Patients who had both mild cognitive impairment (MCI) and clinically relevant depressive symptoms before unilateral knee or hip replacement had substantially higher odds of postoperative delirium than the comparison groups in a prospective cohort study. The association remained significant after researchers adjusted for age, sex, education, body mass index and baseline health conditions. A Firth-corrected analysis produced an odds ratio of 8.455, with a 95% confidence interval from 3.483 to 22.167 and P < 0.001. The unadjusted odds ratio was 18.562, with a 95% confidence interval from 8.534 to 40.767. These figures describe an association in the study population. They do not show that the combined condition caused delirium.
Postoperative delirium, abbreviated POD, occurred in 203 of the 1,464 included patients, an overall incidence of 13.9%. The study classified 1,261 patients as NPOD, meaning they did not develop the recorded outcome. Of 1,625 people screened, 161 were excluded before the analysis.
How the study was run
Researchers followed patients aged 40 to 90 at Qingdao Municipal Hospital who were undergoing unilateral knee or hip replacement under combined spinal-epidural anesthesia. Enrollment ran from July 2020 through September 2021. Baseline cognition was assessed with the Mini-Mental State Examination. After surgery, the team performed twice-daily Confusion Assessment Method checks from postoperative day 1 until discharge or day 7, and cases meeting that definition formed the POD group.
Patients with a preoperative MMSE score below 24 were excluded, so the findings do not address people below that threshold. Depression was defined using a 17-item Hamilton Depression Rating Scale total score of at least 8. The manuscript says this operational category represents clinically relevant depressive symptoms and is not the same as a diagnosis of major depressive disorder.
Cerebrospinal fluid was also analyzed. Researchers collected 2 mL, processed it by centrifugation at 2,000 × g for 10 minutes, stored it at minus 80°C and used blinded ELISA testing. The markers included Aβ42, T-tau, P-tau, Aβ42/T-tau and Aβ42/P-tau.
Biomarkers followed a similar pattern
In logistic analyses, higher T-tau and P-tau were associated with higher odds of POD, while Aβ42 and the two Aβ42-to-tau ratios showed inverse associations. In group comparisons, patients with both MCI and depressive symptoms had lower Aβ42 and higher T-tau than controls and selected single-condition groups. Their P-tau was higher than in all three comparison groups specified by the study.
Statistical mediation models identified T-tau, P-tau, Aβ42/T-tau and Aβ42/P-tau as mediators of the reported association. The estimated mediated proportions ranged from 17.9% to 49.9%. That result means the markers helped account for part of the pattern in the model. It does not establish that tau or amyloid changes are a biological cause of delirium, particularly because the data came from a cohort study.
Useful signal, but not a finished clinical tool
The researchers also built a clinical prediction model using age, gender, MMSE, hypertension, diabetes, CHD, alcohol history and MCI with comorbid depression. Its area under the receiver operating characteristic curve, a measure of how well the model separates patients with and without the outcome, was 0.884, with a 95% confidence interval of 0.860 to 0.908. At the reported cutoff, sensitivity was 0.847 and specificity was 0.817, while the false-positive rate was about 0.152. This was a within-cohort result and does not establish how the model would perform in other settings.
Over 36 months, 29 people in the POD group died, or 14.3%, compared with 58 deaths, or 4.6%, in the NPOD group. The POD group had lower three-year survival, with P < 0.001. Among patients who developed POD, however, survival did not differ significantly across the preoperative cognitive and depression subgroups, with P = 0.79. The study therefore did not establish a mortality difference within the POD group.
What the findings can and cannot show
The study's design and population set clear boundaries around the findings. It was a prospective cohort at one hospital, involving people aged 40 to 90 undergoing unilateral knee or hip replacement under combined spinal-epidural anesthesia. The cohort design means the results describe association rather than proof of cause. The exclusion of MMSE scores below 24 also limits the findings to participants above that threshold, and the depression measure was a symptom category rather than a diagnosis of major depressive disorder. The prediction model's reported performance should likewise be treated as a result from this cohort, not as established performance elsewhere.
Paper data and sources
Original title: Association of mild cognitive impairment and comorbid depression with postoperative delirium in patients undergoing knee/hip arthroplasty: a prospective cohort study
Authors: Lin X, He Q, Wang K et al.
Journal/Repository: Frontiers in medicine
Status: Peer-reviewed
First online: 2026-10-07
DOI: Not available
Original paper · Full text