668 - Every Week Counts: Time to Presentation Prolongs Recovery After Pediatric Sports Concussion

Publication Number: 2651.668

Julian Javier, Icahn School of Medicine at Mount Sinai, New York, NY, United States; Reginald Brewster, Icahn School of Medicine at Mount Sinai, New York, NY, United States; Ramone Brown, Icahn School of Medicine at Mount Sinai, New York, NY, United States; Jordan Bryan, Icahn School of Medicine at Mount Sinai, New York, NY, United States; Eric W. Small, Mount Sinai, New York, NY, United States

Background: Recovery after pediatric sports-related concussion varies widely, and clinicians need simple, clinic-ready predictors at the initial and 2-week visits.

Objective: To identify baseline and early-visit factors associated with longer recovery and to quantify "from-today" prognosis at 14 days using a landmark approach.

Design/Methods: IRB-approved, single-center retrospective cohort of patients ages 9-17 with sports-related concussion, 4/2021-4/2025 (n=135). Exposures included age, sex, sport, mechanism, time to presentation (TTP), baseline symptoms, and exam (balance, gait, convergence, saccades, cognition). The primary outcome was time to symptom resolution. Secondary outcomes were prolonged recovery (≥21, ≥28 days) and, among those symptomatic at 14 days (n= 75), remaining time symptom resolution. We fit accelerated failure-time (log-normal) models and translated effects to days using a median anchor (T₀=15d); logistic models estimated thresholds. Models adjusted for age, sex, and TTP with FDR control.

Results: Cohort medians were as follows: age 15y (IQR 14-16), TTP 7d (3-13), time to resolution 15d (7.5-27.5), RTP 17d (12-30); sex for 135: 53 female (39%), 82 male (61%)
Each 1-week delay to presentation was associated with about 3.2 additional days to recovery (95% CI 2.1-4.3; p=1.6×10⁻⁸). Female sex was associated with about 6.7 more days (p=2.7×10⁻⁴). By sport, basketball injuries were about 11.7 days longer and soccer about 4.8 days shorter (both p< 0.01); mechanisms (e.g., ball-to-head, fall) were not associated. On exam, balance difficulty was associated with about 12.3 more days (p=0.0068). Convergence distance added about 5.0 days per inch (p=0.0037). Individual baseline symptoms were not associated after multiplicity control.
Prolonged recovery occurred in 56/135 (41.5%) ≥21d and 34/135 (25.2%) ≥28d; odds increased with TTP (≥21d OR 1.10, 95% CI 1.05-1.16; ≥28d OR 1.12, 1.07-1.19). Among those symptomatic at 14 days, each additional week of TTP was linked to about 20% longer remaining time from Day-14 (p=2.6×10⁻⁴; ~+2.6 days/week), given median remaining time 13d. At Day-14, effects of sex, sport, and exam were not significant.

Conclusion(s): Time to presentation is a modifiable, clinic-ready predictor of longer concussion recovery-both from the time of injury and, in those still symptomatic at 14 days, from that visit onward. Balance findings and sport add counseling context, whereas mechanism does not stratify prognosis. Findings underscore the need for timely access to concussion evaluation and support tracking TTP as a quality metric. Single-center design warrants prospective validation.

Figure 1. Adjusted differences (days) in time to symptom resolution by predictor
Accelerated failure-time (log-normal) models were fit; effects were translated to days using a median anchor (T0 = 15 d). Models adjust for age, sex, and time-to-presentation (TTP); false discovery rate (FDR) controlled within predictor families. TTP is displayed as a 1-week delay for interpretability. Convergence (+1 in NPC) denotes per +1-inch increase in near-point of convergence (exploratory, n=7). Points = estimates; lines = 95% CIs.