women between 20 and 90 years of age, and their results were consistent with lifetime age-related increases in femoral neck cross-sectional area of 7% and 13% in men and women, respectively. Ahlborg et al.(28) used single-photon absorptiometry to image the distal radius in a cohort of elderly women over a 15-year period and reported a 0.7% annual rate of increase in radial width. Although our study setting differs in multiple aspects from that of aging, our longitudinal results contribute to this general area by showing that the bone loss in a load-bearing skeletal site is associated with an increase in bone size after resumption of weight bearing. Although we observed increases in bone size measures and partial recovery of trabecular vBMD, these increases did not translate to recovery of bone strength indices, potentially because the indices depend roughly quadratically on BMD, and the newly formed cortical bone may not yet be fully mineralized at the end of 1 year.

Periosteal apposition is considered an adaptive geometric response to preserve bone strength in the context of bone loss. However, it protects bone strength with respect to the loading forces encountered in normal mechanical use of the proximal femur and may not protect against the forces exerted during a traumatic event such as a fall. Using high-resolution pQCT to image cadaveric femora harvested from donors 20–95 years of age, Mayhew et al.(29) characterized the local stresses associated with ambulation and falling around the femoral neck cortex. They found that the stresses associated with normal ambulation were sufficient to induce local thickening of the inferior cortex with age but insufficient to prevent thinning of the supero-lateral cortex. Based on the local thinning of the superior cortex, they proposed buckling as an potential mechanism of failure for the femoral neck during a fall. A similar situation may occur after disuse atrophy, with resumed ambulation resulting in replacement of lost bone in the inferior but not the superior portion of the cortex, resulting in increased instability with respect to buckling. Because trabeculae may protect against this instability,(30) ultimate failure to recover trabecular bone, or a delay in its recovery, may exacerbate this problem after a very long spaceflight or repeated long-duration spaceflight.

This study has both fundamental strengths and weaknesses. It is the first systematic longitudinal study to use vQCT to characterize changes in proximal femoral BMD, geometry and strength indices associated with response to weight bearing after a sustained period of skeletal unloading. As far as we know, it is the first clinical research study using imaging to directly observe an increase in the cross-sectional area of the femoral neck in response to bone atrophy, either by disuse or by aging. The weaknesses of our study design include the small sample size (inherent to almost all human spaceflight studies) and the fact that the some of the postflight measurements took place up to 3 weeks after the flight. Because some bone recovery took place before the postflight measurement, this would tend to underestimate the extent of both the inflight and postflight changes in the proximal femoral measures. However, in our cohort, the observed inflight and postflight changes in those subjects with the largest elapsed time between landing and postflight showed no trend to be smaller than those subjects with short times between landing and postflight measurements. In addition to limitations in the study design, partial volume averaging is a source of technical error that must be considered, particularly for cortical bone measurements. Partial volume averaging errors affect measurement of structures that are small compared with the spatial resolution of the imaging device. For thin cortices, such as the superomedial aspect of the femoral neck, this results in underestimation of BMD and overestimation of volume or thickness measurements, with smaller errors for estimation of mass.(31) Our hip cortical region of interest is primarily composed of the inferomedial cortex of the proximal femur, for which thickness values exceeding 3–4 mm have been reported in humans, but also includes the thin superomedial cortex (thickness ≈ 0.3 mm).(32) In a study based on a helical CT scanner similar to our own, Prevrhal et al.(33) reported that detection of changes in cortical thickness is possible for cortices with thicknesses >1.2 mm. Thus, it is likely that the cortical volume changes observed in this study primarily reflected changes occurring in the thick inferomedial cortex. Another potential source of error is trabecular vBMD variability caused by changes in the fat content of the bone marrow.(34,35) If the fat content of the proximal femoral bone marrow increased over the course of the flight, this would artifactually decrease the apparent postflight BMD, resulting in larger apparent changes. Similarly, if the marrow fat content decreased during the recovery period, this would produce a larger apparent BMD value and larger apparent increases during the recovery period. Although our single energy vQCT measurement cannot take proximal femoral marrow fat changes directly into account, a previous MRI study of 17-day missions found no significant change in the cellular composition of the vertebral marrow, a finding that was confirmed for long-duration missions.(36) Another limitation in our study is that the various bone strength indices have not been validated as predictors of fracture risk and should not be taken as such. In this study, we have used these indices as interpretive tools to relate changes in bone density and geometry to estimate the potential impact of the spaceflight on whole bone strength in bending and compression.

In conclusion, we have observed the recovery of bone mass after prolonged skeletal unloading in long-duration spaceflight involves increases in measures of bone size such as cross-sectional area and cortical volume. However, the newly formed bone is either less mineralized or more porous than the material lost during the disuse period, resulting in a slower recovery of vBMD. Although reductions in strength indices are not recovered at the end of 1 year, the combination of increased bone size and enhanced formation of new bone in the cortical envelope of the proximal femur are consistent with a pattern that provides the most efficient mode of increasing bending strength for minimal addition of new bone mass.


The authors acknowledge the organizational and logistic assistance of Karin Bergh and the excellent technical assistance of Terry Schuelke. This study was supported by NASA (Contract NAS-9-99055 and Grant NNJ04HC7SA from NASA Johnson Spaceflight Center).