Stem cell transplantation in traumatic spinal cord injury: a systematic review and meta-analysis of animal studies

Abstract

Spinal cord injury (SCI) is a devastating condition that causes substantial morbidity and mortality and for which no treatments are available. Stem cells offer some promise in the restoration of neurological function. We used systematic review, meta-analysis, and meta-regression to study the impact of stem cell biology and experimental design on motor and sensory outcomes following stem cell treatments in animal models of SCI. One hundred and fifty-six publications using 45 different stem cell preparations met our prespecified inclusion criteria. Only one publication used autologous stem cells. Overall, allogeneic stem cell treatment appears to improve both motor (effect size, 27.2%; 95% Confidence Interval [CI], 25.0%-29.4%; 312 comparisons in 5,628 animals) and sensory (effect size, 26.3%; 95% CI, 7.9%-44.7%; 23 comparisons in 473 animals) outcome. For sensory outcome, most heterogeneity between experiments was accounted for by facets of stem cell biology. Differentiation before implantation and intravenous route of delivery favoured better outcome. Stem cell implantation did not appear to improve sensory outcome in female animals and appeared to be enhanced by isoflurane anaesthesia. Biological plausibility was supported by the presence of a dose-response relationship. For motor outcome, facets of stem cell biology had little detectable effect. Instead most heterogeneity could be explained by the experimental modelling and the outcome measure used. The location of injury, method of injury induction, and presence of immunosuppression all had an impact. Reporting of measures to reduce bias was higher than has been seen in other neuroscience domains but were still suboptimal. Motor outcomes studies that did not report the blinded assessment of outcome gave inflated estimates of efficacy. Extensive recent preclinical literature suggests that stem-cell-based therapies may offer promise, however the impact of compromised internal validity and publication bias mean that efficacy is likely to be somewhat lower than reported here.

Figure 1. Summary of data included in meta-analysis of use of stem cells to treat spinal cord injury with individual comparisons ranked according to their effect on (A) % improvement in motor score and (B) % improvement in sensory score.

The shaded grey bar represents the 95% confidence limits of the global estimate. The vertical error bars represent the 95% confidence intervals for the individual estimates.

Figure 2. Assessment of publication bias shown with (A) Funnel plot and (B) Egger regression.

Figure 3. Study characteristics which account for heterogeneity of total motor dataset.

(A) Behavioural test used, (B) location of injury, (C) sex of animals, (D) immunosupressant used, (E) type of Injury, (F) stem cell source, and (G) effect of blinding. The shaded grey bar represents the 95% confidence limits of the global estimate. The vertical error bars represent the 95% confidence intervals for the individual estimates.

Figure 4. Study characteristics that account for heterogeneity of motor data subanalysis when only data from rats implanted with allogeneic stem cells after injury created with an impactor at the midthoracic level and assessed by BBB.

(A) Anaesthetic used, (B) immunosupressant used, and (C) influence of additional behavioural testing on BBB. The shaded grey bar represents the 95% confidence limits of the global estimate. The vertical error bars represent the 95% confidence intervals for the individual estimates.

Figure 5. Study characteristics that account for heterogeneity in sensory score.

(A) Type of manipulation of stem cells prior to implantation, (B) dose–response relationship, (C) route of stem cell delivery, (D) anaesthetic used, and (E) sex. The shaded grey bar represents the 95% confidence limits of the global estimate. The vertical error bars represent the 95% confidence intervals for the individual estimates.