Statistical Analysis
All patients follow-up was done by outpatient service and telephone. All patients without outpatient follow-up records within 6 months were confirmed by telephone follow-up. Follow-up was updated as of 30th September, 2020.Complete blood counts, Hb composition analysis, transfusion volume were compared before and after treatment. Chi square tests were used to calculate comparisons between groups. Data were shown as mean (95% confidence interval) or median (range).Changes in HbF level were compared using Kruskal-Wallis H test between different groups of treatment response. One-way ANOVA analysis of variance or Kruskal-Wallis H test was performed to determine factors that predicted efficacy, including start dosage of thalidomide, thalidomide dosage after third month visit and age. Cochran-Mantel-Haenszel Chi-square test was performed to compare the distribution of variants of sex, genotype category, transfusion interval before thalidomide treatment, whether Hb decreased during the first month of treatment between different groups of treatment response.To evaluate the performance of independent risk factors in predicting major responders, the receiver-operating characteristic (ROC) curve analysis was performed and the optimal cut-off values were obtained from the Youden’s index. Diagnostic accuracy was expressed as the sensitivity, specificity, positive predictive value, negative predictive value, and the area under the ROC curve (AUC).Statistical analysis was performed by using SPSS 22.0 (SPSS Institute, Cary, NC). P values <0.05 were considered significant.