Extracorporeal membrane oxygenation in neonates and critically ill adults demands more than technical skill — it requires precise clinical judgment about which patients will benefit most and which carry unacceptable risk. That is exactly why ecmo scoring systems like the PRESERVE score ECMO were developed: to give bedside clinicians a structured, evidence-based framework for predicting survival and guiding cannulation decisions before a patient is committed to this resource-intensive therapy.
Extracorporeal membrane oxygenation in neonates and critically ill adults demands more than technical skill — it requires precise clinical judgment about which patients will benefit most and which carry unacceptable risk. That is exactly why ecmo scoring systems like the PRESERVE score ECMO were developed: to give bedside clinicians a structured, evidence-based framework for predicting survival and guiding cannulation decisions before a patient is committed to this resource-intensive therapy.
The PRESERVE score ECMO — an acronym for Prediction of Survival on ECMO Therapy — was derived from a large multicenter cohort of adults receiving venovenous extracorporeal membrane oxygenation for severe respiratory failure. It assigns points across eight variables including age, body mass index, immunocompromised status, prone positioning history, days of mechanical ventilation before ECMO, lung compliance, and the presence of nitric oxide therapy. Total scores range from 0 to 14, and higher scores correlate with markedly lower hospital survival, giving intensivists a reproducible number to anchor conversations with families and ethics committees.
Understanding the extracorporeal membrane oxygenation procedure from a scoring perspective means recognizing that patient selection is arguably as important as circuit management. Studies consistently show that centers with disciplined selection protocols — using tools like PRESERVE, RESP, SAVE, and neonatal-specific indices — achieve better outcomes than centers that apply ECMO broadly. The score does not replace clinical judgment; rather, it sharpens it by quantifying risk factors that might otherwise be weighed inconsistently across providers and shifts.
For extracorporeal membrane oxygenation treatment to deliver its full life-saving potential, the right patient must receive it at the right time. Delayed cannulation in a deteriorating neonate or an adult with refractory ARDS allows secondary organ injury to accumulate, worsening the prognosis regardless of how well the extracorporeal membrane oxygenation circuit subsequently functions. Scoring systems help teams avoid both under-utilization — leaving salvageable patients without support — and over-utilization, where patients with near-zero survival probability are subjected to invasive cannulation and prolonged suffering.
This article explores the major ECMO scoring tools in clinical use today, explains what each variable measures and why it matters, and provides practical guidance for ECMO specialists, intensivists, perfusionists, and respiratory therapists who encounter these scores on certification exams and in daily practice. Whether you are preparing for the Certified ECMO Specialist (CES) examination or managing a complex ICU patient, a thorough grasp of these risk stratification frameworks is essential professional knowledge.
The landscape of ECMO scoring has expanded considerably since the first tools were introduced in the 1990s. Today separate validated instruments exist for neonatal respiratory failure, pediatric cardiac ECMO, adult respiratory ECMO, and adult cardiac ECMO, each reflecting the distinct physiology, comorbidity burden, and outcome determinants of those populations. Knowing which score to apply — and how to interpret its result — is a core competency for anyone working in a modern ECMO program.
Throughout this guide, real numerical thresholds, score breakdowns, and published survival data are provided so readers can internalize not just the names of these tools but their actual clinical meaning. By the end, you will understand how to calculate a PRESERVE score at the bedside, how the RESP score differs for adults with respiratory failure, how the SAVE score applies to cardiogenic shock, and how neonatal-specific indices guide decisions for the smallest and most vulnerable patients on extracorporeal membrane oxygenation.
Predicts hospital survival for adults on venovenous ECMO for severe respiratory failure. Eight variables including age, BMI, immunocompromise, and ventilation duration. Scores 0–14; a score ≥6 corresponds to less than 20% predicted survival in the derivation cohort.
Respiratory ECMO Survival Prediction score uses 12 pre-ECMO variables to classify patients into five risk categories. Derived from the ELSO registry with over 2,000 patients; scores range from −22 to +15 with higher scores predicting better survival.
Survival After Veno-Arterial ECMO score applies specifically to cardiogenic shock patients on VA ECMO. It incorporates diagnosis, organ function markers, and pre-ECMO cardiac arrest status. Validated across multiple international registries with strong discrimination.
Neonatal-specific tools — including the oxygenation index and alveolar-arterial oxygen difference — guide ECMO candidacy for congenital diaphragmatic hernia, meconium aspiration syndrome, and persistent pulmonary hypertension of the newborn.
