DR CHATREE CHAI-ADISAKSOPHA (Orcid ID : 0000-0001-7209-9722)

PROF. ALFONSO IORIO (Orcid ID : 0000-0002-3331-8766)

Accepted Article
Article type : Original Article - Clinical Haemostasis and Thrombosis

Association of body weight with efficacy and safety outcomes in phase III

randomized controlled trials of direct oral anticoagulants: a systematic review

and meta-analysis

Body weight and outcomes in DOACs trials

Kochawan Boonyawat, MD*,**, Francois Caron, MD , Ang Li, MD , Chatree Chai-

Adisaksopha, MD*, Wendy Lim, MD*, Alfonso Iorio, MD, PhD*, Renato D. Lopes, MD, PhD,

MHS , David Garcia, MD and Mark A. Crowther, MD*

*
Department of Medicine, McMaster University, Hamilton, Ontario, Canada, Population

Health Research Institute, Hamilton, Ontario, Canada, University of Washington, Fred

Hutchinson Cancer Research Center, Duke Clinical Research Institute, Duke Medicine,

Durham, North Carolina, Department of Medicine, University of Washington, **Department of

Medicine, Ramathibodi Hospital, Mahidol University, Thailand

Correspondence: Kochawan Boonyawat, MD, Department of Medicine, Ramathibodi

Hospital, Mahidol University, Thailand, 270 Rama 6th Road, Rachathewi, Bangkok, Thailand

10400; Phone +662-201-1392email: kochawan.boonyawat@medportal.ca

This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process, which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1111/jth.13701
This article is protected by copyright. All rights reserved.
 Key words: Meta-analysis, Anticoagulants, Body weight, Thromboembolism, Hemorrhage
Accepted Article
Essentials

- The association of body weight and patient-important outcomes remains unknown.

- Phase III randomized controlled trials of direct oral anticoagulants (DOACs) were

searched.

- Risk of outcomes varying among body weight subgroups is not attributable to

anticoagulant type.

- Dose adjustment of DOACs, outside that recommended, is unlikely to improve the

outcomes.

Summary

Background: Concerns have arisen in direct oral anticoagulant (DOAC)-treated patients about

safety and efficacy in extremes of body weight. The aims of this systematic review were to

investigate the association of body weight and patient-important outcomes in patients treated

with DOACs or warfarin, and to demonstrate the fixed-dose effect of DOACs.

Methods: MEDLINE, EMBASE were searched until November 2016. Phase III randomized

controlled trials (RCTs) using DOACs in atrial fibrillation (AF) and acute venous

thromboembolism (VTE) were included. Relative risk and 95% confidence interval were

calculated. The pooled estimates were performed using Mantel-Haenszel random effects

model.

Results: A total of 11 phase III RCTs were included. Low body weight was associated with

increased risk of thromboembolism compared with non-low body weight (RR, 1.57; 95% CI,

This article is protected by copyright. All rights reserved.

 1.34-1.85). High body weight was not associated with risk of thromboembolism compared with

non-high body weight (RR, 0.88; 95%CI, 0.63-1.23). The subgroup of AF patients with high
Accepted Article
body weight had a lower risk of thromboembolism compared with non-high body weight (RR,

0.43; 95%CI, 0.28-0.67). Bleeding outcomes were comparable for all body weight

comparisons. There were no clear interactions between types of anticoagulant in all outcomes.

Conclusion: The pooled effect of both the DOAC and comparison arms were likely

attributable to differences in baseline thrombotic risk in each body weight category, rather than

an effect of the types or dose of DOAC used in each indication. Dose adjustment of DOACs,

outside that recommended in the package insert, is unlikely to improve safety or efficacy.

Background

Direct oral anticoagulants (DOACs) and warfarin are widely used for the prevention of

systemic embolism in AF and in the treatment of venous thromboembolism (VTE). As DOACs

are given in fixed doses without regard for body weight, concerns have arisen about the safety

and efficacy of DOACs in patients at the extremes of weight-specifically those with extremely

high or low body weight. Clinicians may hesitate to initiate DOACs in these patients due to a

perception that the fixed dose might not be sufficiently effective in the high body weight

individuals, or might increase the bleeding risk in low body weight individuals. Although there

are pharmacokinetic studies [1-4] that have documented drug exposure when DOACs are given

to extremely high or low body weight patients, a study that documents the rate of patient-

important outcomes in this population is lacking. A previous meta-analysis of this subject

performed such analyses separately for each body weight category and demonstrated

comparable efficacy and safety of DOACs for acute VTE compared with warfarin.[5] In our

study, we used a different approach to explore the association of extreme body weight relative

This article is protected by copyright. All rights reserved.

 to normal body weight on thromboembolic and bleeding outcomes in patients treated with

DOACs and warfarin for AF and acute VTE, and to demonstrate the fixed dose effect of
Accepted Article
DOACs. We hypothesized that if the fixed dose effect does not hold true, patients with higher

body weight would have a higher rate of thrombosis compared with non-high body weight

while taking DOACs, but not while taking warfarin (due to INR driven dosing adjustment).

Similarly, low body weight patients would have a higher risk of bleeding compared with non-

low body weight patients, while taking DOACs, but not in warfarin treated patients.

Methods

Data source and search strategy

This systematic review and meta-analysis was conducted and reported according to the

PRISMA statement.[6] A literature search was performed via the Ovid platform through

PubMed, MEDLINE (1946 to present) and EMBASE (1996 to present). The Cochrane Central

Register of Controlled trials (CENTRAL) and www.ClinicalTrial.gov were searched for

unpublished or ongoing registered trials. The U.S. Food and Drug Administration (FDA)

reviews of new drug applications and a pharmaceutical industry trials registry (Boehringer

Ingelheim, Bristol-Myers Squibb, Bayer and Daiichi Sankyo) were searched for grey literature

and unpublished studies. Conference proceedings or abstracts that had been identified from the

databases were assessed for eligibility. In addition, we searched online abstracts and

conference proceedings (from January 2008 to January 2016) from the congresses of the

American Society of Hematology, International Society of Thrombosis and Hemostasis,

American College of Cardiology and American Heart Association. Lists of references of

potentially relevant articles were manually reviewed and screened for eligibility.

