Left atrial (LA) volume is a more accurate measure of LA size than LA diameter (1- 2), although both LA size parameters have been shown to be markers of cardiovascular risk (3- 12). We are unaware of any prospective studies that compare the utility of the two LA size parameters for the prediction of cardiovascular outcomes. Additionally, although LA volume has been demonstrated to be a robust predictor of cardiovascular outcomes among patients with sinus rhythm in a number of retrospective studies (5- 6,13- 14), its prognostic utility for patients with atrial fibrillation (AF) is unknown. In this prospective study, we assessed the clinical and echocardiographic correlations of biplane LA volume and M-mode LA dimension and compared the utility of LA volume, area, and dimension for the prediction of age-related cardiovascular outcomes in patients with sinus rhythm and those with AF.

This study was approved by the Mayo Foundation Institutional Review Board. Patients age ≥50 years referred for a general medical consultation were invited to participate if they had no history of congenital heart disease, treatment with pacemaker implantation, valvular surgery, or cardiac transplantation.

Age, gender, height, weight, brachial blood pressure, cardiac rhythm, and history of comorbid conditions were recorded at enrollment. Definitions for all covariates have been previously published (15).

Baseline cardiac rhythm was considered sinus if the patient was in sinus rhythm at the time of echocardiography and had no prior history of atrial arrhythmias. Paroxysmal AF was defined by a history of AF episodes with intervening sinus rhythm. Because of the small number of paroxysmal AF patients, we did not try to differentiate whether conversion to sinus rhythm in these patients was spontaneous or not. Permanent AF was defined by AF at baseline and the persistence of the arrhythmia without intervening sinus episodes.

Data for all echocardiographic studies were collected prospectively, and the parameters of interest were specified a priori. Measurements were obtained at least two times for an average if the rhythm was sinus and at least three times if the rhythm was AF. M-mode LA dimension was measured as per the American Society of Echocardiography method (16). Single-plane area was evaluated from the four-chamber view of the left atrium at end-ventricular systole, ensuring that there was no foreshortening of the atrium. The area was then planimetered with the inferior LA border defined as the plane of the mitral annulus, excluding the confluence of the pulmonary veins and the LA appendage.

Maximal biplane LA volume was measured in all patients with a modified biplane area-length method (17). This method as well as the Simpson’s method of disc had both been well-validated (18- 20). Orthogonal apical views, most commonly apical four- and two-chamber views, were obtained for determination of LA area and length (from the middle of the plane of the mitral annulus to the posterior wall). The apical long-axis view was used instead of the two-chamber view if the left atrium in the latter view appeared foreshortened. Specifically, the maximal LA chamber area and length were measured at end ventricular systole, excluding the LA appendage and pulmonary veins. Left atrial volume was calculated on the basis of the algorithm ([0.85 × A1 × A2]/L); where A1 is the four-chamber LA area, A2 is the two-chamber or apical long axis LA area, and L is the average of the two lengths obtained from the orthogonal views) and indexed to body surface area.

Indexed LA diameter, four-chamber LA area, and indexed LA volume were categorized according to current American Society of Echocardiography guidelines (21). Additionally, we also assessed non-indexed LA diameter with 40 mm as the cut-off for normal on the basis of common clinical practice.

Other echocardiographic variables, specified a priori, included left ventricular (LV) dimension at end-systole and end-diastole; LV septal and posterior end-diastolic wall thickness; M-mode LV ejection fraction, mitral inflow filling velocities (peak E and A); mitral inflow deceleration time; mitral isovolumic relaxation time; pulmonary venous systolic and diastolic forward flow, and atrial reversal flow velocities; and mitral annular motion septal velocity by tissue Doppler imaging (E’).

All patients were prospectively followed for development of new outcome events, which included AF, stroke, transient ischemic attack (TIA), myocardial infarction (MI), coronary revascularization, congestive heart failure (CHF), and cardiovascular death.

