To analyze the clinical characteristics of patients with AF, the management of antiarrhythmic treatment, and treatment to prevent embolic events and readmissions, and to evaluate sex differences.

To determine the epidemiological characteristics and possible sex differences in AF patients visiting cardiology outpatient departments. To identify the risk factors for embolism in patients with AF and analyze the prevalence of each factor. To identify the prescribed treatment and adherence to the CPGs, particularly with regard to prescription of anticoagulant and antiarrhythmic drugs.

Patients were included via an online data collection form designed with the assistance of the Spanish Society of Cardiology. The form collected all clinical data, their source, the approach in the emergency department to AF patients, and the treatment received by the patient prior to arrival at the cardiology department. International normalized ratio (INR) values 3 months prior to the hospital visit were collected from patients taking VKAs to calculate the time in therapeutic range (TTR) using the Rosendaal method, measuring the days within the therapeutic range and dividing by the total number of days that the patient was taking anticoagulants. Rhythm on ECG was documented at the time of the hospital visit and the AF was classified according to clinical patterns established by the CPGs. Ischemia was defined as a depressed ST segment or a deeply inverted T wave on electrocardiogram with ischemic characteristics accompanied by typical clinical characteristics. Prior AF episodes or hospital admissions were documented when confirmed by the relevant clinical report. Data were recorded on the physician's use of echocardiography at the clinic in patients with AF, therapeutic recommendations, and referral at patient discharge.

Outpatient care for patients with AF involves various aspects that should be assessed to improve the quality of care: a) referral source and the opinion of the treating cardiologist on the appropriateness of the referral (yes/no); we considered a referral to be appropriate when the patient actually had a clinical problem (regardless of whether the patient attended the emergency department or not) that led to treatment of the AF episode by a cardiologist; we considered a referral to be inappropriate when the patient was referred for a clinical problem that was later ruled out and the cardiologist did not change his or her approach to the AF episode; b) the performance of echocardiography in patients with AF; c) the treatment received by the patient in the cardiology department, and d) the use of treatments for embolism prevention, anticoagulation therapy with VKAs assessed by TTR, and the reasons given by cardiologists for not using these treatments.

In 2013, over 50 000 patients were seen in outpatient cardiology clinics of the 2 hospitals, of whom 24 958 were first visits from primary care, the emergency departments of both hospitals, or referrals from other specialties: In total, 14 183 patients attended hospital A and 10 775 patients attended hospital B. The prevalence of AF was 21% in patients attending our cardiology departments, of whom 23.7% had de novo AF and an unknown percentage of the remainder had decompensated AF. We estimated that 5% of all our patients could be included in the registry.6 A sufficiently large number of patients had to be included to ensure the robustness of the statistical analysis. The sample had an alpha risk < 0.5 and an accuracy of 5% for calculating the size of the sample. Assuming 5% of patients would be lost, it was estimated that the number of patients included in our 2010 study would have to be tripled. Beginning on March 1, 2014, more than 500 patients were included over a 3-month period.

Quantitative variables with a normal distribution are expressed as mean ± standard deviation; the Student t test was used for their analysis. The chi-square test or, if needed, Fisher's exact test was used to compare qualitative variables by sex or type of AF. A P value of < .05 was used as a cutoff for statistical significance. The data were analyzed using SPSS version 15.0 (SPSS, Inc.; Chicago, Illinois, United States).

More than 53% of patients with clinical problems were referred from primary care or other hospital departments, and more than a third were referred from the emergency department. Once the referral was assessed according to predefined criteria, 94% of patients were considered to have been appropriately referred in the opinion of the treating cardiologist. Thus, this care indicator performed beyond expectations and does not need improvement.

Echocardiography is recommended in the CPGs. More than 94% of patients underwent the technique and only 35 (6.5%) patients were not considered to require an echocardiogram after assessment. Notably, 321 of these patients underwent the technique in the same consultation, which demonstrates the high percentage of patients attended using a strategy known as one-stop consultations. Therefore, there was full compliance with this care indicator.

As recommended by CPGs, beta-blockers are the drugs most commonly used to slow heart rate.2 However, ECV was recommended in only 17 (3.2%) of the study patients and pulmonary vein ablation in only 20 (3.8%); more patients should undergo these procedures, as recommended by the CPGs. Importantly, many of the patients were already receiving anticoagulants and could have been recommended to undergo ECV (class I if there are symptoms or hemodynamic instability, and class IIa to begin a long-term rhythm control strategy2). There was no apparent reason for sex differences in indications for ECV. The low use of ECV should be put in context, among which could be poorly controlled anticoagulation therapy, but it is difficult to find a single explanation. Similar findings8,9 have been described in Spanish emergency department registries, in which the proportion of patients with AF who underwent ECV was only 6%, and in European registries, in which the proportion was only slightly higher (9.7%). This proportion is also similar to that described in clinical trials of DAs. The RE-LY trial reported that 9.14% of patients had received DAs,10 the ROCKET trial reported 1%,11 and the ARISTOTLE trial reported 4%.12 Furthermore, recent studies have shown that the use of DAs in patients undergoing ECV is safe and effective, which should certainly encourage their use and contribute to an increase in indications for ECV in the future.13,14

Although it may appear that there was little use of pulmonary vein ablation, no explanation can be offered for the difference found between sexes. In the present study, the use of this procedure was comparable to that of European registries (4.4%) and Spanish registries, which reported 2201 pulmonary vein ablation procedures in 2013. Currently, few suitable patients undergo this technique, and there is much room for improvement in these figures.9,15 Clearly, the fact that 24.3% of our patients were younger than 65 years should serve to increase indications for this technique in the future.

