Although BNP is a well-established screening tool for LVSD and AER is routinely used for cardiovascular risk estimation, it is not known whether these two biomarkers are clinically useful for screening of LVDD and LVH in patients with type 2 diabetes with no prior CVD. The present study has demonstrated that echocardiographic screening of asymptomatic diabetic subjects without known CVD may identify significant numbers of patients with asymptomatic LVDD and LVH. We also show that BNP is independently related to moderate to severe LVDD whereas AER is directly and independently related to LVM.
Prevalence of subclinical heart disease in a diabetic population
Our data shows a prevalence of 49% of LVDD (29% mild, 19% moderate and 0.7% severe LVDD) in presumably heart healthy asymptomatic patients with type 2 diabetes. These findings are well in line with earlier studies i.e. data from Poirier et al were 46 men with type 2 diabetes who were aged 38-67 years; without evidence of diabetic complications, hypertension, coronary artery disease, congestive heart failure, or thyroid or overt renal disease; and with a maximal treadmill exercise test showing no ischemia, were studied. Twenty-eight of the subjects (60%) were found to have LVDD, of whom 13 (28%) had moderate LVDD and 15 (32%) had mild LVDD .
In our study, LVH was observed in only 9.4% of the patients. Fang et al. reported that in asymptomatic patients with diabetes mellitus without known cardiac disease, which underwent clinical evaluation and detailed echo assessment, approximately 20% fulfilled the criteria for LVH.
It has recently been published data by Srivastava and coworkers, showing a high prevalence of LVDD and LVH (59% and 70% respectively)  in patients with type 2 diabetes. As in our study, the prevalence of LVSD was fairly low (16%). The prevalence of LVH reported in that study is higher compared to our findings (70% v s 9.4%). However in that study population, 19% had known macrovascular disease. Interestingly, increasing age, diabetes duration and BMI were the only independent predictors of cardiac abnormalities. No differences regarding cardiac abnormalities between male and females were reported for .
BNP and correlations to echo abnormalities
We have previously reported that patients with type 2 diabetes have higher levels of BNP compared to patients without diabetes, indicating a high prevalence of asymptomatic heart disease . However, since no echo registrations were done in this study , no information regarding what underlying heart disease an elevated BNP value would represent could be obtained. For screening purposes BNP have been proven useful to rule out heart failure (HF) with impaired LVSD . BNP have also been shown useful for the assessment of LVDD at the population level  whereas its role for LVDD screening has not previously been investigated in type 2 diabetes patients with no prior CVD. We found that BNP is an independent determinant of moderate to severe LVDD. This is illustrated in figure 1 were BNP appears as good discriminator for the detection of moderate to severe LVDD in patients with type 2 diabetes. Henceforth, plasma BNP levels above 1.5 pmol/l indicates coexisting moderate diastolic dysfunction and further examination with echo would be warranted to further confirm this.
On the other hand, we did not found any significant relationship between BNP and mild LVDD or LVM, respectively. In accordance with the latter finding, screening with BNP for the presence of LVH in hypertensive patients has showed limited value .
AER and correlations to echocardiographical abnormalities
Whereas BNP was not associated to LVH, our data suggests that AER could be useful as a marker for LVH, as estimated by LVM, given its independent and direct relationship with LVM. An independent relationship between LVM and AER has not been reported earlier in type 2 diabetes and further emphasizes increased CVD risk and need of antihypertensive treatment in type 2 diabetes patients with microalbuminuria. Thus, this finding might serve as an explanatory model for the known risk of CVD in patients with microalbuminuria , since LVH is well associated with increased morbidity and mortality in CVD .
Male gender and left ventricular hypertrophy
The correlation between male gender and LVM is in line with the study by Concardy et al in 2004, in which it was shown that the LVH estimation without sex-specific criteria underestimated the prevalence of LVH in women and overestimates it in men, and even if sex specific definitions for LVH was used male gender still contributed to the prevalence of LVH in this study . Furthermore, an independent influence of increasing LVM and age were reported by Concardy et al and this influence was only seen in the male population of the study. Even if the patients studied by Concardy et al were hypertensive peers with higher prevalence of LVH (52.2%) compared to our diabetic population with a LVH prevalence of 9.4%, the data are in concordance with our findings and might represent a gender specific mechanism for the development of LVH. This is further supported by data from the Tromsö study in which male sex was found to be an independent predictor of LVH in a multivariate logistic regression analysis .
Female gender and diastolic dysfunction
Another interesting finding from the present study is the increased risk for moderate LVDD for female study subjects. The mechanism behind such a gender dependent pathophysiology is unclear. However, the results of the Framingham study revealed that between the ages of 45 and 75, men and women with diabetes had a two-fold- and five-fold increase in risk of developing heart failure, respectively, compared to patients without diabetes. This risk persisted even after considering age, blood pressure, cholesterol, weight, and history of coronary artery disease . Further studies are indeed warranted regarding this matter.
Screening for cardiac abnormalities in asymptomatic patients with type 2 diabetes
In our study almost half of the study population had LVDD and nearly ten percent had LVH. The treatment of LVH and LVSD is well established and has proved to reduce cardiovascular complication [4, 5]. However, the importance of diagnosing LVDD has just been acknowledged and the presence of LVDD provide essential prognostic information . Although the most effective treatment of LVDD in patients with type 2 diabetes is unknown, it is well established that an effective aggressive multifactorial approach indeed reduces micro- and macrovascular complications . Furthermore, recently published data suggests cardiovascular benefits of more aggressive blood pressure lowering therapy amongst patients with type 2 diabetes . Henceforth, an early detection of cardiac abnormalities such as LVSD, LVDD and LVH is crucial and might guide the clinicians to more intense risk factor management and aggressive pharmacological treatment of this high-risk population. Screening with echo would be considered optimal for such a conduct, but is not always eligible due to lack of accessibility and high cost. Our data suggest that the measurement of BNP and AER might be used as a screening tool in order to select patients for further work up with echo. However, the study sample size of a total of 153 subjects, which indeed is a respectable number for an echocardiographic study, might not be large enough to be definitive or inform clinical practice and therefore the results from the multivariate analysis needs to be confirmed in larger study samples.