Continuous variables are presented as means ± standard deviation. Categorical variables are displayed as percentages. Differences between group means were evaluated with t tests (continuous variables) or chi-square analyses (categorical variables), as appropriate. Spearman correlations were used to assess the relation between LA size and other clinical and echocardiographic variables. Receiver-operator characteristic curves were generated to assess the overall performance of the various LA size parameters for the prediction of cardiovascular events. Differences in rates of cardiovascular events by LA size categories were examined with log-rank tests. Cox proportional hazards modeling was used to determine the association between LA dimension, LA area, and LA volume with future cardiovascular events, after adjusting for age, gender, and other significant covariates.

A total of 423 patients (mean age 71 ± 8 years, 56% men) met all study criteria and consented to participate. Clinical indications for medical consultation included: dyspnea (24%); chest discomfort (23%); palpitations, presyncope, and syncope (16%); and other reasons (11%). Of these, 317 (74%) were in sinus rhythm and had no prior history of AF, 49 (12%) had a history of paroxysmal AF, and 57 (13%) had permanent AF. Characteristics of the study population are outlined in (Table 1), stratified by the baseline cardiac rhythm.

Clinical and echocardiographic correlates of LA dimension or indexed LA volume are listed in (Table 2). Some parameters were significantly associated with one LA size parameter but not the other. Many baseline characteristics correlated significantly but not closely with LA size. For both LA dimension and indexed LA volume, LV mass indexed to height and permanent AF were the two strongest correlates.

Over a mean follow-up period of 3.5 ± 2.3 years, 62 patients had 90 new events (23 new AF, 8 strokes, 6 TIAs, 12 MIs, 14 developed CHF, 26 coronary revascularizations, and 1 cardiovascular death). Of 317 subjects in sinus rhythm at baseline, 26% had a LA diameter <40 mm and therefore considered “normal.” According to the American Society of Echocardiography guidelines (22), there were significant differences in the categorization of the LA size abnormality, depending on which LA size parameter was used (Table 3). Patients who developed cardiovascular events had larger LA size at baseline, irrespective of the method of quantitation (Table 1).

When the overall sensitivity and specificity of different methods of LA size quantitation using receiver operator characteristic curves were compared, the area under the curve for indexed LA volume was greater than that of the other two methods, and its sensitivity was superior throughout the range of LA size (Figure 1), even if coronary revascularization was excluded as an end point (area under curve for LA volume 0.73 versus 0.70 and 0.66 for LA area and diameter, respectively). There was a stepwise increase in risk of cardiovascular events with each increment of LA volume category (p < 0.0001) (Figure 2). Although there was an overall association between indexed LA dimension and LA area and future cardiovascular events (both p < 0.0001), a graded association between increasing indexed LA diameter or LA area and risk of events was not present. After adjustment for age, gender, and other covariates, larger indexed LA volume and LA dimension predicted higher risks of cardiovascular events (Table 4).

Among the 106 patients with paroxysmal or permanent AF at baseline, 35 had 43 new events (10 strokes, 4 TIA, 5 MI, 16 CHF, and 8 coronary revascularization procedures) after a mean follow-up period of 3.0 ± 2.4 years. Subjects with AF at baseline were more likely than those with sinus rhythm to develop new events (36.8% vs. 19.6%, p < 0.0001). Specifically, AF patients had a higher incidence of stroke (9.4% vs. 2.5%, p = 0.002), TIA (3.8% vs. 1.9%, p = 0.27), and CHF (15.1% vs. 4.4%, p < 0.0001), but there was no difference between subgroups in the incidence of MI (4.7% vs. 3.8%, p = 0.67) or coronary revascularization (7.6% vs. 8.3%, p = 0.83). The difference in LA diameter, area, or indexed volume between individuals with and without a cardiovascular event during follow-up did not reach significance. For those with and without events, baseline median (interquartile range) LA dimension was 51 mm (47 to 57 mm) versus 51 mm (45 to 57 mm) (p = 0.84); four-chamber LA area 30 cm² (26 to 35 cm²) versus 28 cm² (24 tot 36 cm²) (p = 0.49); and indexed LA volume 57 ml/m² (46 to 78 ml/m²) versus 53 ml/m² (37 to 68 ml/m²) (p = 0.34). The overall performance of LA diameter or indexed volume to distinguish AF patients who did or did not have subsequent cardiovascular events was poor, irrespective of whether coronary revascularization was included as an end point. In the AF subgroup, none of the LA size parameters was independently predictive of the combined outcome when adjusted for age, gender, hypertension, diabetes, heart failure, and myocardial infarction.

