European Journal of Cardiovascular Medicine

Extracorporeal membrane oxygenation (ECMO) Veno-Venous is for supporting and replacing pulmonary function in acute respiratory failure not responding to conventional measures. The Extracorporeal Life Support Organization reported an increasing trend in ECMO usage as a rescue therapy with an overall survival of 58%¹. The prevalence of ARDS is high worldwide. According to data from the seminal LUNG SAFE (Large Observational Study to Understand the Global Impact of Severe Acute Respiratory Failure) study, which involved intensive care units (ICUs) in 50 countries, ARDS is responsible for 23.4% of patients who require invasive mechanical ventilation and 10.4% of all ICU admissions. These numbers highlight how common the syndrome is and how it affects critical care units all over the world on a daily basis. Although incidence estimates differ by location, they always indicate a serious public health concern. An incidence of 27.6 cases per 100,000 patient-years was found in a prospective Canadian study, whereas the incidence in the United States is estimated to be between 64 and 78 cases per 100,000 person-years.²

Despite advances in critical care, including the widespread adoption of lung-protective ventilation strategies, mortality from ARDS remains unacceptably high. A key determinant of outcome is the severity of hypoxemia at presentation, as categorized by the Berlin definition. Multiple large-scale studies and systematic reviews published in the 2020-2025 timeframe consistently report in-hospital mortality rates for severe ARDS ranging from 40% to 60%.³ We have reasonable guidelines for the VA ECMO weaning process, but there is no published report to outline a standard approach for weaning of VV ECMO. The process requires a thorough understanding of ventilator management and ECMO physiology. Weaning involves balancing the risk of premature decannulation, which could lead to catastrophic respiratory collapse, with the potential for prolonged ECMO-related complications.⁴ With strong evidence to support its use in high-volume centres, veno-venous extracorporeal membrane oxygenation (VV-ECMO) has emerged as a key component of rescue therapy for severe acute respiratory distress syndrome (ARDS).  The best time and techniques for weaning patients off the circuit, however, are still poorly understood and not well standardised, despite the fact that the reasons for starting VV-ECMO are becoming more and more evidence-based. There is a substantial evidence gap at a critical stage of patient treatment since current weaning methods are mostly dependent on expert opinion and clinician preference rather than data from thorough clinical trials.⁵

Continuing ECMO therapy in a resource-limited setting to near-complete physiological recovery causes exuberant cost and resource utilisation. The primary objective of weaning is to confirm that the native lung has recovered sufficiently to sustain adequate gas exchange without causing lung damage from patient effort or a ventilator mechanism while weaning off ECMO. This assessment requires a holistic assessment of oxygenation, ventilation, respiratory mechanics, and the patient's intrinsic respiratory effort to identify the early candidate for ECMO weaning.⁵ This current study evaluates parameters that can predict weaning failure.

We conducted a retrospective observational study on adult patients (≥18 years) who underwent venovenous extracorporeal membrane oxygenation (VV ECMO) as rescue therapy for severe acute respiratory distress syndrome (ARDS) with refractory hypoxemia in the critical care unit between September 2022 and October 2024. Patients who were withdrawn from ECMO for futility or transitioned to alternative ECMO configurations (VVA or VA ECMO) were excluded.

At ECMO initiation, the following parameters were recorded: demographic data, APACHE II score, comorbidities, baseline hemodynamic and ventilatory parameters, and serum creatinine. During the ECMO course, changes in respiratory mechanics, blood gas values, and ventilatory settings were analysed. The requirement for sedation and neuromuscular blockade was also assessed during sweep gas titration. Institutional Review Board approval was obtained. The need for informed consent was waived owing to the retrospective and observational nature of the study.

Definition of Unsafe Weaning

Weaning from ECMO was considered unsafe if any of the following occurred:

• Requirement for recannulation due to worsening pulmonary mechanics.
• Escalation of ventilatory support (conversion to controlled mode) with deteriorating compliance, defined by an increase in dynamic driving pressure (ΔP) ≥5 cm H₂O.
• Use of rescue therapies such as new initiation of neuromuscular blockade, increased sedation, or prone positioning.

