DISCUSSION
The relevant results of the present study were:
(1) The relevance of the ramp test during follow up for the hemodynamic optimization in patients implanted with Left ventricular support.
(2) Right ventricle function remains a limiting factor in patients with LVAD.
(3) The current parameters usually adopted for setting the RPM are not exhaustive to make the patient’s hemodynamic profile optimal.
(4) This newly designed tool for hemodynamic optimization that we called hemodynamic index, can support clinicians to easily identify the hemodynamic profile of individual patients and to optimize medical treatment aiming to prevent events of right failure and related rehospitalizations. Based on this model, we recommend maintaining an HI above 60, during pump speed optimization, to preserve good right ventricle function while ensuring optimal left ventricle unloading. HI could provide a useful guide during ramp test, avoiding an excessive increase in rpm.
Ramp testing during right heart catheterization to optimize the RPM is recommended in the guidelines of the Society of Heart and Lung Transplantation to optimize the rpm of the device during the postoperative course. The current guidelines encourage the use of the echocardiography as an integral part in determining that adequate unloading of the left ventricle is obtained while maintaining central positioning of the septum and minimal mitral regurgitation (class I recommendation)[8]. Setting the RPM to ensure intermittent opening of the aortic valve is currently a class IIb recommendation, to prevent the development of aortic regurgitation or aortic leaflets fusion. These recommendations are somewhat vague and not standardized, and the application of right heart catheterization in conditions of clinical stability is not well defined. Right heart catheterization is currently recommended (class I) to cope with specific situations, such as when symptoms of heart failure occur, for evaluation of pulmonary hypertension in patients eligible for heart transplantation and in the event of right ventricular failure. When the explant of LVAD is planned, the hemodynamic evaluation is also recommended (class IIa) to obtain more data confirming myocardial recovery. Diagnosis of LVAD outflow obstruction or suspected device thrombosis is another indication for performing the ramp test during right heart catheterization [9]. Uriel et al. observed that an LV end-diastolic dimension slope less than an absolute value of 0,16 during a ramp test is a strong predictor of thrombosis in HeartMate II patients, and evidence not confirmed in patients implanted with HVAD [10-11]. There is no broad consensus on the routine use of right heart cathetherization, and it is currently not recommended in the guidelines. Uriel et al. highlighted in their study that many patients have abnormal hemodynamic profiles at the set RPM level, despite no signs or symptoms of heart failure were identified [12]. Suwa et al. have recently demonstrated that 57% of clinically stable patients had a significant increase in CVP and PCWP at baseline LVAD speed [13]. These findings were also confirmed by our study, which estimated a PCWP value greater than 12 mmHg in 30 % of patients. In both groups, we achieved a significant reduction in PCWP after speed optimization. As highlighted by Table 2, the pre- and post- speed optimization PCWP values were respectively 17.15 ± 4.93 mmHg and 12.55 ± 2.21 mmHg (p <0.001) respectively for the RVF- group. The PCWP values ​​pre and post speed optimization in RVF + patients were 14.75 ± 3.46 mmHg and 10.16 ± 2.51 mmHg (p <0.001) respectively. Although PF and CO increases were achieved in both patient groups by improving the LVAD speed, the CO increased significantly only in the RVF - group. This finding suggests that current approaches for setting the optimal LVAD speed are insufficient and that a hemodynamic evaluation provides important additional information about the clinical condition of the patient [14]. The hemodynamic profile obtained with ramp tests during right catheterization might be used to better tailor drug therapy so that a better hemodynamic profile and a better quality of life can be achieved. Although further studies are needed, Jung et al. have highlighted how changes in hemodynamic parameters during RPM changes can have important clinical implications. Jung pointed out that a reduction in PCWP during ramp tests in patients implanted with HM II correlated with a lower NYHA class and that an increase in CO was related to a better quality of life [15]. In addiction, data are emerging concerning the effect of pump speed optimization on long-term outcomes. Sarswat et al. conducted a two-year prospective observational study of 62 LVAD patients after performing hemodynamic optimization by invasive ramp test. The rate of hospital readmission was lower in patients with an optimized hemodynamic profile when compared with patients in whom the hemodynamic condition had not been optimized. Couperus also highlighted the effectiveness of speed optimization on the function of the right ventricle [16]. Our results showed the ramp test assists in the unloading of the left ventricle but was not so effective in improving the right heart function. A significant reduction of CVP was observed only in RVF- (Tab 2). If we consider that only in RVF - patients there was a significant increase in CO, we can conclude that despite optimal unloading of the left ventricle, hemodynamic optimization is substantially conditioned by the residual intrinsic function of the right ventricle. Despite the results obtained by Couperus and Coll, the correlation between hemodynamic optimization and a reduced incidence of right ventricular failure and better survival remains to be validated. An important aspect to highlight is that many studies focus on the incidence of early right failure with minimal focus on late right failure, which however represents an important complication during support