Non-pharmacological treatment and prevention of bone loss after spinal cord injury: a systematic review

Abstract

Objective:

Review the literature on non-pharmacological prevention and treatment of osteoporosis after spinal cord injury (SCI).

Methods:

PubMed, EMBASE and the Cochrane Controlled Trials Register were searched. All identified papers were read by title, abstract and full-length article when relevant. Hand search of the articles’ sources identified additional papers. For included studies, the level of evidence was determined.

Results:

No studies conclusively showed an effective intervention. However, there are few randomized controlled trials (RCTs), and those that exist assess interventions and outcome measures that could be improved. Five studies on weight-bearing early post-injury are conflicting, but standing or walking may help retain bone mineral. In the chronic phase, there was no effect of weight bearing (12 studies). One study found that an early commencement of sports after SCI improved bone mineral, and the longer the period of athletic career, the higher the (leg) bone mineral. Early after SCI, there may be some effects of electrical stimulation (ES) (five studies). Chronic-phase ES studies vary (14 studies, including mixed periods after injury), but improvement is seen with longer period of training, or higher frequency or stimulus intensity. Improvements correspond to trabecular bone in the distal femur or proximal tibia. Impact vibration and pulsed electromagnetic fields may have some positive effects, whereas pulsed ultrasound does not. Six studies on the influence of spasticity show inconsistent results.

Conclusions:

Bone mineral should be measured around the knee; the length and intensity of the treatment should be sufficiently long and high, respectively, and should commence early after SCI. If bone mineral is to remain, the stimulation has to be possibly continued for long term. In addition, RCTs are necessary.

Introduction

Significant osteoporosis develops, in particular, in the long bones of the lower extremities in the first months and years after spinal cord injury (SCI),^{1, 2, 3} and may well continue years after SCI, with trabecular bone after a log curve levelling off from 1 to 3 years after SCI, whereas the cortical bone sites appear to decrease progressively beyond 10 years after SCI.⁴

Factors leading to more bone loss include the level of injury (tetraplegics more than paraplegics), complete lesions and longer duration after injury. Aging may also be a contributing factor.^{3, 4, 5, 6}

The sublesional bone loss means an increased risk of fragility fractures, in particular in the distal femur and the proximal tibia.^{7, 8} Unfortunately, criteria for assessing fracture risk in the SCI population are lacking.^{4, 9}

In contrast to the long bones in the lower extremities, there is no obvious loss of bone in the lumbar spine, which may be because of the weight-bearing function, although degenerative changes may influence the high values found.³

There have been recent reviews on bone loss and osteoporosis in individuals after SCI.^{3, 9, 10, 11, 12} The present systematic review will concentrate on the evidence related to the treatment and prevention possibilities with non-pharmacological-related interventions.

Methods

For identification of articles for the review, PubMed and EMBASE were searched with no language restrictions using the following words: osteoporos^*/osteopen^*/bone mineral/bmc/bmd and spinal cord inj^*/spinal cord lesion^*/spinal cord dis^*/parapleg^*/tetrapleg^*/quadripleg^*. Searches were performed in the Cochrane Controlled Trials Register using osteoporosis/bone mineral and spinal. The PEDro database was hand searched under the key words: Osteoporosis and Spinal Cord Injury.

All identified papers were initially read by title, and, if relevant, by abstract. When abstracts indicated results within the purpose of the review, the full-length articles were found and reviewed.

In addition, a hand search from the other articles’ sources identified additional papers that could be included.

To determine the level of evidence, the studies were judged according to the five levels used in SCIRE (Spinal Cord Injury Rehabilitation Evidence)¹³ (Box 1)

Box 1 Levels of evidence—SCIRE¹³

Two of the authors (FB-S and BH) independently evaluated all studies, and those that were judged differently were discussed for possible agreement. If the first two authors were still in conflict, a third author (BL) evaluated the same paper independently to adjudicate a final decision.

Results and discussion

The Pedro Search identified nine articles for individual review (including review articles for bibliographic searching), whereas the EMBASE review revealed 415 references and the Medline revealed 349 references for abstract review. A total of 45 studies were explored in detail (Tables 1, 2, 3, 4, 5, 6 and 7). Seventeen studies analyzed the effect of weight bearing by standing and walking, of which five included data from the first year after injury. Three studies looked at the effect of exercise, 19 studies analyzed electrical stimulation, 3 looked at other physical interventions and 6 studies made some reference to the effect of spasticity/spasms. Two studies only showed adequate randomization, allocation concealment and blinding (PEDro score ⩾6).^{16, 54} Both studies showed no effect, used a within-subjects study design and dual-energy X-ray absorptiometry but analyzed different outcome measures. Ben et al.¹⁶ used bone mineral density (BMD) at the proximal femur and Warden et al.⁵⁴ used bone mineral content (BMC) at the heel. The intervention group for Ben et al.¹⁶ (Table 1) was standing weight bearing on one leg on a tilt table for 30 min, thrice weekly for 12 weeks, whereas Warden et al.⁵⁴ (Table 6) carried out a low-intensity pulsed ultrasound on one heel for 20 min day⁻¹, 5 days/week over 6 weeks. None of these considered bone changes around the knee, where the majority of bone loss occurs after SCI.

