Hydrocortisone Linked to Higher Mortality in Bloodstream Infection Cohort

Among matched intensive-care patients with hospital-acquired bloodstream infection, 50% of hydrocortisone recipients died by day 28, compared with 42% of patients not receiving it. The association remained after statistical adjustment, but the study cannot establish that hydrocortisone caused the difference.

Evidence

The EUROBACT-2 hydrocortisone analysis, published September 15, was an ancillary observational analysis of a prospective cohort: patients were followed over time, but treatment was not randomly assigned. It included 2,401 adults whose bloodstream infections were managed in intensive care. Hospital-acquired infection was defined by a positive blood culture—a test for organisms in blood—from a sample taken more than 48 hours after hospital admission.

Of these patients, 608 received hydrocortisone. One-to-one propensity-score matching, which pairs treated and untreated patients with similar measured characteristics, produced 521 pairs. Researchers then used regression to adjust for multiple recorded factors, with the statistical analysis accounting for matched pairs.

Deaths occurred in 260 of 521 hydrocortisone recipients and 220 of 521 controls. The adjusted odds ratio for death was 1.448, with a 95% confidence interval of 1.096–1.914. Odds express the probability of dying relative to the probability of surviving; this estimate does not mean a 45% increase in the probability of death. The interval reflects statistical uncertainty.

Hydrocortisone exposure was also associated with fewer days free of blood-pressure-support medicines and mechanical breathing support by day 28, and less favorable changes in an organ-dysfunction score from diagnosis through day 7. These were exploratory outcomes measured after baseline.

The separate Manchester inclusive-research strategy study, published September 16, describes co-production: developing a strategy jointly with research teams, public contributors and other stakeholders. Iterative workshops and individual consultations informed a strategy structured around five principles. Research inclusivity was organised into three functions: Equality, Diversity and Inclusion; Patient and Public Involvement and Engagement; and Inclusive Research Methods. An Inclusive Research Oversight Board provides strategic governance across the Manchester Biomedical Research Centre and Clinical Research Facility.

Analysis

Interpretation: These studies make different contributions. The infection analysis identifies a treatment-outcome association that needs stronger testing. The Manchester study describes how responsibilities, stakeholder engagement and governance can be organised for inclusive research; it does not demonstrate improved clinical outcomes.

The studies do not corroborate one another. The organisational framework neither tests hydrocortisone nor explains the mortality difference. Its value here is as an implementation account, distinct from evidence about treatment effects.

Limitations

The infection study did not record why hydrocortisone was prescribed or exactly when it started relative to blood sampling or infection diagnosis. Matching can balance measured characteristics, but differences in unmeasured factors may remain. Uncertain timing also leaves possible time-dependent bias: when treatment begins during an evolving illness can distort the comparison. The exploratory organ-support outcomes do not resolve these causal limitations.

The Manchester study provides no comparative clinical-effect estimates. Stakeholders differed in their understanding of inclusive research, and implementation requires continuing evaluation and adaptation. Practical, linguistic, financial, cultural and trust-related barriers remain relevant to whether an organisational strategy translates into more inclusive participation.

What to watch

For Manchester, useful evaluation would measure who participates and whether accessibility or acceptability improves against a defined comparison, rather than documenting governance alone.

For hydrocortisone, a more informative study would record treatment rationale and exact initiation times, align follow-up between groups and address treatment timing explicitly. Whether the day-28 mortality association persists or substantially weakens under that design would change confidence in the finding.