Severe sepsis has had a 35-45% mortality rate in clinical trials. Gagan Kumar et al use national observational data to suggest that while population rates of severe sepsis are increasing, survival has likely improved, with mortality falling from 39% to 27%, 2000-2007. However, most of the new survivors are not going home, but rather to long-term acute care and other skilled nursing facilities. (Their outcomes after that are unclear.)
Kumar and team queried the National Inpatient Database (representing ~20% of U.S. hospitals, including community and academic centers) for ICD-9 codes related to severe sepsis. This method showed that discharges with a severe-sepsis-qualifying diagnosis increased from 300,270 in 2000 to 781,725 in 2007. This translated into an increase of 143 per 100,000 to 343 per 100,000, or a 160% increase.
- Mortality fell from 39 to 27% during the observed period, 2000-2007.
- Mean length of hospital stay fell from 17 to 15 days.
- The proportion of people discharged home without assistance remained stable at 20%.
- However, more people (35% in 2007) were discharged to LTACs or nursing homes than in 2000 (27%). Because of the increased number of cases, that meant 276,000 people per year went to skilled nursing facilities, up from 82,000.
I'm a bit skeptical of trends in this kind of data, because they seem highly subject to changes in physician/administrator behavior, culture, and coding practices over time. Authors acknowledge this limitation.
However, I think the most current data is probably highly reflective of reality -- one can only expect coding for severe sepsis to become more accurate over time, as standard diagnostic criteria for severe sepsis have become more widely accepted and practiced (especially by the cohort of younger physicians trained on the SIRS criteria, etc).
If you believe them (as I do), these data are gnarly:
- That was 82,000 discharges to skilled nursing facilities in 2000, and 276,000 per year in 2007.
- 62% of people with severe sepsis died or went to nursing homes/LTACs in 2007, overall.
- No information on what proportion went to LTACs and which to nursing homes, or happened to them after that -- but it would make a big difference in interpreting this data. For instance, we know that one-year survival after after LTAC admission post-critical illness is ~50%. Those authors (Kahn et al in JAMA) suggested LTAC costs were $1 billion / year in 2006.
Age is the major determinant of survival from severe sepsis (did you know that 90% of otherwise healthy young people survive septic shock?), but we only get teasers on that relationship here. Authors report 61% of the total sample were older than 65, and 35% of the total sample got mechanical ventilation. There was no stratification by age on what the mortality rates or discharge destinations were. But we know that among Medicare beneficiaries (~76 years old) undergoing mechanical ventilation (for any cause of respiratory failure), ~84% have died or become significantly disabled after one year.
I'll be most interested to hear some 6- to 12-month data on older folks who survive severe sepsis. With that, we could give these families good, reality-based expectation-setting advice as they make the tough decisions for their loved ones in these stressful and emotional situations.
Kumar G et al (the MICOR investigators). Nationwide Trends of Severe Sepsis in the 21st Century (2000-2007). CHEST 2011;140:1223-1231.