Pain, Impairment Linger Months After ECMO for Most

Many extracorporeal membrane oxygenation (ECMO) patients experienced long-term neuropsychiatric symptoms, cognitive impairment, or functional impairment, a meta-analysis found.

Fully 41% of ECMO patients had neuropsychiatric symptoms at 6 months or longer follow-up, with pain or discomfort (52%), sleep disturbance (37%), and anxiety (36%) as the most common symptoms, reported Sung-Min Cho, DO, MHS, of Johns Hopkins Hospital in Baltimore, Maryland, and colleagues.

Long-term follow-up at 6 months and beyond also showed the secondary outcomes of neurocognitive impairment in 38% and functional impairment in 52%.

"For ECMO patients, these outcomes are especially critical because they are associated with long-term mortality, quality of life, and economic impact," Cho and colleagues wrote in Neurology.

ECMO use is on the rise, but "the outlook of ECMO survivorship remains uncertain," according to an invited commentary in Neurology by Imad Khan, MD, of the University of Rochester Medical Center in New York, and Nicholas Morris, MD, of the University of Maryland in Baltimore.

"ECMO is associated with direct neurologic injury, including both macrohemorrhages and microhemorrhages, infarcts, and diffuse hypoxic-ischemic brain injury that likely contribute to long-term outcomes," they wrote. However, they added, "In patients who do not experience acute brain injury, we have not established whether ECMO directly contributes to poor long-term neurologic outcomes or merely serves as a surrogate for severe critical illness."

They pointed to a number of limitations of the current body of evidence, such as the more than 10 different data collection instruments used across the studies in Cho's analysis and inadequate controls to adjust for confounding.

The findings emphasize that "clinical standardization in assessment, outpatient care, and multi-specialty consultation is necessary in the clinic," Cho added in an email to MedPage Today.

The editorialists agreed, noting that a "contribution of this study may be in assisting power calculations for clinical trials of interventions to improve long-term patient-centered outcomes," such as neuromonitoring, early mobilization, and sedation reduction.

"As survival increases with improvements in care delivery, refinements in patient selection, and technological advances, the mission shifts from helping patients survive to helping survivors thrive," Khan and Morris concluded.

The meta-analysis and systematic review included 59 studies with a variety of designs (10 prospective cohort, three retrospective case-control, 45 retrospective cohort, and one randomized controlled trial). Of the 3,280 patients analyzed, 69% were male, and the median age was 54.

Most of the patients had venoarterial ECMO (VA-ECMO, 86%), with the remaining 14% getting venovenous (VV) ECMO. The most common indication for VA-ECMO was extracorporeal cardiopulmonary resuscitation; the most common indication for VV-ECMO was acute respiratory failure, and these patients were typically supported longer (median 12.5 vs 3.7 days).

The studies measured neuropsychiatric, neurocognitive, and functional outcomes, evaluated postdischarge by questionnaire. Median follow-up lasted 15.8 months.

Functional impairment, typically defined by the European Quality of Life 5 Dimensions Questionnaire, most commonly impacted daily activities (54%), followed by physical activity or mobility (41%), and personal or self-care (21%).

Favorable neurologic outcomes were reported in 23% at 3 months, 25% at 6 months, and 28% at 1 year and beyond when including nonsurvivors and defined as a modified Rankin scale of 3 or less, cerebral performance category score of 2 or less, or a Glasgow Outcome Scale score of 4 or greater. When excluding those who died during follow-up, the rates were 59%, 84%, and 90% across the three time points.

Neurocognitive impairment -- which encompassed memory, attention, and reasoning issues -- was numerically but not significantly more common with VV-ECMO (50% vs 24%, P=0.30). More VV-ECMO patients had neuropsychiatric symptoms than VA-ECMO (55% vs 32%, P=0.01).

Study authors acknowledged that they were limited by heterogeneity in instruments used to measure neuropsychiatric symptoms and neurocognitive impairment, as well as a broad range of follow-up endpoints and diverse outcomes. Most of the studies were retrospective, which could have introduced recall bias, and patient characteristics were not consistent across studies. The most common indication for VA-ECMO, accounting for most of patients studied, might not be the most common indication for ECMO outside the study, the researchers cautioned.

Comment