MethodsWe screened a hospital-based cohort of Chinese patients with sporadic ALS without C9orf72 repeat expansions and neurologically healthy controls for C9orf72 GGGGCC and AXTN2 CAG repeat length to compare the frequency of possible detrimental length alleles using several thresholds. Furthermore, the clinical features of ALS were compared between patients with ALS subgroups using different length thresholds of maximum C9orf72 and ATXN2 repeat alleles, such as sex, age of onset, diagnostic delay, and survival.
ResultsOverall, 879 sporadic patients with ALS and 535 controls were included and the repeat lengths of the C9orf72 and ATXN2 were both detected. We found significant survival differences in patients using a series of C9orf72 repeat length thresholds from 2 to 5, among which the most significant difference was at the cutoff value of 2 (repeats 2 vs. >2: median survival 67 vs. 55 months, log-rank p = 0.032). Furthermore, Cox regression analysis revealed the role of age of onset [hazard ratio (HR) 1.04, 95% CI 1.03–1.05, p < 0.001], diagnostic delay (0.95, 0.94–0.96, p < 0.001), and carrying C9orf72 repeat length of 2 (0.72, 0.59–0.89, p = 0.002) in the survival of patients without C9orf72 repeat expansions. In addition, bulbar onset was associated with poorer survival when the patients carried the maximum C9orf72 repeat allele over 2 (1.81, 1.32–2.48, p < 0.001). However, no survival difference was found when applying a series of continuous cutoff values of ATXN2 or stratified by C9orf72 repeats of 2.