Haloperidol benefit in ICU delirium by subtype, a screening tool predicts survival in older cancer patients, and virtual reality does not beat cognitive stimulation for dementia symptoms

The edition in brief

A post hoc Bayesian analysis of the AID-ICU trial (987 adults, 54.7% hypoactive and 45.3% hyperactive delirium) found moderate to high probabilities of benefit from haloperidol in both subtypes on days alive out of hospital and 90-day mortality, with about a 73% to 77% probability of a larger effect in hyperactive delirium; credible intervals were wide (difference in mortality risk difference -3.7 percentage points, 95% CrI -15.3 to 7.8) and the authors call for confirmation. A meta-analysis of 36 studies (7,110 patients) found moderate accuracy of the Vulnerable Elders Survey-13 for abnormalities on comprehensive geriatric assessment (diagnostic OR 1.93) and higher VES-13 scores predicted worse overall survival (HR 1.68, 95% CI 1.53-1.84) and more adverse events (pooled OR 4.29, 95% CI 1.39-13.23). A secondary analysis of 161 adults aged 70 or over with advanced cancer found sarcopenia in 57% and, in adjusted models, an association with severe non-haematological toxicity (OR 2.24, 95% CI 1.12-4.47) and worse one-year survival on first-line therapy (HR 2.12, 95% CI 1.14-3.93). A randomised trial of 56 nursing home residents with behavioural symptoms of dementia found no significant difference between group virtual reality and active cognitive stimulation in symptom reduction.