The Extracorporeal Life Support Organization registry provides outcome benchmarks that inform center-specific quality metrics. Risk-adjusted survival comparisons help programs audit whether their patient selection and management match expected performance.
The PRESERVE score ECMO was formally published in 2013 by Schmidt and colleagues after analysis of a 140-patient cohort from two European referral centers. The eight variables it evaluates were selected through multivariate logistic regression as independent predictors of in-hospital mortality. Understanding each variable individually gives clinicians insight into the biological mechanisms that determine ECMO outcomes, not just a number to look up in a table before making a decision.
Age contributes one point for patients aged 18–60 and two points for those older than 60. This reflects the well-documented observation that older patients on extracorporeal membrane oxygenation treatment carry higher baseline comorbidity burden, reduced physiological reserve, and greater vulnerability to the complications of anticoagulation, immobility, and prolonged critical illness. Body mass index above 30 kg/m² also earns one point; severe obesity complicates oxygenation, increases sedation requirements, and raises the technical difficulty of cannulation and patient positioning during the extracorporeal membrane oxygenation procedure.
Immunocompromised status — defined as active malignancy, solid organ transplant, or chronic immunosuppressive therapy — adds two points, reflecting catastrophically high mortality in this group historically. Prone positioning prior to ECMO adds one point, which may seem counterintuitive since prone positioning is a guideline-recommended intervention for ARDS; the scoring rationale is that patients who required and received proning before ECMO cannulation had particularly severe disease burden, not that proning itself is harmful.
The number of days on mechanical ventilation before ECMO initiation is one of the most clinically actionable variables in the score. Zero to three days earns zero points; four to seven days earns one point; and more than seven days earns two points. This steep gradient reinforces the message that early referral to an ECMO-capable center — before ventilator-induced lung injury and secondary organ dysfunction accumulate — substantially improves predicted outcomes. Teams should resist the temptation to exhaust every conventional option before calling for ECMO consultation.
Lung compliance below 20 mL/cmH₂O on the day before ECMO adds one point, and the absence of vasopressors adds a negative one point, meaning hemodynamically stable patients fare better. Nitric oxide use adds one point. The composite PRESERVE score is then interpreted as follows: 0–2 points corresponds to predicted survival greater than 70%; 3–5 points corresponds to 30–70%; and 6 or more points corresponds to less than 20%. These thresholds have been externally validated in multiple European and North American cohorts, though some variability exists across settings, reinforcing that the score should inform rather than dictate clinical decisions.
A critical practical point often missed on certification exams is that the PRESERVE score was derived specifically for venovenous extracorporeal membrane oxygenation — patients with isolated respiratory failure whose cardiac function is sufficient to maintain perfusion without additional support. Applying it to venoarterial ECMO patients with cardiogenic shock introduces systematic error because the underlying pathophysiology, comorbidity profile, and outcome drivers in cardiac ECMO differ substantially. The SAVE score is the validated instrument for that population.
Comparing PRESERVE with the RESP score reveals meaningful differences in scope and design. RESP uses 12 variables and a larger derivation cohort from the ELSO registry (n = 2,355), giving it greater statistical power but also greater complexity.
It classifies patients into five risk classes: Class I (score ≥6, predicted survival 92%) through Class V (score ≤−6, predicted survival 18%). PRESERVE, with only eight variables and a simpler scoring algorithm, is somewhat faster to calculate at the bedside, which matters in time-pressured retrieval situations. Many programs use both tools in concert, treating concordant high-risk scores as a strong signal to involve palliative care and ethics consultants early.
Extracorporeal membrane oxygenation in neonates relies primarily on the oxygenation index (OI) and the alveolar-arterial oxygen difference (AaDO₂) as the central candidacy thresholds. An OI greater than 40 on two consecutive measurements, or an AaDO₂ exceeding 600 mmHg for more than four hours, is widely accepted as meeting the threshold for ECMO referral. Congenital diaphragmatic hernia, meconium aspiration syndrome, and persistent pulmonary hypertension of the newborn together account for the majority of neonatal ECMO cases, each with disease-specific survival benchmarks published by ELSO.
Unlike adult tools, neonatal ECMO scoring does not aggregate a multivariate point total into a single survival probability. Instead, clinicians interpret the trajectory of the OI alongside gestational age, birth weight, echocardiographic findings, and the severity of associated anomalies. Gestational age below 34 weeks and weight below 2 kg remain relative contraindications at most centers because the risk of intraventricular hemorrhage during systemic anticoagulation is prohibitively high in the most premature infants. Birth defects incompatible with meaningful survival are absolute contraindications regardless of oxygenation indices.