For MEDLINE and EMBASE, search term included: dabigatran, rivaroxaban, apixaban,

edoxaban, direct oral anticoagulant*, target-specific oral anticoagulant*, novel oral

This article is protected by copyright. All rights reserved.

 anticoagulant*, non-vitamin K oral anticoagulant*, DOAC*, NOAC*, TSOAC*, AF, atrial

flutter, venous thrombosis, venous thromboembolism, deep vein thrombosis, pulmonary

Accepted Article
embolism, and clinical trials. All terms were searched in indexed (subject heading and

explosion) and keyword using the OVID platform (Figure S1). We limited the search to RCTs

comparing the effect of therapy using filter terms recommended by the Health Information

Research Unit, McMaster University which have a sensitivity of 99% [7, 8]. The search was

limited to humans. There were no language restrictions. The search was performed on February

21th, 2016. Additional searches on MEDLINE and EMBASE was performed on November

20th, 2016 to update the results of the analysis.

Study selection

The studies were eligible if they were subgroups of phase III RCTs investigating DOACs

including dabigatran, rivaroxaban, apixaban and edoxaban for the prevention of stroke and

systemic embolism in AF and in acute VTE treatment, or sub-studies or subgroup analysis of

the phase III RCTs. All included studies reported thromboembolic or bleeding outcomes by

body weight or body mass index (BMI). Low body weight was defined by the investigators in

each study but had to be less than or equal to 60 kg. High body weight was similarly defined

for each study but was a minimum of 100 kg. Normal body weight was defined as being

between the cut-offs of high and low body weight in each study. Obesity was defined as a BMI

of more than 30 kg/m2. Studies examining DOACs for primary prevention of VTE in

orthopedic surgery, medically ill patients or extended treatment of VTE or other indications

(acute coronary syndrome, atrial thrombus, peri-operative management, anti-phospholipid

syndrome) were excluded.

Two reviewers (K.B and F.C) independently screened titles and abstracts of the retrieved

studies using predefined inclusion and exclusion criteria. After a calibration exercise, we

developed additional inclusion criteria to improve agreement.

This article is protected by copyright. All rights reserved.

 Full text eligibility including the supplementary indexes of the studies was assessed

independently by two reviewers (K.B and F.C.). For highly relevant conference proceedings or
Accepted Article
abstracts, we contacted the authors for additional information. Disagreement was resolved by

discussion. If this disagreement could not be resolved, a third adjudicator (C.C.-A) was

contacted to make a final decision. The Kappa statistic was calculated to determine the inter-

rater agreement between two reviewers.

Data abstraction and quality assessment

Two reviewers (KB, FC) independently abstracted data using a study-specific data

extraction sheet. Thromboembolic outcomes including stroke and/or systemic embolism in AF

studies and symptomatic recurrent VTE or VTErelated death in VTE studies were recorded

by subgroups of body weight or BMI as provided by the studies. Bleeding outcomes including

major bleeding as defined by the International Society on Thrombosis and Hemostasis

(ISTH)[9] and/or clinically relevant non major bleeding (CRNMB) were recorded for each

subgroup of body weight or BMI as provided by the studies. For studies which reported the

outcome only as a rate (%/year), the rate was converted to the number of events assuming a

constant risk during the follow-up period.

The risk of bias was assessed using the modified New-Castle-Ottawa scale. This was

appropriate because we treated a single arm derived from RCTs as a prospective observational

study.[10] This scale evaluates the risk of bias in the selection of the study population,

confidence in exposures, comparability of the cohort, the adequacy of follow-up and outcome

assessment. It has an additional assessment of prognostic factors and impact of co-intervention

on the study conclusions. The overall risk of bias for each outcome was assessed by two

independent reviewers (K.B. and F.C.). Publication bias was not formally assessed by a funnel

plot due to the small number of included studies. The quality of evidence of each outcome was

This article is protected by copyright. All rights reserved.

 assessed using the Grading of Recommendation Assessment Development, and Evaluation

score [11] by two independent reviewers (KB and FC).

Accepted Article
Statistical analysis

Study characteristics were summarized. The effect measures for each study were calculated

and reported as a relative risk (RR) with 95% confidence intervals (CI).

Meta-analysis was performed using a random effects model and the effect measures of the

included studies were pooled using a Mantel-Haenzel method. A random effects model was

used based on the assumption that there was heterogeneity in the individual studies because of

variation in types of anticoagulants, indication for treatment, DOAC type, and duration of

treatment. We performed meta-analyses separately for the DOAC and warfarin arms. When no

significant difference in the effect estimate between groups was observed, we pooled the data

from the DOACs and warfarin arms together to increase the power of the analysis. The results

of the separate analyses are presented in the supplementary material (Figure S7). The meta-

analyses were performed using Review Manager 5.3 software. Heterogeneity was explored

based on a priori hypotheses using visual inspection and the Cochrane Q test. A p-value < 0.05

was considered statistically significant. The I2 statistic was calculated to provide a quantitative

estimate of heterogeneity.

Subgroup analysis

We specified four a priori hypotheses including types of anticoagulant (DOACs vs.

warfarin), indication (AF vs.VTE), type of DOAC (direct thrombin inhibitor vs. factor Xa

inhibitor) and studies with low or standard doses of DOACs. During data abstraction, we found

that 4 studies had adjusted DOAC dose in patients with low body weight or those with

impaired renal function (creatinine clearance 30-50 ml/min), age >80 years or concomitant use

This article is protected by copyright. All rights reserved.

 of potent P-glycoprotein inhibitors. Thus, an additional post-hoc subgroup analysis in the

DOAC studies with dose adjustment versus studies without dose adjustment was performed. If
Accepted Article
there was only one study in one of the subgroups, the analysis was not performed.

