Comparative assessment of hemodynamic changes and outcomes in ventilator weaning

Mechanical ventilation is a vital parameter in intensive care medicine. It has served as a bridging therapy for compromised patients to alleviate unneeded pulmonary exhaustion and allowed for smooth neurological, cardiac, and pulmonary recovery with minimal complications [1]. However, the disadvantages, including the hemodynamic effects of mechanical ventilation, have been studied thoroughly, and the emphasis on preload, afterload, and contractility affection by positive pressure ventilation (PPV) can't be overstated [2,3].

Mechanical ventilation can be delivered through many interfaces, whether invasive or non-invasive [4,5]. Controlled modes of mechanical ventilation occur when the ventilator administers a predetermined volume or pressure, irrespective of the patient's inspiratory efforts. Assist-control ventilation suits individuals who are not deeply sedated and can initiate breathing but not sustain spontaneous respiration [6]. However, controlled modes are associated with fixed volumes and/or pressures, which can increase intrathoracic pressure (ITP) and impede venous return and, thus, cardiac output. Also, patient-ventilator desynchrony is a significant side effect [7].

Variable pressure support ventilation (VPS) is a form of assisted ventilation that adjusts the level of pressure support on a breath-by-breath basis to re-establish the physiological variability of respiratory activity [8]. VPS emulates the delicate diversity of regular respiration by producing random fluctuations in inspiratory pressure. The primary potential outcomes of this mild modification are enhanced pulmonary function and a diminished risk of ventilator-associated lung injury [9]. Patients also feel more comfortable with minimal adjustments to the ventilator settings: more homogenous ventilation, surfactant production, and pulmonary perfusion without adverse hemodynamic effects of elevated intrathoracic pressures [7].

Transesophageal echocardiography (TEE) has emerged as a valuable diagnostic and monitoring instrument in critical care environments, mainly where transthoracic echocardiography is challenging to execute [10]. It offers essential information concerning mechanical ventilation in specific populations, such as obese individuals and those with surgical dressings and chest tubes. TEE can also be performed in a prone position [11,12]. It also enables visualization of the superior vena cava, which can aid in predicting fluid response in mechanically ventilated patients [7].

Transthoracic echocardiography (TTE) is among the most frequently conducted cardiac assessments. It offers extensive information regarding heart structure and function, aiding diagnosis and therapeutic guidance, and is no longer exclusive to specialist cardiology departments [13]. While TTE is mainly utilized to investigate heart function, its applications in hemodynamic evaluations are gaining popularity. Estimates of cardiac output, proper atrial pressure, and assessments of patients' fluid status are prevalent applications of the TTE [14].

This study will use TEE and TTE to compare the hemodynamic profiles of VPS and Pressure Support (PS) modes to those of the Controlled mode of mechanical ventilation, volume-controlled ventilation (VCV), and compare weaning outcomes and survival in both.

The mean age of the enrolled patients was 55.55±28.7 years, where age showed a statistical difference between both arms of patients with a mean value of 62.2 in arm-A and 48.9 in arm-B with a p-value of 0.028. As shown in Table 1, males were predominant, representing 65% of the included individuals, while 35% were females. There was no statistical significance regarding gender between the two arms (p-value=1). The prevalence of comorbidities is represented in Table 1; only Diabetes Mellitus showed a statistical significance between arm-A and arm-B with a p-value of 0.047.

Throughout the hemodynamic assessment (using TTE and TEE), both vital signs, heart rate, and mean arterial pressure changes didn't show statistical differences between the two arms. For the echocardiographic data, the left ventricular outflow tract (LVOT) diameter didn't show any statistical significance between the two arms, with a mean value of 2.05 cm in arm-A and 1.93 cm in arm-B with a p-value of 0.2. Comparing the two arms in terms of hemodynamic parameters under VCV (stroke volume SV, velocity time integral VTI, and cardiac output CO) measured by TTE and TEE, there was no statistically significant difference, as shown in Table 1.

The inter-statistical comparison between the three modes revealed that in arm-A, there was statistical significance between VCV and VPS modes with drops in VTI, SV, and CO with p-values of TEE VTI, and TEE SV were 0.02 and 0.03. TTE VTI and TTE SV were 0.04,0.05, respectively. CO drop was also significant, with a p-value of 0.01 (Table 2). Statistical analysis depicted a decline of 13% in mean TEE and mean TTE VTI when the change from VCV to VPS was done, which translated to subsequent drops in mean SV and mean CO by 12.1 % and 13%, respectively (Figure 1).

Fig. 1.

Intercomparison of hemodynamic parameters between the VTI, SV, and CO. CO: cardiac output; SV: stroke volume; TEE: transesophageal echocardiography; TTE: transthoracic echocardiography; VCV: volume-controlled ventilation; VPS: variable pressure support; VTI: velocity time integral

In group B, there was statistical significance between VCV and PS modes with drops in VTI, SV, and CO with p-values of TEE VTI and TEE SV were 0.04 and 0.03. TTE VTI and TTE SV were 0.05,0.02, respectively. CO drop was also significant, with a p-value of 0.04 (Table 2). Statistical analysis depicted a decrease of 6% in mean TEE and mean TTE VTI when the change from VCV to PS was done, which translated to subsequent drops in mean SV and mean CO by 6 % and 3.8%, respectively (Figure 1).