Even when warfarin therapy at baseline was considered, a relationship between LA size and future events was not detected. Of the 106 patients with AF, 52 (49%) were receiving warfarin at baseline (compared with 18 of 317 [6%] of patients with sinus rhythm, p < 0.0001). In multivariable models for the prediction of events in the AF group, warfarin therapy was not independently related to the combined end point. For instance, in the multivariate model with clinical parameters (hypertension, diabetes, congestive heart failure, coronary artery disease, and ejection fraction), there seemed to be a trend toward higher risk of events associated with warfarin use (hazard ratio 1.89, p = 0.063).

Indexed LA volume was a more sensitive risk marker of future cardiovascular events than LA diameter or area for patients in sinus rhythm. The predictive utility of LA size for future cardiovascular events was poor in AF patients, regardless of the method used for quantitating LA size.

Dilatation of the left atrium, in the absence of organic mitral valve disease or history of AF, has been shown to reflect the burden of cardiovascular disease (22- 24). During ventricular diastole, the left atrium is exposed to the pressures of the left ventricle. With increased stiffness of the left ventricle, LA pressure rises to maintain adequate LV filling, and the increased atrial wall tension leads to chamber dilatation and stretch of the atrial myocardium.

Although M-mode LA dimension is easy to acquire, its validity has recently been challenged (1). Because the left atrium is an asymmetrical cavity, LA size is more accurately reflected by a measurement of volume rather than area or a linear dimension. Furthermore, LA dilatation might not be evenly distributed in all planes, and measurement of anteroposterior dimension is likely to be insensitive to changes in LA size.

In the presence of AF, however, our findings suggest that LA enlargement needs to be interpreted with caution. The mechanism responsible for the lack of an association between LA size and cardiovascular events in the presence of AF cannot be fully determined in this study. In addition to the various mechanisms that lead to increased LV filling pressures and thus LA enlargement in sinus patients, AF patients might also develop progressive and refractory LA dilation from tachycardia-induced atrial myopathy (20) with advanced remodeling of the atrium with replacement fibrosis, independent of increased LV filling pressures. The more prevalent use of warfarin in the AF group could not explain the lack of association between LA size and risk for development of cardiovascular events. Rather than exerting a protective effect with lowering the risk of combined events, warfarin use seemed to be associated with a nonsignificant trend toward a higher risk for such events. This was most likely secondary to a selection process; that is, patients receiving warfarin at baseline were at higher risk for cardiovascular events. Admittedly, the relatively small sample of AF patients in this study might not have provided sufficient power to detect an association between LA size and outcomes in this subgroup. Nonetheless, LA enlargement in AF patients does not seem to have the same robustness for prediction of cardiovascular outcomes as it does in sinus rhythm.

The study population is referral-based, and the extent to which the data can be generalized to other population groups is not known. The contribution of medical therapy to change in LA size was not considered in this study, given the relatively small sample size, the diverse range of reasons for which the patients were seen, and the differences in dosage and duration of various therapies.

Indexed LA volume is a more robust cardiovascular risk marker than LA area or diameter in patients who are in sinus rhythm; however, in patients with AF, the predictive utility of LA size for future cardiovascular events seemed unsatisfactory, regardless of the method of LA quantitation. Future studies to include larger number of AF patients for further evaluation will be warranted.