Conventional NMA as a better standard for evaluating elastic network models

Normal mode analysis (NMA) is an important tool for studying protein dynamics. Because of the complexity of conventional NMA that uses an all‐atom model and a semi‐empirical force field, many simplified NMA models have been developed, some of which are known as elastic network models. The quality of these simplified NMA models was assessed mostly by evaluating their predictions against experimental B‐factors, and rarely by comparing them with the original NMA. In this work, we take the effort to create a publicly accessible dataset of proteins with their minimized structures, NMA modes, and mean‐square fluctuations. Then, for the first time, we evaluate the quality of individual normal modes of several widely used elastic network models by comparing them with the conventional NMA. Our results demonstrate that the conventional NMA presents a better and more complete evaluation measure of the quality of elastic network models. This realization should be very helpful in improving current or designing new, higher quality elastic network models. Moreover, using the conventional NMA as the standard of evaluation, a number of interesting and significant insights into the elastic network models are gained. Proteins 2015; 83:259–267. © 2014 Wiley Periodicals, Inc.     

Network Meta-Analysis Using R: A Review of Currently Available Automated Packages

Network meta-analysis (NMA) – a statistical technique that allows comparison of multiple treatments in the same meta-analysis simultaneously – has become increasingly popular in the medical literature in recent years. The statistical methodology underpinning this technique and software tools for implementing the methods are evolving. Both commercial and freely available statistical software packages have been developed to facilitate the statistical computations using NMA with varying degrees of functionality and ease of use. This paper aims to introduce the reader to three R packages, namely, gemtc, pcnetmeta, and netmeta, which are freely available software tools implemented in R. Each automates the process of performing NMA so that users can perform the analysis with minimal computational effort. We present, compare and contrast the availability and functionality of different important features of NMA in these three packages so that clinical investigators and researchers can determine which R packages to implement depending on their analysis needs. Four summary tables detailing (i) data input and network plotting, (ii) modeling options, (iii) assumption checking and diagnostic testing, and (iv) inference and reporting tools, are provided, along with an analysis of a previously published dataset to illustrate the outputs available from each package. We demonstrate that each of the three packages provides a useful set of tools, and combined provide users with nearly all functionality that might be desired when conducting a NMA.     

Bridging between normal mode analysis and elastic network models

Normal mode analysis (NMA) has been a powerful tool for studying protein dynamics. Elastic network models (ENM), through their simplicity, have made normal mode computations accessible to a much broader research community and for many more biomolecular systems. The drawback of ENMs, however, is that they are less accurate than NMA. In this work, through steps of simplification that starts with NMA and ends with ENMs we build a tight connection between NMA and ENMs. In the process of bridging between the two, we have also discovered several high‐quality simplified models. Our best simplified model has a mean correlation with the original NMA that is as high as 0.88. In addition, the model is force‐field independent and does not require energy minimization, and thus can be applied directly to experimental structures. Another benefit of drawing the connection is a clearer understanding why ENMs work well and how it can be further improved. We discovered that can be greatly enhanced by including an additional torsional term and a geometry term. Proteins 2014; 82:2157–2168. © 2014 Wiley Periodicals, Inc.     

The shaping of intrinsic membrane potential oscillations: positive/negative feedback,ionic resonance/amplification,nonlinearities and time scales

The generation of intrinsic subthreshold (membrane potential) oscillations (STOs) in neuronal models requires the interaction between two processes: a relatively fast positive feedback that favors changes in voltage and a slower negative feedback that opposes these changes. These are provided by the so-called resonant and amplifying gating variables associated to the participating ionic currents. We investigate both the biophysical and dynamic mechanisms of generation of STOs and how their attributes (frequency and amplitude) depend on the model parameters for biophysical (conductance-based) models having qualitatively different types of resonant currents (activating and inactivating) and an amplifying current. Combinations of the same types of ionic currents (same models) in different parameter regimes give rise to different types of nonlinearities in the voltage equation: quasi-linear, parabolic-like and cubic-like. On the other hand, combinations of different types of ionic currents (different models) may give rise to the same type of nonlinearities. We examine how the attributes of the resulting STOs depend on the combined effect of these resonant and amplifying ionic processes, operating at different effective time scales, and the various types of nonlinearities. We find that, while some STO properties and attribute dependencies on the model parameters are determined by the specific combinations of ionic currents (biophysical properties), and are different for models with different such combinations, others are determined by the type of nonlinearities and are common for models with different types of ionic currents. Our results highlight the richness of STO behavior in single cells as the result of the various ways in which resonant and amplifying currents interact and affect the generation and termination of STOs as control parameters change. We make predictions that can be tested experimentally and are expected to contribute to the understanding of how rhythmic activity in neuronal networks emerge from the interplay of the intrinsic properties of the participating neurons and the network connectivity.     

