Relative adrenal insufficiency in sepsis has been extensively debated on; however, accurate diagnosis and therapeutic intervention remain controversial. The authors aimed to evaluate adrenocorticotropic hormone (ACTH), salivary cortisol, total cortisol and estimated plasma-free cortisol, cholesterol, and lipoproteins as predictors of adrenal insufficiency in patients within 24 h of septic shock diagnosis. This prospective study evaluated all hospitalized patients older than 18 years who developed septic shock and were using vasoactive drugs within 24 h of diagnosis. Blood and saliva samples were drawn at baseline and 60 min (T60) after 250 μg tetracosactide intravenous injection. Patients were divided into two groups: responders (Δ [T60 minus baseline] total cortisol >9 μg/dL) and nonresponders (Δ total cortisol ≤9 μg/dL or baseline total cortisol <10 μg/dL). The latter group was considered to have adrenal insufficiency. A total of 7,324 hospitalized patients were monitored, and 34 subjects with septic shock were included in the analysis. Adrenal insufficiency was found in 32.4%. Total cholesterol, high-density lipoprotein cholesterol, triglycerides, and salivary cortisol did not differ between groups. Estimated plasma-free cortisol was not better than total plasma cortisol in estimating adrenal function. Baseline endogenous ACTH was higher in nonresponders than responders (55.5 pg/mL vs. 18.3 pg/mL, respectively; P = 0.01). The cutoff ACTH value that discriminated patients with adrenal insufficiency was 31.5 pg/mL. Thus, endogenous ACTH measured within 24 h of septic shock diagnosis could predict adrenal response to tetracosactide.