Studies that reported various aspects of AF including epidemiology, co-morbidities, risk factors, management and/or clinical outcomes in the Middle East or South Asia were included. Exclusion criteria included studies on sub-clinical AF, post-operative AF, genetics, cost-effectiveness, and smart devices. The countries in this review were categorized according to the world regions as per the World Health Organization (WHO), i.e., South-East Asia and Eastern Mediterranean regions [6]. For the purpose of this review, the terms Middle East(ern) and South Asia(n) will be used.

A comprehensive electronic literature search using MEDLINE, EMBASE, and CENTRAL was performed on September 26, 2020. Boolean terms (“OR” and “AND”), Medical Subject Headings (MeSH), Emtree and broad keywords were used. The search terms were combined and included: “atrial fibrillation”, “Middle East”, “South Asia”, “Arab”, individual country and nationality. Search limitations were not applied. Reviews and meta-analyses, international registries and the references’ lists of the retrieved articles were manually screened to identify additional relevant studies. The detailed search strategy is described in Supplementary Table 1.

All records (N = 8995) were reviewed on the titles and abstracts levels. Duplicate publications, posters, abstracts, irrelevant studies or those that did not meet the inclusion criteria were excluded. Relevant abstracts (n = 305) were screened and 84 of them were retrieved in full texts. After excluding 29 studies (Supplementary Table 2), the search strategy resulted in 55 studies [7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61] (Figs. 1,2). However, 17 studies [45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61] were related to three registries [30, 33, 41] (Supplementary Table 3) and were used when provided additional data or analysis. The included studies were tabulated, and their data were extracted for the study design, setting, enrolment period, sample size, demographics, prevalence or incidence of AF, co-morbidities, risk factors, AF types and symptoms, management, outcomes, and risk determinants (Supplementary Tables 3,4,5,6,7).

Fig. 1.

Flowchart of search strategy and identification of relevant studies with summary of the inclusion and exclusion process.

Fig. 2.

The selected studies. Visual summary of the studies: AF or non-AF registry (Top) and single- or multi-center non-registry (Bottom) studies. The studies enrolled either AF patients only or specified subgroups which included proportion of patients with AF. The studies which are represented in an oval shape have an enrollment period of two years or less. Abbreviations/acronyms: AF, atrial fibrillation; COOL-AF, Cohort of Antithrombotic Use and Optimal INR Level in Patients With Non-Valvular Atrial Fibrillation in Thailand; GARFIELD-AF, Global Anticoagulant Registry in the FIELD-Atrial Fibrillation; Gulf RACE-2, second Gulf Registry of Acute Coronary Events; Gulf SAFE, Gulf Survey of Atrial Fibrillation Events; HEARTS, Hearts Function Assessment Registry Trial in Saudi Arabia, HP-RF/RHD, Himachal Pradesh- Rheumatic Fever/Rheumatic Heart Disease; IHRS-AF, Indian Heart Rhythm Society-Atrial Fibrillation; IRAF, Iranian registry of atrial fibrillation; RealiseAF, Real Life global Survey Evaluating patients with atrial fibrillation; SAS, Saudi Atrial Fibrillation Survey.

The incidence and prevalence of AF was reported in association with other diseases. The overall incidence of AF during hospitalization was 1.8% in patients admitted with acute myocardial infarction (AMI) in Qatar [48], with the incidence in Arabs being higher than that in South Asians (3.1% vs. 0.8%). From the earlier data of the same Qatari registry (1991 to 2002) [45], the overall annual incidence among Qataris was higher than that of other nationalities (12% vs. 8.0%). However, the rate (8.0–9.3%) remained consistent over years for the overall patient population, i.e., those hospitalized with acute cardiac illness [30, 46, 47, 50]. The incidence of AF in the Gulf RACE-2 registry was 2.7% among patients hospitalized with ACS [27]. Two studies from Pakistan [15, 18] reported similar frequency (9.1%) in patients hospitalized with AMI. The prevalence of AF among general hospital admissions was 3.4% in Bahrain [13], 4.2% in Kuwait [19], and 5.8% in Pakistan [24]. Whereas, a lower prevalence (2.8%) was reported in the only study that was conducted in a primary health care center in Iran [14]. The prevalence of AF was 7.8% among patients on maintenance HD in Jordan [22], and 10.6% in patients with chronic heart failure in Morocco [29]. In Iran, the prevalence of AF among patients with stroke was 11.1% [20], while in Qatar, it was 6.4%. In the latter Qatari study, 2.6% of stroke patients were newly diagnosed with AF [28]. The highest AF prevalence reported was 17.8% in patients hospitalized with heart failure [26], and 23.9% in patients with RHD [39] (Table 2 and Supplementary Table 3).