For adults with refractory ARDS, the PRESERVE and RESP scores provide complementary perspectives on predicted survival. The Berlin Definition criteria for severe ARDS — PaO₂/FiO₂ ratio below 80 mmHg with FiO₂ of 1.0 and PEEP ≥5 cmH₂O — serve as the physiological entry criterion, but scoring tools filter candidates further by accounting for reversibility, comorbidities, and ventilation duration. The EOLIA trial demonstrated a 35% absolute reduction in the composite outcome of death or ECMO crossover when ECMO was initiated early in severe ARDS, validating the strategy of early referral before score trajectories worsen.
Venovenous extracorporeal membrane oxygenation dominates the adult respiratory ECMO landscape because it supports gas exchange while preserving native cardiac output. The circuit drains deoxygenated blood from the right atrium, passes it through an oxygenator membrane, and returns oxygenated blood to the right atrium or femoral vein — allowing the injured lungs to rest at lung-protective settings. The extracorporeal membrane oxygenation circuit for VV configuration typically uses a single bicaval dual-lumen cannula or a two-cannula femoral-jugular approach, and scoring systems help determine which patients have sufficient reversibility to justify either configuration.
Extracorporeal membrane oxygenation COVID cases represented the most dramatic expansion of adult ECMO since the H1N1 pandemic of 2009. At the height of the pandemic, ELSO registry data captured over 4,000 COVID-19 patients on ECMO globally, with in-hospital mortality around 37% — comparable to non-COVID severe ARDS outcomes when selection criteria were rigorously applied. The challenge was that traditional PRESERVE and RESP thresholds had not been validated in COVID-19 specifically, prompting several centers to publish COVID-adapted scoring guidance emphasizing duration of high-pressure ventilation and inflammatory marker trajectories as additional risk signals.
Multiple studies found that COVID ECMO patients cannulated within seven days of intubation had substantially better outcomes than those cannulated later, reinforcing the universal ECMO scoring principle that time on mechanical ventilation before cannulation is one of the most modifiable determinants of survival. Some centers developed local COVID-specific triage algorithms overlaying PRESERVE and RESP scores with ferritin, D-dimer, and IL-6 thresholds, though none has achieved the formal validation status of the original tools. Recognizing this limitation remains important when interpreting scoring results for unusual etiologies of respiratory failure.
Of all eight PRESERVE score variables, only ventilation days before ECMO cannulation can be directly influenced by care team decisions. Patients ventilated for fewer than four days before ECMO earn zero points on this variable; those ventilated for more than seven days earn two points — a swing that alone can shift a patient from intermediate to high-risk category. Early ECMO consultation and timely cannulation, rather than exhausting every conventional measure over many days, is the single highest-yield intervention teams can make to improve predicted outcomes before the score is even calculated.
Applying ECMO scoring tools to patient selection in real clinical practice requires understanding not just the arithmetic of each score but the broader decision-making framework in which scores operate. No validated scoring system designates a hard cutoff above which ECMO is categorically withheld; rather, scores inform a conversation that must also incorporate the patient's wishes, the center's resource availability, the reversibility of the underlying condition, and the quality of available supportive alternatives.
In practice, most experienced ECMO programs use a tiered approach. Patients with PRESERVE scores of 0–2 or RESP Class I–II are typically offered ECMO without significant hesitation provided they meet physiological entry criteria. Patients in the intermediate range — PRESERVE 3–5 or RESP Class III — enter a more intensive multidisciplinary discussion that often includes palliative care consultation and documented family meetings. Patients with PRESERVE ≥6 or RESP Class IV–V are counseled that current evidence suggests a high likelihood of non-survival even with ECMO, and goals-of-care conversations become central to decision-making.
The extracorporeal membrane oxygenation machine price and resource infrastructure also influence how scoring thresholds are applied institutionally. A single ECMO run for a critically ill adult costs between $50,000 and $300,000 in direct hospital charges, and each circuit requires around-the-clock specialist staffing. During periods of high demand — such as the COVID-19 pandemic or winter respiratory virus surges — some centers developed formal scoring-based triage criteria to allocate scarce ECMO resources. These protocols relied heavily on PRESERVE and RESP thresholds as objective, defensible allocation criteria.
