For decades, the left ventricular ejection fraction has been used to predict cardiovascular outcomes. More recent data have shown that maximum left atrial volume (maxLAV) indexed for body surface area (maxLAVi) is an independent predictor of survival after myocardial infarction and congestive heart failure [1‐5]. It has also been suggested that maxLAVi is less liable to fluctuate with short-term changes in loading conditions compared with Doppler filling indices and, therefore, maxLAVi has been proposed as a more specific marker of a persistent increase in left ventricular (LV) filling pressure [6]. The fact that maxLAVi carries independent prognostic information and seems to add indispensable information to the Doppler filling indices for assessing LV filling highlights the importance of measuring maxLAVi. Far less attention has been paid to the possibility that elevated filling pressure may cause a reduction in left atrial (LA) contractility, with a corresponding enlargement of minimum left atrial volume (minLAV). Consequently, minLAV indexed for body surface area (minLAVi) may have an additive value to the maxLAVi and Doppler filling indices in the evaluation of LV diastolic function [7].

In the guidelines of the American Society of Echocardiography/European Association of Cardiovascular Imaging (ASE/EACVI) [8], the given upper limit of maxLAVi is based on small cross-sectional studies performed in relatively young patient-based subsets during the 1980s [9, 10], in a small subgroup of 44 young patients [11], and in a group of 92 healthy volunteers [12]. In our clinical experience, the ASE/EACVI guidelines’ recommended upper reference limit for maxLAVi is frequently encountered in adults without known or apparent cardiovascular disease. We set out to examine maxLAVi and minLAVi in a well-characterized population-based sample of people without apparent cardiovascular disease or other factors potentially associated with LA enlargement.

In the present population-based study, the distributions of maximum and minimum LA volumes were studied in 168 randomly selected subjects without apparent cardiovascular disease or comorbidity known to be related to LA enlargement. The current data confirm that maxLAVi was independent of age and sex, and displayed an upper normal limit of 37 mL/m². In contrast, minLAVi was correlated with age, and exhibited estimated upper reference limits of 15 mL/m² and 23 mL/m² in subjects aged 40 and 80 years, respectively.

LA enlargement has been proposed as a barometer of diastolic burden and as a predictor of stroke, heart failure, and cardiovascular death. Therefore, it is important to include maxLAVi in the standard echocardiographic examination [1‐6, 16]. However, this apparently simple assignment presents several problems. Several imaging techniques, such as echocardiography, cardiomagnetic resonance imaging, and high-resolution computed tomography, have been used to obtain LA volumes [10, 17‐21]. The geometric models and methods used for LA measurement vary between single or biplane methods, 3D models, area-length models, and disc models. The differences in LA volume measurements between different studies and methods are disturbingly large and emphasize the need for proper population-based reference values for the actual method used.

In the present study, the upper limit of the reference interval for maxLAVi was 37 mL/m², as measured using the apical 4-chamber single-plane disk method. In comparison, the mean + 2 SD of biplane Simpson indexed LA volumes reported in the 1980s by two smaller studies (n < 60) of young healthy volunteers (mean age <40 years) were slightly lower, at 34 and 32 mL/m², respectively [9, 10]. In a healthy subgroup (n = 44) with a mean age of 39 years and free from possible diastolic dysfunction, Tsang et al. [11] reported an upper reference limit (mean + 2 SD) of 32 mL/m². In a patient population free from apparent cardiovascular disease, Thomas et al. [12] observed upper reference limits of 39 and 35 mL/m² in the subgroups younger (n = 47) and older (n = 45) than 50 years, respectively. Identical to our finding, Kou et al. [22] reported an upper reference limit of 37 mL/m² in a very recent large multicenter study of healthy volunteers (n = 734). Interestingly, their estimated upper reference limit of maxLAVi from the single-plane 4-chamber view was exactly the same (i.e., 37 mL/m²) as their biplane estimate.

In the ASE/EACVI guidelines, the upper limit of maxLAVi is reported as ≤28 mL/m² [8]. Obviously, the recommendation is not only derived from cross-sectional studies of healthy subjects, but is also based on the estimated risk related to LA volume and on expert opinion. According to previous cross-sectional studies, the guideline-recommended reference limit of 28 mL/m² would approximately correspond to a population mean + 1 SD in an apparently healthy population. The use of 1 SD to define normality is notable, as it surely will categorize otherwise-healthy individuals as diseased. Certainly, LA volumes ≥28 mL/m² have been reported to be predictive of future congestive heart failure [4] and compound cardiovascular events [23]. An even lower LA volume of ≥27 mL/m² has been reported as being predictive of future atrial fibrillation in patients with hypertrophic cardiomyopathy [24]. Most biological risk markers measured on a continuous scale are likely to display a risk continuum without an exact point above which the risk suddenly increases, and an increased risk can even be detected at levels that are considered to be within the distribution of the healthy population (e.g., serum cholesterol [25] and blood pressure [26] ). In fact, the concept of “optimal” prognostic thresholds has recently been questioned [27]. Risk stratification is not straightforward, as it requires multifactorial considerations in a clinical context and cannot be simplified axiomatically into a threshold of a single parameter. In our opinion, the reference limits should be thresholds dedicated to inform the clinician about the distribution of a marker in the healthy population, separate from prognostic or therapeutic decision limits.

