This was a prospective cohort study. We consecutively included patients aged 65 years and older with suspected sepsis (as judged by ED nurses or doctors), admitted to the EDs of two Norwegian general hospitals; Bærum Hospital (BH) Vestre Viken Hospital Trust in the Oslo region, and Haraldsplass Deaconess Hospital (HDH) in Bergen. Patients were included during the periods from October 23rd 2017 to May 14th 2018 at BH and June 18th to October 14th 2018 at HDH. Patients, for whom an International Classification of Diseases 10th Revision (ICD-10) code corresponding to an infectious disease was not registered for the current hospital stay, were excluded from the analysis. A list of the infectious disease ICD-10-codes used for this purpose, is provided as Supplementary material (Supplementary table 1).
Assessments
4AT. The 4 'A's test assesses four core features of delirium [18]. A total 4AT score of 1–3 indicates possible cognitive impairment, and a score of four or above indicates possible delirum. In patients with a 4AT score of 0, delirium or severe cognitive impairment are unlikely. The 4AT consists of four subscores, with each representing one of the four ‘A’s; Alertness, Abbreviated Mental Test – 4 (AMT4), Attention, and Acute change or fluctuating course:1) Alertness (Is the patient fully alert and not markedly drowsy or agitated?). Patients with altered alertness during bedside assessment, are scored with four points. Patients who are not agitated, and fully alert, or have mild sleepiness for less than 10 seconds, are scored with 0 points. 2) AMT4. This abbreviated cognitive assessment tests if the patient is oriented. The patient is asked to tell her/his age, date of birth, the name of the hospital or building, and the current year. While one mistake is scored with one point, and two or more mistankes with two points, correct answers give a score of 0. 3) Attention. The patient is asked to list the months of the year in backward order. Patients who manages less than seven months, are scored with one point. Patients who are not testable due to drowsiness or disturbed attention, are scored with two points. Patients managing seven or more months correctly, are scored with 0. 4) Acute change or fluctuating course. If there is evidence of change or fluctuation in alterness, cognition or other mental functions that have arisen over the last two weeks and are still present within the last 24 hours, the patient is scored with four points. If not, the patient is scored with 0 points. This subscore often requires information from a next of kin. When performed by delirium experts or researchers, the 4AT detects delirium in acutely ill patients with a sensitivity of 76% and a specificity of 95% [12]. We used the Norwegian translation of the 4AT in this study [19]. 4AT was conducted by ED nurses and doctors without any previouis experience with the screening tool, except from a 45-minute introduction lecture. The screening was performed in the ED within two hours of admission. The operating nurse or doctor registered the number of minutes used on the delirium screening, and if they found the screening tool useful or not in each case.
qSOFA. The qSOFA score assesses the risk of poor outcome in patients with infections. Systolic blood pressure ≤ 100 mm Hg, respiratory rate of ≥ 22/min, and altered mental status are scored with one point each. The most recent consensus on sepsis management (Sepsis-3) suggests that one point for altered mental status should be given when the Glasgow Coma Scale score is < 15 [1]. A qSOFA score ≥ 2 should prompt clinicians to further investigate for organ dysfunction, initiate sepsis therapy and consider increased monitoring. A recent meta-analysis reported that the qSOFA predicts mortality with a pooled sensitivity of 42% and a pooled specificity of 88% in ED patients with infections [17].
Delirium. We diagnosed delirium retrospectively during a thorough review of the patients’ hospital records, and according to The Diagnostic and Statistical Manual of Mental Disorders (DSM) 5 [20]. We used a chart-based method aiming to extract evidence for each of the diagnostic criteria, from hospital records, including daily notes by both doctors, nurses, and other staff. This method has been used in previous studies [21], and shows acceptable validity when performed by delirium experts or delirium researchers [22, 23]. We classified cases with evidence of an acute change in altertness, cognition, or other mental functions, as delirium [11].
Sepsis. We defined sepsis according to existing guidelines, as an evident infection, based on the history and a thorough clinical evaluation, including laboratory and radiological findings, with an acute (within 24 hours) change in total Sequential Organ Failure Assessment (SOFA) score of ≥ 2 points consequent to the infection [1]. SOFA grades impairment by organ system in patients with infection, and accounts for clinical interventions [24]. A change in SOFA score ≥ 2 points indicates organ dysfunction [25]. Septic shock was defined as sepsis identified by SOFA with a mean arterial pressure less than 65 mm Hg despite vasopressor therapy, and hyperlactatemia (> 2 mMol/L, 18 mg/dL) after volume resuscitation [1]. We also registered ICD-10 codes for sepsis (A39.2, A40, A41, I33.0, G00, R65) given at discharge from the hospital.
Infection diagnoses. Infections were defined based on ICD-10 codes from the actual hospital stay.
Mortality. In-hospital mortality was defined as death from any cause during the hospital stay.
Other measures. Age, length of hospital stay, and discharge destination were retrieved from hospital records.
Ethical considerations
This was a study without interventions that could harm the patients. Delirium screening with 4AT was rapid and did not delay sepsis treatment.
The HDH part of the study was approved by the Regional Committee of Research Ethics in Western Norway (2018/909/REK vest).
The BH part of the study was approved as a quality study by the hospital officer for data protection (16/00117 − 85).
Since only routine clinical data were collected from the electronic health records, the requirement for informed consent was waived.
A letter with information about the study was sent by post to all patients after discharge, allowing the patient to withdraw their data. The study complies with the Declaration of Helsinki.