Patients With Moderate to Severe Dementia Who Undergo Elective TKA Face Elevated Risks of Death, Delirium, and Intensive Interventions: A Large, Propensity-weighted Database Study.

Background: Although TKA is one of the most commonly performed and successful orthopaedic interventions, its efficacy may be reduced in individuals living with dementia. Prior studies have not effectively accounted for confounding by indication or the use of intensive interventions, or considered the fact that cognitive impairment exists on a continuum and that patients with mild dementia may have different risk profiles when compared with those with moderate to severe dementia.

Questions/purposes: (1) Are there differences in the risk of death after TKA in patients based on the severity of dementia? (2) Are there differences in intensive interventions and delirium after TKA in patients based on the severity of dementia?

Methods: We used Medicare data to allow for complete surveillance of all medical encounters and events that transpired before and after TKA surgery. We retrospectively identified Medicare beneficiaries living with dementia who underwent primary elective TKA between January 1, 2017, and June 30, 2018. We used the Claims-based Frailty Index (CFI) to stratify dementia severity. Patients with CFI scores of 0.25 to 0.28 and an ICD-10 code for dementia were recorded as living with mild dementia. Patients with an ICD-10 code for dementia and CFI scores of ≥ 0.28 were classified as having moderate to severe dementia. Our cohort included 156,596 patients, with 98% (152,728 of 156,596) classified as not having dementia, 1% (2123 of 156,596) having mild dementia, and 1% (1745 of 156,596) having moderate to severe dementia. We had complete surveillance of outcomes and events after surgery for this cohort of patients. The primary outcome was death within 180 days of surgery. Intensive interventions (such as feeding tube insertion, intubation, resuscitation) and delirium were considered secondarily. We used inverse probability weights to account for confounding and compared outcomes between cohorts using Cox proportional hazards models, generalized estimation equations, and Fine and Gray models.

Results: After adjusting for potentially confounding variables such as age, gender, and comorbidities, we found that individuals with mild dementia (HR 1.74 [95% confidence interval (CI) 1.12 to 2.70]; p = 0.01) and moderate to severe dementia (HR 3.05 [95% CI 1.80 to 5.17]; p < 0.001) both demonstrated elevated hazards of death compared with patients without dementia. Patients with mild dementia (HR 4.25 [95% CI 2.59 to 5.03]; p < 0.001) and moderate to severe dementia (HR 6.40 [95% CI 5.18 to 7.92]; p < 0.001) exhibited elevated hazards of delirium, and those with moderate to severe dementia were found to have greater hazards of one or more intensive interventions (HR 3.24 [95% CI 1.76 to 5.96]; p < 0.001).

Conclusion: We observed elevations in the risk of death and delirium in patients after TKA, irrespective of the severity of dementia and marked elevations in the likelihood of intensive interventions after surgery for those with moderate to severe dementia. We believe that TKA in patients with mild dementia should only proceed after clear communication about the increased risk of death, as well as effective management of chronic medical conditions and prophylactic enhancement of cognitive reserves. Performing TKA in patients with moderate to severe dementia might only be indicated in exceptional humanitarian circumstances.

Level Of Evidence: Level III, therapeutic study.