Primary venous thromboembolism prophylaxis in patients with solid tumors.

Abstract

Our meta-analysis: ‘‘Primary venous thromboembolism prophylaxis in patients with solid tumors’’, was published on line in November 2013 [1]. We are writing to respond to comments made in a letter to the editor referencing our work. The abstracted and analyzed information in our metaanalysis was obtained independently by 2 reviewers. In the year since our initial abstraction and analysis, similar studies have found equivalent results to the question of primary thrombosis prevention in outpatients with cancer. Most recently, in a meta-analysis by DiNisio et al., including patients with all types of cancer the use of preventive low molecular weight heparin (LMWH) was associated with a reduction in venous thromboembolism (VTE) (RR 0.53, 95 % CI 0.38–0.75) [2]. This estimate is concordant with ours (OR 0.53; 95 %CI 0.41–0.70) despite small differences in methodology [1]. Note the risk ratio and odd ratio pooled point estimate were the same. Indeed, odd ratio is known to be an adequate estimate of risk ratio when the events are not very common [3]. In meta-analyses, due to mathematical advantages odds ratio are often used over risk ratios to standardize results of different trials [4], as we elected to do in ours. Although we acknowledge that there are important differences in the interpretation of risk ratios and odds ratios, and that the appropriateness of which to use depends on study design, our use of odd ratios was adequate. Tun adequately noted that 2 studies included arterial events in the definition of vascular thromboembolic events. Arterial events represented the minority of outcomes [5, 6]. However, if we exclude from our pooled analysis FRAGMEN and PROTECHT, the overall interpretation still translates into a significant reduction of VTE among patients with cancer who receive LMWH (OR 0.6; 95 %CI 0.42–0.9). Tun also argues that it was not appropriate to include a trial including upper extremity catheter associated thrombosis. Most clinicians will agree that an upper extremity deep vein thrombosis (DVT) is a concerning event regardless of the triggering event, as they convey a risk of pulmonary embolism, recurrent DVT and postthrombotic syndrome. Indeed, in the study by Young et al. two of the catheter related events were pulmonary embolisms, and four involved the superior vena cava [7]. In addition, we did provide estimates excluding catheter associated thrombosis. Although the trial by Young et al. was designed to evaluate a prophylactic regime to prevent catheter associated thrombosis (CAT) in cancer patients, they report all thrombotic events CAT and non-CAT, and those were the events included in our meta-analysis. Other studies included in our meta-analysis did not exclude CAT in the definition of their outcomes, for this reason and considering that CAT are prevalent in cancer patients we did not exclude them either. Finally, in patients with cancer it is recommended to treat catheter associated thrombosis with the same agents, intensity and duration as any other thrombosis [8]. This is a response to Letter to the Editor (doi: 10.1007/s11239-0141138-6). Original Article to this reply can be found at doi: 10.1007/ s11239-013-1014-9.