Sensitivity analysis

Five sensitivity analyses were performed: 1) exclusion of studies that adjusted DOACs

dose according to body weight, 2) comparison of high vs. normal body weight instead of high

vs. non-high body weight 3) comparison of low vs. normal body weight instead of low vs. non-

low body weight (excluded data from ARISTOTLE and ENGAGE TIMI-48 because we were

unable to define a normal body weight group), 4) comparison of BMI>35 kg/m2 and < 30

kg/m2 , and 5) performing the analyses on the thromboembolic and bleeding outcomes for the

studies with a similar body weight cutoff.

Results

We identified 6,587 articles through title and abstract screening: 2,614 articles from

MEDLINE, 3,511 articles form EMBASE, 212 from PubMed and 250 articles from other

sources. 1,058 duplicate articles were excluded. A total of 5,169 articles were excluded after

title and abstract screening leaving 360 articles for full text review. A further 336 studies were

excluded: 204 articles were abstracts, RCTs or sub-studies of RCTs that did not report

outcomes by subgroups of body weight or BMI, 110 articles were not RCTs, 10 articles

examined an ineligible population and 12 articles were duplicates. Twenty-four articles

remained for data abstraction. These included 12 RCTs, 3 FDA briefing documents, 3 highly

relevant abstracts and 6 sub-studies of RCTs. We received more information from the authors

of 2 abstracts. During data abstraction, 8 articles were excluded: 5 sub-studies of RCTs and 1

abstract with less informative data, 1 RCT and 1 sub-study of an RCT with insufficient

This article is protected by copyright. All rights reserved.

 information (Figure S2). The unweighted Kappa statistics for full text eligibility between the

two reviewers was 0.89 (95% CI, 0.88 to 0.99).

Accepted Article
Study characteristics

A total of 11 of RCTs were included. Characteristics of the included studies are shown in

Table 1. Two pairs of RCTs [12-15] were reported as pooled studies, thus they were treated as

2 unique studies (RECOVERI, II and EINSTEIN DVT, PE).[12, 16] Four studies adjusted the

dose of DOACs in patients with low body weight under selected circumstances (ARISTOTLE,

ENGAGE-TIMI-48, Hokusai study, AVERROES). One of the 9 studies did not report

thromboembolic outcomes by subgroups of body weight.[17] Two of the 9 studies did not

report bleeding outcome by subgroups of body weight.[12, 17] Three studies did not report

thromboembolic outcome by subgroups of BMI [17-19] and 3 studies did not report bleeding

outcome by subgroups of BMI.[12, 18, 19] The low body weight cut-off level was 60 kg in 4

studies and 50 kg in 4 studies. The high body weight cut-off was 100 kg in 5 studies and 110

kg in 1 study. Two studies did not have a high body weight cut-off and thus were not included

in the high body weight comparison (ARISTOTLE, ENGAGE-TIMI-48).[17, 19].

Risk of bias assessment

The assessment for risk of bias using the modified New-Castle Ottawa scale is summarized

in the supplementary material (Table S1). The quality of evidence assessed by GRADE was

very low for all outcomes. Since our analysis used single arms derived from RCTs as

observational studies, the quality of evidence began at low quality, and the quality was

downgraded due to lack of baseline adjustment and indirectness.

This article is protected by copyright. All rights reserved.

 Outcomes
Accepted Article
Low body weight vs. Non-low body weight patients

Thromboembolic outcomes

Eight studies with a total of 98,244 patients were included in the meta-analysis.[12-16, 19-

24] Thromboembolic events occurred in 278 of 6,497 patients (4.28%) in the low body weight

and in 2,513 of 91,747 patients (2.74%) in the non-low body weight groups (RR, 1.57; 95% CI,

1.34 to 1.85, P <0.0001, I2=29%) (Figure 1); absolute difference in risk [low body weight

minus non-low body weight], 0.01; 95% CI, 0.0057 to 0.0187, p=0.0002). Subgroup analyses

based on type of anticoagulant, indication, type and dose of DOACs, and studies with dose

adjustment were performed and provided similar results to the primary analysis (Figure S3).

Bleeding outcomes

Seven studies with a total of 95,223 patients were included in the meta-analysis.[13-16,

19-24] Bleeding outcomes occurred in 502 of 8,420 patients (5.96%) in the low body weight

and in 5,280 of 86,803 patients (6.08%) in the non-low body weight groups (RR, 0.97; 95% CI,

0.86 to 1.10, p=0.67, I2 =41%) (Figure 2). Subgroup analyses based on type of anticoagulant

(Figure 3), indication, dose of DOAC and studies with dose adjustment provided similar results

to the primary analysis (Figure S3).

This article is protected by copyright. All rights reserved.

 High body weight and non-high body weight patients

Thromboembolic outcomes
Accepted Article
Six studies with a total of 59,079 patients were included in the meta-analysis.[12-16,

20-23] Thromboembolic outcomes occurred in 196 of 8,623 patients (2.27%) in the high body

weight and in 1,453 of 50,456 patients (2.88%) in the non-high body weight groups (RR, 0.88;

95% CI, 0.63 to 1.23, p =0.46, I2 =77%) (Figure 4).

Subgroup analysis based on type of anticoagulant (Figure 5), type of DOAC provided

similar results to the primary analysis (Figure S4). Subgroup analysis based on indication

demonstrated a significant interaction (p for interaction <0.0001). The risk of

thromboembolism was lower in high body weight compared with non-high body weight

patients in AF patients (RR, 0.43; 95%CI, 0.28 to 0.67, p =0.0002), but not in VTE patients

(RR, 1.23; 95% CI, 1.00 to 1.53, p =0.05) (Figure 6). Subgroup credibility was assessed using

criteria proposed by Sun et al.[25] (Table S2). The credibility of subgroup effect was likely

plausible.

