Fewer ICU patients with sepsis in UK died in 2017-2019 vs. 1988-1990

Key takeaways:

• As time progressed from 1988, hospital mortality among patients with sepsis in the U.K. went down.
• A matched analysis and an analysis of only ICUs in the same catchment area had similar mortality results.

In the U.K., the proportion of ICU patients with sepsis who died in a hospital was smaller in 2017 to 2019 vs. 1988 to 1990, according to results published in American Journal of Respiratory and Critical Care Medicine.

“Over the last 30 years, hospital mortality for patients admitted to an ICU with sepsis in the United Kingdom decreased substantially,” Hallie C. Prescott, MD, MSc, associate professor in pulmonary and critical care medicine at the University of Michigan and staff physician at the Ann Arbor Veterans Affairs Healthcare System, and colleagues wrote.

In a retrospective cohort study, Prescott and colleagues analyzed 426,812 ICU patients (median age, 66 years; 55.6% men) with nonsurgical sepsis in the U.K. between 1988 and 1990, as well as between 1996 and 2019, to see how presentation and outcomes in this patient population changed over 30 years.

Within the total cohort, 60.9% had a respiratory site of infection, whereas a smaller proportion of patients had a genitourinary (11.5%) or gastrointestinal (10.3%) infection site.

Between the earlier years (1988 to 1990; n = 801) and the more recent years (2017 to 2019; n = 99,948), researchers found differences in median age (63 years vs. 66 years), median Acute Physiology and Chronic Health Evaluation II acute physiology scores (20 vs. 14) and genitourinary sepsis infection prevalence (2% vs. 13.4%).

Among those admitted in 2017 to 2019 vs. in 1988 to 1990, there was also a reduction in two types of mortality: ICU mortality (24.5% vs. 46.6%) and hospital mortality (32.4% vs. 54.6%; difference, 22.2%). Admittance in more recent years was linked to a lower adjusted likelihood for hospital mortality (adjusted OR = 0.64; 95% CI, 0.54-0.75).

Between the two time periods, an 8.8% adjusted absolute reduction was observed in terms of hospital mortality, according to researchers.

A greater proportion of the total reduction in hospital mortality was explained by changes in case mix, which researchers said is potentially “driven by earlier presentation of patients to the hospital, earlier recognition of sepsis, changing threshold or timing of ICU admission, and/or changing underlying demographics of patients at risk for sepsis,” vs. not explained by changes in case mix (60%; 13.4 percentage points vs. 40%; 8.8 percentage points). Researchers further wrote that the 40% “may be a result of improvements in ICU management.”

A similar pattern of reduction in 2017 to 2019 vs. 1988 to 1990 was observed in median hospital length of stay (14 days vs. 15 days) and median hospital length of stay among survivors (17 days vs. 25 days).

In an additional analysis, researchers assessed the same outcomes in a matched cohort (n = 6,912) and again found lower ICU and hospital mortality in 2017 to 2019 (ICU, 27.3%; hospital, 32.8%) vs. 1988 to 1990 (ICU, 46.4%; hospital, 54.3%). Admittance in more recent years was linked to a reduced adjusted likelihood for hospital mortality (aOR = 0.76; 95% CI, 0.61-0.95).

Between the two time periods in the matched analysis, an adjusted absolute reduction in hospital mortality of 5.8 percentage points (27% of total reduction) was observed.

Researchers also found fewer hospital days among survivors in 2017 to 2019 vs. 1988 to 1990 (median, 19 days vs. 25 days) during this analysis.

A third analysis only factored in patients from ICUs in the same catchment area (n = 27,628), and as previously seen in the two analyses above, those admitted in 2017 to 2019 had fewer deaths vs. those admitted in 1988 to 1990 in the ICU (25.3% vs. 47.2%) and the hospital (32% vs. 54.8%). Admittance in more recent years was linked to a lower adjusted likelihood for hospital mortality (aOR = 0.65; 95% CI, 0.51-0.82).

Between the two time periods in the same-ICU analysis, an adjusted absolute reduction in hospital mortality of 8.7 percentage points (38% of total reduction) was observed.

Similar to the main analysis, patients in 2017 to 2019 vs. 1988 to 1990 had a shorter median hospital stay (14 days vs. 17 days) and a shorter median hospital stay among survivors (17 days vs. 27.5 days).

“Determining which patient, physician and hospital factors were key to mediating the observed mortality reduction was outside the scope of this study but represents an important area of future work,” Prescott and colleagues wrote.