with LVAD. As described by Burke et al., late RVF could be mostly related to intrinsic myocardial function, or it could be secondary to various etiologies such as ventricular arrhythmia, the progression of tricuspid insufficiency and pulmonary hypertension [17]. Identifying these risk factors for development of late RVF is clinically very relevant since planning biventricular support can result in a better outcome, especially for high-risk BTT patients [18]. Takeda et al. found that comparing the non-RHF and RHF group, similar hemodynamic values were found, including CVP and CVP / PCWP ratio [19]. These variables are commonly representative markers of intrinsic right ventricular dysfunction. Similarly Kormos et al. reported CVP and CVP/PCWP ratio values similar in those patients who did not develop RHF when compared to patients who developed late RHF. on the other hand, patients developing early RHF showed significantly higher CVP and CVP/PCWP ratio values when compared with data from the non-RHF population [20]. From these results, it emerges that the preoperative hemodynamic evaluation is not sensitive enough to identify patients at risk of late RVF after LVAD implantation. Once the right ventricle adapts to the new physiological state guaranteed by the continuous flow pump support, other factors like the intrinsic right ventricular dysfunction can result in right hear failure. Hence the purpose of our study was to develop a new hemodynamic index calculated on hemodynamic parameters obtained after speed optimization at follow-up. Since the normal parameters such as CVP and CVP / wedge, as previously anticipated, are not adequate to identify patients at risk of RVF, we developed a parameter that integrates the filling pressures of the right and left sections, the afterload represented by the MAP, and the ratio between the optimized RPM and the maximum RPM available. The latter parameter was intended as a correction factor and expressed the support level of the device. The main purpose of this hemodynamic index is to identify a reference parameter to guide the RPM setting during ramp tests in order to optimally balance all the variables involved, thus reducing the risk of right failure. Suboptimal unloading of the left ventricle and right dysfunction are the major determinants of long-term mortality in patients with LVAD [21-22]. The LVAD pump speed can influence both factors. A high LVAD speed increases the unloading of the left ventricle and increases the cardiac output and exercise capacity [23]. However, elevated speed has been associated with aortic valve dysfunction. In particular, a reduction in the opening of the valve due to the increased pump speed results in the fusion of the valve leaflets, thrombus formation and valve insufficiency which reduces survival in patients with long-term LVAD support [24-25]. LVAD speed also has an important effect on the function of the right ventricle, the degree of unloading of the VS, pulmonary arterial pressures, and the right ventricle afterload. A higher level of pump speed might improve right function by reducing the afterload on the right ventricle. However, excessive LVAD speed can also compromise right ventricular function because of increased preload, the leftward shift of the interventricular septum, and modification of the right ventricle’s geometry. Therefore, the optimal hemodynamic balance during support with LVAD is the result of several variables that interact with each other. When considered individually, these variables fail to precisely identify a patient’s hemodynamic profile and showed poor correlation with the incidence of right ventricular failure. The hemodynamic balance of patients with continuous-flow LVAD support is much more complex, and more than a single variable should be considered. The HI attempts to summarize the effects of the different variables by combining them into a single parameter that could represent a reference for the hemodynamic optimization of LVAD patients. By retrospectively evaluating our 38 patients,  a HI of 51,66 ± 5,28 was found in those who required re-hospitalization with the diagnosis of late RVF and resulted significantly lower when compared with the group of patients who had not experienced a right failure episode, HI 80,10 ± 13,45 (p<0,001). In the RVF + group, the lower HI, according to our formula, was the result of a lower PCWP / CVP ratio, a reduced MAP, and a lower ratio between the RPM set/ RPM max. The reduced ratio between PCWP / CVP at a reduced RPM / RPM max ratio is due to a poor function of the right ventricle, illustrating a clear relationship between this hemodynamic feature and high risk of right failure. When a low hemodynamic index was found in conjunction with a low MAP, then the right dysfunction could have been explained by excessive unloading of the left ventricle and significant leftward shift of the septum and modified geometry of the right ventricle. In this case, attention should be paid to reduce antihypertensive therapy to restore normal afterload. According to our results, a ROC analysis identified a cut off level for HI of 55, a level significantly predictive of late right failure.
Our study shows that this new hemodynamic index can be used for various purposes:
1) Hemodynamic and pump speed optimization by aiming for a minimum HI of at least 55.
2) Identification of different hemodynamic profiles with stratification of patients according to the risk of late RVF. In these patients, a different strategy should be adopted by anticipating the inclusion in the transplant list or proceed with an emergency transplant.
3) Medical therapy optimization in patients with a very high HI. A very high HI is an expression of high PAM, a high value of the PCWP / CVP ratio, and high RPM set/max RPM ratio, indicative of poor left ventricular unloading. In such patients, an increase in antihypertensive therapy to reduce the afterload of the ventricle is certainly recommended (see algorithm Fig 4).