Rate control was the strategy reported in more than half of the patients. Rhythm control was approached in up to one-third of the patients using amiodarone in up to 40% of them. Beta-blockers were used in more than 50% of the patients, digoxin in 30% and calcium channel blockers in approximately 20% of them. Only two AF registries [31, 32] reported the use of ablation. In the Saudi registry (SAS), 3.8% of patients had been ablated [32]. The Iranian registry (IRAF) reported ablation use in 34.0% and 29.8% of patients with familial and non-familial AF, respectively [31]. The use of vitamin K antagonists (VKAs) across the studies ranged from about 40% to 70%, and more than 90% in two studies [34, 41]. Of the 14 studies [17, 20, 24, 26, 31, 32, 34, 35, 40, 41, 43, 44, 46, 59] that reported anticoagulation agents use, seven studies [20, 31, 35, 40, 41, 43, 44] recruited patients in the non-vitamin K oral anticoagulants (NOACs) era, i.e., beyond 2010. Of the seven studies, two did not report NOACs use [20, 35], four South Asian AF registries [40, 41, 43, 44] reported use between 1.9% to 9.1%, and the IRAF registry reported higher average NOACs than VKAs use (45.7% versus 23.0%) [31]. Across the AF registries, aspirin and clopidogrel use ranged from 19.5% to 54.4% and 11.0% to 19.5%, respectively (Table 5, Ref [32, 55, 59] and Supplementary Table 6).

The main adverse outcomes reported by the studies included mortality (in-hospital, one-year), cerebrovascular events, and bleeding. In-hospital mortality was reported in 0.79% of patients in Bahrain [13] and was doubled in Thailand (1.6%) [37]. The rate was 3.6% in Qatar among patients admitted with general cardiac conditions [30], with a rate of 4.8% in Middle Eastern and 1.2% in South Asian patients as defined in the study [47]. The rates were higher in patients with acute heart failure (6.7%) [26], chronic kidney disease (CKD) (11.7%) [51], ACS (14.7%) [27], and acute ischemic stroke (9.0% [20] and 26.8% [36]). One-year mortality rate ranged from 6.5% to 18.1% in four South Asian studies [37, 40, 43, 44]. Gulf SAFE registry reported a rate of 15.0% [58] with a significantly lower mortality rate in patients who were compliant with the AF Better Care (ABC) pathway (7.3% versus 13.1%, p = 0.033) [55]. In the same registry, diabetic patients had significantly higher one-year mortality rate than non-diabetics (14.4% versus 9.6%, p = 0.003) [53]. Similarly, diabetic patients who were compliant with the ABC pathway had lower mortality rate as well (5.8% versus 15.9, p = 0.0014) [54]. Higher one-year mortality rates ($>$20%) were reported in AF patients presenting with ACS [27] or acute heart failure [26]. The observed rate of cerebrovascular accident (CVA) and/or transient ischemic attach (TIA) in the South Asian AF registries ranged from 0.85% to 3.0% [43, 44, 60], while a single-centre study found much higher rates; 17.2% in non-valvular and 23.1% in valvular AF patients [35]. The Gulf SAFE reported a rate of 6.4% [58], while the Qatari registry reported a lower rate of 1.1% in patients admitted with cardiac condition [46]. Additionally, the rate was higher in Middle Eastern (1.4%) than South Asian (0.6%) population in the Qatari registry [47]. The rates of CVA/TIA in special AF patient populations were 2.3%, 2.0%, 1.8–2.8%, and 6.7% in CKD [51], acute heart failure, [26] AMI/ACS [27, 48], and RHD [39], respectively. Two single-center studies from Iran reported higher rates in patients presenting with AMI (12.5% [18] and 13.6% [15]). Finally, the rate of major bleeding in studies from the Middle East [26, 27, 58] ranged from 1.1% to 1.7% including those in patients presenting with ACS [27] and acute heart failure [26]. However, one-single center study from KSA reported a rate of 35.6% in those with high risk of bleeding [7]. Studies from South Asia reported rates ranged from 0.31% to 4.4% [35, 43, 44, 60]. A study from UAE found that the use of inappropriate NOACs doses subjected the patients to higher ischemic stroke and major bleeding rates compared to those on appropriate doses (9.9% versus 3.1%, p 0.01 and 11.7% versus 6.0%, p = 0.04; respectively) [8] (Supplementary Table 7).

In hypertensive Thai patients, older age, male gender, and presence of CKD increased the prevalence of AF. On a multivariate analysis, factors of strongest association with increased prevalence were age of 65 years or above (adjusted odd ratio (adj OR) 2.86, 95% confidence interval (CI), 1.89–4.33; p 0.001) followed by male gender (adj OR, 1.98, 95% CI, 1.35–2.91; p 0.01), lower low-density lipoprotein-cholesterol (adj OR, 1.87, 95% CI, 1.28–2.73; p 0.01) and high uric acid levels (adj OR, 1.48, 95% CI, 1.02–2.17; p 0.05). The presence of combined factors, further increased the risk. Odds increased to 2.46 (95% CI, 1.66–3.63; p 0.001) when two risk factors were present and to 8.99 (95% CI, 5.02–16.07; p 0.001) when four factors were combined [38]. In Indian patients with RHD, the following variables were independent determinants of AF, age (adj OR, 1.04, 95% CI, 1.03–1.06), mitral stenosis (MS) (moderate MS (adj OR, 2.4, 95% CI, 1.7–3.4) or severe MS (adj OR, 2.1, 95% CI, 1.5–3.0)), left atrial size (adj OR, 1.10, 95% CI, 1.08–1.11), New York Heart Association classes (II, III, IV), all degrees of tricuspid regurgitation (TR), and combined mitral regurgitation and MS with or without TR [39]. In patients hospitalized with AMI, the only independent factor of AF was age (adj OR, 1.06, 95% CI, 1.04–1.08; p = 0.001) in the Qatari registry [48].

