This exploratory study aimed to investigate whether combining preoperative intranasal insulin with pre-emptive analgesia is associated with a lower incidence of postoperative delirium (POD) in elderly patients undergoing lower extremity fracture surgery.

This prospective, randomized, exploratory trial involved 186 patients (aged ≥65 years and classified as ASA class II–III), who were allocated to one of three groups ( n  = 62): Group C (control: intranasal saline, 0.5 mL); Group I (insulin: intranasal insulin, 20 IU [0.5 mL]) and Group IP (insulin and pre-emptive analgesia: intranasal insulin as in Group I, plus a preoperative ultrasound-guided iliac fascia block with 0.375% ropivacaine [30 mL]). Intranasal administrations were given twice daily for two preoperative days and 20 min before anesthesia induction (five doses in total). The outcomes included NRS pain scores (at baseline [P1], 30 min after the intervention [P2] and 30 min after the third intranasal dose on day two [P4]), SRSS sleep scores (at baseline [P0] and on the night of the intervention [P3]), the incidence of postoperative delirium (assessed using the 3D-CAM on postoperative days one [T1], three [T2] and five [T3]), and serum biomarkers (insulin, glucose, insulin resistance [IR] and p-tau) at baseline [P1] and on postoperative day 1 [T1]).

The incidence of POD was significantly lower in Group IP (17.74%) than in Group C (38.71%; corrected P  = 0.009), while no significant difference was observed between Group I (27.42%) and Group C (corrected P  = 0.181). For secondary exploratory outcomes, Group IP showed lower NRS scores at P2 and P4 (both nominal P   <  0.001 vs. Group C) and lower SRSS scores at P3 (nominal P   <  0.001 vs. Group C), suggesting potential improvements in pain and sleep. These patterns were not observed in Group I. Preoperative biomarkers were comparable across groups. Postoperatively, Group IP showed lower glucose (nominal P  = 0.028), IR (nominal P  = 0.001), and p-tau levels (nominal P   <  0.001) compared with Group C, as well as lower insulin (nominal P   <  0.001), IR (nominal P  = 0.008), and p-tau (nominal P   <  0.001) compared with Group I. Longitudinal analysis (T1 vs. P1) indicated an increase in p-tau in Group C and Group I (both nominal P   <  0.001), whereas p-tau levels remained relatively stable in Group IP. A reduction in IR was observed only in Group IP (nominal P   <  0.001).

In this exploratory study, preoperative intranasal insulin alone showed no clear association with reduced POD or improved sleep. However, its combination with pre-emptive analgesia was associated with lower POD incidence, better pain and sleep profiles, and favorable biomarker changes. These hypothesis-generating findings warrant confirmation in future trials.