Association of multiple sclerosis with mortality in sepsis: a population-level analysis

Abstract

Background: Multiple sclerosis (MS) is associated with increased risk of sepsis and higher sepsis-related mortality, compared to the general population. However, the evidence on the prognostic impact of MS in sepsis has been scarce. We aimed to evaluate the population-level association of MS with short-term mortality in sepsis.

Methods: We performed a retrospective population-based cohort study using a statewide data set to identify hospitalizations aged ≥ 18 years in Texas with sepsis, with and without MS during 2010-2017. Multilevel logistic models were fit to estimate the association of MS with short-term mortality among all sepsis hospitalizations, and for sensitivity analyses among hospitalizations with septic shock and those admitted to ICU.

Results: Among 283,025 sepsis hospitalizations, 1687 (0.6%) had MS. Compared to sepsis hospitalizations without MS, those with MS were younger (aged ≥ 65 years 35.0% vs 56.8%), less commonly racial/ethnic minority (36.2% vs 48.1%), and had lower mean Deyo comorbidity index (1.6 vs 2.7). The rates of septic shock and ICU admission were similar for sepsis hospitalizations with and without MS (58.7% vs 59.6% and 46.7% vs 46.0%, respectively). The unadjusted short-term mortality among sepsis hospitalizations with and without MS for the whole cohort, among those with septic shock, and among ICU admissions were 20.2% vs 31.3%, 25.6% vs 40.0%, and 24.0% vs 34.8%, respectively. On adjusted analyses, MS was associated with 17% lower odds of short-term mortality (adjusted odds ratio [aOR] 0.828 [95% CI 0.723-0.947]). Similar findings were observed on sensitivity analyses of patients with septic shock (aOR 0.764 [95% CI 0.651-0.896]), but MS was not associated with mortality among sepsis hospitalizations admitted to ICU (aOR 0.914 [95% CI 0.759-1.101]).

Conclusions: MS was associated with lower short-term mortality among septic patients, with findings consistent among the subset with septic shock. Among septic patients admitted to ICU, MS was not associated with mortality.

Keywords: Intensive care unit; Mortality; Multiple sclerosis; Sepsis; Septic shock.

Fig. 1

Subgroup analyses of short-term mortality of sepsis hospitalizations with and without multiple sclerosis. Analyses were done using multilevel logistic regression. The odds ratios and 95% confidence intervals have not been adjusted for multiplicity and should not be used to infer definite effects. ^aMS indicates multiple sclerosis