Table 2 The effect of weight bearing by standing and walking on bone in individuals with SCI within the first year of injury

Table 3 The effect of weight bearing by standing and walking on bone in individuals with SCI more than a year after injury

Table 4 The effect of exercise on bone in individuals with SCI

Table 5 The effect of electrical stimulation on bone in individuals with SCI early after injury

Table 6 The effect of electrical stimulation on bone in individuals with SCI longer time after injury or including mixed periods after injury

Table 7 The effect of other physical interventions on bone in individuals with SCI

Table 8 The influence of spasticity/spasms on bone in individuals with SCI

The effect of weight bearing by standing and walking

The studies on weight bearing from the early period (within the first year) after SCI (Table 1) are conflicting. The single level 1 study by Ben et al.¹⁶ showed that early-period weight bearing did not increase BMD; however, the therapy period was relatively short and only involved simple static weight bearing compared with the other two studies showing positive results within this group. In addition, BMD was measured at the proximal femur only. The study by De Bruin et al.¹⁴ indicates that early mobilization led to no or insignificant loss of trabecular bone, whereas the immobilized individuals showed a marked decrease when monitored for 25 weeks. In addition, the recent prospective study by Alekna et al.¹⁸ found that standing, particularly after 2 years, gave significantly higher BMD in legs, pelvis and the total body. The long follow-up carried out in this study is unique. The other studies in Table 1 were of shorter duration and/or less weight-bearing stimulation.

For the chronic phase (Table 2), the picture seems more uniform, with very little evidence for any gain in BMD when the first year after injury has passed. All these studies were of evidence levels 4 and 5.

The ineffectiveness of shorter term (<3 months), less-aggressive early intervention therapy and the overall ineffectiveness of the chronic-phase studies raise the hypothesis that if a weight-bearing intervention is to be considered, it should be more aggressive and should intervene in the early period after injury if a treatment effect is to be found.

The effect of exercise

The quality of evidence available for evaluation is poor. The most interesting study in this group with only three investigations of level 5 evidence (Table 3) is by Miyahara et al.³² They found that the earlier the athlete started sports after injury, the higher the BMD of the legs, body trunk and the entire body. Further, a longer period of athletic career after restarting was significantly related to higher leg BMD. This result is encouraging, and supports the hypothesis of early intervention after injury (suggested in the weight-bearing analysis) if a treatment effect is to be found in future studies.

The effect of electrical stimulation

Of the four level 2 studies performed early after SCI (Table 4), three showed that there may be some effect of electrical stimulation with stimulation of 5 days/week. The single-case long-time prospective investigation by Dudley-Javoroski and Shields³⁷ additionally suggests the importance of where the muscles stress the bone. This further emphasizes the value of specific methods for BMD measurement, that is in most studies, the dual-energy X-ray absorptiometry^{35, 36, 42, 44, 45, 46, 47, 48, 49, 50} is used as it is considered to be the ‘gold standard,’ but peripheral quantitative computed tomography has the advantage of being able to differentiate the cortical from the trabecular bone, and assess both bone geometry and volumetric density.^{14, 15, 37, 51} A quantitative ultrasound has also been used, but probably is still not suitable to monitor the effect of intervention on BMD,⁵⁸ whereas magnetic resonance imaging also seems to have a potential in future studies in assessing bone geometry.⁵⁹

For the chronic phase (Table 5), the studies are conflicting, but those investigations that show improvement seem to be those with a longer period of training, that is 12 months,⁴⁶ or higher frequency, that is 5 times/week^{47, 48} or stimulus intensity.⁴⁵ It is also evident that all these studies measured their improvement corresponding to the trabecular bone, in particular in the distal femur or the proximal tibia. The challenge may be to provide sufficient mechanical stimulus without increasing the risk of fracture.^{12, 60}

It is also observed that the positive effect of electrical stimulation on the bone mass only remains if the stimulation is continued and in sufficient amounts.^{46, 48} Therefore, the recommendation following the experience gained from the studies is that the electrical stimulation should be at least 2–3 times per week, and probably has to be continued for the long term if the bone mass is not to decline further.

The effect of other non-pharmacological interventions

Table 6 shows that the only randomized study by Warden et al.⁵⁴ did not indicate any benefit from low-intensity pulsed ultrasound. Potentially, both impact vibration⁵² and pulsed electromagnetic fields⁵³ showed positive results, but the studies are of poor quality. These results could support the hypothesis that a more dynamic loading may be necessary to elicit a detectable treatment effect.

The influence of spasticity/spasms

The studies evaluating the possible influence of spasticity (Table 7) show inconsistent results. Demirel et al.⁵⁶ found that those with spasticity had higher BMD when compared with flaccid individuals, and Eser et al.²⁸ showed a significant correlation between the degree of spasticity measured with the modified Ashworth scale and BMD.

The available studies are generally of a low level of evidence, and do not support the hypothesis that spasticity maintains bone mineral in individuals with SCI.

Concluding remarks

The level of evidence is important and in this study area, there is no conclusive indication of any effective intervention. This review illustrates that there are issues important in the design of future clinical trials in this area. This includes where the BMD is measured (which is most relevant to be performed around the knee), and the length and intensity of the non-pharmacological intervention carried out. It is important to be aware that the stimulation used has to be maintained if the positive effect on BMD is to remain.

It has to be acknowledged that longitudinal randomized controlled investigations over very long periods on osteoporosis in individuals with SCI are difficult to carry through because of the relatively small number of persons with SCI available for the studies and a potentially high drop out rate over a long follow-up period. At the same time, adequate matching by gender, age, level and completeness of lesion, and time after injury is a challenge to the potential number of individuals recruited.⁴ Unfortunately, if the hypothesis regarding aggressive, high-intensity therapies is true, the ability to practically design and perform within patient evaluations to reduce sample size requirements as conducted by Ben et al.¹⁶ and Warden et al.⁵⁴ becomes more challenging.

The detection of an effective clinical intervention in this study area will require rigorous minimization of bias through a randomized controlled clinical trial design and likely require the involvement of multiple centers. It will also be dependent on appropriate measurement, with adequate intensity and duration of the intervention to detect a treatment effect. Any prospective intervention is likely to benefit from early timing after acute SCI.