Clinical and Decisional Trajectories Preceding Early Postprocedural Death Among Patients with Preexisting Code-Status...

To develop a typology of recurrent clinical and decisional trajectories preceding early postprocedural death among patients with preexisting code-status limitations.

Patients with do-not-resuscitate orders experience higher postoperative mortality, but the clinical and decisional processes underlying early deaths are poorly understood. Mortality and code status alone cannot determine whether these deaths reflect constrained rescue, unsuccessful treatment, progression of terminal illness, or reassessment-driven redirection of care.

We conducted a retrospective qualitative trajectory analysis of electronic health record documentation across five academic and community hospitals from March 2024 through June 2025. Eligible adults had an active preprocedural code-status limitation and died within 3 calendar days of a procedure requiring anesthesia, before hospital discharge. Using an adapted sequence-based composite approach, 2 reviewers independently reconstructed cases as ordered clinical and decisional sequences and reached consensus on recurrent trajectory families; illustrative composites were then generated. A third reviewer independently classified cases while blinded to the consensus assignments; agreement was assessed using unweighted Cohen κ.

Among 2,833 patients with preexisting code-status limitations who underwent a procedure, 44 (1.6%) died within 3 days; the 42 who died before hospital discharge were included. Four trajectory families were identified: established terminal trajectory (5 [12%]), abrupt terminal event (5 [12%]), comfort-focused redirection after reassessment (29 [69%]), and sustained rescue attempt (3 [7%]). Redirection followed progressive deterioration or persistent critical illness despite treatment in 20 cases and a discrete event or finding in 9; rescue-oriented treatment frequently continued or escalated before reassessment and redirection. In sustained-rescue trajectories, intensive treatment continued through death, including in 2 patients whose limitations on CPR remained in effect. The independent reviewer agreed with the consensus classification for 39 of 42 cases (92.9%; κ, 0.86; 95% CI, 0.70-1.00).

Early postprocedural deaths among patients with preexisting code-status limitations followed distinct clinical and decisional trajectories. Mortality and code status alone could not distinguish an established terminal trajectory, abrupt death, reassessment-driven redirection, or sustained but unsuccessful rescue. A longitudinal approach that considers procedural intent, evolving prognosis, treatment, and reassessment provides a more clinically meaningful basis for interpreting mortality and structuring perioperative decision making.

Among patients with preexisting code-status limitations who died within 3 days of a procedure, four recurrent trajectories differed in relationships among baseline illness, clinical developments, prognostic reassessment, rescue-oriented treatment, and changes in treatment focus. These longitudinal dynamics were essential to interpreting early postprocedural death.