Search

BackgroundNonmajor orthopedic surgery of the lower limbs that results in transient reduced mobility places patients at risk for venous thromboembolism. Rivaroxaban may be noninferior to enoxaparin with regard to the prevention of major venous thromboembolism in these patients. MethodsIn this international, parallel-group, randomized, double-blind, noninferiority trial, we randomly assigned adult patients undergoing lower-limb nonmajor orthopedic surgery who were considered to be at risk for venous thromboembolism on the basis of the investigator's judgment to receive either rivaroxaban or enoxaparin. The primary efficacy outcome of major venous thromboembolism was a composite of symptomatic distal or proximal deep-vein thrombosis, pulmonary embolism, or venous thromboembolism-related death during the treatment period or asymptomatic proximal deep-vein thrombosis at the end of treatment. A test for superiority was planned if rivaroxaban proved to be noninferior to enoxaparin. For all outcomes, multiple imputation was used to account for missing data. Prespecified safety outcomes included major bleeding (fatal, critical, or clinically overt bleeding or bleeding at the surgical site leading to intervention) and nonmajor clinically relevant bleeding. ResultsA total of 3604 patients underwent randomization; 1809 patients were assigned to receive rivaroxaban, and 1795 to receive enoxaparin. Major venous thromboembolism occurred in 4 of 1661 patients (0.2%) in the rivaroxaban group and in 18 of 1640 patients (1.1%) in the enoxaparin group (risk ratio with multiple imputation, 0.25; 95% confidence interval, 0.09 to 0.75; P

International audience; Introduction: Over the last decade, antitumor necrosis factor (anti TNF) therapy became the cornerstone management for immune mediated diseases such as inflammatory bowel disease or rheumatoid arthritis. Such conditions predominantly affect women of childbearing age, and a growing number of studies have been published on the safety of anti-TNF therapy during pregnancy. Meta-analyses (MA) could help synthesize findings as well as increase statistical power for rare outcomes such as congenital malformations (CM). Thus, the aim of our work was to assess the risk of CM (major and minor) associated with first trimester exposure to anti TNF using a meta-analytic approach. Material and methods: We conducted a systematic review and MA. Published reports and meeting abstracts were searched for in MEDLINE up to December 2018. We used both MeSH and free text terms to identify relevant articles. Bibliographic reference lists of all identified studies, MA and reviews were hand searched in order to identify additional eligible articles. All comparative cohort and case-control studies reporting data on malformation, and use of anti-TNF during the first trimester were included. An experimented pharmacologist evaluated the eligibility of studies, and extracted data using a proprietary collaborative WEB-based meta-analysis platform (metaPreg.org). The risk of CM between anti-TNF exposed and non-exposed pregnancies, and its 95% confidence interval were estimated using a random effects model. Results: From 1,014 records initially found, 15 studies met eligibility criteria. In addition, the MA was restricted to studies that included diseased unexposed women as a comparator. Ten cohort studies (5 prospective and 5 retrospective), reporting 1,407 exposed, and 6,196 unexposed pregnancies, were included. We included 3 new studies (909 exposed pregnancies) compared to the last published MA on the risk of congenital malformations associated with first trimester exposure to anti-TNF in 2017 (1). First trimester exposure to anti-TNF was not statistically significantly associated with the risk of CM (odd ratio of 0.98 , 95% CI 0.75-1.28). No heterogeneity between studies was observed (I2=0%). Funnel plot and asymmetry test were not suggestive of publication bias. Discussion / Conclusion: We did not find an association between exposure to anti-TNF in pregnancy and an increased risk of CMs. However, a lack of power cannot be ruled out. The risk of CMs after exposure to anti-TNF would be increased at most by by 30%. Reference(s): 1. Komaki et al. J Autoimmun.2017

Dear Sir Whether or not maintenance chemotherapy might improve survival in small cell lung cancer (SCLC) still remains a debated and unanswered question. We have been very much interested by the review of the literature recently published in the present journal by J.P. Sculier et al. [1]. In fact, our group has also recently planned to carry out a meta-analysis on this topic, based on the published data from 15 randomised trials. Interestingly enough, although the parameters we selected to assess the feasibility of such a study were quite different, we came to the same conclusion that it is unpractical because of a too high rate of missing data. Sculier et al. [1] have used a quality evaluation based on the use of two scores, to assess the methodological validity of a trial (Chalmers score and European Lung Cancer Working party score). A large part of their article is devoted to the description and analysis of the respective scores obtained by each of the 15 trials. Despite the fact that they have sometimes been used in meta-analysis for the selection of the trials, quality scores are currently discarded. The relationship between a quality score and the degree to which a trial is free from bias is not obvious [2]. Moreover, the majority of the items contributing to the quality scores are not relevant for the purpose of the meta-analysis and a low score based on these kinds of items do not have any repercussion on the quality of the result of the meta-analysis process. For example, an incorrect statistical analysis is not detrimental in a meta-analysis where only the crude data are used. The lack of primary endpoint cannot challenge the inclusion of the trial because the meta-analysis defines its own primary endpoint. The absence of an a priori determination of sample size does not lead to a biased result but only affects the power of the comparison that will be enhanced by the pooling of several trials in the meta-analysis. A study by Schulz et al. [1] has provided empirical evidence that mostly it is a poor allocation concealment that is associated with bias [2]. Thus, some recent guidelines for meta-analysis [4] propose that trial selection is based only to the following items: the concealment of the randomisation, the blinding of the follow-up (if applicable) and the rate of lost to follow-up. Following this approach, we conclude that 13/15 of the trials were sufficiently valid to be pooled in a metaanalysis. As far as the precise question of the feasibility of a meta-analysis is concerned, the negative con* Corresponding author. Medical Oncology Unit, Centre Hospitalier Lyon Sud, Pierre Benite, 69495 Cedex, France

article