Skip to main content

Table 2 Efficacy and safety review characteristics and main findings

From: Oral anticoagulants: a systematic overview of reviews on efficacy and safety, genotyping, self-monitoring, and stakeholder experiences

Sterne et al. (2017)[20] review section

Methods and study details

Primary outcomes (number of studies)

Main comparator

Main findings

Prevention of AF-related stroke

Search: March 2014, updated September 2014

Included: 23 RCTs on AF (41 articles); 94,656 participants

Published: 1989 to 2014

Quality tool: Cochrane Risk of Bias

Efficacy: stroke or systemic embolism (15); ischaemic stroke (13); myocardial infarction (15)

Safety: major bleeding (18); clinically relevant bleeding (12); intracranial bleeding (6); all-cause mortality (18)

Main comparator: warfarin

The analyses suggested that direct OACs were better than warfarin for most efficacy and safety outcomes

Apixaban (5 mg twice daily) was likely to be one of the best options for almost all outcomes.* For example, all-cause mortality (OR 0.88, 95% CI 0.79 to 0.98; versus warfarin, INR 2 to 3); and expected incremental net benefit £7533 (95% CI 489.9 to 18,228; at a willingness-to-pay threshold of £20,000 per quality-adjusted life-year)

VTE primary prevention (mainly in hip and knee surgery)

Search: March 2014, updated September 2014

Included: 43 RCTs on VTE (46 articles); 77,563 participants

Published: 1996 to 2012

Quality tool: Cochrane Risk of Bias

Efficacy: symptomatic VTE (29);a symptomatic DVT (25); symptomatic PE (35)

Safety: myocardial infarction (9); major bleeding (39); clinically relevant bleeding (27); all-cause mortality (28)

Main comparator: low-molecular-weight heparin

There was no strong evidence to support direct OACs; no direct comparisons with warfarin and few events

Warfarin (INR 2 to 3) was likely to be best (p > 0.9)* for major bleeding (OR 0.57, 95% CI 0.39 to 0.82), and low-molecular-weight heparin for clinically relevant bleeding (p > 0.6)

Rivaroxaban was most likely to be cost-effective, but very uncertain; incremental net monetary benefit (INMB), total replacement of hip £453, 95% CI − 485 to 1312; knee £16, 95% CI − 406 to 329; £20,000 threshold

Acute treatment of VTE

Search: March 2014, updated September 2014

Included: 9 RCTs on VTE (10 articles); 28,803 participants

Published: 2007 to 2014

Quality tool: Cochrane Risk of Bias

Efficacy: symptomatic VTE (8);b symptomatic DVT (9); symptomatic PE (9); myocardial infarction (5)

Safety: major bleeding (9); clinically relevant bleeding (8); all-cause mortality (8)

Main comparator: warfarin

Analyses suggested that direct OACs were no better than warfarin, but apixaban (5 mg twice daily, e.g., major bleeding OR 0.33, 95% CI 0.18 to 0.56) and rivaroxaban (15 mg twice daily then 20 mg once daily, e.g., major bleeding OR 0.55, 95% CI 0.37 to 0.80) may be better for avoiding bleeding

Apixaban (5 mg twice daily) was likely to be one of the best for most outcomes (e.g., p > 0.9 for major bleeding; INMB £710, 95% CI − 1322 to 2185; £20,000 threshold)*

Secondary prevention of VTE

Search: March 2014, updated September 2014

Included: 10 RCTs on VTE (11 articles); 10,390 participants

Published: 1999 to 2013

Quality tool: Cochrane Risk of Bias

Efficacy: symptomatic VTE (10); symptomatic DVT (9); symptomatic PE (9)

Safety: myocardial infarction (5); major bleeding (10); clinically relevant bleeding (6); all-cause mortality (9)

Main comparator: warfarin

Inconsistent evidence suggested; apixaban (2.5 mg twice daily) was worse than warfarin for symptomatic PE (OR 10.1, 95% CI 1.66 to 102), but better for avoiding bleeding (HR 0.24, 95% CI 0.09 to 0.61); dabigatran (150 mg twice daily) was also better for bleeding (HR 0.54, 95% CI 0.41 to 0.71)

There were not enough data for the authors to calculate the likelihood of being the best option

None of the treatments was cost-effective, except possibly aspirin (INMB £623, 95% CI − 6404 to 4602; £20,000 threshold)

  1. aIn the analyses, 28 trials were included for VTE, 20 for DVT, 30 for PE, nine for myocardial infarction, 34 for major bleeding, 25 for clinically relevant bleeding, and 24 for mortality
  2. bTable 107 in Sterne’s report shows eight studies with this outcome, while the summary (p171) only mentions seven
  3. *These probabilities of being best are from rankograms, where a higher probability indicates a higher likelihood of being the best option. See https://methods.cochrane.org/cmi/glossary for the definition of a rankogram
  4. AF atrial fibrillation, RCT randomised controlled trial, OR odds ratio, INR international normalised ratio, CI confidence interval, DVT deep vein thrombosis, PE pulmonary embolism, INMB incremental net monetary benefit, HR hazard ratio