The effectiveness of laceback ligatures during initial orthodontic alignment: a systematic review and meta-analysis.

Sir, We read the systematic review with great interest regarding the effectiveness of canine laceback by Fleming et al. (2012). The authors aimed to critically appraise the current evidence to determine the effectiveness of canine laceback in controlling incisor proclination during the alignment stage. We would like to congratulate the authors for their efforts in this valuable systematic review. After carefully reading this article, we would like to share a few concerns with readers of European Journal of Orthodontics. The authors included two studies (Usmani et al., 2002; Irvine et al., 2004) in the meta-analysis to determine the comparison between laceback and control regarding the changes (before and after treatment) in sagittal position of incisors (Figure 2). The two original studies had contradictory results: Usmani et al. (2002) revealed that laceback was effective in controlling incisor proclination [mean difference: −0.86; 95% confidence interval (CI): −1.57, −0.15], whereas Irvine et al. (2004) did not (mean difference: −0.09; 95% CI: −0.90, 0.72). However, the authors then performed a statistical pooling (mean difference: −0.50; 95% CI: −1.25, 0.25) and concluded that laceback was ineffective in controlling sagittal position of incisors. After carefully reading the two original studies, we found that several critical heterogeneities existed, which should have prevented the authors from the statistical pooling. Firstly and most importantly, Usmani et al. (2002) investigated incisor proclination on upper incisors, whereas Irvine et al. (2004) on lower incisors. Secondly, Usmani et al. (2002) employed reflex metrography, whereas Irvine et al. (2004) used lateral cephalometry for measurements. The two modalities differ significantly regarding their accuracy, with measurement error being 5 mm for lateral cephalometry (Gribel et al., 2011) and 0.2 mm for reflex metrography (Speculand et al., 1988). However, the authors performed meta-analysis without acknowledging these critical heterogeneities, which may be misleading and would bias the results. Thus, the conclusion that canine laceback is ineffective in controlling incisor proclination would be questioned. Moreover, in a systematic review, results in meta-analysis rely largely on the validity of included studies, and an assessment of risk of bias is essential (Higgin and Altman, 2008). In this systematic review, the authors draw the conclusion with confidence that the risks of bias were low in included studies. Specifically, as displayed in Table 2, although both studies had a dropout rate of 12.7 and 17 per cent, the authors evaluated both articles to be low risk of bias for the item ‘free of incomplete data’ due to balanced number of missing data across groups in both studies. However, Cochrane collaboration’s tools for assessing risk of bias state that even if the number of missing data was balanced across groups, bias could be introduced if reasons for missing outcome differed (Higgin and Altman, 2008). However, no evidence was mentioned regarding reasons for missing outcome in both studies. Thus, with such high dropout rate, this item would be assessed at most to be unclear risk of bias. Moreover, as reported previously, the baseline mesiodistal angulation of canine would influence incisor proclination (Usmani et al., 2002) and confound the evaluation of laceback effectiveness. However, the initial canine tip was not assessed in Irvine et al. (2004), and we are unable to know whether this confounding factor was well balanced between groups in this study. Thus, for the item ‘other apparent bias’ in the evaluation of risk of bias, it should be evaluated to be ‘unclear risk of bias’ rather than ‘low risk of bias’ for this study. Considering these risks of bias, the results from the included studies may not be reliable, which would further decrease the credence of the results in this systematic review. Therefore, with regards to inappropriate statistical pooling and unclear risks of bias in included studies, an alternative conclusion—whether canine laceback is effective in controlling incisor proclination cannot be determined based on current evidence—would be more appropriate.