Pre-ICU hospital length of stay and 30-day mortality in the very old: a national cohort of 315 042 ICU admissions

Background The number of very old (≥80 years) patients (VIPs) admitted to intensive care units (ICUs) is rising. Hospital length of stay before ICU admission (pre-ICU LOS) is associated with short-term mortality, but whether this association is uniform across adult age is unclear. We examined how age modifies the association between pre-ICU LOS and 30-day mortality. Methods We included all adult first ICU admissions in the Swedish Intensive Care Registry 2005–2016, with comorbidity and vital status from national registers. We fit a logistic regression on 30-day mortality with restricted cubic splines for age, pre-ICU LOS, and Charlson comorbidity index, with interactions. Marginal predicted risks and within- and between-age risk differences (RDs) were from the model. Two sensitivity analyses included Simplified Acute Physiology Score (SAPS3) functional form of pre-ICU LOS and age, and adjustment for acute physiology at admission (SAPS3 Box III) with multiple imputation. Results Among 315 042 patients, 30-day mortality rose with pre-ICU LOS in every age decile, but the magnitude of this gradient varied with age. In VIPs, the within-age RD from day 0 to day 60 was 12.8 percentage points (95% CI 10.0 to 15.6) against a day-0 baseline of 34.6 %, compared with 17.5 (16.0 to 18.9) in 60–69-year-olds. Findings were consistent across sensitivity analyses. Conclusions Pre-ICU hospital length of stay is a weaker prognostic marker in VIPs than in the majority of ICU patients, although the extent to which this reflects biology versus age-dependent ICU triage remains unresolved. ClinicalTrials.gov identifier NCT07571213 (retrospectively registered).