Twenty-one RCTs (n=1,479) were included. Data from 1,204 patients were used in the analyses.
All clinical and psychometric tests combined (21 RCTs).
The meta-analysis showed that Alcar significantly increased the ES compared with placebo (ES 0.201, 95% CI: 0.107, 0.295). Significant statistical heterogeneity was found (P=0.03). After excluding one RCT with a very high ES, heterogeneity was no longer statistically significant and the pooled ES remained significant (ES 0.191, 95% CI: 0.10, 0.29). CGI alone (11 RCTs).
The meta-analysis showed that Alcar significantly increased the ES compared with placebo (ES 0.32, 95% CI: 0.18, 0.47). Significant statistical heterogeneity was found (P<0.001).
Clinical tests alone.
Significant statistical heterogeneity was found for the meta-analysis of clinical test outcomes (P<0.001). Heterogeneity remained statistically significant after excluding one RCT with a very high ES.
Psychometric tests.
No significant statistical heterogeneity was detected for all psychometric test outcomes combined. A statistically significant effect of Alcar was found in the meta-analyses for memory, attention or performance, and higher intellectual function outcomes. The ES was 0.29 (95% CI: 0.132, 0.449) for memory (10 RCTs), 0.221 (95% CI: 0.090, 0.352) for attention or performance (13 RCTs), and 0.373 (95% CI: 0.173, 0.573) for intellectual function (7 RCTs). The effect of Alcar was not statistically significant for drawing (6 RCTs) or learning (5 RCTs).
Time course (16 RCTs): the ES for clinical and psychometric outcomes increased from 3 to 6 months, but clinician-rating outcomes (CGI) did not increase. Alcar dose: no significant relationship was found between the Alcar dose and ES.
Side-effects: the type and severity of side-effects and adverse events were similar between the treatment groups. Some studies reported more frequent gastrointestinal events, insomnia, agitation and appetite increase with Alcar. The number of early drop-outs was small and similar between treatment groups.
The funnel plot showed no evidence of publication bias.