ECMO Weaning Protocol

All patients were initiated on VV ECMO after failing conventional rescue therapies for severe respiratory failure. For the first 24 hours, deep sedation and neuromuscular blockade were administered to facilitate a resting lung strategy. Thereafter, respiratory mechanics were closely monitored. Sweep gas weaning was attempted once there was evidence of improved lung compliance and enhanced PaCO₂ clearance.

• During sweep gas reduction, the following were assessed:
• Vital signs: heart rate, SpO₂, blood pressure, end-tidal CO₂ (ETCO₂), and inspiratory effort.
• Ventilatory parameters: tidal volume, minute ventilation, and respiratory rate.
• Arterial blood gas (ABG) parameters: pH, PaCO₂, PaO₂, bicarbonate, and lactate (measured pre- and post-sweep gas off trial [SGOT]).
• When sweep gas flow was reduced to ≤1 L/min, an SGOT was performed. Parameters of interest included:
• Gas exchange indices: pH, PaO₂, PaCO₂, and P/F ratio.
• Hemodynamics: heart rate.
• Sedation and paralysis requirements.
• Ventilatory mechanics: tidal volume, minute ventilation, and ventilatory ratio.

Statistical Analysis

Categorical variables were compared using the chi-square test or Fisher’s exact test and expressed as numbers with percentages. Continuous variables were reported as medians with standard deviations. Statistical significance was defined as p < 0.05. All analyses were performed using SPSS softwarea

During the study period, 25 consecutive patients received Veno-venous ECMO for severe respiratory failure. Of these, 5 patients were excluded (2 patients were withdrawn from ECMO, and 2 patients were switched to another configuration).

ECMO therapy: ECMO was initiated in severe hypoxic respiratory failure identified by mechanical ventilation for less than seven days, PaO2/FiO2 less than 80 mmHg, a Murray score of 3 to 4, or hypercapnia with PaCO2 more than 80 mmHg and failure to achieve target inflation pressures (Pplat ≤ 30 mmHg H2O) as a rescue measure³. In the current study, the average time from starting conventional mechanical ventilation to initiating ECMO was less than 1.5 days. ECMO therapy is a supportive therapy that serves as a bridge to recovery. Our cannulation strategies involved placing the access cannula in the right femoral vein and the return cannula in the right internal jugular vein (IJV). We, critical care specialists, perform ECMO cannulation at bedside in the critical care unit using a percutaneous technique with ultrasound guidance (USG), and we request confirmation of the cannula placement through USG and chest X-ray (CXR). The cannula positions are depicted in Fig. no. 1. All ECMO cannulation was done for bridge recovery, and the family was counselled for the same. At our centre, we prefer high blood flow, usually 3.5-4 Liters/min. This is shown in Figure 1.

Figure No. 1 CXR shows extensive bilateral lung parenchymal infiltration, along with the confirmed presence of both an access cannula and a return cannula.

Ventilator management⁴ is crucial in lung resting component of ECMO administration. All patients received ECMO for protective ventilation with FiO₂ less than 50%, positive end expiratory pressure (PEEP) of 10 cmH₂O, and respiratory rate of 10-14/min. Tidal volume (TV) titrated to driving pressure less than 15 cmH₂O. The preferred mode of ventilation was pressure control ventilation (n=18). The goal was to achieve dead space ventilation, with a tidal volume (TV) of less than 1 to 2 ml/kg and moderate PEEP. Patient recovery is monitored indirectly through improving lung compliance, which is evident from the increasing tidal volume (TV) and confirmed by monitoring chest X-rays (CXRs) to ensure that the whiteout lung fields are resolving. Sedation and paralysis were kept minimal and titrated to respiratory mechanics to lung resting ventilation, and sweep gas flow was adjusted while weaning sedation and paralysis. Optimal intravascular volume was targeted using bedside volume assessment with 2D echocardiography and lung ultrasound.

The median age of the population in the current study was 39.5 (27.5-54) years, and 50% (10/20) of them had diabetes mellitus. Community-acquired pneumonia (80%) was the reason for VV ECMO, with an expected mean mortality rate of 88%. (Table no: 1, 2)  Median duration (IQR) of ECMO was 9 (11-15) days with a survival rate of 60%. The observed complications were bleeding (n=15) and coagulopathy, which includes cannula site bleeding (n=12), intra-abdominal bleeding (n=3), postoperative site bleeding, and intracerebral hemorrhage (n=2). This is shown in Figure 2. The following factors observed in non-survivors were a higher APACHE II score, the presence of Acute kidney injury (AKI), and a higher SOFA score, but they did not show statistical significance in the outcome analysis. This is shown in Table 1,2 and 3.