As expected, the maxLAV was larger in men than in women in the present study. However, this sex difference disappeared after adjusting for body surface area. Some studies have not found a relationship between maxLAVi and age [10, 12, 19], whereas others have done so [28]. The present data showed no significant correlation between age and maxLAVi.

In the early nineties, Appleton et al. [7] noticed that a minLAV of ≥40 mL predicted a pulmonary wedge pressure >12 mmHg with a sensitivity and specificity of 82 and 98 %, respectively. Until recently, minLAV was almost ignored. The left atrial function has traditionally been described as modulating LV filling in three phases: the reservoir phase (during atrial relaxation and LV systole, ending up in maxLAV), and the conduit and the atrial contraction phases (during LV filling, ending up in minLAV). As opposed to that observed during the reservoir phase, during the conduit and atrial contraction phases, the LA is directly exposed to LV pressure. An increase in LV filling pressure may directly affect LA pressure and, therefore, LA size. Consequently, minLAV may be a more sensitive marker of an increased LV filling pressure than maxLAV. Supportive of this hypothesis, Russo et al. [29] recently found a considerably stronger association between E/e′ ratio and minLAV than between E/e′ ratio and maxLAV.

The present study confirmed previous findings that minLAVi increase with age [28, 30]; however, we did not assess the pre-atrial contraction volume and may, therefore, only speculate on this age dependency. In contrast with the LA reservoir function, previous studies have shown that the conduit function deteriorates with age [28, 30]. The observed decrease in passive atrial emptying is most likely related to changed LV filling properties associated with age. However, the LA contraction function seems to be maintained or even amplified with age [28, 30]. The observed increase in LA active pump function with age may be a compensation for the impaired early filling and is possibly mediated by the Frank-Starling mechanism. However, the fact that our data and those of others showed a rise in minLAVi with age suggests that increased atrial contractility may not fully compensate for the age related decrease in passive LA emptying.

The intra- and interobserver reproducibility of the LA volume measurements in the present study were very similar to the recent findings of Aune et al. [19], including a slightly poorer reproducibility for minLAV compared with maxLAV. As the LA is at its smallest size at minLAV, the echo broadening in atrial septum increases because of an augmentation in atrial septal thickness, rendering it harder to identify the true wall echoes, which may explain the difference in reproducibility observed between minLAV and maxLAV.

Because LA enlargement is observed frequently in subjects with LV diastolic dysfunction, a condition that presents frequently in elderly individuals, we made an effort to exclude the possibility of a systematic bias, i.e., that several of the participants had preclinical diastolic dysfunction. Therefore, conventional Doppler indices, tissue Doppler filling indices, and NT-proBNP levels were analyzed. The distributions of diastolic Doppler measurements in the present study population were within what could be expected in a population without apparent cardiovascular disease compared with findings from a large Scandinavian study of healthy subjects [31]. In addition, based on an EACVI-guideline-recommended algorithm [13], an analysis suggested a low burden of possible preclinical diastolic dysfunction in the present study population.

The present population-based subset of men and women were all of northern European descent; therefore, the extent to which the data can be extrapolated to other ethnic groups is not known. Only 41 (24 %) of the participants were women, which may have reduced the statistical power. The main reason for the sex difference was that, for each patient included in the VaMIS study, a control subject matched for age and sex was recruited from the general population. Consequently, the sex distribution of the control subjects who were enrolled in the present study reflects the sex difference in patients hospitalized for myocardial infarction.

Unfortunately, the acquisition and storage of loops dedicated to LA planimetry were only obtained in the 4-chamber view, and not in the 2-chamber view in the present study. Thus, the guideline-recommended biplane assessment of LA volumes [8] was not possible, which represented a limitation of the study. However, biplane planimetry is quite often unfeasible because of suboptimal image quality in the apical 2-chamber view, preventing adequate visualization of the LA anterior wall [21]. Lester et al. [32] demonstrated that the mean ± SD of the absolute difference between the single-plane and biplane disc methods was 6 ± 5 mL, indicating a strong agreement between the two methods. Although bi- and single-plane assessment of LA volumes are not interchangeable, previous studies have suggested that the single-plane method may be acceptable for clinical use if reference limits that are specific for the method are available [21, 32].

The intra- and interobserver variability was obtained from the same set of images, i.e., only the measurement variability was tested, and not inconsistencies caused by variations in imaging planes or beat-to-beat variations.

The strengths of the present study included the population-based design and the well-characterized participants without signs of cardiovascular disease or risk factors and who were not taking any medications. The evaluation of LV diastolic function and NT-proBNP concentration was important considering the well-known association between LV filling function and LA volumes [13].

In the present population-based random sample of middle-aged and older subjects without known or apparent cardiovascular disease or comorbidity, the maximum and minimum LA volumes were studied using the single-plane disc method. The current data confirmed that maxLAVi is independent of age and sex, and showed that it displayed an upper normal limit of 37 mL/m². In contrast, minLAVi was correlated with age and exhibited estimated upper reference limits of 15 and 23 mL/m² in subjects aged 40 and 80 years, respectively.