To compare elderly (≥ 80 yr), older (65-79 yr), and younger (< 65 yr) ICU admissions in Scotland in relation to trends in admission rates, regional variation in admissions, ICU treatment intensity, and ICU and 1-year mortality.
National 5-year cohort study of ICU first admissions (January 1, 2005, to December 31, 2009).
All admissions to ICUs and combined units (level 2/3 care) in Scotland captured by the Scottish Intensive Care Society Audit Group database, linked with hospital discharge data and death records.
A total of 40,142 patients: 3,865 were 80 years old or older (9.6%), 13,904 (34.6%) were 65-79 years old; and 22,373 were younger than 65 years (55.7%).
Between 2005 and 2009, elderly admission rates decreased from 36.6/10,000 (95% CI, 34.0-39.2) in 2005 to 28.7/10,000 (95% CI, 26.5-30.9) in 2009 (p < 0.001; relative decrease, 22.0%); older admission rates also decreased, but less steeply (31.1 [95% CI, 29.9-32.2] to 26.1 [95% CI, 25.1-27.1] per 10,000 population; p < 0.001; relative decrease, 16.1%). Rates were static for younger patients. Restricted to mechanically ventilated elderly patients, rates ranged from 13.9 to 30.1/10,000 between healthboard administrative regions (p < 0.001). Emergency surgical diagnoses were more prevalent for elderly patients (elderly, 39.8%; older, 25.1%; younger, 20.3%; p < 0.001). Subgroup analyses limited to pneumonia admissions (elderly, n = 242; older, n = 1,226; younger, n = 1,836) indicated similar acute physiology scores, but fewer preexisting comorbidities among elderly patients (p = 0.007), who received a shorter duration of organ support and ICU stay. Mortality rates were higher in elderly patients at ICU discharge (elderly, 26.5%; older, 25.0%; younger, 17.0%; p < 0.001; confounder adjusted odds ratio elderly vs younger, 2.33 [95% CI, 2.11-2.58]; p < 0.001). Differences persisted at 1 year (elderly, 52.2%; older, 43.8%; younger, 27.6%; adjusted odds ratio elderly vs younger, 3.72 [95% CI, 3.42-4.06]; p < 0.001).
In Scotland, elderly and older ICU admission rates are decreasing, with regional geographic variation. Although limited by an absence of a measure of frailty, patient characteristics and treatment intensity suggest selection of less comorbid elderly patients, indicating possible rationing based on chronologic age.