Acute Anterior Pituitary Dysfunction and In-Hospital Mortality in Moderate-to-Severe Traumatic Brain Injury: A Prospective Observational Study

Abstract Traumatic brain injury (TBI) disturbs the anterior pituitary acutely, but whether early hormone concentrations predict survival is unclear. We examined the acute anterior pituitary profile and in-hospital mortality. In a prospective Strengthening the Reporting of Observational Studies in Epidemiology-compliant study, 101 adults with moderate-to-severe TBI (postresuscitation Glasgow Coma Score [GCS] ≤ 13; 6 had a GCS of 13) had anterior pituitary hormones measured on days 1 and 7. The primary outcome was in-hospital mortality. Day 1 concentrations were compared between survivors and nonsurvivors with correction for multiplicity; a multivariable logistic model was internally validated by bootstrapping, with prespecified sensitivity analyses. Nineteen patients (18.8%) died, all after day 7. At least one abnormality was present in 82% on day 1, most often gonadotropic suppression, low free triiodothyronine (FT3), and hypercortisolemia. Day 1 growth hormone (GH) was higher in nonsurvivors (4.83 vs 1.07 ng/mL; Bonferroni-adjusted p = 0.005), as was FT3 (4.91 vs. 2.20 pg/mL; adjusted p = 0.015); both persisted on day 7. Adjusted for GCS, GH (odds ratio [OR] 1.41 per ng/mL, 95% confidence interval [CI] 1.09–1.82) and FT3 (OR 1.67, CI 1.11–2.49) independently predicted mortality; discrimination was good (area under the curve 0.87; optimism-corrected 0.86) and calibration was acceptable. Findings held in penalized and log-transformed models and in a sensitivity analysis restricted to GCS ≤ 12; after multiple imputations GH was attenuated (OR 1.19, CI 0.98–1.45). Acute anterior pituitary disturbance is common after moderate-to-severe TBI. Elevated day 1 GH is independently associated with in-hospital mortality and is a candidate early prognostic marker, although attenuated under multiple imputations, and requires external validation. The thyroid-axis association ran counter to the chronic low-T3 literature and is hypothesis-generating.