General Discussion

The risk for fragility fractures after SCI has been established and low bone mass is an important factor to be considered. In 2002, the Canadian Medical Association published clinical practice guidelines for prevention and treatment of bone health (Brown et al. 2002). Currently, these guidelines do not specifically address persons with SCI. While, they do provide a resource for osteoporosis diagnosis, prevention and treatment, the lack of SCI-specific, consensus-based guidelines for SLOP, has resulted in diverse SLOP screening, prevention, and treatment practices among SCI clinicians (Morse et al. 2009a; Ashe et al. 2009). Hopefully, future national guidelines will provide recommendations for people who have SCI and diverse impairments that lead to reduced weight-bearing, muscle activity and physical activity levels. Recently a decision guide has been published for rehabilitation professionals on the identification and management of bone health-related issues for people with SCI (Craven et al. 2008, Craven et al. 2009).

In this review, we note some support for pharmacological agents, but less support for rehabilitation modalities for the prevention and management of bone health in people with SCI. Our results have some similarities with the recent systematic review by Bryson and Gourlay (2009). Our results for the non-pharmacological treatment of bone health are consistent with the review by Bering-Sorensen and colleagues (2009) highlighting promise with some modalities. However, this review differs by reporting evidence for early (acute) and late (>12 months) intervention with rehabilitation modalities and therefore describes the results based on whether the goal of therapy is prevention or treatment of SLOP. In the past 40 years, there have been several interventions (both pharmacological and rehabilitation modalities) aimed to maintain or slow down bone mass decline after SCI yet consistent methodological oversights have emerged including: small sample sizes and broad inclusion criteria that do not always account for sex, TPI or impairment differences between participants.

The pharmacological interventions (either prevention or treatment interventions) discussed here report stronger methodologies— all except one were RCTs with PEDro scores ranging from 6-10 indicating moderate to high quality. In contrast, the studies employing rehabilitation modalities had low numbers of participants and only 3 of the 31 studies were RCTs. These factors contribute to difficulties drawing generalizable conclusions regarding the impact of rehab interventions on bone parameters.

Nonetheless, despite the lack of evidence to establish the effectiveness of these rehab modalities on bone parameters, it does not negate these treatments as beneficial to other body systems. For example, NMES and FES-cycling may have small effects on bone but have been shown to have large effects on muscular and cardiovascular health (Jacobs & Nash 2004).

There is a large body of studies that suggest an infinite number of combinations of NMES/FES types and stimulation parameters that could be used during FES/NMES-training. This may be one of the causes of the large heterogeneity of the results when using these techniques. It has long been suggested that the individual capacity of generating relevant torque amplitudes during electrically-evoked contractions is the main factor to maximize training effectiveness (Lieber et al. 1991). As elegantly discussed by Maffiuletti et al. (2017), future NMES-based studies should reduce their focus on NMES types and parameters and increase the emphasis on clinically desired outcomes taking into consideration individual- and impairment-specific requirements. Moreover, the participant’s tolerance to NMES should be taken into consideration since not everyone can tolerate or increase their tolerance over time (i.e., responders vs non-responders). It was suggested that a minimum of >15% of a maximal voluntary contraction is necessary to reach the therapeutic window range (Maddocks et al. 2016).

There are a few key points to consider when interpreting the results from interventions designed to maintain and/or improve bone parameters after SCI. These include biological differences in bone development and maintenance between men and women, the natural decline in bone mass with ageing and the selected primary outcome measure. Age-related changes in bone mass affect both women and men but the pattern of change is different because estrogen plays such a dominant role in bone remodelling. Consequently, in women, the loss of estrogen at menopause initiates a rapid loss of bone that eventually slows but continues throughout life. Men generally do not experience the rapid phase of BMD decline with ageing rather, only a slower phase of BMD decline is observed. Therefore, keeping in mind that bone mass declines over time, a study that reports no significant difference in BMD between two time periods 6 months apart may be interpreted as positive because of the anticipated loss.

Due to the diversity of primary outcomes (BMD by DPA, DXA or pQCT, urine or blood markers and NMES/FES/vibration types and parameters), it is difficult to pool the results from multiple studies. When measuring parameters such as urine or blood biomarkers, studies of short duration may yield significant results. However, using imaging, cortical bone remodelling can take at least 9 months to observe changes within participants over time. Consequently, investigations that did not maintain an intervention for at least 6 months may not show changes, and the results cannot be interpreted as negative. Importantly, all primary outcomes for bone health after SCI are surrogate measures, that is, there has yet to be a study published in this area that investigates the effect of an intervention (either pharmacological or non-pharmacological) on fracture reduction. Fracture reduction studies are somewhat infeasible due to cost and the large number of participants that would be needed and followed longitudinally. Consequently, the clinical significance of the interventions based on fractures for this population remains to be determined. Prospective multicentre intervention studies using common interventions and outcome assessments are urgently needed.


There is a significant risk for fragility fractures after SCI. The risk increases for women, people with motor complete injuries (AIS A and B), longer duration of injury, and with use of benzodiazepines, heparin, or opioid analgesia. Early assessment and ongoing monitoring of bone health are essential elements of SCI care.

There is Level 1 evidence for the prevention and treatment of BMD decline using medications. However, non-pharmacological evidence for preventing a decline in bone mass and treating low bone mass is poor. Interpretation and pooling of bone health studies are limited by small samples, diverse treatment protocols, heterogeneous samples (regarding impairment and injury duration) and short treatment durations given the time required to detect improvements in bone parameters and variability associated with different imaging technologies. As noted in two publications (Craven et al. 2008, Ashe et al. 2009), a consensus regarding the ideal duration of therapy and choice of outcome measures would advance the field.