Six studies in total were included: 4 RCTs (2 double-blind, 1 single-blind, 1 open) and 2 observational studies.
Mean change in Kurtzke's DSS scores measured at 6 and 12 months: 3 RCTs (209 patients) and 1 observational study (20 patients).
Mean change in Kurtzke's DSS scores at 36 months: 1 RCT (112 patients) and 1 observational study (20 patients).
Relative odds of neurological decline in treatment versus the comparison group: at 6-month follow-up, 3 RCTs (209 patients) and 1 observational study (20 patients); at 12-month follow-up 3 RCTs (209 patients) and 2 observational studies (40 patients); at 36-month follow-up, 1 RCT (112 patients) and 1 observational study (20 patients).
Relative odds of neurological improvement: at 6 and 12 months, 3 RCTs (209 patients) and 2 observational studies (40 patients); at 36 months, 1 RCT (112 patients) and 1 observational study (20 patients).
Mean change in Kurtzke's DSS scores at 6 months (treatment minus comparison): all studies, -0.171 (95% confidence interval, CI: -0.377, 0.039; Q statistic, P>0.05).
Randomised studies: -0.177 (95% CI: -0.386, 0.032; Q statistic, P<0.05). Sensitivity analysis after excluding 4 patients: -0.149 (95% CI: -0.359, 0.061; Q statistic, P>0.10).
Mean change in Kurtzke's DSS scores at 12 months: all studies, -0.212 (95% CI: -0.440, 0.017; Q statistic, P>0.05).
Randomised studies: -0.204 (95% CI: -0.437, 0.030; Q statistic, P<0.05). Sensitivity analysis after excluding 4 patients: -0.167 (95% CI: -0.400, 0.065; Q statistic, P>0.10).
Mean change in Kurtzke's DSS scores at 36 months: two studies, -0.104 (95% CI: -0.441, 0.232).
Relative odds of neurological decline in treatment versus comparison group at 6 months: all studies, 0.746 (95% CI: 0.328, 1.696; Q statistic, P>0.25). Randomised studies: 0.879 (95% CI: 0.373, 2.074; Q statistic, P>0.5).
Relative odds of neurological decline in treatment versus comparison group at 12 months: all studies, 0.436 (95% CI: 0.219, 0.871; P<0.05; Q statistic, P>0.25). Randomised studies: 0.441 (95% CI: 0.210, 0.929, P<0.05; Q statistic, P>0.25).
Relative odds of neurological decline in treatment versus comparison group at 36 months: two studies, 0.520 (95% CI: 0.259, 1.044).
Relative odds of neurological improvement in treatment versus comparison group at 6 months: all studies, 1.981 (95% CI: 1.024, 3.833, P<0.05; Q statistic, P>0.5). Randomised studies: 2.321 (95% CI: 1.116, 4.828, P<0.05; Q statistic, P>0.5).
Relative odds of neurological improvement in treatment versus comparison group at 12 months: all studies, 2.129 (95% CI: 1.032, 4.390, P<0.05; Q statistic, P>0.5). Randomised studies: 2.258 (95% CI: 1.091, 5.860, P<0.05; Q statistic, P>0.5).
Relative odds of neurological improvement in treatment versus comparison group at 36 months: 0.433 (95% CI: 0.102, 1.836).