We read the recent published paper in this journal of J Gastrointest Oncol by Zhang and colleagues entitled "Secondary colon cancer in patients with ulcerative colitis: a systematic review and meta-analysis" (1). They performed a systematic review and meta-analysis to assess the correlation between ulcerative colitis (UC) and colon cancer. We appreciate Zhang et al. (1) for the valuable study, however, after a careful learning of the literature, several limitations should be noticed.First, in the results section of the abstract, Zhang et al.(1) performed the meta-analysis by random-effect model because of statistical heterogeneity (Chi 2 =103.10; I 2 =90%; P<0.00001) and found that there were no significant differences between colon cancer in patients with UC and patients without colon carcinoma (Z =12.44; P<0.00001). However, we believe the interpretation of the results was false. There should be significant difference due to P<0.00001.Second, in the statistical methods section of this article, Zhang et al. (1) stated that the odds ratio (OR) was used as an effect size for dichotomous variables. Whereas, in this meta-analysis, the effect size actually was relative risk (RR) showed in figures 5,6 and the OR was not reported in the study. Therefore, we believe the irrelevant effect size depicted would lead to misunderstanding.

ObjectiveThe combination of mifepristone and misoprostol is an established method for induction of early first trimester abortion, but there is no consensus about the best evaluation of treatment outcome. We evaluate serum Angiopoietin-2(Ang-2) and β human chorionic gonadotropin (β-hCG) in women who had undergone a medical abortion as markers of prolonged uterine bleeding (PUB).MethodsProspective trial involving 2843 women attending an gynecology outpatient clinic who following a medical abortion with mifepristone and misoprostol, the study cohort was divided into women with duration of uterine bleeding >14 days (PUB) and women with duration of uterine bleeding ≤14 days (normal uterine bleeding, NUB). Serum determinations of Ang-2 levels by ELISA and β-hCG levels by electrochemiluminiscence immunoassay. Receiver Operating Characteristics (ROC) analyses were calculated and plotted for the diagnostic accuracy of serum β-hCG and Ang-2 concentration to discriminate PUB and NUB.ResultsBaseline characteristics for both groups were similar, Only duration of bleeding showed a significant difference between the PUB group and NUB group. Ang-2 serum levels moderately correlated with serum β-hCG levels with statistically significant correlation coefficients of 0.536. Serum β-hCG and Ang-2 levels on day 7 and on day 14 after medical abortion were signifcantly higher in PUB group than in NUB group. Plotted as ROC curves, β-hCG area under curve (AUC) was 0.65 (95% CI, 0.53–0.76) on day 7, rising to AUC = 0.83 (95% CI, 0.75–0.92) on day 14. Using Ang-2 on day 7 and day 14 as predictive parameter resulted in an analogous AUC (AUC = 0.61 on day 7, AUC = 0.78 on day 14).ConclusionsBoth parameters are clinically useful as a diagnostic test in predicting PUB after medical abortion, and can be helpful in uncertain clinical situations, but should be considered as supplementary to a general clinical evaluation.

No abstract