For extracorporeal membrane oxygenation for adults with cardiac arrest, the landscape of scoring is somewhat different. Emergency cannulation for extracorporeal CPR (eCPR) leaves little time for comprehensive pre-ECMO scoring. Instead, rapid eligibility screening — typically age below 65–70, witnessed arrest, bystander CPR, initial shockable rhythm, no-flow time under five minutes — serves as the de facto scoring framework. Post-cannulation, the SAVE score and SOFA score provide outcome estimation once the patient is stabilized on ECMO and a more complete clinical picture emerges.
Neonatal ECMO centers face a distinctly different challenge because the available evidence base is smaller, the population is more homogeneous in age but more heterogeneous in diagnosis, and the stakes of delaying or denying ECMO are extraordinarily high in a population where many conditions are completely reversible with appropriate support. The oxygenation index remains the most widely used threshold, but experienced neonatal ECMO programs recognize that OI alone is insufficient. Echocardiographic assessment of right ventricular function, pulmonary arterial pressure, and ductal shunt direction provides critical physiological context that numbers alone cannot capture.
An underappreciated aspect of ECMO scoring is its role in post-ECMO outcome prediction and decannulation planning. The same variables that predicted mortality at cannulation continue to influence outcomes during the ECMO run. A patient whose lung compliance fails to improve despite two weeks of ECMO and lung rest, whose PRESERVE score variables have not resolved, and who is developing new multi-organ dysfunction faces a prognosis very different from a patient showing steady improvement.
Serial reassessment — not a single pre-ECMO snapshot — is the gold standard approach, and ECMO teams that incorporate scoring into daily rounds rather than treating it as a one-time pre-cannulation exercise report better alignment between care trajectories and patient and family goals.
The extracorporeal membrane oxygenation diagram that appears in most ECMO training curricula — showing drainage cannula, pump, oxygenator, heat exchanger, and return cannula — maps directly onto the scoring variables that predict outcomes. When the oxygenator efficiency declines due to clot formation, when the pump generates hemolysis, when the circuit triggers an inflammatory cascade — all of these mechanical events are reflected in the clinical trajectory that scoring tools are designed to measure. Understanding the circuit physiology and understanding the scoring framework are not separate disciplines; they are deeply intertwined aspects of comprehensive ECMO expertise.
Preparing for ECMO certification examinations requires not only memorizing score variable names and thresholds but developing the ability to apply scoring logic to clinical vignettes under time pressure. The Certified ECMO Specialist (CES) examination administered by the American Board of Cardiovascular Perfusion and credentialing bodies associated with the Extracorporeal Life Support Organization includes questions testing knowledge of candidacy criteria, scoring tools, and the physiological rationale behind each variable. Candidates who understand why ventilation duration matters — not just that it does — perform consistently better on these analytical questions.
A systematic approach to studying ECMO scoring for certification begins with mastering the PRESERVE score in detail: all eight variables, their respective point values, and the survival probability bands corresponding to each score range. Next, candidates should learn the RESP score structure — particularly its 12 variables and five risk classes — well enough to identify which risk class a given patient falls into from a vignette description. The SAVE score for cardiac ECMO should then be studied, with special attention to the diagnosis categories and the way pre-ECMO cardiac arrest status dramatically increases predicted mortality.
Neonatal scoring knowledge for certification purposes centers on oxygenation index calculation — OI = (mean airway pressure × FiO₂ × 100) / PaO₂ — and the conventional OI thresholds that trigger ECMO candidacy evaluation. Candidates should also know the AaDO₂ formula and the threshold values used by most neonatal ECMO programs. Practice questions frequently present ABG values and ventilator parameters and ask candidates to calculate the OI and determine whether ECMO criteria are met, making arithmetic fluency with these formulas an examination necessity.
Beyond individual score knowledge, certification questions test understanding of how scores integrate into the overall extracorporeal membrane oxygenation procedure workflow. Candidates must know the sequence: identify physiological ECMO criteria met, calculate relevant scores, present to multidisciplinary team, obtain informed consent, select cannulation strategy, and document the clinical decision basis. Questions about contraindications — absolute versus relative, reversibility assessments, gestational age thresholds in neonates — frequently appear alongside scoring questions because they represent the complementary framework within which scores operate.