Impact of reporting bias in network meta-analysis of antidepressant placebo-controlled trials

Background

Indirect comparisons of competing treatments by network meta-analysis (NMA) are increasingly in use. Reporting bias has received little attention in this context. We aimed to assess the impact of such bias in NMAs.

Methods

We used data from 74 FDA-registered placebo-controlled trials of 12 antidepressants and their 51 matching publications. For each dataset, NMA was used to estimate the effect sizes for 66 possible pair-wise comparisons of these drugs, the probabilities of being the best drug and ranking the drugs. To assess the impact of reporting bias, we compared the NMA results for the 51 published trials and those for the 74 FDA-registered trials. To assess how reporting bias affecting only one drug may affect the ranking of all drugs, we performed 12 different NMAs for hypothetical analysis. For each of these NMAs, we used published data for one drug and FDA data for the 11 other drugs.

Findings

Pair-wise effect sizes for drugs derived from the NMA of published data and those from the NMA of FDA data differed in absolute value by at least 100% in 30 of 66 pair-wise comparisons (45%). Depending on the dataset used, the top 3 agents differed, in composition and order. When reporting bias hypothetically affected only one drug, the affected drug ranked first in 5 of the 12 NMAs but second (n = 2), fourth (n = 1) or eighth (n = 2) in the NMA of the complete FDA network.

Conclusions

In this particular network, reporting bias biased NMA-based estimates of treatments efficacy and modified ranking. The reporting bias effect in NMAs may differ from that in classical meta-analyses in that reporting bias affecting only one drug may affect the ranking of all drugs.     

Meta-Analysis of a Complex Network of Non-Pharmacological Interventions: The Example of Femoral Neck Fracture

Background

Surgical interventions raise specific methodological issues in network meta-analysis (NMA). They are usually multi-component interventions resulting in complex networks of randomized controlled trials (RCTs), with multiple groups and sparse connections.

Purpose

To illustrate the applicability of the NMA in a complex network of surgical interventions and to prioritize the available interventions according to a clinically relevant outcome.

Methods

We considered RCTs of treatments for femoral neck fracture in adults. We searched CENTRAL, MEDLINE, EMBASE and ClinicalTrials.gov up to November 2015. Two reviewers independently selected trials, extracted data and used the Cochrane Collaboration’s tool for assessing the risk of bias. A group of orthopedic surgeons grouped similar but not identical interventions under the same node. We synthesized the network using a Bayesian network meta-analysis model. We derived posterior odds ratios (ORs) and 95% credible intervals (95% CrIs) for all possible pairwise comparisons. The primary outcome was all-cause revision surgery.

Results

Data from 27 trials were combined, for 4,186 participants (72% women, mean age 80 years, 95% displaced fractures). The median follow-up was 2 years. With hemiarthroplasty (HA) and total hip arthroplasty (THA) as a comparison, risk of surgical revision was significantly higher with the treatments unthreaded cervical osteosynthesis (OR 8.0 [95% CrI 3.6–15.5] and 5.9 [2.4–12.0], respectively), screw (9.4 [6.0–16.5] and 6.7 [3.9–13.6]) and plate (12.5 [5.8–23.8] and 7.8 [3.8–19.4]).

Conclusions

In older women with displaced femoral neck fractures, arthroplasty (HA and THA) is the most effective treatment in terms of risk of revision surgery.

Systematic Review Registration

PROSPERO no. CRD42013004218.

Level of Evidence

Network Meta-Analysis, Level 1.     

