A recent study investigated the prevalence, determinants, and prognostic implications of pulmonary hypertension (PH) in patients with acute respiratory distress syndrome (ARDS) supported with veno-venous extracorporeal membrane oxygenation (V-V ECMO). PH, defined as a mean pulmonary artery pressure (mPAP) greater than 20 mmHg, is a recognised complication of ARDS and contributes to right ventricular (RV) dysfunction and mortality. Although V-V ECMO improves oxygenation and carbon dioxide clearance and may theoretically reduce pulmonary vascular load, PH frequently persists during ECMO support. The authors aimed to characterise daily pulmonary haemodynamics during ECMO, identify independent determinants of mPAP, and assess whether early changes in mPAP were associated with hospital mortality.
The study included 225 adult ARDS patients receiving V-V ECMO and only the first ECMO run per patient was analysed. The median age was 51 years, 33% were female, and the most common ARDS aetiology was pneumonia (88%), predominantly viral. Before ECMO initiation, patients had severe hypoxaemia and impaired respiratory compliance. Median ECMO duration was 14 days, hospital length of stay was 35 days, and hospital mortality was 30%.
Daily data were collected from ECMO initiation until decannulation or death. Measurements included systemic and pulmonary haemodynamics, ventilator settings, arterial and mixed venous blood gases, and intrapulmonary shunt fraction (Qs/Qt). In total, 4,433 ECMO days were analysed, with mPAP available for over 99% of timepoints.
Pulmonary hypertension was highly prevalent. The median mPAP during ECMO was 27 mmHg, and 91% of all measurements exceeded 20 mmHg. Overall, 98% of patients had at least one mPAP value above the PH threshold. Pulmonary vasodilators were administered in 22% of patients, typically when mPAP was markedly elevated (median 34 mmHg at initiation).
Independent determinants of higher mPAP included longer ECMO duration, lower venous pH, higher intrapulmonary shunt fraction, higher pulmonary artery occlusion pressure (PAOP), higher positive end-expiratory pressure (PEEP), higher cardiac output, higher mean arterial pressure, and lower respiratory system compliance. In contrast, higher venous partial pressure of oxygen (PvO₂) and higher mixed venous oxygen saturation (SvO₂) were associated with lower mPAP. Venous carbon dioxide tension (PvCO₂) was not independently associated with mPAP. The model showed good performance, with a marginal R² of 0.41 and a conditional R² of 0.74.
In a subgroup of 55 patients with PAC monitoring before ECMO initiation, mPAP significantly decreased from 31 mmHg pre-ECMO to 27 mmHg on day 1 of ECMO, suggesting partial haemodynamic improvement with extracorporeal support. Nevertheless, pulmonary pressures often remained elevated throughout the ECMO course.
The prognostic analysis focused on two exposure variables: mPAP on the first day of ECMO and the trajectory of mPAP over the first five ECMO days. The trajectory was calculated as the patient-specific slope of mPAP over this period. In univariable analysis, day-1 mPAP was not associated with hospital mortality. In contrast, an increasing mPAP trajectory was significantly associated with higher mortality. In multivariable logistic regression adjusted for age, Respiratory ECMO Survival Prediction (RESP) score, Sequential Organ Failure Assessment (SOFA) score, and pre-ECMO respiratory compliance, the mPAP trajectory remained the strongest independent predictor of hospital mortality. Each 1 mmHg/day increase in mPAP slope was associated with an 89% increase in the odds of death. Day-1 mPAP was not independently predictive. Higher RESP scores and better pre-ECMO compliance were associated with improved survival, whereas age and SOFA score were not independently significant.
The authors interpret these findings as evidence that dynamic changes in pulmonary vascular load during early ECMO are more prognostically informative than static baseline measurements. Physiologically, elevated mPAP reflected multiple mechanisms: hypoxic pulmonary vasoconstriction linked to shunt and low venous oxygenation, acidosis-induced vasoconstriction, mechanical effects of high PEEP and reduced compliance compressing pulmonary vasculature, increased flow-related pressure with higher cardiac output, and possible post-capillary contributions from elevated PAOP.
In conclusion, pulmonary hypertension is highly prevalent in ARDS patients receiving V-V ECMO and is driven by both respiratory and haemodynamic factors. While baseline mPAP is not predictive of outcome, an early upward trajectory in mPAP independently identifies patients at increased risk of hospital mortality. Serial pulmonary artery pressure monitoring may therefore offer important prognostic insight and help guide management in selected high-risk patients, though routine PAC use in all ECMO patients is not advocated. Further prospective studies are warranted to clarify the role of pulmonary haemodynamic monitoring in this setting.
Source: Critical Care
Image Credit: iStock
