Background <p>Lower mean Hounsfield unit (mHU) values, indicating greater computed tomography (CT) hypodensity of perihematomal edema (PHE), have been proposed as a novel quantitative imaging marker in intracerebral hemorrhage (ICH). We explored its evolution and prognostic importance in a post hoc analysis of the Intracerebral Hemorrhage-Deferoxamine trial (NCT02175225).</p> Methods <p>We included participants with primary supratentorial ICH who had available CT scans at baseline and follow-up after 72–96&#xa0;h and 90-days and/or 180-days outcome data. The primary exposure variable was the mHU of PHE measured on the follow-up CT scan. We investigated (1) its change from baseline and (2) its association with unfavorable outcome (modified Rankin Scale score 3–6) in adjusted mixed-effects models, accounting for between-center and between-participant variability.</p> Results <p>Among 273 of 293 Intracerebral Hemorrhage-Deferoxamine trial participants eligible for analysis (median age 61&#xa0;years, 39% female), the median (interquartile range) mHU of PHE was 30.3 (28.3–32.7) at baseline and 26.9 (24.6–29.2) at follow-up. Τhe mHU of PHE decreased from baseline to follow-up scan by an average of 3.6 (95% confidence interval [CI] 3.2–4.0, <i>p</i> &lt; 0.001). There was no association between the mHU of follow-up PHE with unfavorable outcome at 90&#xa0;days (<i>n</i> = 273; odds ratio 1.05, 95% CI 0.95–1.17, <i>p</i> = 0.32), or at 180&#xa0;days (<i>n</i> = 261; odds ratio 1.01, 95% CI 0.92–1.11, <i>p</i> = 0.81).</p> Conclusions <p>Perihematomal edema after ICH tends to grow more hypodense on CT by day 3–4 compared with baseline. The degree of PHE hypodensity was not associated with long-term clinical outcomes in the setting of a multicenter randomized trial, challenging its utility as a radiological marker in ICH research.</p> Clinical Trial Registration <p>ClinicalTrials.gov: NCT02175225.</p>