Limitation of vWf meta-analysis in LMWH comparison.

Differences in clinical efficacy between unfractionated heparin (UFH) and low molecular weight heparin (LMWH) but also between several LMWH preparations are a hot topic in research due to differences in study results: enoxaparin proved superiority over UFH (1,2), whereas dalteparin and nadroparin showed only equivalence (3,4). As there is a lack of trials directly comparing LMWHs, some investigators tried to answer the question for the most effective LMWH preparation with data from comparative meta-analysis: Montalescot et al. (5) published a post hoc analysis in JACC in which they postulated differences between dalteparin and enoxaparin concerning their action on circulating von Willebrand factor (vWf) levels in patients with unstable angina. This could be of clinical importance as vWf has been shown previously to be a predictor of outcome in acute coronary syndrome (6). However, in scrutinizing the analytical method in Montalescot’s work, serious drawbacks are revealed. Montalescot et al. (5) presented a post hoc analysis that investigated, in part, results from two prospective trials with enoxaparin (1,2), one prospective trial with PEG-hirudin (PEGHIRUD 022) and one registry with dalteparin (USIC registry). A point of criticism is the comparison of prospective studies with a registry especially if randomized prospective data are available (3). In addition, the levels of vWf on UFH treatment in this meta-analysis were obtained by pooling data from the mentioned studies. These pooled UFH data were compared with the LMWH results of the individual studies and the registry, respectively. This is a questionable procedure and weakens the results. The proper way would have been to compare the pooled UFH data with pooled results from all cited LMWH studies. Finally, p values were only calculated for the comparisons of vWf levels between enoxaparin and, respectively, dalteparin and UFH, but not directly between the LMWHs themselves, which is, of course, impossible regarding the chosen analytical method. Only direct comparisons (also of other surrogate markers) could provide this data. In our opinion, it is not correct from a scientific point of view to draw any conclusions concerning the relative efficacy of dalteparin and enoxaparin from these results. Although a proper post hoc analysis can be a helpful tool in certain cases, the question for the most effective LMWH should preferably be answered by head-tohead studies. If these studies are not available (e.g., because of cost reasons), the investigation of surrogate markers of hemostasis could be feasible to predict clinical outcome and to compare different LMWH preparations.