Investigating meta-approaches for reconstructing gene networks in a mammalian cellular context

The output of state-of-the-art reverse-engineering methods for biological networks is often based on the fitting of a mathematical model to the data. Typically, different datasets do not give single consistent network predictions but rather an ensemble of inconsistent networks inferred under the same reverse-engineering method that are only consistent with the specific experimentally measured data. Here, we focus on an alternative approach for combining the information contained within such an ensemble of inconsistent gene networks called meta-analysis, to make more accurate predictions and to estimate the reliability of these predictions. We review two existing meta-analysis approaches; the Fisher transformation combined coefficient test (FTCCT) and Fisher's inverse combined probability test (FICPT); and compare their performance with five well-known methods, ARACNe, Context Likelihood or Relatedness network (CLR), Maximum Relevance Minimum Redundancy (MRNET), Relevance Network (RN) and Bayesian Network (BN). We conducted in-depth numerical ensemble simulations and demonstrated for biological expression data that the meta-analysis approaches consistently outperformed the best gene regulatory network inference (GRNI) methods in the literature. Furthermore, the meta-analysis approaches have a low computational complexity. We conclude that the meta-analysis approaches are a powerful tool for integrating different datasets to give more accurate and reliable predictions for biological networks.     

Network quantification of EGFR signaling unveils potential for targeted combination therapy

The epidermal growth factor receptor (EGFR) signaling network is activated in most solid tumors, and small‐molecule drugs targeting this network are increasingly available. However, often only specific combinations of inhibitors are effective. Therefore, the prediction of potent combinatorial treatments is a major challenge in targeted cancer therapy. In this study, we demonstrate how a model‐based evaluation of signaling data can assist in finding the most suitable treatment combination. We generated a perturbation data set by monitoring the response of RAS/PI3K signaling to combined stimulations and inhibitions in a panel of colorectal cancer cell lines, which we analyzed using mathematical models. We detected that a negative feedback involving EGFR mediates strong cross talk from ERK to AKT. Consequently, when inhibiting MAPK, AKT activity is increased in an EGFR‐dependent manner. Using the model, we predict that in contrast to single inhibition, combined inactivation of MEK and EGFR could inactivate both endpoints of RAS, ERK and AKT. We further could demonstrate that this combination blocked cell growth in BRAF‐ as well as KRAS‐mutated tumor cells, which we confirmed using a xenograft model.     

Efficacy of Supplementation with B Vitamins for Stroke Prevention: A Network Meta-Analysis of Randomized Controlled Trials

Background

Supplementation with B vitamins for stroke prevention has been evaluated over the years, but which combination of B vitamins is optimal for stroke prevention is unclear. We performed a network meta-analysis to assess the impact of different combinations of B vitamins on risk of stroke.

Methods

A total of 17 trials (86 393 patients) comparing 7 treatment strategies and placebo were included. A network meta-analysis combined all available direct and indirect treatment comparisons to evaluate the efficacy of B vitamin supplementation for all interventions.

Results

B vitamin supplementation was associated with reduced risk of stroke and cerebral hemorrhage. The risk of stroke was lower with folic acid plus vitamin B6 as compared with folic acid plus vitamin B12 and was lower with folic acid plus vitamin B6 plus vitamin B12 as compared with placebo or folic acid plus vitamin B12. The treatments ranked in order of efficacy for stroke, from higher to lower, were folic acid plus vitamin B6 > folic acid > folic acid plus vitamin B6 plus vitamin B12 > vitamin B6 plus vitamin B12 > niacin > vitamin B6 > placebo > folic acid plus vitamin B12.

Conclusions

B vitamin supplementation was associated with reduced risk of stroke; different B vitamins and their combined treatments had different efficacy on stroke prevention. Folic acid plus vitamin B6 might be the optimal therapy for stroke prevention. Folic acid and vitamin B6 were both valuable for stroke prevention. The efficacy of vitamin B12 remains to be studied.     

The performance of fine‐grained and coarse‐grained elastic network models and its dependence on various factors

In a recent work we developed a method for deriving accurate simplified models that capture the essentials of conventional all‐atom NMA and identified two best simplified models: ssNMA and eANM, both of which have a significantly higher correlation with NMA in mean square fluctuation calculations than existing elastic network models such as ANM and ANMr2, a variant of ANM that uses the inverse of the squared separation distances as spring constants. Here, we examine closely how the performance of these elastic network models depends on various factors, namely, the presence of hydrogen atoms in the model, the quality of input structures, and the effect of crystal packing. The study reveals the strengths and limitations of these models. Our results indicate that ssNMA and eANM are the best fine‐grained elastic network models but their performance is sensitive to the quality of input structures. When the quality of input structures is poor, ANMr2 is a good alternative for computing mean‐square fluctuations while ANM model is a good alternative for obtaining normal modes. Proteins 2015; 83:1273–1283. © 2015 Wiley Periodicals, Inc.     

