Timing of Recovery After Severe Traumatic Brain Injury: What Long-Term Outcomes Reveal About Early Prognostication

Prognosis after severe traumatic brain injury remains one of the most difficult challenges in acute neurologic care. Even with modern imaging and standardized neurologic examinations, early assessments are limited by sedation, physiologic instability, and evolving injury. Despite these limitations, decisions about withdrawal of life-sustaining therapy are often made within the first several days after injury.

A recent single-center analysis from the Department of Neurological Surgery at UPMC and the University of Pittsburgh School of Medicine provides new data on how long meaningful neurologic recovery can remain unapparent and how often recovery occurs later than early prognostic timelines typically allow.

The study, “A Propensity-Matched Comparison of Capacity for Functional Recovery by 24 Months Post-traumatic Brain Injury in Patients Who Died After Withdrawl of Life-Sustaining Therapy,” published online in January 2026 in Neurosurgery, was led by Shawn Eagle, PhD, assistant professor of neurological surgery at the University of Pittsburgh School of Medicine. David Okonkwo, MD, PhD, professor of neurological surgery and clinical director of neurotrauma at UPMC, was the senior author. Using longitudinal data from patients enrolled at UPMC through TRACK-TBI multicenter registry, Dr. Eagle and colleagues examined recovery patterns among patients with severe traumatic brain injury over two years.

“If I had to distill the findings from this study down to one sentence, it would be that the data support more optimism for potential recovery from these devastating injuries,” Dr. Eagle says. “Not everyone is going to recover to independence or return to their pre-injury status, but the data support more optimism than is often reflected in early prognostic discussions,” Dr. Eagle says.

Why Timing Matters in Severe Traumatic Brain Injury Prognosis

A central issue addressed by the study is the mismatch between the timing of prognostic decisions and the biology of neurologic recovery. Prior research has shown that meaningful neurologic improvement may not be apparent early after injury, even in patients who later experience favorable long-term outcomes. This creates uncertainty when early clinical assessments are used to guide irreversible decisions.

“Patients who go on to have good recoveries frequently do not start following commands until close to two weeks after injury,” Dr. Eagle says. “At the same time, national data suggest that withdrawal of life-sustaining therapy often occurs within the first three days,” Dr. Eagle says.

The analysis followed patients for up to 24 months. This allowed assessment of recovery over time rather than at a single early endpoint.

Recovery Continues Beyond Early Follow-Up

Among patients who were fully dependent at three months after injury, more than one-third achieved at least partial independence by two years. Recovery was not limited to early phases of care. Many patients improved between six and twenty-four months and even after 12 to 24 months.

“For severe traumatic brain injury, recovery does not stop at six months,” Dr. Eagle says. “Two years is a realistic timeframe for understanding what functional recovery can look like for some patients,” Dr. Eagle says.

Even patients with higher-risk clinical profiles demonstrated meaningful recovery in some cases. Late deterioration was uncommon across the cohort. Stability or improvement predominated.

Limits of Early Prognostic Tools

The study examined commonly used prognostic models when applied to individual patients. One such tool is the International Mission on Prognosis and Analysis of Clinical Trials (IMPACT) model, which estimates the likelihood of mortality or severe disability after traumatic brain injury based on early clinical and imaging data. While useful for population-level analysis, the model showed limitations when applied to individual patients. Some patients with high predicted risk later achieved functional independence.

“These models were never designed to make individual patient decisions,” Dr. Eagle says. “They are useful for research and for understanding populations, but they are not precise enough to determine who does or does not have recovery potential,” Dr. Eagle says.

This limitation is relevant when prognostic estimates are used to guide discussions about withdrawal of life-sustaining therapy early in the hospital course.

“There is no bigger decision than whether someone is given the opportunity to recover,” Dr. Eagle says. “The data show that some patients predicted to do very poorly still recover meaningful function,” Dr. Eagle says.

Severity on Imaging Does Not Preclude Recovery

Outcomes were also examined across different CT injury patterns. Patients with mass lesions, including those requiring surgical intervention, demonstrated recovery trajectories similar to patients with less severe imaging findings.

“Some of the most severe-appearing injuries still had meaningful recovery over time,” Dr. Eagle says. “Imaging severity alone should not define long-term expectations,” Dr. Eagle says.

Late functional improvement remained possible across imaging subgroups. Early appearance did not reliably define long-term outcome.

Implications for Withdrawal of Life-Sustaining Therapy

The study does not argue against withdrawal of life-sustaining therapy when clinically appropriate. Decisions surrounding withdrawal are influenced by medical status, patient wishes, family values, and broader ethical considerations that are not captured in observational datasets.

“There are absolutely situations where withdrawal of life-sustaining therapy is the right decision,” Dr. Eagle says. “What this study supports is allowing appropriate time when there are clinical signals that recovery may still be possible,” Dr. Eagle says.

By quantifying how often recovery occurs and how late it can emerge, the findings highlight the risk of relying too heavily on early neurologic assessments or prognostic estimates when making irreversible decisions in the acute phase of severe traumatic brain injury care.

Read a commentary on the study from UPMC neurosurgeon, John Kanter, MD.

Reference

Eagle SR, Yue JK, Shanahan RM, Shim J, Puccio AM, Okonkwo DO. A Propensity-Matched Comparison of Capacity for Functional Recovery by 24 Months Post-traumatic Brain Injury in Patients Who Died After Withdrawl of Life-Sustaining Therapy. Neurosurgery. 2026; 00: 1-10 Online ahead of print.