Table 1: BASELINE CHARACTERISTICS

Table 2: Baseline Mean Variables:

Figure No. 2: Observed complications during ECMO therapy.

Table 3: Baseline characteristics of patient before initiating ECMO

Out of 20 VV ECMO patients 12 were successfully weaned and 8 were unsuccessful. All patients received two or more SGOTs for 30 minutes.

ECMO weaning: once we observed clinical recovery with lung mechanics, a daily ECMO weaning trial was attempted^5–8. As the lung compliance improves on ECMO, with better oxygenation and CO₂ removal, ECMO sweep gas was titrated. While titrating ECMO sweep gas flow, respiratory mechanics, pulse oximetry (SpO2) >88%, and normocarbia were maintained. The first step involved titration of sweep gas with close monitoring of changes with respiratory effort. If increasing respiratory effort with more inspiratory drive for TV weaning was aborted to the lowest sweep gas flow that patient tolerated. Weaning Sweep gas continued every day as the lung showed improvements with compliance, oxygenation, and ventilation. After administering 1 liter of sweep gas, the trial to discontinue sweep gas was initiated. If patient respiratory effort and TV changes were noted, sedation was increased to RASS -2, and SGOT continued for 2-4 hours. Once successful SGOT ECMO decannulation is completed, ECMO FiO₂ titration is not a routine practice in our protocol. SGOT was terminated if the following details were observed. If patient desaturation is <85%, or worsening hypotension or increasing ETCO2, sweep gas weaning is abandoned.

During SGOT, if the patient shows a change in inspiratory efforts in the form of tachypnoea (>24–27/min), higher minute ventilation (>8 l/min), or tidal volume (>7 ml/kg), sedation or, if required, a low dose of paralysis was restarted, and SGOT was continued with close monitoring of the patient.

Fifty percent of SGOT (n=19/38) were successful. Of the trials, 32 involved sedations, while 16 used half paralytics because the patients exhibited increased inspiratory effort during ECMO weaning. Prone ventilation was attempted for two patients during the trials. Neither mortality nor recannulation occurred after decannulation.

A total of 38 SGOT parameters were included for analysis (Table no: 4). SGOT was given to all patients before decannulation once they reached a low sweep gas flow of less than 1 L/min. Various parameters The analysis included heart rate, ventilator parameters such as peak inspiratory pressure (PIP), tidal volume (TV), minute ventilation (MV), ventilatory ratio, PaO2, FiO2, P/F ratio, PaCO2, pH, and the use of sedation and paralysis. The following variables—PaO2, PaCO2, minute ventilation, peak airway pressure, and heart rate (Table:4) were shown statistical difference in successful SGOT on the other hand outcome benefit was analysed with un paired T test except P/F Ratio, PaCo2, Heartrate the other variables failure to show outcome benefit. During SGOT (n=38) usage sedation (n=32) and partial paralysis dose (n= 16) did not show significance in outcome (Fisher’s Exact test).

Table no. 4: Variables around the SGOT trial and its statistical significance:

In this study we investigated SGOT, patient respiratory drive, ventilatory parameters, oxygenation, and ventilation. We considered weaning earlier and attempted ECMO decannulation after a 2-hour sweep gas-off trial. During the SGOT procedure, sedation was used to control the inspiratory drive, which is an important predictor of weaning failure due to secondary ventilatory lung injury. ⁴

The decision to discontinue VV ECMO support is made at the discretion of the clinical team. This requires clinicians to assess lung recovery, which was well indicated by chest radiography and lung compliance (which can be unreliable during spontaneous respiratory efforts), and oxygenation levels.

Due to uncertainty in ECMO discontinuation Various patient safety screening strategies were used while attempting weaning ECMO^6-12

• Low sweep gas flow < 3L¹⁰ or < 1L ⁹
• The sweep gas flow off duration varies between 30 minutes and 26 hours.⁵
• Mean time to decannulate was 2 days from first-pass SGOT¹³
• Having an ECMO free trial with safety screening¹⁰ the following were considered high risk for ECMO weaning: use of neuromuscular blockers, high doses of vasopressors, SG>3L, ventilator FiO₂>60%, pH <7.35, PaO₂<60 mmHg, SpO₂<90%, and HR>120 BPM. And it is like looking for the perfect scenario, stringent screening vs. liberal screening protocol, which would be time-bound and involve exorbitant costs in developing countries.