Bayesian meta-analysis for longitudinal data models using multivariate mixture priors

We propose a class of longitudinal data models with random effects that generalizes currently used models in two important ways. First, the random-effects model is a flexible mixture of multivariate normals, accommodating population heterogeneity, outliers, and nonlinearity in the regression on subject-specific covariates. Second, the model includes a hierarchical extension to allow for meta-analysis over related studies. The random-effects distributions are decomposed into one part that is common across all related studies (common measure), and one part that is specific to each study and that captures the variability intrinsic between patients within the same study. Both the common measure and the study-specific measures are parameterized as mixture-of-normals models. We carry out inference using reversible jump posterior simulation to allow a random number of terms in the mixtures. The sampler takes advantage of the small number of entertained models. The motivating application is the analysis of two studies carried out by the Cancer and Leukemia Group B (CALGB). In both studies, we record for each patient white blood cell counts (WBC) over time to characterize the toxic effects of treatment. The WBCs are modeled through a nonlinear hierarchical model that gathers the information from both studies.     

A Network Meta-Analysis of Efficacy and Evaluation of Safety of Subcutaneous Pegylated Interferon Beta-1a versus Other Injectable Therapies for the Treatment of Relapsing-Remitting Multiple Sclerosis

Subcutaneous pegylated interferon beta-1a (peginterferon beta-1a [PEG-IFN]) 125 μg every two or four weeks has been studied in relapsing-remitting multiple sclerosis (RRMS) patients in the pivotal Phase 3 ADVANCE trial. In the absence of direct comparative evidence, a network meta-analysis (NMA) was conducted to provide an indirect assessment of the relative efficacy, safety, and tolerability of PEG-IFN versus other injectable RRMS therapies. Systematic searches were conducted in MEDLINE, Embase, and the Cochrane Library, and conference proceedings from relevant annual symposia were hand-searched. Included studies were randomized controlled trials evaluating ≥1 first-line treatments including interferon beta-1a 30, 44, and 22 μg, interferon beta-1b, and glatiramer acetate in patients with RRMS. Studies were included based on a pre-specified protocol and extracted by a team of independent reviewers and information scientists, utilizing criteria from NICE and IQWiG. In line with ADVANCE findings, NMA results support that PEG-IFN every 2 weeks significantly reduced annualized relapse rate, and 3- and 6-month confirmed disability progression (CDP) versus placebo. There was numerical trend favoring PEG-IFN every 2 weeks versus other IFNs assessed for annualized relapse rate, and versus all other injectables for 3- and 6-month CDP (6-month CDP was significantly reduced versus IFN beta-1a 30 μg). The safety and tolerability profile of PEG-IFN beta-1a 125 μg every 2 weeks was consistent with that of other evaluated treatments. Study limitations for the NMA include variant definitions of relapse and other systematic differences across trials, assumptions that populations were sufficiently similar, and inability to perform NMA of adverse events. With similar efficacy compared to other RRMS treatments in terms of annualized relapse rate and 3- and 6-month CDP, a promising safety profile, and up to 93% reduction in number of injections (which may improve adherence), PEG-IFN every 2 weeks offers a valuable alternative treatment option for patients with RRMS.     

Flexible protein docking refinement using pose-dependent normal mode analysis

Modeling conformational changes in protein docking calculations is challenging. To make the calculations tractable, most current docking algorithms typically treat proteins as rigid bodies and use soft scoring functions that implicitly accommodate some degree of flexibility. Alternatively, ensembles of structures generated from molecular dynamics (MD) may be cross-docked. However, such combinatorial approaches can produce many thousands or even millions of docking poses, and require fast and sensitive scoring functions to distinguish them. Here, we present a novel approach called "EigenHex," which is based on normal mode analyses (NMAs) of a simple elastic network model of protein flexibility. We initially assume that the proteins to be docked are rigid, and we begin by performing conventional soft docking using the Hex polar Fourier correlation algorithm. We then apply a pose-dependent NMA to each of the top 1000 rigid body docking solutions, and we sample and re-score multiple perturbed docking conformations generated from linear combinations of up to 20 eigenvectors using a multi-threaded particle swarm optimization algorithm. When applied to the 63 "rigid body" targets of the Protein Docking Benchmark version 2.0, our results show that sampling and re-scoring from just one to three eigenvectors gives a modest but consistent improvement for these targets. Thus, pose-dependent NMA avoids the need to sample multiple eigenvectors and it offers a promising alternative to combinatorial cross-docking.     