In the Qatari registry, ACS was associated with increased risk of in-hospital mortality (adj OR, 4.36, 95% CI, 1.77–10.74; p = 0.001) [30, 47] and not the age alone as a correlate with mortality. In a multiple logistic regression analysis, beta-blocker administration upon hospital admission was a good correlate with reduced in-hospital mortality (adj OR, 0.36, 95% CI, 0.15–0.87; p = 0.02) [30]. Gender [46] or ethnicity [47] were not associated with poor outcomes. A multivariate analysis found the following as correlates with in-hospital mortality, cardiogenic shock (adj OR, 285, 95% CI, 84.0–970; p = 0.001), diabetes mellitus (adj OR, 1.52, 95% CI, 1.03–2.26; p = 0.04), and chronic renal impairment (adj OR, 1.18–3.68; p = 0.001) [46]. In Thailand, correlates with mortality after AF hospitalization included CKD (hazard ratio (HR), 2.01, 95% CI, 1.68–2.41; p 0.0001), high CHA${}_2$DS${}_2$-VASc score (e.g., $\ge$6; HR, 2.23, 95% CI, 1.34–3.70; p 0.002), and RHD (HR, 1.72, 95% CI, 1.53–1.92; p 0.0001) [37]. In Thai patients who experienced AF-associated acute ischemic stroke on admission or during hospitalization, age ($\ge$75 year) and acute kidney injury were independent risk factors for three-month mortality (adj OR, 3.08, 95% CI, 1.17–8.13; p = 0.019 and adj OR, 6.38, 95% CI, 1.96–20.76; p = 0.001, respectively). Baseline oral anticoagulation therapy was associated with 87% risk reduction (adj OR, 0.13, 95% CI, 0.03–0.64; p = 0.002) [36].

In the Gulf SAFE registry, in-hospital or emergency room (ER) mortality was strongly associated with congestive heart failure (adj OR, 2.64, 95% CI, 1.79–3.89; p 0.001), peripheral vascular disease (adj OR, 2.26, 95% CI, 1.01–5.08; p = 0.048), and reason for ER visit (cardiac and noncardiac). Warfarin and antiplatelet use upon discharge were independent protective factors (adj OR, 0.51, 95% CI, 0.32–0.83; p = 0.006 and adj OR, 0.55, 95% CI, 0.35–0.87; p = 0.01, respectively) [59]. When comparing patients with and without diabetes mellitus, in-hospital or ER mortality was associated with only the presence of CKD (relative risk (RR) 6.38, 95% CI 2.73–14.87) in diabetic patients. Whereas, in non-diabetics, several factors were involved, age $\ge$75 years (RR, 3.04, 95% CI, 1.69–5.50), CKD (RR, 7.41, 95% CI, 3.69–14.92), heart failure (RR, 2.05, 95% CI, 1.21–3.68), and stroke (RR, 3.00, 95% CI, 1.51–5.98). Correlates with one-year mortality in both diabetics and non-diabetics, were older age ($\ge$75 year), heart failure, and CKD. Digoxin use increased mortality risk by 74% in the absence of diabetes (RR, 1.74%, 95% CI, 1.04–2.90; p = 0.035). In contrast, 55% mortality risk reduction was associated with beta-blocker use in diabetic patients (RR, 0.45, 95% CI, 0.24–0.84; p = 0.012) [53]. History of previous stroke or TIA increased the risk of one-year stroke event rate (OR, 2.81, 95% CI, 1.82–5.00; p = 0.001). Similarly, antithrombotic therapy (ATT) undertreatment (OR, 2.76, 95% CI, 1.42–5.35; p = 0.003) but not ATT overtreatment, was associated with one-year stroke risk. However, ATT overtreatment was the only factor associated with significantly higher one-year major bleeding event rate (OR, 3.29, 95% CI, 1.51–7.15; p = 0.003) [57]. In patients with non-valvular AF, smoking, cardiac reasons for ER visit, and both CHADS${}_2$ and CHA${}_2$DS${}_2$-VASc scores of $\ge$2 were independent correlates with one-year stroke or TIA. Warfarin prescription upon discharge reduced the risk by about 60% (OR, 0.38, 95% CI, 0.17–0.83; p = 0.015) [59]. Adherence to the ABC pathway was significantly associated with lower six-month and one-year mortality either in the entire registry population (OR, 0.31, 95% CI, 0.13–0.77; p = 0.013 and OR, 0.46, 95% CI, 0.25–0.86; p = 0.015, respectively) [55], or in diabetic population (OR, 0.18, 95% CI, 0.42–0.75; p = 0.019 and OR, 0.29, 95% CI, 0.11–0.76; p = 0.012, respectively) as compared to the non-ABC adherent group [54].