- Top of page
Bone loss during immobilization is well documented. Currently, the only means of studying this in human beings is bed rest, which is resource intensive and inconvenient for the subjects. Unilateral lower limb suspension (ULLS) has been suggested as an alternative, but has not previously been demonstrated to cause bone loss. The main aim of our study was to test the hypothesis that ULLS would cause bone loss determined by peripheral quantitative computed tomography (pQCT). We investigated eight young healthy volunteers (19.1 ± 0.7 years; body mass index, 22.4 ± 2.6 kg m−2), who underwent ULLS for 24 days; their right foot was suspended with a strap attached to the shoulder so the knee angle was 10 deg and they wore a left shoe with a 7.5 cm sole to allow clearance of the right foot and used bilateral crutches to perambulate. Bone scans were obtained by pQCT from the distal epiphyses and from the diaphyses of the tibia in each leg twice before suspension, at days 7, 14 and 21 of the ULLS, and at days 4, 9, 35 and 90 of recovery. After 21 days of ULLS, bone mineral content of the peripheral portion of the epiphysis of the suspended tibia was reduced by 0.89 ± 0.48% (from 280.9 ± 34.5 to 278.4 ± 34.2 mg mm−1, P < 0.001) but no changes were observed in its central portion or in the unsuspended tibia. In the peripheral epiphyseal portion, significant bone loss (by 0.32 ± 0.54%, P= 0.045) occurred as early as day 7 of ULLS. We have demonstrated, for the first time, that in humans bone is lost during ULLS at rates comparable to those seen with bed rest, without alteration in limb fluid volumes thus validating the technique and raising important questions about the mechanisms involved.
Immobilization-induced bone loss is a major concern in hospitalized patients, especially the elderly, because losses in bone mineral content (BMC) and bone mineral density (BMD) result in decreased bone strength (Ebbesen et al. 1997) and increased risk of fracture as well as side effects such as hypercalcaemia and kidney stones (Watanabe et al. 2004). At present it is not certain to what extent the adult skeleton can recover bone losses (LeBlanc & Schneider, 1991), which underlines the importance of their prevention whenever possible.
Although known to elicit a pronounced loss of bone mineral (Andersson & Nilsson, 1979), whole-limb casting, as formerly used in volunteer subjects, can no longer be recommended for ethical reasons (Ingemann-Hansen & Halkjaer-Kristensen, 1977). The study of bone loss in clinical conditions (stroke (Jorgensen et al. 2000), spinal cord injury (Eser et al. 2004), critical illness requiring intensive care (Van den Berghe et al. 2003) and reconstruction of the anterior cruciate ligament (Leppala et al. 1999)) is complicated by the involvement of organ systems beyond bone. Hence, the only currently available technique is bed rest either horizontal or with 6 deg head-down tilt (Kakurin et al. 1976), the latter inducing fluid shifts similar to those observed during space flight, with a reduction of venous and interstitial fluid pressure in the lower extremities (Moore & Thornton, 1987).
However, bed rest studies are resource-intensive and require hospitalization of the volunteers. Also bed rest has extensive and complicated physical and mental effects. Hence, unilateral lower limb suspension (ULLS) has been suggested as an alternative for the study of the effects of immobilization on the human musculoskeletal system (Tesch et al. 2004).
It is surprising that possible bone loss in response to ULLS has never previously been investigated. This is despite the facts that limb suspension is a common clinical treatment, and that, in the context of bone physiology, ULLS is a particularly interesting model, as it allows immobilization in the presence of orthostasis (the latter not occuring during bed rest), thereby allowing the influence of immobilization per se to be assessed without a reduction in venous pressure. Fluid flow within the bone canaliculi appears to play an essential role in the sensing and signalling of strains by bone cells (Bacabac et al. 2004). Based on the idea that interstitial fluids flow from the medullary space towards a periosteal drainage, Bergula et al. (1999) have demonstrated that increasing the intramedullary pressure by partial ligation of the femoral vein counteracts bone loss in hindlimb-suspended rats. It has therefore been suggested that the reduction in lower limb venous pressure may hamper the efficiency of exercise programmes during bed rest and in space (Turner, 1999).
However, there is ample evidence in clinical cases of immobilization, as in stroke or after spinal cord injury, that bone is readily lost from the immobilized limbs (Eser et al. 2004) despite a more or less normal orthostasis. Thus, we hypothesized that bone loss does indeed occur in ULLS and set about testing our hypothesis.