Evaluating the Quality of Evidence from a Network Meta-Analysis

Systematic reviews that collate data about the relative effects of multiple interventions via network meta-analysis are highly informative for decision-making purposes. A network meta-analysis provides two types of findings for a specific outcome: the relative treatment effect for all pairwise comparisons, and a ranking of the treatments. It is important to consider the confidence with which these two types of results can enable clinicians, policy makers and patients to make informed decisions. We propose an approach to determining confidence in the output of a network meta-analysis. Our proposed approach is based on methodology developed by the Grading of Recommendations Assessment, Development and Evaluation (GRADE) Working Group for pairwise meta-analyses. The suggested framework for evaluating a network meta-analysis acknowledges (i) the key role of indirect comparisons (ii) the contributions of each piece of direct evidence to the network meta-analysis estimates of effect size; (iii) the importance of the transitivity assumption to the validity of network meta-analysis; and (iv) the possibility of disagreement between direct evidence and indirect evidence. We apply our proposed strategy to a systematic review comparing topical antibiotics without steroids for chronically discharging ears with underlying eardrum perforations. The proposed framework can be used to determine confidence in the results from a network meta-analysis. Judgements about evidence from a network meta-analysis can be different from those made about evidence from pairwise meta-analyses.     

Activation of the Ghrelin Receptor is Described by a Privileged Collective Motion: A Model for Constitutive and Agonist-induced Activation of a Sub-class A G-Protein Coupled Receptor (GPCR)

Three homology models of the human ghrelin receptor (GHS-R1a) have been generated from the available X-ray structures of rhodopsin (RHO model), opsin (OPS model) and beta-2 adrenergic receptor (B2 model). The latter was used as a starting point for combined molecular dynamics simulation (MDS) and full atom normal modes analysis (NMA). A low-frequency normal mode (mode 16) perfectly reproduced the intracellular motions observed between B2 and RHO models; in the opposite direction along the same mode, the generated structures are closer to the OPS model, suggesting a direct link with GHS-R1a activation. This was in agreement with motions of the seven transmembranous segments, increase of the solvent accessibility of the 140-ERY-142 sequence, and flip of the Trp276 (C WLP) residue, some features related to GPCRs activation. According to our model, His280 was proposed to stabilize Trp276 in the active state; this was verified by site-directed mutagenesis and biochemical characterization of the resulting H280A and H280S mutants, which were fully functional but sharing an important decrease of their basal activities. Docking performed with short ghrelin derivatives Gly-Ser-Ser _[octa]-Phe-NH ₂ and Gly-Ser-Ser _[octa]-Phe-Leu-NH ₂ allowed the identification of a robust position of these peptides in the active site of the receptor. This model was refined by MDS and validated by docking experiments performed on a set of 55 ghrelin receptor ligands based on the 1,2,4- triazole scaffold. Finally, NMA performed on the obtained peptide-receptor complex suggested stabilization of the Trp276 residue and of the whole receptor in the active state, preventing the motion observed along mode 16 computed for the unbound receptor. Our results show that NMA offers a powerful approach to study the conformational diversity and the activation mechanism of GPCRs.     

A Coarse-Grained Elastic Network Atom Contact Model and Its Use in the Simulation of Protein Dynamics and the Prediction of the Effect of Mutations

Normal mode analysis (NMA) methods are widely used to study dynamic aspects of protein structures. Two critical components of NMA methods are coarse-graining in the level of simplification used to represent protein structures and the choice of potential energy functional form. There is a trade-off between speed and accuracy in different choices. In one extreme one finds accurate but slow molecular-dynamics based methods with all-atom representations and detailed atom potentials. On the other extreme, fast elastic network model (ENM) methods with C_α−only representations and simplified potentials that based on geometry alone, thus oblivious to protein sequence. Here we present ENCoM, an Elastic Network Contact Model that employs a potential energy function that includes a pairwise atom-type non-bonded interaction term and thus makes it possible to consider the effect of the specific nature of amino-acids on dynamics within the context of NMA. ENCoM is as fast as existing ENM methods and outperforms such methods in the generation of conformational ensembles. Here we introduce a new application for NMA methods with the use of ENCoM in the prediction of the effect of mutations on protein stability. While existing methods are based on machine learning or enthalpic considerations, the use of ENCoM, based on vibrational normal modes, is based on entropic considerations. This represents a novel area of application for NMA methods and a novel approach for the prediction of the effect of mutations. We compare ENCoM to a large number of methods in terms of accuracy and self-consistency. We show that the accuracy of ENCoM is comparable to that of the best existing methods. We show that existing methods are biased towards the prediction of destabilizing mutations and that ENCoM is less biased at predicting stabilizing mutations.     

