Research output per year
Research output per year
Research output: Contribution to journal › Comment/debate › Research › peer-review
In the article, “Use of direct oral anticoagulants in the first year after market entry of edoxaban: A Danish nationwide drug utilization study”, the authors have regrettably identified an error in the analysis reported in their paper. The issue was caused by an error in the specification of the study population in the initial data retrieval from the registries and have led to some users of edoxaban not being identified and included in the study. The results for the other direct oral anticoagulants (DOAC), dabigatran, rivaroxaban, and apixaban, were not affected, except for a slight increase in the proportion with an atrial fibrillation diagnosis (≈2%; see the corrected Table), owing to the accrual of diagnosis data used for this classification. Indication for use of direct oral antagonists initiated between June and December 2016, based on registry-based hospital discharge diagnoses (Table presented.) *Edoxaban is not approved for use in relation to knee and hip surgery in Europe. The authors noted that the in the corrected results, they identified 948 new edoxaban users (Table) compared to 609 users in the original report. Consequently, the corrected monthly rate of new use of edoxaban among all Danish adults now increases to 3.1 per 100 000 person months in June 2017 (Figure, left panel). Although the general characteristics of edoxaban users have not changed (Table), and they are thus still comparable to users of other DOACs, the proportion of anticoagulant-naïve edoxaban users have increased from 4.8% in the original report to 38.8% in the corrected results (Table). This change relates directly to the underlying error in data specification, with almost all of the newly identified users being anticoagulant-naïve. While the proportion of anticoagulant-naïve edoxaban users is still markedly lower compared to users of the other DOACs (38.8% vs. 56.5–68.6%), the author note that this weakens one of the principle conclusions of the original report, i.e. that edoxaban users were almost always switched to edoxaban from other anticoagulant therapy. With this noticeable exception, remaining results and conclusions, including that of the abstract, final paragraph of the discussion and ‘key points’, remain unaltered by the correction. The authors provided the corrected figure and tables below. The Supplementary Tables only concerned the other DOACs and are thus not provided. Baseline characteristics of direct oral antagonist (DOAC) users between June and December 2016 (Table presented.) NOTES: AC = oral anticoagulant; IQR = interquartile range; GI = gastrointestinal; TIA = transient ischaemic attack; NSAID = non-steroidal anti-inflammatory drug; P2Y12-antagonists = clopidogrel, ticagrelor and prasugrel. *Defined as the time from inclusion (index prescription) to the recent prescription for any oral anticoagulant (VKA or DOAC): current (<120 days), recent (120 days to 2 years), distant (>2 years) and never-use. **Dose is based on the prescribed capsule- or tablet strength, classified as high (60 mg edoxaban; 150 mg dabigatran; 20 mg rivaroxaban; 5 mg apixaban), low (30 mg edoxaban; 110 mg dabigatran; 10-15 mg rivaroxaban; 2.5 mg apixaban) or very low (15 mg edoxaban; 75 mg dabigatran; 2.5 mg rivaroxaban). Baseline characteristics of new users of edoxaban between June and December 2016, stratified by indication for use (Table presented.) NOTES: AC = oral anticoagulant; IQR = interquartile range; GI = gastrointestinal; TIA = transient ischaemic attack; NSAID = non-steroidal anti-inflammatory drug; P2Y12-antagonists = clopidogrel, ticagrelor and prasugrel. *Defined as the time from inclusion (index prescription) to the recent prescription for any oral anticoagulant (VKA or DOAC): current (<120 days), recent (120 days to 2 years), distant (>2 years) and never-use. **Dose is based on the prescribed tablet strength. 1 (Figure presented.) The monthly rate of new adult users of edoxaban (per 100 000 person-months) and, for comparison, dabigatran, rivaroxaban, and apixaban from June 2016 to June 2017 CHA2DS2-VASc and HAS-BLED scores among users of edoxaban, dabigatran, rivaroxaban, and apixaban, restricted to those with atrial fibrillation (Table presented.) The authors apologize for any confusion, which these errors may have caused.
Research output: Contribution to journal › Journal article › Research › peer-review