Background and Objectives: Obstructive sleep apnea (OSA) is biologically heterogeneous, and neurohormonal change is a recognized dimension of this heterogeneity. Findings on cortisol in OSA are inconsistent, possibly because studies sample different parts of the diurnal rhythm. We examined salivary cortisol and its diurnal slope in relation to OSA severity, collapse pattern at drug-induced sleep endoscopy (DISE), and screening questionnaire performance. Materials and Methods: This cross-sectional analysis of baseline data from a prospective cohort study included 73 patients with moderate (n = 41) or severe (n = 32) OSA undergoing DISE. Salivary cortisol was collected at 07:00–08:00 and 22:00–23:00, and the diurnal slope was calculated as morning minus evening cortisol. Groups were compared with the Mann–Whitney U test and by analysis of covariance (ANCOVA) adjusted for age, BMI, and arterial hypertension. Discrimination of severe OSA by the STOP-BANG questionnaire and the Epworth Sleepiness Scale (ESS) was assessed by ROC analysis. Results: Morning cortisol (median 8.99 vs. 11.98 ng/mL, p = 0.002) and diurnal slope (6.29 vs. 9.89 ng/mL, p = 0.001) were lower in severe than moderate OSA, while evening cortisol did not differ (p = 0.649). After adjustment for age, BMI, and hypertension, severe OSA remained associated with lower morning cortisol (p = 0.005) and a flatter slope (p = 0.005), and the AHI correlated inversely with both (partial ρ = −0.330 and ρ = −0.359). Patients with complete concentric collapse had higher evening cortisol than those with other patterns (p = 0.045; p = 0.028 after adjustment). The STOP-BANG questionnaire discriminated severe OSA (AUC = 0.667, p = 0.023); the ESS did not (AUC = 0.615, p = 0.118). Conclusions: Severe OSA is associated with lower morning cortisol and a flattened diurnal rhythm than moderate OSA, independent of age, BMI, and hypertension. The association between concentric collapse and evening cortisol is novel but exploratory and requires replication.