METHODS

The study was carried out in Kuopio University Hospital, Finland. Written informed consent for the study was obtained from the parents or subjects. The study was approved by the research ethics committee of the hospital.

Study population

Altogether 169 healthy children and young adults, 85 male and 84 female, were enrolled in the study. Weight and height of all subjects were recorded and body surface area was calculated.¹⁶ No subject had heart disease, as judged by history, clinical examination, and echocardiography. For the analysis the subjects were divided into five groups according to body surface area: 0.5–0.75 m², 0.75–1.0 m², 1.0–1.25 m², 1.25–1.5 m², and over 1.5 m². The clinical details of the subjects are shown in table 1.

Echocardiographic examination

Transthoracic echocardiographic examination was undertaken with the patient lying supine or in the left lateral semi-recumbent position. Sedation was not used. Examinations were carried out by a single observer (TP) using a GE Vingmed System Five ultrasound scanner (Horten, Norway) and saved in digital form on the hard disk of the ultrasound scanner. The transducer frequencies used were 2.2 MHz octave, 3.5 MHz, and 5 MHz. Standard parasternal, apical, and subcostal views were used. 3DE was done using rotational image acquisition from the parasternal long axis view for left atrium and from the apical view for left ventricle, with ECG gating and without respiratory gating. The three dimensional echocardiographic method used has been described in detail in our earlier reports, which showed closely similar left atrial and left ventricular volumes and left ventricular mass by 3DE and MRI.^14,15

Echocardiographic analysis

The echocardiographic data were analysed by a single observer (TP). Previous studies showed good intraobserver and interobserver repeatability of the measurements.^14,15

The three dimensional datasets were analysed with a detached computer. Left atrial cyclic volume change was calculated as the difference between maximum and minimum left atrial volumes. The atrial stroke volume (active emptying) is the volume reduction from the onset of atrial systole to the minimum volume at the end of diastole. The last 15% of the heart cycle was chosen to represent this time period of diastole. The passive emptying of left atrial volume was defined as the left atrial maximum volume minus the left atrial volume at the onset of atrial systole. The passive emptying percentage (PE% = PE/[left atrial maximum volume − left atrial minimum volume]*100) and active emptying percentage (AE% = AE/[left atrial maximum volume − left atrial minimum volume]*100) were calculated.

Left ventricular ejection fraction was calculated from end diastolic volume (EDV) and end systolic volume (ESV). Peak ventricular ejection rate (PER) and peak filling rate (PFR) represent the greatest volume change per unit time during three frame intervals in 3DE recording (on average 40 ms) in the ejection phase and in the filling phase of the time–volume curve. The rates of the left atrial and left ventricular volume changes were normalised by left atrial cyclic volume change and left ventricular stroke volume, respectively.

Mean time–volume curves of left atrium and left ventricle were calculated to obtain a visually informative representation of the volume changes in the subjects, as previously described.^14,15

Statistical analysis

Differences between means were assessed with Student’s paired or unpaired t test or one way analysis of variance. The Kolmogorov–Smirnov test with was used to check the normal distribution of variables. In case of non-normality the non-parametric Mann–Whitney/Wilcoxon test was used. Pearson’s correlation coefficient was used. The data are given as mean (SD).

RESULTS

All 3DE examinations were carried out successfully, and manual tracing of images was possible. The acquisition time for a 3DE dataset was 10–15 seconds, depending on the frame rate used and the heart rate. Image transfer to the detached computer took 2–3 minutes. Tracing of the left atrial and left ventricular volumes took 60–110 minutes per patient, depending on image quality and the number of frames.

The clinical details of the subjects are presented in table 1. All subjects were in good clinical condition with normal exercise tolerance. Seven children had asthma without symptoms at the time of examination. Three of the children with asthma were on treatment with inhaled corticosteroids.

Echocardiography

Figure 1 shows the time–volume curves of left atrium and left ventricle in each body surface area group. The different phases in left atrial and left ventricular volume changes and their dependency on heart rate are clearly seen in the curves.

