Author + information
- Pierfrancesco Agostoni, MD* ()
- Giuseppe G.L. Biondi-Zoccai, MD and
- Antonio Abbate, MD
We read with interest the study by Moreno et al. (1), which demonstrated, through ameta-analytic technique, the significant reduction of restenosis due to stent incomparison to percutaneous transluminal coronary angioplasty in small-vessel coronaryartery disease. However, the investigators failed to emphasize that substantial statisticalheterogeneity was present, as shown in their overall analysis in Figure 1 (p forheterogeneity = 0.019).
We believe that this constitutes a methodological flaw in the study. Indeed, substantialheterogeneity is considered by several investigators to be a contraindication to quantitative poolinganalysis (2), although other authorities recommend the need for summary measures forthe best estimate of the impact of an intervention, albeit that correct methodologicaltechniques are used to investigate the differences among single trials (3).
Furthermore, Moreno et al. (1) tried to assess, by means of linear regression analysis, thepossible association of several angiographic variables with the benefit of stents,finding a significant inverse relationship between reference vessel diameter (RVD) and the risk reduction of angiographic restenosis after stent placement with respect to balloon angioplasty (i.e., the smaller the vessel, the larger the benefit of stenting). However, we believe the correct methodological tool to address whether a covariable mayhave a significant effect on the outcome in a meta-analysis is meta-regression (4).Using this technique (which weights each study according to its statistical weight [i.e.,the inverse of the variance]) to analyze the very same data presented in their study, wedid not find any significant relation between RVD and risk reduction of angiographicrestenosis after stent placement.
Moreover, another recent trial, published only as an abstract (5), did not confirm theinvestigators' hypothesis. Indeed, in front of the smallest mean RVD of all the trials on small-vesseldisease (2.17 mm), there was a nonsignificantly increased risk of restenosis (relativerisk 1.14 [95% confidence interval 0.89 to 1.48]) after stenting.
Finally, meta-analytic techniques allow quantitative assessment of treatment effects from pooleddata. With the widening of their use and because of potential errors, improper analysesmay result in misleading conclusions; thus, optimal methodological procedures shouldbe utilized to validate findings.
- American College of Cardiology Foundation
- Moreno R.,
- Fernandez C.,
- Alfonso F.,
- et al.
- Petitti D.B.
- Hausleiter J.,
- Kastrati A.,
- Mehilli J.,
- et al.