Clinical Scorecard: A substantial share of excess deaths among immunocompromised patients with viral severe acute respiratory infection occurs after ICU discharge

At a Glance

Key Highlights

• In the cited cohort, 78 of 780 immunocompromised ICU survivors died later in hospital (10.0%), compared with 136 of 3,432 immunocompetent survivors (4.0%).
• The authors calculate that post-ICU excess mortality was 2.5-fold, compared with 1.5-fold excess mortality during ICU admission; approximately half of the absolute disparity in hospital mortality arose after ICU discharge.
• The article raises the possibility that treatment limitations, late secondary infection, or progression of underlying disease may contribute to post-ICU deaths, but does not establish their roles.
• The authors identify potential case-definition and missing-immune-status selection issues, and note that COVID-19 incidental-admission exclusions were numerous.

Guideline-Based Recommendations

Diagnosis

• Clarify which version of the incidental-admission exclusion criterion was applied, including whether the primary reason for ICU admission was considered.
• Report, by pathogen, the proportion of immunocompromised patients among admissions excluded as incidental.
• Assess the impact of the 1,862 adults with unknown immune status through outcome reporting or an analysis weighted by the probability that immune status was known.
• In a linked analysis, consider restricting admissions to those with a respiratory APACHE III diagnosis, as proposed by the authors.

Management

• If the post-ICU mortality signal is confirmed, the authors propose documenting goals of care at ICU discharge and providing structured outreach review during the ward phase.
• Consider linking the cohort with the ANZICS Adult Patient Database to re-estimate associations by baseline severity and examine patients admitted under treatment limitation separately.

Monitoring & Follow-up

• Present ICU deaths and deaths after ICU discharge as distinct outcomes, with time from ICU discharge to death, ICU readmission, and discharge destination.
• For competing-risks analyses, report the cause-specific discharge hazard alongside cumulative incidence functions, or report days alive and out of hospital through day 30.
• Consider a discharge-model sensitivity analysis restricted to covariates known at ICU admission.

Risks

• Treatment received may affect SARI case eligibility differently by immune status, particularly when residual SARS-CoV-2 RNA and antiviral use are present.
• Stratifying by invasive ventilation conditions on a decision influenced by immune status and goals of care; persistence of an association across strata cannot by itself distinguish immune dysfunction.
• A Fine and Gray subdistribution hazard ratio for discharge may be reduced when death is more common as a competing event.
• Adjusting discharge models for post-admission treatments may partly reflect length of stay, because patients must remain hospitalized to receive those treatments.

Patient & Prescribing Data

The cited SPRINT-SARI analysis included 4,703 adults with laboratory-confirmed viral SARI admitted to 46 Australian ICUs; 780 immunocompromised and 3,432 immunocompetent patients were discharged alive from the ICU.

Antivirals were used more frequently in immunocompromised patients than immunocompetent patients (67.3% versus 57.7%). The article cautions that pathogen distribution and the case definition may affect this comparison.

Clinical Best Practices

• Distinguish deaths during ICU admission from deaths after ICU discharge when reporting hospital mortality.
• Report goals of care at ICU discharge and, if the post-ICU signal is confirmed, consider structured ward outreach review, as proposed by the authors.
• Clarify case eligibility criteria and examine whether incidental-admission exclusions differ by immune status and pathogen.
• Separate mortality from delayed discharge when interpreting 30-day discharge outcomes in the presence of competing risks.

Related Resources & Content