Bleeding outcomes

Five studies with a total of 54,165 patients were included in the meta-analysis.[14-16,

20-23] Bleeding outcomes occurred in 511 of 7,810 patients (6.54%) in the high body weight

and in 3,179 of 46,355 patients (6.86%) in the non-high body weight groups (RR, 0.93; 95%CI,

0.82 to 1.06, p=0.31, I2=43) (Figure 7). Subgroup analyses based on type of anticoagulant, and

indication provided similar results to the primary analysis (Figure S4).

This article is protected by copyright. All rights reserved.

 Obesity and non-obesity

Thromboembolic outcomes
Accepted Article
Six studies with total of 68,870 patients were included in the meta-analysis.[12, 14-16,

20, 21, 24] Thromboembolic outcomes occurred in 564 of 24,594 obese patients (2.29%) and

in 1,286 of 44,276 non-obese patients (2.90%) (RR, 0.81; 95% CI, 0.71-0.93, p =0.002, I2

=39%; absolute difference in risk [obesity minus non-obesity], -0.0055, 95% CI -0.0088 to -

0.0022, p=0.001).

Subgroup analyses based on types of anticoagulant and types of DOACs provided

similar results to the primary analysis. Subgroup analyses based on indication demonstrated a

significant interaction (P =0.03). The risk of thromboembolism was lower in obese compared

with non-obese patients with AF (RR, 0.73; 95%CI, 0.65 to 0.82, p <0.001, but not in VTE

patients (RR, 1.02; 95%CI 0.78 to 1.33,p =0.13) (Figure S5).

Bleeding outcomes

Six studies with a total of 66,697 patients were included in the meta-analysis.[14-17,

20, 21, 23, 24] The aspirin arm of the AVERROES study was not included in the analysis.

Bleeding outcomes occurred in 2,487 of 41,672 obese patients (5.97 %) and in 1,344 of 25,025

non-obese patients (5.37%) (RR, 1.06; 95% CI, 0.99 to 1.14, p=0.12, I2=17%). Subgroup

analyses based on type of anticoagulant and indication provided similar results to the primary

analysis.

Sensitivity analysis

Five sensitivity analyses were performed as defined previously. In all cases the analysis

supported the results of the primary analyses (Figure S6).

This article is protected by copyright. All rights reserved.

 Discussion

Our systemic review explores the association between body weight and
Accepted Article
thromboembolic or bleeding outcomes in DOAC and warfarin treated patients across a range of

body weights and indications. The results of our systematic review do not support our

hypothesis that patients with high body weight would have a higher rate of thrombosis or low

body weight patients would have a higher risk of bleeding while taking DOACs, but not while

taking warfarin. We found that low body weight patients treated with DOACs and with

warfarin have an increased risk of thromboembolism but no increase in bleeding risk compared

to that in non-low body weight patients, irrespective of indication. On the other hand, high

body weight AF patients treated with DOACs and warfarin have a favourable thromboembolic

outcome compared to that in non-high body weight. Our counterintuitive findings were likely

influenced by the variation in baseline characteristics affecting thromboembolic risk in low or

high body weight patients; plasma levels of DOACs may be less important than we

hypothesized.

The observation that lower body weight is associated with a paradoxical increase in

morbid outcomes has been reported in both VTE and AF patients [26-28]. Several factors

could explain this observation. Low body weight could be a consequence of chronic illness

such as malignancy which is also associated with an increased risk of venous thrombosis. We

also observed that heavier patients have a favourable thromboembolic outcome. This

observation, the so called obesity paradox, has been reported in patients with AF, congestive

heart failure, coronary heart disease and end-stage kidney disease. [29-32]. A post-hoc analysis

of the ARISTOTLE trial demonstrated a significant risk reduction in all-cause mortality and

the composite endpoint of stroke and systemic embolism, myocardial infarction and death in

obese compared with non-obese patients[33]. There is no clear explanation for this

observation; perhaps higher body weight might receive optimal cardiovascular medications

This article is protected by copyright. All rights reserved.

 more often than in non-high body weight patients resulting in a lower risk of thromboembolic

and cardiovascular outcomes [33, 34].

Accepted Article
In our study, low body weight patients treated with DOACs had no difference in

bleeding compared with non-low body weight patients. Likewise, high body weight and obese

patients had no difference in bleeding risk compared with non-high, or non-obese patients.

Subgroup analysis of bleeding outcomes did not demonstrate a significant interaction between

studies with dose adjustment and those with no dose adjustment. The results were confirmed in

a sensitivity analysis where we excluded patients who had their DOAC dose adjusted based on

body weight or other characteristics. If one follows approved prescribing information, these

findings support that decreasing doses in individuals with lower weight may not be required.

The modest changes in dabigatran, rivaroxaban and apixaban exposure seen in early

pharmacokinetic studies suggests that it is unlikely to require dose adjustment according to

body weight.[1-4] In a phase II RCT of edoxaban in AF patients, body weight less than 60 kg

was an independent risk factor for major bleeding in the edoxaban 60 mg group compared to

the warfarin group. [35] Thus, in subsequent RCTs, the doses of edoxaban were reduced by

50% if patients had any of the three criteria: body weight less than 60 kg, impaired renal

function, or concomitant use of potent P-glycoprotein inhibitors. Our study supports the

previous PK studies of DOACs [1-3, 35] showing body weight has only a modest effect on the

pharmacokinetic profile of DOACs as we did not observe a higher rate of thromboembolism in

high body weight patients or higher rate of bleeding in low body weight patients.

