Figure 3.

Segregation of clinical symptoms with mitochondrial function. The 95% CIs were calculated with control values for each mitochondrial outcome significantly different between controls and carriers of the premutation (citrate synthase activity and coupling). A) According to these cutoffs, all subjects were classified into 3 clusters, defined as no MD, mild MD, and severe MD. B) Means ± sem of citrate synthase activity and RCRu for each group. The P values (obtained by ANOVA followed by Tukey’s post hoc test for multiple comparisons) are as follows: 0.032 (c); 0.016 (a); and <0.0001 (b, d, e). C) The frequency of each clinical symptom (shown in Table 1) was counted within each of the 3 clusters (no MD, with mild MD, or with severe MD; A) and shown as a percentage of symptomatic individuals within any given class. Low IQ was set at <90. Executive function was considered impaired if ≤14. Statistical analyses were performed by the 1-tailed χ² test with Yates correction. Statistically significant clinical outcomes that segregated with MD were (in decreasing order of significance) low executive function (as judged by BDS-2; P = 0.0008), low IQ (P = 0.006), tremors (P = 0.045), and presence of FXTAS symptoms (P = 0.049). For these outcomes, the comparison of proportions between presence and absence of individual clinical symptoms was also computed. P values are as follows: 0.050 (d); 0.049 (g); 0.047 (e); 0.045 (h); 0.024 (c); 0.015 (f); 0.006 (a); and 0.0007 (b). D) Means ± sem of CGG repeats, IQ, and executive function scores (as judged by BDS-2) in the 3 groups. The P values (obtained by ANOVA followed by Tukey’s post hoc test for multiple comparisons) are as follows: 0.028 (a); 0.02 (e); 0.006 (d); 0.004 (c); and 0.002 (b).