Left atrial volumes

Data on left atrial volumes by 3DE are shown in table 2. Left atrial maximum and minimum volumes correlated well with body surface area (r = 0.88 and 0.85, respectively; both p < 0.001) and height (r = 0.84 and 0.82; both p < 0.001). Left atrial volume per unit body surface area increased with growth. Figure 2 shows the percentages of left atrial passive emptying (LAPE%) and left atrial active emptying (LAAE%) in relation to body surface area. The proportion of active emptying increased with body surface area and age (r = 0.39 and 0.34; p < 0.001). The simultaneous decrease in heart rate showed a similar correlation for LAAE% (r = −0.28; p < 0.001). In a linear regression model including both age and heart rate the effect of age for LAAE% was significant (p = 0.004), while the effect of heart rate remained non-significant. No differences between boys and girls were found in left atrial maximum and minimum volumes.

Left ventricular volumes

The 3DE findings for the left ventricle are presented in table 3. Estimates for EDV and ESV correlated well with body surface area (r = 0.91 and r = 0.89, respectively; both p < 0.001) and height (r = 0.89 and r = 0.86; p < 0.001). Left ventricular volumes/body surface area increased with increasing age. EDV and ESV estimates tended to be greater in boys than in girls, but significance was achieved only in the body surface area group over 1.5 m² (p < 0.001 for both), and in the group 0.75–1.0 m² for EDV alone (p = 0.047).

Left atrial and left ventricular PER and PFR

The results of systolic and diastolic indices calculated for left atrium and left ventricle are presented in table 4. Left atrial PER, PFR, PER/cyclic volume change, and PFR/cyclic volume change correlated with heart rate (r = 0.50, −0.48, −0.43, and 0.50, respectively; all p < 0.001). Left ventricular PER, PFR, PER/stroke volume, and PFR/stroke volume showed similar correlations with heart rate (r = 0.57, −0.55, −0.47 and 0.46, respectively; all p < 0.001).

DISCUSSION

In this study left atrial and left ventricular volumes were measured by 3DE throughout the heart cycle and the respective time–volume curves were reconstructed in 169 healthy children and young adults. Changes related to increasing body surface area and age and simultaneously decreasing heart rate were seen both in left atrial and left ventricular volumes.

Left atrial and left ventricular volumes calculated per unit of body surface area increased with growth. Estimates of left atrial volume obtained in this study exceeded the normal values measured by cross sectional echocardiography in children.² The underestimation of left atrial volumes measured by cross sectional echocardiography compared with 3DE and MRI has been reported recently.⁷ Also, validation studies have shown a good correlation between 3DE and MRI measurements of left atrial volumes.^7,14,17 In an angiographic study of 21 normal children, estimates of left atrial volumes normalised to body surface area were smaller in infants than in older children.¹⁸ Our study confirms these findings. However, estimates of left atrial cyclic volumes by both cross sectional echocardiography and angiography are crude owing to the distinctly changing shape of the left atrium in different phases of the heart cycle. The suitability of 3DE for studies in outpatients favours its use over MRI, the other method capable of three dimensional presentation. We thus believe that our results serve as ample reference data for further use.

The effect of age on left atrial function has recently been studied by cross sectional acoustic quantification in a large group of normal subjects aged 3 to 79 years.¹⁹ Active left atrial emptying increased from 24% to 56% and passive left atrial emptying decreased from 76% to 44% between the first and eighth decades. In our study active emptying percentage (AEPE%) of the left atrium increased with age, from 19% (mean body surface area 0.6 m², age 2.8 years) to 35% (mean body surface area 1.9 m², age 17.9 years), which is consistent with earlier findings.^19,20 The increase in left atrial active emptying with age was associated with a decrease in heart rate and a simultaneous increase in left atrial volume/body surface area with growth. The increase in left atrial contractile function may in part reflect greater left atrial volumes at the onset of left atrial contraction (Starling effect). It has been reported earlier that left atrial size at the onset of atrial contraction is a major determinant of left atrial ejection force.²¹ A decreasing heart rate might in part explain the increase in left atrial active emptying. Heart rate is the strongest determinant of atrial filling fraction and atrial velocities, as shown by Doppler echocardiography in children.²² There is also a steady state level in early peak filling velocity in children, suggesting that the diastolic properties of the left ventricle do not change greatly during childhood.