As our study included different types of DOACs, the results require cautious

interpretation. In the dabigatran VTE studies [12, 13] heavier patients treated with either

dabigatran or warfarin had a significant increase in their risk of thrombosis; this contrasts with

a non-significant difference in the risk with rivaroxaban, apixaban and edoxaban. The

This article is protected by copyright. All rights reserved.

 observation that both dabigatran and warfarin had an unfavourable outcome suggests the

patient populations enrolled in the dabigatran studies differed systematically from those in the
Accepted Article
other studies. Four studies, including ARISTOTLE, AVERROES, ENGAGE-TIMI-48, and the

Hokusai study, employed empiric dose reductions. The ARISTOTLE and the AVERROES

studies reduced doses if patients had any two of the three criteria: a body weight of less than 60

kg, impaired renal function, or advancing age. In the ENGAGE-TIMI-48 and the Hokusai

study, dose reductions were applied to patients with any of the three criteria: a body weight of

less than 60 kg, impaired renal function, or concomitant use of potent P-glycoprotein

inhibitors.

Although none of the included RCTs had an upper weight limit in their inclusion

criteria, 3 studies have reported the proportion of extremely high body weight patients (body

weight >120 kg or BMI >40 kg/m2) participating in the study. In a post-hoc analysis of the

pooled EINSTEIN studies, 222 of 8,271 patients (3.67%) had body weight between 120-140

kg and 81 of 8,271 had body weight greater than 140 kg (1%).[16] In a post-hoc analysis from

the ARISTOTLE study, 1006 of 18,107 patients (5.55%) had BMI > 40 kg/m2.[36] As only a

few studies reported outcomes in extremely high body weight patients, we could not perform a

separate analysis in this population.

To our knowledge, there are no published systematic reviews and/or meta-analyses

specifically investigating the association of outcomes across a broad range of weights in

DOAC and warfarin treated patients. Strengths of this systematic review include the large

number of patients including more than 90,000 patients across the two largest indications for

DOAC use. Second, all included studies were large well-conducted RCTs. Last, the body

weight threshold used in each study was representative of the low body weight and high body

weight population.

This article is protected by copyright. All rights reserved.

 Our study has limitations. First, owing to the study-level meta-analysis of subgroups,

we were unable to adjust for baseline characteristics between body weight subgroups. This
Accepted Article
could be further addressed by meta-regression or an individual patient level meta- analysis.

Second, the associations we observed do not demonstrate causation and should not be

interpreted as such. Third, although low body weight patients were well represented, very few

extremely high body weight patients were included; thus our findings do not apply to

extremely high body weight patients (e.g. those weighing more than 120 kg). Fourth, body

weight was measured at baseline for all studies, thus we did not account for the change in body

weight during the study. Fifth, the results of our study can only be applied with caution to those

DOACs that used dose adjustment per body weight in their constituent trials. Sixth,

inconsistent reporting on body weight or BMI for each outcome between studies resulted in

varying number of studies included in the meta-analyses for each outcome. Finally, we

observed a heterogeneity in thromboembolic outcome in the high body weight comparison.

However, we explored the heterogeneity based on a priori hypotheses and found that the

subgroup effect was significant. In addition, subgroup credibility was assessed and the effect

was likely plausible.

In conclusion, in AF and VTE patients treated with DOACs or with warfarin, we found

that patients with low body weight had a paradoxical increase in the risk of thromboembolism

compared to non-low body weight patients. The subgroup of AF patients with a high body

weight had a favourable thromboembolic outcome compared with AF patients with a non-high

body weight. These differences in thrombosis risk according to body weight were seen in both

the DOAC and the comparison arms were likely to be attributable to differences in other

patient characteristics known to affect the baseline risk of thromboembolism, rather than the

type of anticoagulant. We conclude that dose adjustment of DOACs, outside that

recommended in the package insert, is unlikely to improve safety or efficacy.

This article is protected by copyright. All rights reserved.

 Acknowledgements
Accepted Article
M. A Crowther is a Career Investigator of the Heart and Stroke Foundation and holds the Leo

Pharma Chair in Thromboembolism Research at McMaster University.

Addendum

K. Boonyawat, A. Iorio, and M. A. Crowther designed the methods; K. Boonywat and F. Caron

performed study selection, data extraction, study quality assessment, and analysis; K.

Boonyawat drafted the manuscript; and A. Li, W. Lim, C. Chai-Adisaksopha, D. Garcia, R. D.

Lopes, and M. A. Crowther critically revised the manuscript.

Disclosure of Conflict of Interests

M. A. Crowther discloses having sat on advisory boards for Octapharma, Bayer, Boehringer

Ingelheim, Janssen, Pfizer, Leo Pharma, and Portola; discloses having participated in study

Steering committees, or other research-related activities for projects involving AKP America,

Daichii, Bayer, and Leo Pharma. M. A. Crowther discloses his institutions (McMaster

University and/or St Josephs Healthcare) have received funding for research projects from the

Heart and Stroke Foundation of Canada, Leo Pharma, and Bayer for work in which M. A.

Crowther is involved. M. A. Crowther discloses having received funding for preparation of

educational materials and/or presentations from Alexion, Ortho Clinical Diagnostics, BMS-

Pfizer alliance, Leo Pharma, Abbvie, Bayer, Celgene, Shire, and CSL Behring. A. Iorio has

served as a consultant for Pfizer, NovoNorkdisk, and Bayer via McMaster University. R. D.

Lopes has served as a consultant for Bayer, Boehringer Ingelheim, Bristol-Myers Squibb,

This article is protected by copyright. All rights reserved.

 Glaxo Smith Kline, Merck, Pfizer, and Portola. R. D. Lopes' institution has received funding

for research projects from Bristol-Myers Squibb and Glaxo Smith Kline. D. Garcia has served
Accepted Article
as a consultant for Genzyme, Incyte, Alexion, Janssen, Pfizer, Bristol-Myers Squibb, and

Boehringer Ingelheim. The other authors state that they have no conflict of interest.

Reference

1 Kubitza D, Becka M, Zuehlsdorf M, Mueck W. Body weight has limited influence on

the safety, tolerability, pharmacokinetics, or pharmacodynamics of rivaroxaban (BAY 59-

7939) in healthy subjects. J Clin Pharmacol. 2007; 47: 218-26.