This section presents 4 relevant findings. Firstly, among the 339 patients with a recommendation for VKA use, acenocoumarol was used much more frequently than warfarin: 249 (46.7%) vs 90 (16.9%) patients. To our knowledge, there is no reason for not using warfarin; furthermore, no clinical trial has demonstrated the efficacy of acenocoumarol in these patients. However, acenocoumarol is widely used throughout Spain.

Secondly, the TTR of 48.4% ± 37% in the 3 previous months is not a good indicator. Given that anticoagulation therapy is not monitored in cardiology departments, it is a care indicator that cannot be optimized in that setting; however, its use should be optimized in those departments where it is administered. However, our TTR data are similar to those of an American study reporting a TTR of 53.7% in a population of 138 319 patients. However, this figure is far lower than the 63.8% reported in the CALIFA study, which included 1056 Spanish patients, or the mean TTR of 66% in Spanish patients included in the RE-LY trial.16–18 Action should be taken in collaboration with primary care and hematology services to increase the TTR in order to optimize anticoagulation therapy. Similarly, if it becomes clear that the TTR cannot be optimized, VKAs should be replaced by DAs.

Thirdly, in 2014 the approach of cardiologists in the same departments was compared with that in 2010. We found that the percentage of patients receiving anticoagulation therapy had increased (from 60.0% to 79.8%); a justification was found for the 20.2% of the patients not recommended this treatment.6 This improvement may have been influenced by the CPGs and by the important clinical information obtained in clinical trials published in this field, which has increasingly convinced cardiologists of the need to use embolism prevention strategies in patients with AF. Regarding the levels of risk on the CHADS2 and CHA2DS2-VASc scales, the use of VKAs and DAs significantly increased in relation to the increased risk of embolism, which differed from that observed in 2010. The analysis of anticoagulation therapy in patients with nonvalvular AF showed that, as in 2010, 60% of patients with CHADS2 score = 0 and CHA2DS2-VASc score = 0-1 received anticoagulation therapy, which is not indicated in such patients according to the CPGs. A possible explanation for this approach is that most of these patients had been recently diagnosed and would be treated for electrical or pharmacological cardioversion; however, since only 14.9% received an antiarrhythmic agent and only 3.4% were referred for ECV, and thus this explanation is incorrect. The Val-FAAP study found that 48.9% of patients with a CHADS2 score= 0 received anticoagulation therapy.19 The reasons for not prescribing anticoagulation therapy to 20.2% of the patients with AF were similar to those described in other studies. The AFABE study showed that 24% of patients did not receive anticoagulation therapy. The most common reasons reported were the lack of indications based on the CHADS2 score (5.1%), cognitive impairment (3.6%), and the risk of bleeding (2.9%).20 The present study observed a significant reduction in the use of aspirin (18.1%) compared with its use in 2010 (37.1%). In another study, aspirin was prescribed in 25 of the 425 patients who had indications for oral anticoagulation therapy and in 72 of the 108 patients without an indication.6 The Val-FAAP study, which was also conducted in 2010, showed that at that time aspirin use ranged from 31.9% of AF patients with a CHADS2 score = 0 to 19.3% of AF patients with a CHADS2 score ≥ 2.19

Finally, only 15.7% of patients receiving anticoagulation therapy received DAs due to poor INR control. This aspect is clearly novel, because these drugs were not used in 2010. Poor INR control was the underlying reason for the increased use of DAs from 9% to 15.7% in patients attending the cardiology department. However, this increase is very low and is far from optimal. Administrative issues and therapeutic inertia were the 2 main reasons given by cardiologists to explain this problem, despite a high percentage of patients with poor INR control (TTR, 48.4% ± 37%). The enormous administrative problems can be broken down into 2 issues: firstly, the inability of the cardiologists to access patient data on INR control, although this information was available in this study and therefore cannot be used to explain the nonprescription of DAs. Secondly, physicians in Spain cannot follow the recommendations of the CPGs that highlight the superiority of DAs over VKAs, but can only follow the criteria recommended by the Spanish Agency for Medicines and Health Products, which prohibits the use of these drugs in patients with de novo AF and limits their use almost exclusively to patients with poor INR control.

Attention should be drawn to therapeutic inertia because, given the observed TTR, there has been a marked improvement in adherence by cardiologists to the recommendations of the CPGs. It is very likely that the strategy of using DAs will strongly increase in the future.

The study was affected by selection bias because the patients were referred from other departments in our hospitals. The diagnostic criteria are also another possible source of bias, given that they may have led to patient exclusion due to diagnostic errors. The data collected from each patient may have suffered bias, which may have led to errors in their interpretation. The study could have been optimized by the inclusion of the symptoms reported by the patients. It would have been useful to have compared the results obtained in 2014 with those before the publication of the CPGs, using a statistical design with the same population as in 2010. However, this was not done and different populations were studied, thus preventing such an analysis. The HAS-BLED (hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio, elderly [> 65 years], drugs/alcohol concomitantly) scale was not used to analyze the risk of bleeding, which would have completed the assessment of the patients with AF. Poor practice was shown by the finding that nearly 60% of patients with no indication for anticoagulation therapy received this treatment; this is an outstanding issue that needs to be addressed and resolved.