Introduction

Left ventricular (LV) dysfunction may develop as a result of LV dyssynchrony and/or inappropriate atrioventricular (AV) delay in some patients after single chamber, ventricular pacing. Even after dual-chamber sequential pacing, maintenance of AV synchrony is necessary to preserve cardiac function and to achieve a better prognosis [1, 2]. AV delay optimization is, therefore, important to maintain better cardiac function and a favorable long-term outcome after sequential pacing [3, 4] or cardiac resynchronization therapy [5, 6]. Although several echo-Doppler- [7–13] as well as electrocardiogram- [14–17] based methods to optimize AV interval have been proposed, routine or systematic use of AV optimization remains controversial [5, 6, 18–20]. Transmitral flow (TMF) by transthoracic Doppler echocardiography is commonly used to optimize AV delay. However, the advantage of echo-Doppler-based AV optimization over fixed AV delay or a commercially available AV optimization algorithm based on electrocardiogram has not been proven yet.

Theoretically (based on the Frank–Starling law), AV delay should be optimized to achieve maximal LV filling without deterioration of LV function [2]. Because TMF alone does not reflect both systolic function and LV filling pressure, TMF-based AV optimization may not provide enough advantage over the other methods. A previous echo-Doppler study demonstrated that the difference between the duration of pulmonary venous flow reversal (PVa) and mitral forward flow during atrial systole (A) reliably estimates LV filling pressure [21]. We hypothesized that a combination of TMF and PV flow measurements may be beneficial to further optimize AV delay to achieve better cardiac function with adequate LV filling pressure. Therefore, the aim of this study was to assess feasibility of the AV delay optimization by combined use of TMF and PV flow.

Materials and methods

Study population

This study included 32 patients after dual-chamber pacing for complete AV block (n = 26, mean age = 79 ± 8 years; 12 males) or cardiac resynchronization therapy (n = 6, mean age = 65 ± 16 years; 4 males). The exclusion criteria were current atrial arrhythmia and frequent premature ventricular beats. Informed consent was provided by each participant before enrollment in this study.

Study protocol

Echocardiography was performed with a Sonos 5500 and S3 transducer (Philips Medical Systems, Andover, MA, USA). TMF was obtained from apical 3-chamber or 4-chamber views with the sample volume positioned at the tip of the mitral leaflets. TMF consists of 2 distinct flow signals, early or E wave and late or A wave during atrial contraction. PV flow was obtained from an apical 4-chamber view with the sample volume placed in the left superior pulmonary vein. An effort was made to maintain the same position of the pulsed Doppler sample throughout the echo-Doppler examination.

AV delay optimization was performed using TMF and PV flow at rest. Optimal AV delay (AVDOPT) was defined as the AV delay where the duration of PVa minus A was the minimum (=0). Because the onset of the A wave cannot be always detected, the difference between the duration of PVa and A was alternatively measured as (time interval between the onset of the Q wave and the end of the A wave) − (time interval between the onset of the Q wave and the end of the PVa wave). To simplify this method, TMF and PV flow were recorded at the pre-set AV delay. Then, AVDOPT was determined as (pre-set AV delay) + (duration of PVa − duration of A) (Fig. 1). Stroke volume (SV) was measured by a pulsed Doppler method obtained at the LV outflow tract and was used as an index to assess cardiac function during AV optimization. SV obtained at the AVDOPT was compared with SV obtained at shorter (AVDOPT − 50 ms) or longer (AVDOPT + 50 ms) AV delays.

Fig. 1
figure 1

AV delay optimization using TMF and PV flow. a (Step 1) At a pre-set AV delay (=180 ms in this case), both TMF and PV flow signal were recorded. (Step 2) The difference in duration between PVa and A wave was measured (=−30 ms). b (Step 3) Optimal AV delay was calculated as (pre-set AV delay) + (duration of PVa − duration of A). In this case, the optimal AV delay was calculated as 180 ms + (−30 ms) = 150 ms

Full size image

Statistical analysis

The measurements are expressed as mean ± standard deviation. Statistical analyses were performed with one-way analysis of variance (ANOVA) using the Bonferroni post hoc test. Values of P < 0.05 were considered statistically significant.