The present study is the first to report left ventricular volumes and function assessed by 3DE in relation to growth. Previous studies with MRI and 3DE have shown that 3DE is an accurate method for studying left ventricular volumes in children.^8–15 Left ventricular time–volume curves have been generated by MRI,⁵ radionuclide angiography,⁴ and recently by real time 3DE.²³ Accurate determination of PER and PFR requires a temporal resolution of less than 40 ms/frame.²⁴ In this study, temporal resolution was on average 14 ms (mean 55 frames/heart cycle, range 25–85), showing different phases of heart cycle in most of the individual time–volume curves. The mean time–volume curves were calculated to obtain a visually informative way of presenting time–volume data on a large number of subjects. The calculation method used decreased the temporal resolution of the mean time–volume curves as compared with individual curves. Indices for left ventricular ejection and filling were calculated from individual 3DE data in children and adolescents of varying size. Our results were in accordance with earlier studies showing heart rate dependency of PER and PFR.^4,25 Time–volume methods as well as Doppler echocardiography are strongly affected by heart rate, which may change rapidly during the study. It remains to be seen whether assessment of diastolic left ventricular function by 3DE will prove to have any advantages over the previously used methods in pathological conditions.

Limitations of the study

The major limiting factor of 3DE in clinical practice is the time needed for manual tracing of images. In our study, determining the volume of either the left atrium or the left ventricle during the heart cycle took an average of 45 minutes per patient. However, maximum and minimum volumes could be determined in five minutes, and these measurements could be adapted for clinical use even now. Further improvements in technology will provide us with accurate automatic border detection and will shorten the time needed for volume calculations. So far, the use of endocardial border detection has been reported to underestimate left ventricular volumes, while only a moderate decrease in the time needed for analysis was achieved.²⁶

Respiratory gating was ignored in this study in order to shorten the image acquisition time and to allow the examination of small children without sedation. The harmful effect of filled lungs was avoided by using a left semirecumbent position and by finding optimal acoustic windows into the left atrium (parasternal window) and the left ventricle (apical window). The quality of 3DE datasets collected was suitable for analysis even in young adults. Data acquisition during several heart cycles may have led to a loss of data in late diastole because of physiological changes in heart rate. Image acquisition during several heart cycles produced an estimate of averaged volumes during both heart and respiratory cycles.

Ten image planes were chosen in order to obtain an image acquisition time short enough to allow small children to be studied. The number of image planes was predestined and was the same in all the children. There are 3DE studies that have reported acceptable accuracy for left ventricular volumes with precordial acquisition at 12–16° intervals.²⁷ Studies with the rapid apical rotational image acquisition technique have shown accurate quantification of left ventricular volumes with as few as four to six coaxial slices. Our 3DE method was also validated with MRI in 30 children, who are included in this study.

Real time echocardiography is expected to overcome problems related to the technology used in the present study. A matrix probe enables acquisition of instantaneous volumetric images without respiratory or ECG gating.²⁸ Rapid image acquisition also decreases the risk of motion artefact.

The onset of atrial systole was considered to begin with the onset of the P wave on ECG. Although atrial systole starts after the P wave because of the electromechanical delay, this delay is short and left atrial volumes are not expected to change significantly from the onset of the P wave to the beginning of left atrial systole.²⁹

The atrial appendage is a true part of the left atrium. However, its location may vary and the volume of the hooked and tube shaped appendage is not always easy to measure. Studies based on cross sectional echocardiography^2,3,20 or acoustic quantification of cross sectional echocardiography¹⁹ do not discuss the role of the atrial appendage. In an MRI study in adults the appendage was included in left atrial volume¹⁷ and in a 3DE study the appendage was excluded.⁷ As a reliable assessment of the left atrial appendage was not always possible in the present study its volume was excluded from the reported left atrial volumes.

Despite these limitations, the results presented provide a framework for further use of 3DE in children with various heart diseases. On-line 3DE volume measurements will soon be available in modern echocardiographic machines, which will undoubtedly increase the role of 3DE in clinical decision making.