Time management during examination preparation should reflect the weighting of topics in the CES blueprint. Scoring and patient selection consistently appear in the examination framework alongside circuit management, anticoagulation, and complication recognition. Allocating two to three dedicated study sessions to scoring tools — using practice question banks that present vignette-based scenarios requiring score calculation rather than simple recall — produces the most durable retention and the best examination performance.
One effective study technique is to work through published PRESERVE and RESP validation studies, reading the patient characteristic tables and applying the score algorithm to described patients before reading the outcome data. This reverse-engineering approach builds authentic clinical reasoning skills rather than pattern-matching to memorized threshold numbers. Many ECMO certification candidates report that this type of active, case-based study was more valuable than passive review of scoring summaries, particularly for questions that present patients at borderline thresholds where clinical judgment must supplement the numerical result.
Finally, understanding the limitations of each scoring tool — and being able to articulate those limitations in a clinical context — is itself a certification competency. Knowing that PRESERVE was derived from a small European cohort, that RESP may overestimate survival in immunocompromised patients, or that neonatal OI thresholds were established before modern high-frequency oscillatory ventilation protocols became standard are the kinds of nuanced knowledge points that distinguish high-performing candidates.
The goal is not to dismiss these tools but to use them with appropriate confidence intervals, always integrating the score into a broader clinical narrative rather than treating it as a definitive verdict.
Practical tips for ECMO specialists working with scoring tools in daily clinical practice go beyond what certification examinations test. In real ICU environments, scoring calculations are often performed under time pressure, with incomplete data, and in the context of strong emotional stakes for patients and families. Developing a personal workflow for reliable, rapid score calculation is therefore an essential professional skill that complements textbook knowledge.
The most efficient approach is to maintain a laminated pocket reference or a smartphone-accessible scoring app — several free validated ECMO scoring calculators are available from ELSO and from academic medical centers — so that variable inputs and score outputs can be verified quickly without relying on memory alone. Even experienced clinicians benefit from structured prompts that prevent inadvertently skipping a variable, particularly during high-stress situations such as rapid patient transfers or emergent consultations at non-ECMO centers.
Communication of scoring results to families requires particular skill and sensitivity. Presenting a PRESERVE score of 7 — implying less than 20% predicted survival — to a patient's family members is not equivalent to saying their loved one has a 20% chance of survival.
It is a statement about a statistical prediction derived from patients with similar characteristics at the time of ECMO initiation, and it does not account for the individual's unique trajectory, the quality of their ECMO management, or factors not captured by the scoring model. Language that acknowledges uncertainty while conveying the seriousness of high-risk scores is both medically accurate and compassionate communication.
For quality improvement purposes, ECMO programs should track score distributions over time alongside actual outcomes to evaluate whether their institutional experience aligns with published survival predictions. Programs that consistently achieve better-than-predicted outcomes for high-PRESERVE-score patients may have developed specialized management protocols or patient selection refinements worth sharing in the literature. Programs with worse-than-predicted outcomes should investigate whether selection criteria, circuit management practices, or complication rates explain the gap, using scoring data as a quality benchmarking tool.
Multidisciplinary ECMO team training should explicitly include scoring education for all team members — not just physicians. ECMO specialists, perfusionists, and critical care nurses who understand the clinical significance of a PRESERVE score of 3 versus 7 are better equipped to participate meaningfully in daily goals-of-care discussions, recognize when a patient's trajectory is diverging from predicted outcomes, and advocate appropriately for timely palliative care involvement. This shared scoring literacy strengthens team communication and improves the quality of patient-centered decision-making across the ECMO run.
Future directions in ECMO scoring include machine learning models that incorporate real-time physiological trend data — not just static pre-ECMO snapshots — to generate dynamic survival probability estimates that update throughout the ECMO run. Preliminary studies suggest these dynamic models outperform static scores like PRESERVE and RESP in predicting 90-day survival, particularly for patients with prolonged ECMO runs exceeding 14 days. As these tools reach clinical implementation, ECMO specialists who understand the foundations of current scoring systems will be well positioned to evaluate and adopt next-generation predictive tools.
In summary, mastery of ECMO scoring — from the PRESERVE score ECMO variable-by-variable breakdown to neonatal oxygenation index thresholds to the limitations of each validated tool — is an indispensable component of ECMO expertise at every career stage. Whether you are a new ECMO specialist building foundational knowledge, an experienced intensivist refining your patient selection framework, or a candidate preparing for CES certification, investing time in deep understanding of these scoring systems will make you a more effective clinician and a more capable advocate for the critically ill patients who depend on ECMO for survival.