Results

Clinical characteristics of the 32 patients are shown in Table 1. Dual chamber (DDD) pacing was used in 12 patients and ventricular (VDD) pacing in 20 patients. All patients after dual-chamber sequential pacing were in New York Heart Association (NYHA) class I or П. On the other hand, all patients after cardiac resynchronization therapy (CRT) were in NYHA class III. Left ventricular ejection fraction (LVEF) was 55 ± 15 %.

Table 1 Clinical characteristics
Full size table

AV optimization using our current method could be performed in 27 of 32 patients (84 %). In the remaining 5 patients, adequate PV flow signal could not be recorded. The measurements made in all patients are summarized in Tables 2 and 3. Mean AVDOPT was 143 ± 35 ms. As expected, the mean AVDOPT was significantly lower in the VDD than in the DDD mode (133 ± 32 ms vs. 170 ± 37 ms, P = 0.014). SV at AVDOPT was significantly higher than shorter or longer AV delay (63 ± 18 ml vs. 57 ± 15 ml vs. 56 ± 16 ml, P = 0.001) (Fig. 2).

Table 2 Hemodynamic and Doppler echocardiography parameters
Full size table
Table 3 Pacing mode, pacing rate at initial enrollment and TMF A, PVa duration pre and post AV delay optimization
Full size table
Fig. 2
figure 2

Comparison of SV in all patients. SV obtained at the AVDOPT was compared with SV obtained at shorter (AVDOPT − 50 ms) or longer (AVDOPT + 50 ms) AV delays

Full size image

In a subset of patients after sequential dual-chamber pacing for complete AV block, AV optimization could be performed in 22 of 26 patients (85 %). The AVDOPT in VDD mode was 128 ± 38 ms and the AVDOPT in the DDD pacing mode was 177 ± 39 ms. SV with AVDOPT was significantly higher than shorter or longer AV delay (64 ± 19 ml vs. 57 ± 16 ml vs. 56 ± 17 ml, P = 0.0001) (Fig. 3).

Fig. 3
figure 3

Comparison of SV in patients after sequential, dual chamber pacing for complete AV block. SV obtained at the AVDOPT was compared with SV obtained at shorter (AVDOPT − 50 ms) or longer (AVDOPT + 50 ms) AV delays

Full size image

Similarly, in a subset of patients after CRT, AV optimization could be performed in 5 of 6 patients (83 %). The AVDOPT in VDD mode was 128 ± 38 ms and the AVDOPT in the DDD pacing mode was 177 ± 39 ms. SV with AVDOPT was significantly higher than shorter or longer AV delay (61 ± 13 ml vs. 53 ± 11 ml vs. 57 ± 10 ml, P = 0.026) (Fig. 4).

Fig. 4
figure 4

Comparison of SV in CRT recipients. SV obtained at the AVDOPT was compared with SV obtained at shorter (AVDOPT − 50 ms) or longer (AVDOPT + 50 ms) AV delays

Full size image

Reproducibility of PV flow measurements was analyzed. Correlation coefficients were high for repeated measurements by the same observer (r = 0.98 for duration of PVa minus A) and measurements by 2 different observers (r = 0.88 for duration of PVa minus A).

Discussion

This study shows that AV delay optimization based on a new echo-Doppler method using TMF and PV flow is feasible. In addition, increased SV during AV delay optimized by this method may suggest a potential favorable impact on cardiac function and possibly prognosis.