As shown in Table 3, comparing the use of TTE and TEE for the detection of VTI in both groups, there wasn't any statistical significance between using TTE or TEE in detecting VTI with a p-value in VCV, PS, and VPS of 0.313, 0.202, and 0.108, respectively.

For the hemodynamic assessment, the patients were randomly assigned to arm A (left to continue weaning in VPS mode) and arm B (left to continue weaning in PS mode). The analysis revealed no statistical significance between both groups regarding ventilator days (p-value of 0.88). Weaning trial success was also statistically insignificant, showing a p-value of 0.525. Likewise, the length of stay and survival showed no statistically significant difference between the two arms, as shown in Table 4.

Mechanical ventilation provided for each patient can be done in various modes, whether controlled/assisted or spontaneous, and these three general outlines are further subdivided into a diverse panel of modes, each with an advantage to suit a particular patient [26,27]. An assessment of the disadvantages of the different modes that support the compromised state of the typical ICU patient is needed. This study was conducted on 40 mechanically ventilated patients who met the inclusion criteria of this study and then further subdivided into two arms: arm A comprised of patients receiving VPS, and arm B comprised of patients receiving PS. This was done by TTE and TEE of the patients, determining LVOT and VTI to calculate CO.

In our study, a notable difference in baseline characteristics between the treatment arms, particularly in age and the prevalence of diabetes, was observed. The average age of participants in the VPS group was significantly higher than in the control group (mean age 62.2 vs. 48.9 years, p < 0.05), and a greater proportion of participants in the VPS group had a diagnosis of diabetes (50% vs. 20%, p < 0.05). These differences are critical to consider as they may influence the outcomes of interest, including hemodynamic parameters, ventilator days, and overall clinical recovery. There was no statistical significance regarding gender between the two arms (p-value=1). The prevalence of comorbidities is represented in Table 1.

Age is a well-established factor that can affect physiological responses and recovery trajectories in critically ill patients. Older adults often exhibit altered hemodynamic responses and may have a higher prevalence of comorbid conditions that can complicate their clinical course [28,29]. The significant age difference between the groups may contribute to the observed variability in outcomes, particularly in the VPS group, which exhibited greater declines in VTI, SV, and CO. These hemodynamic changes, while statistically significant, didn't affect the clinical relevance in older patients who have different baseline physiological reserves compared to younger patients [30]. The presence of diabetes is another critical factor that may affect recovery and outcomes. Diabetes can lead to microvascular complications and impaired hemodynamic regulation, potentially exacerbating the impact of hemodynamic changes on clinical outcomes [31]. The interaction between diabetes and hemodynamic stability may warrant further investigation [32], as it could explain some of the differences in clinical outcomes observed between the groups. To address these baseline differences, we employed multivariable regression analysis to adjust for age and diabetes status in our primary outcome assessments.

In our study, we found out that there was a drop in VTI (which affected SV and CO) when changing from VCV to VPS mode in the majority of arm A patients, with a significant statistical p-value of 0.02 and was measured to be a 13% drop in LVOT VTI. This agrees with a study by Carlos et al. [33], which showed that PPV can harm the circulatory system, whether on the extremely sensitive right ventricle, a borderline function left ventricle, or in a preload-dependent state [33]. Likewise, Lai et al. [34] deduced that PEEP levels had a deleterious impact when increased in the circulatory system. Also, Vignon et al. [35] revealed that TEE is a valuable, well-tolerated imaging technique used in mechanically ventilated patients to assess left ventricular systolic function in the presence of PEEP [35].

This was controversial to Naik et al. [36], who revealed that VPS (or as alternately named noisy PS) had no adverse effects on hemodynamics [36]. Furthermore, Spieth et al. [37] claimed that VPS mode had no significant adverse hemodynamic effects compared to the controlled mode [37]. Discrepancies between this study and Spieth et al. could be attributed to one of three reasons. Firstly, in our study, maximum variability was adjusted with a range of pressures from 10 cmH2O to 20 cmH2O, which provided a wide range of pressures, some of which can generate supra-normal volumes [37]. Secondly, the baseline volume status differs from patient to patient and could be affected by multiple constantly changing factors. Finally, patients experienced dyspnea when switched from controlled modes to spontaneous ones, which could negatively affect certain patients, especially those with ischemic heart disease, and those constituted 17.5% of the total study population.

This research detected a less significant drop in most arm B patients when transitioning from VCV to PS mode, with a 6% drop in LVOT VTI and a statistically significant p-value of 0.04. These findings were similar to those of Mauri et al. [38], who reported that there is a potential risk of double and reverse triggering during the spontaneous modes of mechanical ventilation, which could lead to the delivery of non-protective volumes to the patient with a massive swing in intrathoracic pressures with a direct effect on the right ventricle [38]. Similarly, Frazier et al. [39] found a drop in MAP and CO when patients transitioned from controlled to spontaneous mechanical ventilation [39].