How regulatory CD25+CD4+ T cells impinge on tumor immunobiology: the differential response of tumors to therapies

Aiming to get a better insight on the impact of regulatory CD25(+)CD4(+) T cells in tumor-immunobiology, a simple mathematical model was previously formulated and studied. This model predicts the existence of two alternative modes of uncontrolled tumor growth, which differ on their coupling with the immune system, providing a plausible explanation to the observation that the development of some tumors expand regulatory T cells whereas others do not. We report now the study of how these two tumor classes respond to different therapies, namely vaccination, immune suppression, surgery, and their different combinations. We show 1) how the timing and the dose applied in each particular treatment determine whether the tumor will be rejected, with or without concomitant autoimmunity, or whether it will continue progressing with slower or faster pace; 2) that both regulatory T cell-dependent and independent tumors are equally sensitive to vaccination, although the former are more sensitive to T cell depletion treatments and are unresponsive to partial surgery alone; 3) that surgery, suppression, and vaccination treatments, can synergistically improve their individual effects, when properly combined. Particularly, we predict rational combinations helping to overcome the limitation of these individual treatments on the late stage of tumor development.     

Polydesigns and causal inference

In an increasingly common class of studies, the goal is to evaluate causal effects of treatments that are only partially controlled by the investigator. In such studies there are two conflicting features: (1) a model on the full cohort design and data can identify the causal effects of interest, but can be sensitive to extreme regions of that design's data, where model specification can have more impact; and (2) models on a reduced design (i.e., a subset of the full data), for example, conditional likelihood on matched subsets of data, can avoid such sensitivity, but do not generally identify the causal effects. We propose a framework to assess how inference is sensitive to designs by exploring combinations of both the full and reduced designs. We show that using such a "polydesign" framework generates a rich class of methods that can identify causal effects and that can also be more robust to model specification than methods using only the full design. We discuss implementation of polydesign methods, and provide an illustration in the evaluation of a needle exchange program.     

Comparison of the Non-VKA Oral Anticoagulants Apixaban,Dabigatran, and Rivaroxaban in the Extended Treatment and Prevention of Venous Thromboembolism: Systematic Review and Network Meta-Analysis

BackgroundHistorically, warfarin or aspirin have been the recommended therapeutic options for the extended treatment (>3 months) of VTE. Data from Phase III randomised controlled trials (RCTs) are now available for non-VKA oral anticoagulants (NOACs) in this indication. The current systematic review and network meta-analysis (NMA) were conducted to compare the efficacy and safety of anticoagulants for the extended treatment of VTE.MethodsElectronic databases (accessed July 2014 and updated April 2016) were systematically searched to identify RCTs evaluating apixaban, aspirin, dabigatran, edoxaban, rivaroxaban, and warfarin for the extended treatment of VTE. Eligible studies included adults with an objectively confirmed deep vein thrombosis, pulmonary embolism or both. A fixed-effect Bayesian NMA was conducted, and results were presented as relative risks (RRs). Sensitivity analyses examining (i) the dataset employed according to the time frame for outcome assessment (ii) the model used for the NMA were conducted.ResultsEleven Phase III RCTs (examining apixaban, aspirin, dabigatran, rivaroxaban, warfarin and placebo) were included. The risk of the composite efficacy outcome (VTE and VTE-related death) was statistically significantly lower with the NOACs and warfarin INR 2.0–3.0 compared with aspirin, with no significant differences between the NOACs. Treatment with apixaban (RR 0.23, 95% CrI 0.10, 0.55) or dabigatran (RR 0.55, 95% Crl 0.43, 0.71) was associated with a statistically significantly reduced risk of ‘major or clinically relevant non-major bleed’ compared with warfarin INR 2.0–3.0. Apixaban also showed a significantly reduced risk compared with dabigatran (RR 0.42, 95% Crl 0.18, 0.97) and rivaroxaban (RR 0.23, 95% Crl 0.09, 0.59). Sensitivity analyses indicate that results were dependent on the dataset, but not on the type of NMA model employed.ConclusionsResults from the NMA indicate that NOACs are an effective treatment for prevention of VTE or VTE-related death) in the extended treatment setting. However, bleeding risk differs between potential treatments, with apixaban reporting the most favourable profile compared with other NOACs, warfarin INR 2.0–3.0, and aspirin.