Atrial Fibrillation during Cerebral Infarction in Brazzaville: Frequency and Predictive Factors

In order to contribute to the improvement of brain infarction management in Brazzaville, a cross-sectional and analytical study with prospective data col-lection was conducted in the cardiology and neurology departments of the Brazzaville University Hospital, from February 1 to July 31, 2018. It included patients hospitalized for cerebral infarction confirmed with imaging, and having done an etiological assessment with at least one electrocardiogram at rest and one of long duration. Among these 138 patients included, 11 had atrial fibrillation, equaling a frequency of 7.9%. The mean age of AF patients was 71 ± 8.8 years. The cardiovascular risk factors found were hypertension in eight cases (72.7%), diabetes in five cases (45.5%), abdominal obesity in four cases (36.4%). AF was permanent in 10 cases (91%), and paroxysmal in one case (9%). It was valvular in three cases (27.3%) and non-valvular in eight cases (72.7%). The cardiopathy involved was hypertensive in seven cases (63.6%), ischemic and valvular in two cases each. The CHA2DS2-VASc score, calculated in eight patients, was an average of 2.2, and ≥2 in more than 80% of patients; HAS-BLED score of 2.4 on average was ≥ 3 in more than 72% of patients. Digoxin was prescribed in seven cases (63.6%) and an anti-vitamin K in eight cases (72.7%). In multivariate analysis, age (OR = 20.10, p = 0.023), arterial hypertension (OR = 23.82, p = 0.011), and dyslipidemia (OR = 2.03, p = 0.032) were the predictive factors found. AF is infrequent during brain infarction in Brazzaville. This systematic research raises the problem of age in our context.


Introduction
Cerebral ischemia is a major public health problem worldwide [1] [2]. Indeed, it is the second leading cause of death in the world and in developing countries, it is responsible for 30% of deaths occurring in the first three weeks of its onset, and also responsible of 30% of cases of acquired permanent physical handicap [3] [4] [5]. Among these etiological mechanisms, cerebral embolisms occupy a prominent place alongside the atherothrombosis of large vessels. Cerebral ischemia of cardio-embolic origin is generally more serious, with a high risk of recurrence, and a high mortality rate, hence it's of great importance of recognizing a cardio-embolic origin in order to ensure better prevention by anticoagulant therapy [6] [7] [8] [9]. These cardiac-origin cerebral embolisms are usually associated with supraventricular arrhythmias, such as atrial fibrillation (AF) or atrial flutter, which are evidenced by several diagnostic methods, including ECG at rest, Holter ECG, continuous monitoring, and implantable devices, each with varying sensitivity and specificity, particularly for the detection of paroxysmal AF [10] [11] [12]. Also, in order to help improve the management of patients with cerebral infarction in our context, we began the present study with a triple objective: to determine the frequency of AF during cerebral infarctions, to list the main etiologies of AF, and identify its predictive factors.

Patients and Methods
It was a cross-sectional prospective data collection, descriptive and analytical, scores, the rate control and anti-thrombotic treatments used. Sampling was done by exhaustive pulling. Thus, out of 1781 hospitalized patients, 261 were for a stroke, of which 147 for a cerebral infarction. Among these ones, 138 files were selected by simple random pulling, and making our study sample.

Statistical analysis
The software CSPro 7.1 and Stata 12.0 allowed the recording, the classification and the analysis of the data. The qualitative and quantitative variables were compared with the Pearson Khi-2 test and the Student's test. Multivariate logistic regression analysis allowed the identification of predictive factors for AF. The threshold of significance was set at a value of p < 0.05.

Ethical clearance
This study has obtained ethical clearance from the Congolese Ethical Comittee for Research in Health Sciences.