2 Upreti VV, Wang J, Barrett YC, Byon W, Boyd RA, Pursley J, LaCreta FP, Frost CE.

Effect of extremes of body weight on the pharmacokinetics, pharmacodynamics, safety and

tolerability of apixaban in healthy subjects. Br J Clin Pharmacol. 2013; 76: 908-16.

3 Lip GY, Agnelli G. Edoxaban: a focused review of its clinical pharmacology. Eur

Heart J. 2014; 35: 1844-55.

4 Ingelheim B. Dabigatran advisory committee briefing document. Available at FDA

website Accessed. 2010.

5 Di Minno MN, Lupoli R, Di Minno A, Ambrosino P, Scalera A, Dentali F. Effect of

body weight on efficacy and safety of direct oral anticoagulants in the treatment of patients

with acute venous thromboembolism: a meta-analysis of randomized controlled trials. Ann

Med. 2015; 47: 61-8.

6 Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic

reviews and meta-analyses: the PRISMA statement. Bmj. 2009; 339: b2535.

7 Haynes RB, McKibbon KA, Wilczynski NL, Walter SD, Werre SR, Hedges T. Optimal

search strategies for retrieving scientifically strong studies of treatment from Medline:

analytical survey. BMJ. 2005; 330: 1179.

This article is protected by copyright. All rights reserved.

 8 Search Filters for MEDLINE in Ovid Syntax and the PubMed translation.

9 Schulman S, Kearon C, Subcommittee on Control of Anticoagulation of the S,

Accepted Article
Standardization Committee of the International Society on T, Haemostasis. Definition of major

bleeding in clinical investigations of antihemostatic medicinal products in non-surgical

patients. J Thromb Haemost. 2005; 3: 692-4.

10 GH G, JW B. A modified version of the Newcastle-Ottawa Scale: a Tool to Assess Risk

of Bias in Cohort Studies.

11 Guyatt G, Oxman AD, Akl EA, Kunz R, Vist G, Brozek J, Norris S, Falck-Ytter Y,

Glasziou P, DeBeer H, Jaeschke R, Rind D, Meerpohl J, Dahm P, Schunemann HJ. GRADE

guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables. J Clin

Epidemiol. 2011; 64: 383-94.

12 Schulman S, Kakkar AK, Goldhaber SZ, Schellong S, Eriksson H, Mismetti P,

Christiansen AV, Friedman J, Le Maulf F, Peter N, Kearon C, Investigators R-CIT. Treatment

of acute venous thromboembolism with dabigatran or warfarin and pooled analysis.

Circulation. 2014; 129: 764-72.

13 Schulman S, Kearon C, Kakkar AK, Mismetti P, Schellong S, Eriksson H, Baanstra D,

Schnee J, Goldhaber SZ. Dabigatran versus warfarin in the treatment of acute venous

thromboembolism. New England Journal of Medicine. 2009; 361: 2342-52.

14 The EINSTEIN Investigators. Oral rivaroxaban for symptomatic venous

thromboembolism. New England Journal of Medicine. 2010; 363: 2499-510.

15 Buller HR, Prins MH, Lensing AWA, Decousus H, Jacobson BF, Minar E, Chlumsky J,

Verhamme P, Wells P, Agnelli G, Cohen A, Berkowitz SD, Bounameaux H, Davidson BL,

Misselwitz F, Gallus AS, Raskob GE, Schellong S, Segers A. Oral rivaroxaban for the

treatment of symptomatic pulmonary embolism. New England Journal of Medicine. 2012; 366:

1287-97.

This article is protected by copyright. All rights reserved.

 16 Prins MH, Nisio MD, Vedovati MC, Riera-Mestre A, Mueller K, Cohen AT, Wells PS,

Beyer-Westendorf J, Brighton TA, Bounameaux H, Schneider J, Lensing AW. Fixed-dose

Accepted Article
rivaroxaban is not associated with increased recurrent venous thromboembolism or major

bleeding in patients with a high or low body weight. Journal of Thrombosis and Haemostasis.

2015; 13: 35-6.

17 Connolly SJ, Eikelboom J, Joyner C, Diener HC, Hart R, Golitsyn S, Flaker G, Avezum

A, Hohnloser SH, Diaz R, Talajic M, Zhu J, Pais P, Budaj A, Parkhomenko A, Jansky P,

Commerford P, Tan RS, Sim KH, Lewis BS, Van Mieghem W, Lip GY, Kim JH, Lanas-

Zanetti F, Gonzalez-Hermosillo A, Dans AL, Munawar M, O'Donnell M, Lawrence J, Lewis

G, Afzal R, Yusuf S, Committee AS, Investigators. Apixaban in patients with atrial fibrillation.

New England Journal of Medicine. 2011; 364: 806-17.

18 Buller HR, Decousus H, Grosso MA, Mercuri M, Middeldorp S, Prins MH, Raskob

GE, Schellong SM, Schwocho L, Segers A, Shi M, Verhamme P, Wells P. Edoxaban versus

warfarin for the treatment of symptomatic venous thromboembolism. New England Journal of

Medicine. 2013; 369: 1406-15.

19 Giugliano RP, Ruff CT, Braunwald E, Murphy SA, Wiviott SD, Halperin JL, Waldo

AL, Ezekowitz MD, Weitz JI, Spinar J, Ruzyllo W, Ruda M, Koretsune Y, Betcher J, Shi M,

Grip LT, Patel SP, Patel I, Hanyok JJ, Mercuri M, Antman EM. Edoxaban versus warfarin in

patients with atrial fibrillation. New England Journal of Medicine. 2013; 369: 2093-104.

20 Agnelli G, Buller HR, Cohen A, Curto M, Gallus AS, Johnson M, Masiukiewicz U, Pak

R, Thompson J, Raskob GE, Weitz JI. Oral apixaban for the treatment of acute venous

thromboembolism. New England Journal of Medicine. 2013; 369: 799-808.

21 Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, Breithardt G, Halperin

JL, Hankey GJ, Piccini JP, Becker RC, Nessel CC, Paolini JF, Berkowitz SD, Fox KA, Califf

This article is protected by copyright. All rights reserved.