On the contrary, EL Gazzar et al. [40] showed a statistically significant reduction in diaphragmatic performance in patients who received controlled mechanical ventilation alone. Conversely, diaphragmatic performance improved when pressure support ventilation (PSV) was combined with cytomegaly virus, with no significant alterations in diaphragmatic mobility parameters. No significant link was observed between echocardiographic measures (left ventricular ejection fraction, right ventricular size, tricuspid annular plane systolic excursion, and proper ventricular systolic pressure) and various mechanical ventilation [40].

No consensus has been reached regarding using TTE or TEE to assess hemodynamics during PPV. For instance, in 2019, Tongyoo et al. used TTE for the evaluation of hemodynamics [41], while back in 1993, Poelaert et al. [42] used TEE in the hemodynamic assessment of mechanically ventilated patients [42]. Further Oh. et al. studied critically ill patients using TEE to assess intravascular volume [43]. The authors found that the predominant indication, observed in 25 individuals (49%), was unexplained hemodynamic instability. TEE indicated cardiovascular issues in 30 patients (59%) that were not distinctly diagnosed by TTE. Meanwhile, TEE enabled the definitive exclusion of suspected anomalies due to its enhanced imaging capabilities. Cardiac surgery was necessitated by TEE results in 12 individuals (24%), all of which were corroborated during the operation [43].

In this study, the authors primarily chose to use TEE due to the presence of patients with poor TTE windows, as in COPD patients, which constituted 12.5% of the patients. Furthermore, obese patients and others with skeletal deformities also had obscure echocardiographic windows and bad images. TEE offered a more detailed and less interpatient variability of echo images. Nevertheless, TTE and TEE didn't show any statistically significant values with almost identical p-values in VCV, PS, and VPS of 0.313, 0.202, and 0.108, respectively. In alignment with this study, Si et al. [44] concluded that TEE alone or in combination with TTE could provide valuable information for diagnosis, which may bring significant therapeutic benefits, as 66% of their patients had the same findings in both TTE and TEE with no change in therapy [44].

After hemodynamic assessment, the patients in arm A(VPS) and arm B (PS) were left for weaning in their respective modes. Both modes didn't show a statistically significant p-value regarding days on mechanical ventilation and liberation success, with p-values of 0.88 and 0.525, respectively. Our results are concordant with those of Mauri et al. [45], who revealed that VPS led to even and homogenous air distribution and was equally effective as PS in weaning [45]. Moreover, Wysocki et al. [46] have deducted that in ICU patients undergoing a spontaneous breathing trial, breathing variability is more significant in patients successfully separated from the ventilator and the endotracheal tube [46]. Furthermore, Kiss et al. [47] revealed that VPS was equally effective in weaning as PS mode [47]. However, our findings were discordant with those of Rolland-Debord et al. [48], who claimed that higher breathing variability and lower complexity were associated with higher survival and lower duration of mechanical ventilation [48].

Moreover, there was no apparent difference between the two groups in terms of total days of ICU and overall survival, with p-values of 0.651 and 0.525, respectively. Agreeing with our data, a currently ongoing large multicenter trial examining VPS vs. PS in terms of weaning duration, named the EVA trial (ClinicalTrials.gov Identifier: NCT00786292, [49]) proposed preliminary results that variable compared to conventional PSV didn't increase discomfort or deteriorate the cardiopulmonary function in mechanically ventilated patients in the ICU [49].

In fact, variable PS significantly reduced the work of breathing and increased comfort in some patients. Therefore, the burden and risks to patients resulting from the intervention are low, and patients assigned to variable PS will be weaned faster from the mechanical ventilator than those assigned to conventional PSV. Finally, more trials need to be conducted to further evaluate VPS in terms of more accurate hemodynamic monitoring tools, aeration, patient comfort, desynchrony, and the overall outcome of mechanical ventilation compared to PS.

Our study found significant hemodynamic effects when transitioning from VCV mode to VPS and standard PS mode, primarily resulting in adverse effects for most patients. Although there was a greater drop in CO among VPS patients compared to those on standard PS, this did not have practical implications for vital signs, particularly mean arterial pressure, or improved clinical outcomes. Additionally, this study concluded that both TTE and TEE typically yield similar results, allowing them to be used interchangeably. Our findings indicate that VPS is as effective as traditional pressure support (PS) in facilitating successful liberation from mechanical ventilation, with no substantial differences in ventilator duration or overall survival rates. The observed hemodynamic stability during VPS, despite some fluctuations, underscores its potential benefits in enhancing patient comfort and reducing the work of breathing. VPS is significantly relevant to the weaning process for mechanically ventilated patients. Lastly, VPS and standard PS led to comparable outcomes regarding ventilator days, liberation success, ICU days, and overall mortality. While more patients were liberated from mechanical ventilation and survived their ICU stay, these differences were not statistically significant. Further studies and trials with large sample sizes are necessary before adopting VPS as a standard method for weaning in conjunction with PSV with an emphasis on more patient-centered endpoints (validated comfort scores).