- Top of page
A significant time effect was found for BMC of the suspended tibia epiphysis (P < 0.001). From Day LS21 onwards, values were significantly lower than at baseline (see Table 1). BMC of the suspended tibia epiphysis was lowest on day R35 (P < 0.001, compared to baseline), implying a loss of 4.7 mg mm−1 or 1.1% of the baseline value (Fig. 2A). No significant time effect was found for BMC of the unsuspended tibia epiphysis, or for ATot for either side (P > 0.45 in all cases).
Table 1. Bone mineral content at the distal tibia epiphysis
Figure 2. Percentage change in bone mineral content (BMC) from baseline at the distal tibia epiphysis Suspension of the right leg lasted from day 0 to day 24. A, total BMC: significant losses were found on the suspended side from day LS21 onwards (see Table 1). B, central BMC, assessed for the inner 45% of the total area. Except for an increase in the right tibia, on day LS14, no significant change was observed. C, peripheral BMC, assessed for the outer 55% of the total area. Significant losses were found on day LS7 and from LS21 onwards (see Table 1).
Download figure to PowerPoint
It is surprising that no significant loss over time was found in TrbDen of the suspended tibia epiphysis, but rather a significant gain was seen on day LS14 (P= 0.002). We therefore analysed BMC_Centr of the suspended tibia epiphysis and substantiated a significant gain on day LS14 (P= 0.04), but no other time effect (see Fig. 2B). However, highly significant time effects were found for BMC_Peri (P < 0.001, see Fig. 2C), with significant losses for all days from LS7 onwards loss, except on LS14 (P= 0.86). BMC_peri in the suspended tibia epiphysis was lowest on R35 (P < 0.001), implying a loss of 4.12 mg mm−1 or 1.5% of the BDC value. No time effect was found for BMC_Centr or BMC_Peri in the unsuspended tibia epiphysis (P= 0.49 and P= 0.24, respectively).
No significant time effect was found for BMC, ATot, RPol or Rho for the tibia diaphysis, either on the suspended or on the unsuspended side (see Table 2).
Table 2. Bone variables at the tibia diaphysis
- Top of page
In this study, we have unequivocally shown for the first time that suspension of a human lower limb leads to bone loss. In a previous study, Caruso et al. (2004) investigated the effects of resistance exercise on areal bone mineral density, with and without albuterol, making the assumption that ULLS did induce bone loss and concluding that the combination of albuterol and resistance training should be able to prevent or possibly reverse immobilization-induced bone loss. However, no control group of subjects without any intervention was studied, so the question of whether bone loss does occur made the crucial question moot.
The present study therefore fills an important gap in our knowledge, especially in the context of information coming from work within the hindlimb-suspended rat. It is important at this point to consider that the cardiovascular effects of human ULLS and rat hindlimb suspension differ in some respects. Whereas the hindquarter is elevated in relation to the hydrostatic indifference point in the rat preparation, this is clearly not the case in human ULLS. In consequence, hindlimb venous pressure will be slightly reduced in the rat preparation, but unchanged or, due to a lack of muscle contractions, even increased in human ULLS. Therefore, considering the bone loss observed in the present study, it seems that increased venous pressure was not able to prevent such losses in humans, as might have been expected from the study by Bergula et al. (1999). It should be mentioned at this point that the increase in venous pressure during human ULLS, if present at all, is probably small in relation to that induced by venous ligation.
On the other hand, the rat preparation and human ULLS most probably involve similar changes in perfusion and vascular resistance. For the rat preparation, Colleran et al. (2000) reported a reduction in hindlimb muscle perfusion by up to 80% in the red vastus lateralis muscle, but a moderate increase in perfusion of the white vastus lateralis fraction. The driving factor behind these changes was an increase in vascular resistance. For human ULLS, a decline by 30% in calf muscle blood flow and a concomitant increase in vascular resistance by 40–50% have been reported by Bleeker et al. (2005). In the rat preparation, bone and bone marrow perfusion was reduced in the suspended limbs, but increased or unchanged in the other bones. In most cases, the main factor for the changes in bone perfusion was a change in vascular resistance. Nothing is known as to the effects of ULLS on bone perfusion in humans, but given the similarity of changes in muscle perfusion in both models (Colleran et al. 2000; Bleeker et al. 2005), it seems reasonable to assume a reduction in bone perfusion in the lower extremity bones in human ULLS too. Future studies should attempt to substantiate this and thus help to elucidate the mechanistic role of bone perfusion in disuse-related bone loss.