A previous randomized, prospective study comparing echo-guided AV delay optimization and an empiric, fixed AV delay of 120 ms demonstrated improved clinical outcome at 3 months in patients with echo-guided AV optimization [19]. In their study, optimal AV delay was defined as the largest aortic velocity–time integral at one of eight tested AV intervals (between 60 and 200 ms). On the other hand, a more recent large-scale randomized prospective multicenter trial (SMART-AV trial) to compare between a fixed empirical AV delay (120 ms), echocardiographically optimized AV delay, and AV delay optimized with SmartDelay electrocardiogram-based algorithm did not show superiority of echocardiography or SmartDelay over a fixed AV delay of 120 ms [18]. In their study, Ritter’s method [10, 22] and/or an iterative method [23] using TMF were used to optimize AV delay as endorsed by the American Society of Echocardiography [23, 24]. Based on their negative results, the authors stated that routine echocardiographic AV optimization using the American Society of Echocardiography recommended method for patients with CRT should be abandoned [18]. However, it is not certain whether all echo-Doppler methods should be abolished.

Ritter et al. [22] first reported an echo-Doppler method to optimize AV delay in patients with complete AV block and a normal LV systolic function. Ritter et al. defined the AV delay with the echo method that provided the longest diastolic filling time without interruption of the A wave. Ritter’s formula, which can be regarded as the current “gold standard” in AV delay optimization [24] requires 2 measurements: (1) QA short = the time interval between the onset of the Q wave and the end of the truncated “A” wave of the TMF at a short (30–60 ms) AV delay; and (2) QA long = the time interval between the onset of the Q wave and the end of the “A” wave of the TMF at a long (200 ms) AV interval. According to the formula, optimal AV delay was calculated as AV long − (QA short − QA long). This method has been used in several clinical trials because it is a simple, non-invasive and reproducible method [20]. On the other hand, Ishikawa et al. used diastolic mitral regurgitation to optimize AV delay. As compared with Ritter’s method in which AV delay was optimized to achieve the highest cardiac output, Ishikawa’s method is to achieve the lowest possible left atrial or LV filling pressure [9, 25]. In our present study, we used both TMF and PV flow to achieve the lowest LV filling pressure and the highest SV.

The concept of Doppler assessment of LV filling pressure using both TMF and PV flow was first reported in 1993 by Rossvoll and Hatle [21]. The difference in duration between PVa of the PV flow and antegrade A wave by the TMF was positively and strongly correlated with LV end-diastolic pressure (r = 0.68, P < 0.001). A longer duration of PVa versus A wave predicted increased (>15 mmHg) LV end-diastolic pressure [21]. The mechanisms for a longer duration of PVa than the A wave was explained by the increased LV end-diastolic pressure as a result of reduced LV compliance. Therefore, an AV delay that does not prolong PVa more than the A wave could be considered as a hemodynamically optimal AV delay. Although our preliminary data suggest that AV optimization based on the TMF and PV flow is feasible, it was not possible for AV optimization to be performed in some patients in whom PV flow could not be detected. This is a possible limitation of this study. Detection of the PV flow signal using the transthoracic approach depends upon the image quality of the echo-Doppler machine. Although the sensitivity of the Doppler measurements for some specific conditions was not sufficient when using old echo-Doppler machines and therefore required contrast enhancement [26, 27], modern echo-Doppler machines have sufficiently sensitive Doppler equipment [28, 29]. Another apparent limitation is that 2 different Doppler measurements are required for our method which appears to be time consuming. However, as compared with Ritter’s method, which requires 2 TMF recordings at 2 different AV delay settings, our method is less time consuming.

Because this is a small pilot study, further investigations will be necessary. First, the hemodynamically favorable acute results should be confirmed by invasive hemodynamic monitoring. Second, the long-term clinical impact of the acute results should be investigated by a serial observation of the study population. Finally, the advantages of the current method should be investigated by comparing it with other echo-Doppler methods or empirical fixed AV delay prospectively.

Conclusions

A novel AV delay optimization method using TMF and PV flow has been shown to be feasible. The usefulness of this method requires further investigation with a larger study population.