Demographics, epidemiology, mortality and difficult-to-treat resistance patterns of bacterial bloodstream infections in the global US Military Health System from 2010 to 2019: a retrospective cohort study

Abstract

Objective: To describe demographics, causative pathogens, hospitalisation, mortality and antimicrobial resistance (AMR) of bacterial bloodstream infections (BSIs) among beneficiaries in the global US Military Health System (MHS), a single-provider healthcare system with 10-year longitudinal follow-up.

Design: Retrospective cohort study.

Setting: Clinical and demographic data collected from the MHS Data Repository and collated with microbiological data obtained from the Defense Centers for Public Health-Portsmouth. Participants: 12 748 MHS beneficiaries diagnosed with 15 357 bacterial BSIs (2010-2019).

Main outcomes and measures: Demographic data and diagnosis codes preceding BSI episodes and during hospitalisations were collected. Inpatient admission data identified acute clinical diagnoses, intensive care unit (ICU) admission and mortality. BSI pathogens were evaluated for AMR, including difficult-to-treat resistance (DTR). Crude mortality trends were assessed.

Results: The decade analysed included 15 357 BSI episodes in 12 748 patients; 6216 patients (48.8%) were≥65 years and 83.7% of episodes had≥1 comorbidity (12 856 of 15 357). Approximately 29% of episodes with hospitalisation required ICU admission and∼34% had concurrent urinary tract infections. Pathogen distribution was 53% and 47% for Gram-positive bacteria and Gram-negative bacilli (GNB), respectively. Inpatient mortality was 4.4%, and at 1 year was 23.4%; 0.5% (16 of 2977) of deaths were associated with DTR GNB. Among an average 8 145 778 individuals receiving care annually in the MHS, annual rates of overall BSI, methicillin-resistant Staphylococcus aureus, vancomycin-resistant Enterococcus spp and DTR GNB BSI were 18.9, 1.30, 0.25 and 0.05 per 100 000 beneficiaries, respectively. Over the decade, annual mortality did not significantly increase for any pathogen and decreased by∼2% for overall BSI (p=0.024) and∼3% for lactose-fermenting GNB BSI (p=0.048).

Conclusions: In the global US MHS, the mortality burden associated with BSI was substantial (approximately one in four dying at 1 year), relatively unchanged over a decade, and associated with older age and comorbidities. First-line treatment options remained available for 99.7% of BSIs. Population-level improvements in BSI survival might be maximally influenced by focusing on prevention, early detection, prompt antibiotics and other novel therapies not contingent on in vitro activity.

Keywords: Adult intensive & critical care; Epidemiology; MICROBIOLOGY.

Figure 1. Flow diagram for: (a) Bloodstream infection (BSI) episodes diagnosed between 2010 and 2019 included in the analysis. (b) Pathogens and antimicrobial susceptibility testing. ‘Top 5’ represents the most numerous and clinically relevant pathogens (ie, E. coli, Acinetobacter spp, K. pneumoniae, P. aeruginosa and Citrobacter spp). Pathogens not in the ‘Top 5’ include Stenotrophomonas spp and Enterobacter spp. GN pathogens assessed for DTR were those that had susceptibility testing for three classes of first-line antibiotics (ie, carbapenems, extended-spectrum beta-lactams or fluoroquinolones). Antibiotic resistance information for GP organisms is in online supplemental figure 2. DTR, difficult-to-treat resistance; GN, Gram-negative; GP, Gram-positive; MHS, Military Health System; QA, quality assurance.

Figure 2. Bacteria associated with bloodstream infections (BSI). (a) Distribution of total bacteria; (b) BSI rate per 100 000 Military Health System (MHS) beneficiaries per year by classification group. There were no DTR Gram-positive bacteria. DTR, difficult-to-treat resistance; GNB, Gram-negative bacilli.

Figure 3. Total number of (a) lactose-fermenting Gram-negative bacilli and (b) non-lactose-fermenting Gram-negative bacilli bloodstream infection (BSI) episodes and deaths, including occurrence of difficult-to-treat resistance (DTR). Of the 16 DTR BSI episodes with deaths, one patient had two different DTR GNB (Pseudomonas aeruginosa and Klebsiella pneumoniae) on the same day, and one patient had a polymicrobial BSI (DTR P. aeruginosa plus three susceptible pathogens: Enterococcus spp, K. pneumoniae and Staphylococcus aureus). GNB, Gram-negative bacilli.