ECMO weaning predictors: ^{11,12,14–18}

Inspiratory drive¹¹Ventilatory ratio correlates with dead space ventilation, and it is calculated as (minute ventilation ml/min  X arterial PaCO₂ mmHg / Predicted body weight 100 X 37.5).Fischbach et al identifies ventilatory ratio as emerging tool as predictor of ECMO weaning failure. Patient with unsafe weaning had high ventilatory ratio ( 1.58 Vs 1.89) with low P/F ratio (243 Vs 169). Our current study of 38 off trial lower mean ventilatory ratio 1.81 Vs 1.77 and mean P/F ratio 218 VS 157 mmHg in unsuccessful trial. The use of sedation (no=16 Vs 16) and partial neuromuscular blockers (n=14 Vs 11) were more with unsuccessful SGOT trial to control the inspiratory drive. The noticeable cause of weaning failure in the physiological cohort is increased work of breathing.

End Tidal PaCO2 to Arterial PCO2: A multi-centre study from Europe to assess VV ECMO weaning predictors, which includes a prospective physiological cohort and a retrospective clinical cohort, concluded that the main reason for weaning failure is high respiratory effort, as evident by an oesophageal pressure swing of more than 15 and a higher PETCO2/PaCO2 ratio¹⁴. Daily ECMO weaning with safety screening tools and same-day decannulation with SGOT will facilitate early identification of patients for decannulation¹⁰.Given the fact that increasing complication with increasing ECMO days² and the resource implication of prolonged ECMO pose a potential risk of MDR infection with prolonged ECMO, it is necessary to identify the patient's early recovery phase and wean them from ECMO by escalating the conventional way of mechanical ventilation with a lung-protective ventilation strategy. We need to identify them safely; shortening ECMO days would be appropriate in resource-limiting settings.

Limitations

This study has several limitations. First, its retrospective, single-center design with a relatively small sample size limits the generalizability of the findings. Second, patient management, including sedation, neuromuscular blockade, and ventilator strategies, was influenced by clinician discretion, which may have introduced variability in outcomes. Third, although we identified physiological parameters associated with successful SGOT, we were unable to adjust for all potential confounders, such as differences in comorbid conditions, infection status, or severity of lung pathology. Finally, long-term outcomes beyond hospital survival were not evaluated.

Future Directions

Future prospective, multicenter studies with larger cohorts are needed to validate SGOT-based protocols and establish standardized criteria for VV ECMO weaning. Incorporating advanced monitoring tools such as esophageal manometry, diaphragmatic ultrasound, and end-tidal to arterial CO₂ ratios may further refine the prediction of weaning readiness. Randomized comparisons between conservative and liberal weaning strategies could help determine optimal timing of decannulation. Additionally, cost-effectiveness analyses are warranted, especially in resource-limited settings, to balance early liberation from ECMO with patient safety.

In this retrospective observational study of adult patients supported with VV ECMO for severe ARDS, nearly 60% were successfully weaned with no need for recannulation or post-decannulation mortality. The sweep gas off trial (SGOT) proved to be a practical and reproducible method to assess readiness for ECMO discontinuation. Parameters such as PaO₂, PaCO₂, minute ventilation, peak airway pressure, and heart rate showed significant differences between successful and unsuccessful SGOTs, highlighting their potential role as physiological markers of weaning outcomes.

Importantly, we observed that excessive inspiratory effort and increased ventilatory ratio were major contributors to unsafe weaning, often necessitating the use of sedation or neuromuscular blockade to prevent secondary lung injury. These findings emphasize that careful monitoring of patient effort, ventilatory mechanics, and gas exchange is essential for safe ECMO liberation. Our results support the feasibility of structured, protocol-driven weaning using SGOT in resource-limited settings, balancing early decannulation against the risks of prolonged ECMO-related complications. Larger multicenter prospective studies are warranted to validate these predictors and to develop standardized, evidence-based guidelines for VV ECMO weaning.