 RM, Investigators RA. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J

Med. 2011; 365: 883-91.

Accepted Article
22 Hokusai VTEI, Buller HR, Decousus H, Grosso MA, Mercuri M, Middeldorp S, Prins

MH, Raskob GE, Schellong SM, Schwocho L, Segers A, Shi M, Verhamme P, Wells P.

Edoxaban versus warfarin for the treatment of symptomatic venous

thromboembolism.[Erratum appears in N Engl J Med. 2014 Jan 23;370(4):390]. New England

Journal of Medicine. 2013; 369: 1406-15.

23 Eikelboom J, Parekh A, Pogue J, Reilly PA, Themeles E, Varrone J, Wang S, Alings

M, Diener HC, Joyner CD, Alings AMW, Amerena JV, Avezum A, Baumgartner I, Brugada J,

Budaj A, Caicedo V, Ceremuzynski L, Chen JH, Commerford PJ, Connolly SJ, Dans AL,

Darius H, Di Pasquale G, Diaz R, Erol C, Ezekowitz MD, Ferreira J, Flaker GC, Flather MD,

Franzosi MG, Gamboa R, Golitsyn SP, Gonzalez Hermosillo JA, Halon D, Heidbuchel H,

Hohnloser SH, Hori M, Huber K, Jansky P, Kamensky G, Keltai M, Kim S, Lau CP, Le

Heuzey JYF, Lewis BS, Liu LS, Nanas J, Oldgren J, Pais PS, Parkhomenko AN, Pedersen KE,

Piegas LS, Raev D, Razali O, Simmers TA, Smith PJ, Talajic M, Tan RS, Tanomsup S,

Toivonen L, Vinereanu D, Wallentin L, Xavier D, Yusuf S, Zhu J. Dabigatran versus warfarin

in patients with atrial fibrillation. New England Journal of Medicine. 2009; 361: 1139-51.

24 Granger CB, Alexander JH, McMurray JJ, Lopes RD, Hylek EM, Hanna M, Al-Khalidi

HR, Ansell J, Atar D, Avezum A, Bahit MC, Diaz R, Easton JD, Ezekowitz JA, Flaker G,

Garcia D, Geraldes M, Gersh BJ, Golitsyn S, Goto S, Hermosillo AG, Hohnloser SH, Horowitz

J, Mohan P, Jansky P, Lewis BS, Lopez-Sendon JL, Pais P, Parkhomenko A, Verheugt FW,

Zhu J, Wallentin L, Committees A, Investigators. Apixaban versus warfarin in patients with

atrial fibrillation. New England Journal of Medicine. 2011; 365: 981-92.

25 Sun X, Briel M, Walter SD, Guyatt GH. Is a subgroup effect believable? Updating

criteria to evaluate the credibility of subgroup analyses. BMJ. 2010; 340: c117.

This article is protected by copyright. All rights reserved.

 26 Barba R, Marco J, Martin-Alvarez H, Rondon P, Fernandez-Capitan C, Garcia-Bragado

F, Monreal M, investigators R. The influence of extreme body weight on clinical outcome of

Accepted Article
patients with venous thromboembolism: findings from a prospective registry (RIETE). J

Thromb Haemost. 2005; 3: 856-62.

27 Hamatani Y, Yamashita Y, Esato M, Chun YH, Tsuji H, Wada H, Hasegawa K, Abe

M, Lip GY, Akao M. Predictors for Stroke and Death in Non-Anticoagulated Asian Patients

with Atrial Fibrillation: The Fushimi AF Registry. PLoS One. 2015; 10: e0142394.

28 Hamatani Y, Ogawa H, Uozumi R, Iguchi M, Yamashita Y, Esato M, Chun YH, Tsuji

H, Wada H, Hasegawa K, Abe M, Morita S, Akao M. Low Body Weight Is Associated With

the Incidence of Stroke in Atrial Fibrillation Patients - Insight From the Fushimi AF Registry.

Circ J. 2015; 79: 1009-17.

29 Badheka AO, Rathod A, Kizilbash MA, Garg N, Mohamad T, Afonso L, Jacob S.

Influence of obesity on outcomes in atrial fibrillation: yet another obesity paradox. Am J Med.

2010; 123: 646-51.

30 Doehner W, Schenkel J, Anker SD, Springer J, Audebert HJ. Overweight and obesity

are associated with improved survival, functional outcome, and stroke recurrence after acute

stroke or transient ischaemic attack: observations from the TEMPiS trial. Eur Heart J. 2013;

34: 268-77.

31 Wang J, Yang YM, Zhu J, Zhang H, Shao XH, Tian L, Huang B, Yu LT, Gao X, Wang

M. Overweight is associated with improved survival and outcomes in patients with atrial

fibrillation. Clin Res Cardiol. 2014; 103: 533-42.

32 Romero-Corral A, Montori VM, Somers VK, Korinek J, Thomas RJ, Allison TG,

Mookadam F, Lopez-Jimenez F. Association of bodyweight with total mortality and with

cardiovascular events in coronary artery disease: a systematic review of cohort studies. Lancet.

2006; 368: 666-78.

This article is protected by copyright. All rights reserved.

 33 Sandhu RK, Ezekowitz J, Andersson U, Alexander JH, Granger CB, Halvorsen S,

Hanna M, Hijazi Z, Jansky P, Lopes RD, Wallentin L. The 'obesity paradox' in atrial
Accepted Article
fibrillation: observations from the ARISTOTLE (Apixaban for Reduction in Stroke and Other

Thromboembolic Events in Atrial Fibrillation) trial. Eur Heart J. 2016.

34 Schenkeveld L, Magro M, Oemrawsingh RM, Lenzen M, de Jaegere P, van Geuns RJ,

Serruys PW, van Domburg RT. The influence of optimal medical treatment on the 'obesity

paradox', body mass index and long-term mortality in patients treated with percutaneous

coronary intervention: a prospective cohort study. BMJ Open. 2012; 2: e000535.

