Comparative Study of Clinical Diagnoses and Autopsy Findings in Intensive Care

• Autopsy (Diagnostic test), also known as anatomopathological reports: Autopsy realised in adult patients during of after a hospitalization in an intensive care unit.

Primary outcomes

• Rate of Major Diagnostic Discrepancies (Goldman Class I and II) between Clinical and Autopsy Diagnoses [Diagnoses assessed at time of autopsy, for patients who died in ICU between January 1, 2008 and December 31, 2025]: Proportion of autopsied ICU patients with a major diagnostic discrepancy (Goldman class I or II) between the antemortem clinical diagnosis and the post-mortem autopsy diagnosis, classified according to the Goldman classification

Secondary outcomes

• Type and anatomical/etiological category of diagnoses missed clinically and identified at autopsy (Goldman Class I-IV) [Diagnoses assessed at time of autopsy, for patients who died in ICU between January 1, 2008 and December 31, 2025]: Distribution of missed diagnoses by organ system and etiological category (infectious, vascular, neoplastic, etc.) among discrepancies classified as Goldman class I to IV.
• Odds Ratios for Patient and Clinical Characteristics Associated With Major Diagnostic Discrepancy (Goldman Class I/II) [Assessed using clinical data collected from ICU admission to death, for patients who died and underwent autopsy between January 1, 2008 and December 31, 2025]: Association between diagnostic discrepancy status (Goldman class I/II vs. class III-VI) and pre-specified patient/clinical characteristics - including age, sex, immunosuppression status, indication for hospitalization, duration of hospital and ICU stay, admission unit, SAPS II score, type of organ failure, presence of septic shock, use of mechanical ventilation, renal replacement therapy, ECMO support, immunosuppressive therapy, and imaging modality - assessed by univariable and multivariable logistic regression, expressed as odds ratios with 95% confidence intervals.
• Autopsy Rate Among ICU Deaths [Calculated over the entire study period, from January 1, 2008 to December 31, 2025]: Proportion of ICU deaths for which an autopsy was performed, calculated as the number of autopsies divided by the total number of deaths occurring after ICU admission, expressed as a percentage
• Concordance Rate Between Antemortem Imaging Findings and Autopsy/Pathology Findings [Assessed at time of autopsy report, for patients who underwent both antemortem imaging and autopsy between January 1, 2008 and December 31, 2025]: Proportion of cases in which findings from the last antemortem imaging examination (CT, MRI, or ultrasound) performed before death are concordant with autopsy and/or anatomopathological findings. Concordance was assessed qualitatively by investigator review of imaging reports compared with autopsy/pathology reports, and classified as concordant, partially concordant, or discordant, without use of a standardized scoring system.

Sex: All

Inclusion Criteria:

* Inclusion of all adult patients who died since 2008 in a department of the Metz-Thionville Regional Hospital or the Nancy University Hospital, who had been admitted to an intensive care unit during the same hospital stay, and who underwent a medico-scientific autopsy followed by an anatomopathological analysis.

Exclusion Criteria:

* Minor patients
* Patients who were not admitted to intensive care
* Forensic autopsies

• Guillaume Louis, MD, Principal investigator, Central Hospital, Metz Thionville, France