Obesity is a major public health problem increasing risk for hypertension, diabetes mellitus, cardiovascular diseases, heart failure, stroke, left atrial enlargement and obstructive sleep apnoea. Conventionally, body mass index (BMI) is used in many studies to diagnose and classify obesity. However, epidemiological studies have shown that waist:hip ratio is a better predictor for myocardial infarction and sudden cardiac death compared with BMI. Nevertheless, some subjects who have ‘healthy obese phenotype’ have a benign cardiometabolic profile due to greater deposition of the fat tissue in the subcutaneous area rather than in the visceral area. Large clinical studies have confirmed ‘obesity survival paradox’ in patients with coronary artery disease and heart failure. ‘Sexual dimorphism’ is another interesting issue about obesity survival paradox that gender should be considered as an important confounding factor when investigating the correlation of obesity with cardiovascular diseases.1 ,2 These findings suggest that metabolic effects of adipose tissue may be different between males and females. It is well known that body fat mass is greater in females and the regional distribution of fat deposits is different between males (apples) and females (pears).

Adipose tissue is accepted as an endocrine organ secreting many molecules called ‘adipokines’. Leptin, resistin, adiponectin, visfatin are some of the well-known adipokines and a broad range of effects of adipokines on endocrine and cardiovascular systems have been reported. In addition, other organs such as bone marrow and mononuclear cells also secrete adipokines such as resistin, provoking the inflammatory response in obesity. However, adipokine levels are affected by age, gender, ethnicity, body fat composition, dietary habits, physical activity and normal ranges for adipokines have not been defined. Adipokine levels may also change during the menstrual cycle. There are several articles investigating the correlation of adipokines with different diseases, but a major problem is inadequate standardisation of the aforementioned characteristics between the study populations.3 ,4

Atrial fibrillation (AF) is the most common arrhythmic condition affecting millions of people and obesity increases risk of AF. The comorbid diseases such as hypertension and diabetes mellitus make it difficult to investigate the direct effect of obesity on risk of developing AF. Inflammation is accepted as a pathophysiological mechanism of AF and obesity may trigger a chronic inflammatory state via adipokines. Resistin is a candidate molecule to explain the link between obesity and AF through its proinflammatory effects.

In their Heart publication, Ermakov et al reported a correlation between higher resistin levels and incident AF in postmenopausal women as part of the Women's Health Initiative investigation. Over a follow-up of 11.1 years, 18% of 4937 women developed AF.5 In multivariable analysis, adiponectin and leptin were not found to predict incident AF in their cohort. Ermakov et al reported novel findings as their population was restricted to female subjects. The other parameters such as age and BMI were comparable to the previous studies. In none of the previous studies, gender subgroup analyses were performed and this article gives clue about the effect of sexual dimorphism on adipokine and AF relation. Gender is an important factor considering body fat composition, dietary habits, total caloric intake and adipokine levels. Leptin, adiponectin and resistin levels were found to be higher in healthy females compared with males.4 Another important finding is that even though the incidence of AF was not different between smokers/non-smokers and current alcohol users and the rest of the study population, patients with higher resistin levels had higher risk for AF in subgroup analyses. The authors concluded that resistin levels may be used for prediction of AF risk in alcohol users and for explaining the mechanism by which alcohol predisposes patients to AF.

There are a few studies investigating the correlation between resistin, adiponectin and leptin levels with risk of AF and the findings are controversial (table 1).5–10 The incidence of AF ranged from 2.7% to 27.8% indicating the heterogeneity of the study populations and follow-up durations. All of the studies reported HRs/ORs adjusted for BMI, thus the reported associations may be accepted to reflect the independent effect of adipokines on AF incidence. The median resistin level was about 12 ng/mL but a cut-off value of resistin for prediction of AF was not reported in any of these reports. All of the studies provided only one measurement of adipokine levels. It is obvious that using mean of more than one measurement for any investigational parameter is more appropriate. For studies following patients for years, it may be more logical to measure annual adipokine levels and correlate the risk of AF with yearly obtained adipokine levels.

It is clear that adipokines have important role in human metabolism and abnormalities of adipose tissue cause metabolic abnormalities through adipokines. Whereas, the normal values and levels predicting disease states are not well established for adipokines and it seems that individual differences have great impact on adipokine levels. The mentioned articles provide novel data about the correlation of adipokines with different diseases but the results are controversial. Especially, the effect of gender and dietary habits should be considered in future studies and results of gender specific subgroup analysis should be included.