{
  "abstract": "Objectives To evaluate the time-dependent predictive performance of inflammatory markers for 30-day all-cause mortality in patients with suspected infection.Design Retrospective cohort study.Setting The ambulatory care division of the department of general medicine or the emergency department of a single acute care hospital in Japan, from April 2015 to March 2017.Participants The participants were 977 consecutive adult patients with suspected infection defined as those who underwent at least two sets of blood culture.Primary outcome measures The primary outcome, ascertained from electronic medical records, was 30-day all-cause mortality. The predictive performance of three inflammatory markers (C-reactive protein (CRP), neutrophil-to-lymphocyte ratio and red cell distribution width (RDW)) was assessed across four time intervals from symptom onset to hospital presentation: ≤6, >6 to ≤24, >24 to ≤72, and >72 hours.Results Time from symptom onset to hospital presentation influenced the predictive performance of inflammatory markers for 30-day all-cause mortality, and CRP ≥15 mg/dL was the most useful for identifying patients at higher risk of mortality within 6 hours (positive likelihood ratio (LR+) 7.7); however, this association weakened over time (LR+ 1.3 at >72 hours). Conversely, RDW ≥16% became more informative later in the course (LR+ 4.1 at >72 hours). For identifying patients at lower risk of 30-day all-cause mortality, CRP <5 mg/dL showed some utility after 6 hours (negative likelihood ratio (LR–) 0.5–0.7), while RDW <14% was more predictive after 72 hours (LR– 0.3).Conclusion These findings suggest the potential utility of CRP for ruling in high-risk patients within 6 hours from symptom onset, and of RDW for both ruling in and ruling out patients after 72 hours, thereby suggesting that time-dependent variations are important when interpreting inflammatory markers in emergency care settings. However, caution is needed when interpreting these findings because of the retrospective design of this study and potential selection bias.",
  "authors": [
    {
      "affiliations": [
        "Department of General Medicine, Shirakawa Satellite for Teaching And Research (STAR), Fukushima Medical University, Fukushima, Fukushima Prefecture, Japan",
        "Department of Emergency and General Medicine, Tokyo Metropolitan Tama Medical Center, Tokyo, Japan"
      ],
      "name": "Moeko Omiya"
    },
    {
      "affiliations": [
        "Department of General Medicine, Shirakawa Satellite for Teaching And Research (STAR), Fukushima Medical University, Fukushima, Japan"
      ],
      "name": "Toshihiko Takada"
    },
    {
      "affiliations": [
        "Department of General Medicine, Shirakawa Satellite for Teaching And Research (STAR), Fukushima Medical University, Fukushima, Japan"
      ],
      "name": "Tetsuhiro Yano"
    },
    {
      "affiliations": [
        "Department of General Medicine, Shirakawa Satellite for Teaching And Research (STAR), Fukushima Medical University, Fukushima, Japan",
        "Hokkaido Centre for Family Medicine, Sapporo, Hokkaido, Japan",
        "Azai Higashi Clinic, Nagahama, Shiga, Japan"
      ],
      "name": "Kotaro Fujii"
    },
    {
      "affiliations": [
        "Department of General Medicine, Shirakawa Satellite for Teaching And Research (STAR), Fukushima Medical University, Fukushima, Japan"
      ],
      "name": "Ryuto Fujiishi"
    },
    {
      "affiliations": [
        "Department of General Medicine, Shirakawa Satellite for Teaching And Research (STAR), Fukushima Medical University School of Medicine, Fukushima, Japan",
        "Section of Clinical Epidemiology, Department of Community Medicine, Kyoto University Graduate School of Medicine Faculty of Medicine, Kyoto, Japan"
      ],
      "name": "Shunichi Fukuhara"
    }
  ],
  "title": "Time-dependent predictive performance of inflammatory markers for 30-day all-cause mortality in patients with suspected infection in a Japanese emergency department: a retrospective cohort study",
  "uid": "810c6f01-4890-5868-8fd8-699f299a5f3a"
}