Results
Of the 138 patients included, 11 had atrial fibrillation (AF), equals a frequency of 7.9%. The mean age of AF patients was 71 ± 8.8 years (range: 53 to 83 years) with a sex ratio of 0.83. The cardiovascular risk factors found were: arterial hypertension in eight cases (72.7%), diabetes in five cases (45.5%) and abdominal obesity in four cases (36.4%). Table 1 presents the main characteristics of patients according to the existence or not of AF. Regarding medical imaging, the middle and anterior cerebral arteries were the most frequent topographies, with 47% and 30% of cases respectively. AF was permanent in 10 cases (91%), and paroxysmal in one case (9%). It was labeled valvular in three cases (27.3%) and non-valvular in eight cases (72.7%). Cardiopathies involved were hypertensive in seven cases (63.6%), ischemic and valvular (mitral stenosis and mitral insufficiency) in two cases each. The CHA2DS2-VASc score, calculated in eight patients was at average at 2.2, and ≥2 in more than 80% of patients; HAS-BLED score of 2.4 on average was ≥ 3 in more than 72% of patients. For the rate control treatment, digoxin was used in seven cases (63.6%), a calcium channel blocking inhibitor in four cases (36.4%), and a beta-blocker in three cases (27.3%). Anti-thrombotic treatment was made of low molecular weight heparin (LMWH) in nine cases (81.8%), an anti-vitamin K in eight cases (72.7%), platelet antiaggregant (acetylsalicylic acid) in two cases (18.2%), and a direct oral anticoagulant in one case (9.1%). Multivariate analysis after logistic regression noted that, of the various factors analyzed, only age (OR = 20.10, p = 0.023), arterial hypertension (OR = 23.82, p = 0.011), and dyslipidemia (OR = 2.03, p = 0.032) were predictive factors for the occurrence of AF during cerebral infarction. Table 2 presents the results of the logistic regression.

Discussion
In the literature, the frequency of atrial fibrillation during cerebral infarctions is   [17]. This great variability is probably due to methodological differences, not only related to the duration of the study and the size of the sample which is the case in our study, but also and especially to the methods used to detect AF. Indeed, if the diagnosis of permanent AF is easy on the standard ECG alone, it is more difficult for paroxysmal AF. In the different studies, there is a clear correlation between the duration of the recording and the detection rate of paroxysmal AF. In these different series, the lowest AF frequencies were obtained with recordings lasting 24 to 48 hours [12] [18] [19] [20], while the highest FA frequencies were recorded with implantable devices, performing long-term recordings of up to 1500 hours on average [21]- [26]. This is the case for Ritter et al. [23], who reported in a comparative study of FMD detection, frequencies of 2% and 17%, depending on whether it was a recording of shorter duration (168 h) or longer duration (1536 h). Thus, the short recording time (24 to 48 hours) may explain the low detection rate of AF, as some authors have shown [18] [19] [20]. In addition, more recent studies have highlighted the importance of certain recording methods in the detection of paroxysmal AF, with much greater sensitivity and specificity than traditional recording methods. This is the case of continuous monitoring in Neurovascular Intensive Care Unit (NICU), by conventional or automated method [22]; this latter makes it possible to detect paroxysmal AF with a high sensitivity from the first 72 hours after admission to NICU. However, despite the ever-increasing number of strokes in our context, the scarcity of these units makes it difficult to exploit this innovative and beneficial technique. In our series as well as in other studies in sub-Saharan Africa [15] [17] [27] [28], patients with cerebral infarction are relatively young, compared to those in Western series [29], the frequency of AF increases with age [30], which may explain the relative rarity of this rhythm disorder during cerebral infarctions in this part of the continent where high blood pressure is the main risk factor for stroke. Regarding factors predicting the occurrence of AF during cerebral infarctions in our series, only advanced age (>70 years old), the presence of arterial hypertension, and the existence of dyslipidemia were identified as factors associated with AF, as other authors have also shown [10] [15]. However, there are a multitude of other factors often involved in the onset of AF, including diabetes mellitus, obesity, sleep apnea syndrome, and many others [31]. The systematic search for these various factors and their effective management could help to reduce the risk of atrial fibrillation and the burden of morbidity and mortality attributable to it.

Conclusion
This preliminary study showed that atrial fibrillation (AF) is relatively infrequent during brain infarction in Brazzaville. It is frequently found in elderly people, mostly hypertensive, constituting its main predictive factors. Its systematic research in the etiological assessment of cerebral infarctions raises the problem of age because of the profitability of the 24 h Holter, its availability and its cost. Hence it's in need of preventive measures based on the effective management of modifiable risk factors (hypertension, diabetes, dyslipidemia, obesity) to reduce the burden of AF and its complications.