Novel oral anticoagulants (NOACs) have become an alternative to vitamin K antagonists (VKAs) in non-valvular atrial fibrillation and venous thromboembolism. Little is known about the efficacy and safety of NOACs when used in adult congenital heart disease (ACHD) patients, since this cohort is not represented in the pivotal trials.
The study published by Freisinger et al. aims to know the safety and efficacy profile of NOACs, compared to VKA, in ACHD patients and in a real-life scenario. It is a retrospective, nationwide register, where a total of 44 097 ACHD patients were identified.6504 ACHD patients with a first anticoagulation prescription between 2010 and 2016 formed the basis of the longitudinal analysis, after the exclusion of patients with atrial septal defects and patients with mechanical heart valves.
The results were striking and worrisome. The authors found a progressively higher use of NOAC in this population and worse outcomes in the cohort treated with NOACs when compared to the one treated with VKAs. Patients with NOACs had higher mortality rates and both higher number of bleeding and thromboembolic events.
The percentage of ACHD patients on any oral anticoagulation increased from 6.3% in 2005 to 12.4% in 2018. The prescription of NOACs gradually increased, accounting for the 45.3% (n=2478) of all prescriptions for oral anticoagulants in ACHD patients. The usage of NOACs in 2018 over warfarin was similar in the different complexity groups (44.9% in simple cardiac defects, 47.8% in moderate complexity disease, and 40.2% in complex ACHD patients). When looking at the supplementary material, there was a high use of NOACs even in really complex pathologies: 44% of patients with transposition of the great arteries, 50% of patients with Eisenmenger syndrome and 22.6% of the patients with univentricular physiology.
Adult congenital heart disease patients on NOACs had higher thromboembolic (3.8% vs. 2.8%), MACE (7.8% vs. 6.0%) and bleeding rates (11.7% vs. 9.0%), after 1 year of therapy compared with VKAs. Thromboembolic events and MACE remained higher in the NOACs cohort on the 2-year rate analysis, while bleeding risk did not differ between both treatment groups beyond the 1st year. There were no differences in the intracranial bleeding overall.
Across the entire orally anticoagulated ACHD cohort, all-cause mortality was 3.0% at 1 year and 5.4% at 2 years of follow-up. All-cause mortality was higher in the NOACs group when compared with VKAs (4.0% vs. 2.8% at 1 year and 6.8% vs. 5.0% at 2 years, p<0.05 for both)
Both groups were not equal. Patients on NOACs were slightly older, more likely to be female (45.1 vs 48.9%) and atrial arrhythmias were more common in ACHD patients prescribed with NOACs (73% vs 63%; P<0.001). However, adjustment of both treatment groups for patients’ individual risk profiles showed consistently increased bleeding risk for NOACs compared with VKAs during follow-up, MACE and all-cause mortality. The risk of major bleeding and thromboembolic events did not differ between both groups after this adjustment.
Regarding to ACHD complexity level, risk of all- cause mortality was relevantly increased in simple (cause-specific HR 1.37; P<0.001) and moderate (cause-specific HR 1.88; P<0.001), but not in complex cardiac defects. MACE, major-thromboembolic events, and overall bleeding risk were persistently increased in all ACHD complexity subgroups. This would suggest that the negative association between NOACs therapy and outcome is likely independent of lesion complexity.
Regarding the type of NOAC used, mortality was higher with rivaroxaban, apixaban and edoxaban, while this effect was not seen in dabigatran. However, is important to point out that rivaroxaban (59.9%) and apixaban (23.2%) were the main NOAC used, while dabigatran (14.2%) and edoxaban (2.7%) were underrepresented.
This is, to date, the largest registry of patients with congenital heart disease treated with NOACs, and one of the very few comparing outcomes with another cohort of patients treated with VKAs. The results are a word of caution on the increasing prescription of NOACs to ACHD patients.
It is Intriguing, however, the discrepancy of the results when compared to others previously published, both in congenital and acquired heart disease. For example in the NOTE registry , the previous largest cohort published of ACHD patients (n=530) treated with NOACs, over a median follow-up of 1 year, the events rate was substantially lower than in the German registry. The thromboembolic event rate was 1.0%; the annualized rate of major and minor bleeding were 1.1% and 6,3% respectively. Mortality was not reported. The authors of the NOTE registry compared their results with historical cohorts treated with VKAs, which showed similar event rates. For example, Khairy et al . reported an annualized thromboembolic event rate of 1.14% in 482 ACHD patients (treated with antiplatelet therapy and/or VKA or neither) and an annualized major bleeding rate of 0.77% in 262 VKA users during a median follow-up of 5.9 years. In the present German study, the event rate was substantially higher, both in patients treated with NOAC or VKA.
This discrepancy may well be explained, at least in part, by differences in the definition of the endpoint variables among studies and by the type of patients included. In the NOTE registry the patients’ mean age was 47 years and predominantly moderate or complex defects (85%) were included, all of them followed in specialized ACHD clinic. In the current study, the mean age of the patients is 69 years old, with 64% of patients being classified in the simple defects group but having a significant number of comorbidities (78% of hypertension, 25% of diabetes, 23% of obesity, 44% of heart failure and 48.8% of chronic coronary syndromes). This could explain the higher rate of events in this population, who might be in many ways more similar to the kind of population followed in a general cardiology clinics than the one followed in a specific ACHD unit. In fact, the authors point out that many of their patients were likely community based, with up to one-third of ACHD patients in Germany only followed by general practitioners and many of the remaining patients only under local general cardiology care. This fact underlines the important role of the specialized ACHD units and the unsolved problem of the lost-to-follow-up patients, a common feature among countries.
However, it still remains elusive why these patients fare worse with NOACs than with VKAs. Especially when NOACs have been considered equal to or better than VKAs, both in pivotal studies and in real-life studies with acquired heart disease patients, who seem similar to the ones included in the German registry. Authors hypothesize that NOACs may have been prescribed liberally especially to ACHD patients with simple lesions potentially not linked appropriately to tertiary care and this may have increased rates of complications observed in their analysis.
However, differences are still found in the small subgroup of high complexity patients, probably younger and similar to the ones studied in the previously mentioned papers and to the ones followed in ACHD clinics. In this particular subgroup of patients NOACs should be used with special caution, since they probably do have a very different physiology when compared to acquired heart disease.
Finally, as the authors claim, we cannot conclude that NOACs should not be used in ACHD patients. The article highlights a clear need for prospective and, ideally, randomised studies in this topic. As pointed out in the editorialiv, the design and financing of such controlled studies will be a challenge in the era of generic NOACs. Scientific societies and healthcare funding agencies will need to come up with an innovative study design to investigate the optimal anticoagulant strategy in the growing group of patients with ACHD. Meanwhile, the use of these therapies should be done carefully and supervised by cardiologists trained in congenital heart disease.
Our mission: To reduce the burden of cardiovascular disease.