Static pre-ECMO scores showed only modest post-cannulation discrimination, and 24-hour lactate clearance ≥65% was independently associated with survival, though the incremental prognostic benefit of adding lactate clearance to existing scores showed mixed signals and requires prospective multicenter confirmation.
Key Findings
Results
All three pre-ECMO prognostic scores demonstrated only modest discrimination for in-hospital mortality when applied post-cannulation.
Three scores evaluated: TiPS65, RESCUE-IHCA, and Survival After Veno-Arterial ECMO (SAVE).
Area under the receiver operating characteristic curve (AUROC) ranged from 0.574 to 0.668 across the three scores.
The study included 158 adults treated with ECPR at a single center from 2015 to 2022.
Scores were originally developed for pre-cannulation risk estimation and are 'frequently repurposed for post-cannulation decisions.'
Results
24-hour lactate clearance ≥65% was independently associated with survival after adjustment for RESCUE-IHCA score.
Analysis performed in a prespecified 24-hour landmark cohort of n=105 patients, with complete cases n=103.
Adjusted OR for survival with LC ≥65% was 3.34 (95% CI 1.45–7.68; P=.004) after adjustment for RESCUE-IHCA.
Lactate clearance (LC) was defined as ≥65% reduction over the first 24 hours post-cannulation.
The association was identified in a retrospective, single-center cohort.
Results
Adding 24-hour lactate clearance to RESCUE-IHCA score produced a modest improvement in discrimination with mixed reclassification signals.
Change in AUROC (ΔAUROC) when adding LC to RESCUE-IHCA was 0.069.
Integrated discrimination improvement (IDI) was 0.037 (P=.032), reaching statistical significance.
Category-based net reclassification improvement (NRI) was 0.178 but did not reach statistical significance (P=.18).
Reclassification signals were described as 'mixed,' indicating inconsistent evidence of incremental benefit across metrics.
Results
Decision curve analysis showed a slightly higher net benefit for the combined model at intermediate threshold probabilities, but with overlapping confidence intervals.
The combined model (RESCUE-IHCA + LC) showed slightly higher net benefit at threshold probabilities of 40%–60% compared to RESCUE-IHCA alone.
Bootstrap confidence intervals for the difference in net benefit included zero across all thresholds examined.
This indicates that the clinical decision-making advantage of adding lactate clearance could not be confirmed statistically.
Conclusions
The authors characterize findings as exploratory and hypothesis-generating, proposing a 2-step prognostic concept requiring prospective multicenter validation.
The proposed concept involves baseline risk estimation using a static pre-ECMO score followed by early dynamic reassessment using lactate clearance.
Authors explicitly state findings require 'confirmation in prospective, multicenter cohorts before any clinical use.'
The study is described as 'hypothesis-generating' rather than practice-changing.
Limitations include single-center design, retrospective methodology, and reduction from 158 to 103 complete cases for the landmark analysis.
What This Means
This research suggests that existing scoring tools used to predict outcomes in patients who receive extracorporeal cardiopulmonary resuscitation (ECPR — a form of life support using a heart-lung bypass machine during cardiac arrest) are not very accurate when applied after the patient is already connected to the machine. The study examined three commonly used scoring systems in 158 patients treated at a single hospital over seven years and found that all three had only modest ability to predict who would survive to hospital discharge once ECPR had already started. This matters because clinicians sometimes use these scores not just before starting ECPR, but also afterward when making ongoing treatment decisions.
The researchers then asked whether measuring how quickly a patient's lactate level (a blood marker of how well the body's tissues are being supplied with oxygen) drops over the first 24 hours could provide additional useful information. They found that patients whose lactate cleared by 65% or more within 24 hours were over three times more likely to survive, even after accounting for one of the pre-ECMO scores. Adding this lactate clearance measure to the score did improve prediction somewhat, but the improvement was inconsistent depending on which statistical method was used, and confidence intervals were wide.
This research suggests a potential two-step approach to prognostication in ECPR: first estimate risk before cannulation using a static score, then reassess dynamically at 24 hours using lactate clearance. However, the authors are careful to label these findings as preliminary and hypothesis-generating only, noting that the single-center, retrospective design and incomplete data for some patients limit conclusions. They call for prospective studies at multiple hospitals before this approach could be considered for actual clinical decision-making.
Kim T, Lee S, Song K, Yoo J. (2026). Post-cannulation prognostic reassessment in ECPR: 24-hour lactate clearance beyond pre-ECMO scores: a retrospective cohort study.. Medicine. https://doi.org/10.1097/MD.0000000000050674