Sepsis and Subsequent Psychiatric Morbidity: A Nationwide Population-Based Matched Cohort Study, 2008-2019.

OBJECTIVES

To quantify the risk of incident psychiatric morbidity after community-acquired sepsis and assess whether new chronic diseases mediate the association.

DESIGN

Nationwide, population-based matched register cohort; hazards estimated with weighted Cox regression.

SETTING

Sweden, linking the National Quality Sepsis Registry, National Patient Register, Prescribed Drug Register, and population registers.

PATIENTS: Ten thousand three hundred eight adults (≥ 18 yr) treated in an ICU for sepsis (2008-2019), matched to 155,705 population controls by sex, age, region, and year. Individuals with a psychiatric diagnosis within 5 years or psychotropic medication within 1 year before index were excluded.

INTERVENTIONS: None.

MEASUREMENTS AND MAIN

RESULTS

The primary outcome, psychiatric event, was first occurrence after index date of either initiation of a psychotropic medication (anatomic therapeutic chemical classification system code N05A, N05BA, N05C, N06A) in the Prescribed Drug Register (capturing prescriptions from primary and specialist care) or a new International Classification of Diseases , 10th Edition mood (F3) or anxiety (F4) diagnosis in specialist care. Weighted Cox models balanced baseline covariates. We used a Landmark approach with risk sets at 0-30, 31-90, 91-365 days; 1-3, 3-5, and greater than or equal to 5 years after the index date. Sepsis was associated with increased hazards of psychiatric events vs. matched controls, with the strongest associations in the first year (0-30 d: adjusted hazard ratio [aHR], 6.2 [5.0-7.7]; 31-90 d: aHR, 7.4 [6.5-8.6]; and 91-365 d: aHR, 2.3 [2.1-2.5]) attenuating over time but remaining elevated through 5 years (1-3 yr: aHR, 1.2 [1.1-1.5]; 3-5 yr: aHR, 1.3 [1.1-1.5]; and ≥ 5 yr: aHR, 1.1 [0.9-1.3]). In mediation analyses considering incident chronic diseases, estimates changed little, suggesting that these conditions did not mediate the association.

CONCLUSIONS

Patients with sepsis had a higher subsequent incidence of psychiatric events compared with matched population controls, with a persistently elevated risk for at least 5 years. This increased risk suggests that sepsis may have a long-term impact on psychiatric health, warranting consideration of preventive strategies.