35 Yamashita T, Koretsune Y, Yasaka M, Inoue H, Kawai Y, Yamaguchi T, Uchiyama S,

Matsumoto M, Ogawa S. Randomized, multicenter, warfarin-controlled phase II study of

edoxaban in Japanese patients with non-valvular atrial fibrillation. Circ J. 2012; 76: 1840-7.

36 Sandhu RK, Ezekowitz J, Andersson U, Alexander J, Granger C, Halvorsen S, Hanna

M, Hijazi Z, Jansky P, Lopes R, Wallentin L. Body mass index and outcomes with apixaban

versus warfarin in patients with atrial fibrillation in the aristotle (apixaban for reduction in

stroke and other thromboem bolic events in atrial fibrillation) trial. Journal of the American

College of Cardiology. 2015; 1): A284.

37 Eikelboom JW, Wallentin L, Connolly SJ, Ezekowitz M, Healey JS, Oldgren J, Yang S,

Alings M, Kaatz S, Hohnloser SH, Diener HC, Franzosi MG, Huber K, Reilly P, Varrone J,

Yusuf S. Risk of bleeding with 2 doses of dabigatran compared with warfarin in older and

younger patients with atrial fibrillation: An analysis of the randomized evaluation of long-term

anticoagulant therapy (RE-LY) Trial. Circulation. 2011; 123: 2363-72.

38 AG BH. Rivaroxaban FDA Advisory Committee Briefing Document. 2011. 2011.

39 Sankyo D. Inc. SAVAYSA (edoxaban tosylate). FDA draft briefing document for the

cardiovascular and Renal Drugs Advisory Committee. NDA 206316. Meeting date October 30,

2014. 2016.

This article is protected by copyright. All rights reserved.

ccepted Articl
Study
Table 1 Characteristic of included studies

Study Indication Type and dose Comparator body weight or Age in body n % of body weight by Outcome
design of DOACS BMI in DOACs DOACs weight category (DOACs) (thromboembolic
(mean body (meanSD) cut-off /bleeding)
weight SD) (Kg)
RECOVER I,II [12, 13] RCT Acute VTE Heparin/ Dabigatran 150 Heparin/ 83.5 18.95 54.8516 50-100 DOACs =2553 Low 1.0 Recurrent symptomatic
mg bid warfarin Range 33-175 Kg Warfarin = 2554 Normal 81.6 VTE or VTE-related
High 18.3 death/
Bleeding events

Pooled EINSTEIN DVT- RCT Acute VTE Rivaroxaban 15 mg bid x3 Enoxaparin/ 82.518.9 5717 50-100 DOACs = 4150 Low 1.8 Recurrent symptomatic
PE [14-16] weeks, then 20 mg OD warfarin Range 33-220 Kg Warfarin = 4131 Normal 71.2 VTE/
High 16.8 MB, CRNMB
AMPLIFY [20] RCT Acute VTE Apixaban 10 mg bid x7 days, Enoxaparin 84.6 19.8 57.2 16 60-100 DOACs = 2691 Low 8.4 Recurrent symptomatic
then /warfarin Warfarin = 2704 Normal 69.5 VTE or VTE-related
5 mg bid High 18.9 death/
MB, CRNMB
Hokusai[22] RCT Acute VTE Heparin/Edoxaban 60 mg Heparin/ 28.3 55.716.3 60-100 DOAC = 4118 Low 12.7 Recurrent symptomatic
warfarin Warfarin = 4122 Normal 72.3 VTE/
daily (25.2-32.1)
High 14.8 MB, CRNMB
RELY [4, 23, 37] RCT AF Heparin/ Dabigatran 110 Warfarin 82.919.9 71.4+-8.6 50-100 Dabigatran 110 Low 2.07 Stroke or systemic
mg bid = 6015 Normal 80.8 embolism/
Heparin/ Dabigatran 150 82.519.4 71.5+-8.8 Dabigatran 150 High 17.1 MB
mg bid Range 32.7- = 6076
222.3 Kg Warfarin = 6022
ROCKET AF [21, 38] RCT AF Rivaroxaban 20 mg daily Warfarin 28.3 ** 50-110 DOAC = 7131 Low 2.2 Stroke or systemic
73 (65-78)
Warfarin = 7133 Normal 89.3 embolism/
(25.2-31.1)
High 7.47 MB
ARISTOTLE [24, 36] RCT AF Apixaban 5 mg bid* Warfarin 82 ** >60 DOACs = 9120 Low 10.9 Stroke or systemic
70 (63-76)
Warfarin = 9081 Other 88.8 embolism/
(70-96)
MB
AVERROSE[17] RCT AF Apixaban 5 mg x bid* Aspirin 285 709 Only BMI DOACs = 2,808 Only BMI provided Stroke or systemic
provided Aspirin =2,791 embolism/
MB
ENGAGE TIMI[19, 39] RCT AF Edoxaban 60 mg OD Warfarin 9.7% ** >60 Edoxaban 60 Low 9.8 Stroke or systemic
72 (64-78)
9.9% = 7,035 Other 89.9 embolism/
Edoxaban 30mg OD
body weight < Edoxaban 30 MB
60 =7,034
Warfarin= 7,034

*reduced to 2.5 mg bid if 2 in 3 of the following were presented: age80 years or, body weight <60 kg or, Cr>1.5 mg/dL
reduced by half, if presence of creatinine clearance = 30-50 or body weight <60 Kg or concomitant use of potent P-gp inhibitors
Median BMI (IQR), Median body weight (IQR), Mean BMI SD, **Median (IQR)
VTE = venous thromboembolism, AF= AF, MB= major bleeding, CRNMB = clinical relevant non major bleeding

This article is protected by copyright. All rights reserved.

Accepted Article

This article is protected by copyright. All rights reserved.

Accepted Article

This article is protected by copyright. All rights reserved.