Our data show quite clearly that, over the entire study, the central and the peripheral portions of the distal tibia epiphysis responded differently to unloading. No significant bone losses were observed from the central portion, and on the day on which it was largest (R35), the bone loss from the central portion accounted for 4.2 mg mm−1 out of the total of 4.7 mg mm−1 for the whole cross section. Thus, approximately 90% of the loss in total BMC was from the peripheral 55% of the epiphysis (Table 1), an unexpected finding, as generally trabecular bone has a faster turnover and seems to react readily to disuse (Rittweger et al. 2005). Of some importance, bone loss from the peripheral epiphysis was significant as early as day LS7 (i.e. when the first measurement was taken during ULLS), but no significant change was observed for the entire cross section at this early stage. During head-down tilt bed rest, there seems to be no significant bone loss from the entire cross section within the first 14 days (Rittweger et al. 2005), but it may well be that bone is lost from the peripheral portion at an earlier stage. To the best of our knowledge, such a regional epiphyseal bone loss has never been reported before. The question therefore arises, whether this phenomenon is specific for bone loss in limb suspension, or whether it always accompanies disuse-related bone loss. To date, most studies into disuse effects on bone have relied on dual-energy X-ray absorptiometry (DXA). However, DXA measures areal BMD and BMC and therefore does not provide any detailed anatomical information, for example distinguishing between trabecular and compact compartments. It may be that the phenomenon we observed has previously escaped the attention of researchers only a few of whom investigated bone loss with the more advanced technique of pQCT (Vico et al. 2000; Eser et al. 2004; Rittweger et al. 2005). It has been reported that exercise-specific effects upon bone geometry are more pronounced in the cortical portion than in the trabecular portion of the epiphyses of the long bones (Nikander et al. 2006), suggesting that the peripheral portion of the epiphyseal bone may, indeed, be more responsive to mechanical stimuli than the central portion.
If indeed the phenomenon is a general one, this peripheral epiphyseal bone loss might only occur during the early stages of immobilization. This might be inferred from the fact that in patients who suffered from spinal cord injury more than 10 years earlier, the bone losses from the distal tibia epiphysis were quite uniform across central and peripheral portions (J. Rittweger, unpublished data; Eser et al. 2004).
No significant changes at all were observed in the diaphyseal variables. However, a closer look at Table 2 suggests that there may very well be a trend in the BMC data of the left tibia diaphysis, with lower values as the study progressed. During bed rest, bone loss is 3–4 times faster from the epiphysis than from the diaphysis (Rittweger et al. 2005). We would therefore argue that diaphyseal bone loss did occur in this study but was too small to be detected. A statistical power calculation (http://calculators.stat.ucla.edu/powercalc) suggests that our study would have detected a loss of 0.5% (β= 0.8), a value that is in excess of any expected value. Conversely, a sample size of n= 27 would have been required to substantiate a significant diaphyseal bone loss on day R35. However, further studies of longer duration are needed to substantiate this notion. No trend, however, could be discerned in Rho, RPol or any other variable that we analysed.
Due to the short study period, caution is needed when comparing these data in relation to results obtained during bed rest. In a long-term study of bed rest (Rittweger et al. 2005), 0.73% of the epiphyseal BMC was lost on day 28 of head-down tilt bed rest (J. Rittweger and D. Felsenberg, unpublished data). In the present study, we found a bone loss of 0.70% on day 21 of ULLS (95% confidence interval, 0–1.36%), and on day 4 of the recovery period (i.e. 28 days after the onset of limb suspension) the loss amounted to 0.57% (95% confidence interval, 0.24–0.90%). It may be that, as in bed rest (Rittweger et al. 2005), bone loss in ULLS appeared to continue during early recovery (see Table 1 and Fig. 2). Taking that into account, the amount of bone loss from the distal tibia epiphysis in bed rest and in ULLS seem to be similar.
Our study had a number of limitations. First of all, due to the technical and ethical implications, the sample size was small, but sufficiently large to allow for our observations. The subjects' adherence to the protocol could only be verified by indirect means (comparing leg temperature). However, ULLS being advocated as an in-extensive alternative to bed rest this uncertainty is quite inevitable. Also there was no dietary control, but it is unlikely that alterations in dietary intake or composition could explain a differential loss in the tibia of the suspended limb either during the period of suspension or afterwards.
In conclusion, we have shown that unilateral suspension of the human leg does induce a loss of bone mineral from the suspended limb. As in bed rest, bone losses were larger from the epiphysis than from the diaphysis, and they continued after reambulation. Our findings suggest that in its extent bone loss in ULLS resembles the bone loss induced by bed rest, and that ULLS may therefore constitute a resource-effective alternative to bed rest for the study of bone.
However, in addition and somewhat surprisingly, bone loss was almost exclusively from the peripheral portion of the epiphysis of the suspended tibia, and not from the central portion. Further studies will be needed to determine whether or not this is a finding unique to ULLS, or is generally found in immobilization.