11 chronic heart failure patients of our clinic were included in this study. All patients had a biventricular ICD (pre-implantation NYHA III-IV, EF < 35%, QRS width > 120 ms). Clinical characteristics are demonstrated in Table 1. Patient exclusion criteria were as follows: atrial fibrillation, pacemaker malfunction and oesophageal diseases, NYHA IV, prosthetic mitral valve replacement.

For the AV delay optimization we used the method proposed by Ismer et al 22.

This approach needs the placement of a bi-polar oesophageal electrode to provide a filtered left-atrial electrogram (LAE). We applied a 5F oesophagus electrode (Osypka TO2/5F, order no. TA12991101, Rheinfelden, Germany). Filtered oesophageal electrogram and telemetric real-time pacemaker markers provided by the programmer's analogue output were superimposed on the display of transmitral flow velocity on the Doppler-echo system (Figure 1). The simultaneous recording of transmitral flow, the left atrial oesophageal electrogram and the real-time sense-event markers, allow determining the components of the optimal AV delay (Table 2, Figure 1 and 2).

Based on these measurements, optimal AV delays were calculated for VDD (atrial-triggered) and DDD (atrial-paced) mode using the equations:

AVD_OPT VDD = MA-LA + LA-EAC_long - Sv-EAC_short

and

AVD_OPT DDD = SA-LA + LA-EAC_long - Sv-EAC_short

Echocardiography to assess dyssynchrony was performed subsequently under three pacemaker settings: optimal AV delay (AVD_OPT), optimal AV delay minus 50 ms (AVD_OPT-50), optimal AV delay plus 50 ms (AVD_OPT+50).

Echocardiography was performed on the Vivid 5 and Vivid 7 Dimension (GE Vingmed Ultrasound, Horton, Norway) machines. The TDI and strain analysis were performed in an off-line work station. The LVEF was assessed by area-length method in the apical four chamber view. The CI and the SV were calculated from the systolic velocities measured by PW-Doppler in the aortic outflow tract. Strain rate, tissue Doppler velocities were measured in the basal segments of the apical four-, three- and two-chamber views.

Values are expressed as mean ± standard deviation (SD). Groups were compared by parametric or non-parametric tests (t-tests and Wilcoxon-Mann-Whitney tests, respectively). Statistical significance was assumed at a value of P < 0.05. Statistical analysis was performed with the SPSS 12 software package (SPSS; Chicago, Ill, USA).

In all patients, we could define an optimal AV delay in the VDD and the DDD modes respectively. The AVD_OPT in VDD mode was 105.5 ± 38.1 ms and the AVD_OPT in the DDD pacing mode was 186.9 ± 52.9 ms. The results are summarized in Table 3. As expected, the mean optimal AV delay was lower in the VDD than in the DDD mode.

Echocardiography was performed subsequently under three pacemaker settings: AVD_OPT, AVD_OPT-50, AVD_OPT+50. All patients had continuous biventricular stimulation even under AVD_OPT+50.

The LVEF with AVD_OPT was 28% (± 12%), with an AVD_OPT-50 20% (± 7%, p= 0.03 compared to AVD_OPT), with an AVD_OPT+50 23% (± 7%, p = 0.11 compared to AVD_OPT). The heart rate did not change significantly in the different settings (AVD_OPT: 65,4/min, AVD_OPT-50: 65,6/min, AVD_OPT+50: 65,8 ms). The hemodynamics (SVI, CI, LVEF) and the TDI derived data are listed in Table 4. There was no significant difference of the amount of segments with dyssynchrony in TSI in the three settings. The maximal delay in the basal segments in the apical two-, three- and four-chamber views measured by TSI and strain did not differ in the AVD_OPT, AVD_OPT+50 and AVD_OPT-50 setting.

To date, Ismer's method for the optimal AV delay was applied to patients with DDD pacemakers and normal left ventricular function 2223. This is the first study to assess the optimal AV delay by Ismer's method in patients with reduced left ventricular function. In our CRT patients, an optimal AV delay according to Ismer's method could be defined. This is the only method that allows separate measurement of the three AV-delay components: i.e., the pacemaker-related interatrial conduction time, the left-atrial electromechanical action, and the left-ventricular latency period. The benefits of this method, however, are offset by the necessity for placement of an oesophageal electrode. This requirement explains why only a few medical centres have applied this method in clinical practice and in most cases for purposes of scientific investigation only.

Our results concerning the AVD_OPT in the VDD mode (105.5 ± 38.1 ms) are in agreement with the results of other studies on AVD_OPT in CRT patients: Butter 16 determined an AVD_OPT of 100 ms in 30 patients, Auricchio 17 an AVD_OPT of 112 ± 33 ms in 41 patients and Kass 18 an AVD_OPT of 125 ± 49 ms. A study that was recently published by Porciani 29 found an AVD_OPT during simultaneous biventricular pacing of 97 + 27 ms.

In the literature, there are no published data on AVD_OPT in DDD mode. Therefore, our AVD_OPT in DDD mode of 186.9 ± 52.9 ms cannot be compared to other studies.

Intra-individually, the patients had the best LVEF under optimal AV-delay compared to the +50 and -50 ms settings. The LVEF is significantly higher in the AVD_OPT setting than in the AVD_OPT -50 setting. Obviously the formation of "cannon waves" seen with a shorter AV interval (AVD_OPT -50) had a more negative hemodynamic effect than the diastolic mitral regurgitation seen with longer AV delays (AVD_OPT +50). The hemodynamically unfavourable effects of "cannon waves" are described since the beginning of pacemaker therapy and are also termed "pacemaker syndrome". It is generally accepted that an adequate pacemaker programming can avoid this 30. Toda et al. 31 could show in his studies that the mean LVEF in AVD_OPT is higher than in prolonged AV delays. However, he found no significant difference.

Changes of dyssynchrony can be seen immediately, as seen in studies that have examined on/off comparisons in CRT patients 32. However, an optimized AV interval does not change the markers of dyssynchrony. The reason for the improved hemodynamic situation under AVD_OPT seems to be the better left ventricular filling and not the altered dyssynchrony.