Infectious Diseases

Predictors for mortality in patients admitted with suspected bacterial infections – A prospective long-term follow-up study

a b s t r a c t

Objective: The aim was to examine predictors for all-cause mortality in a long-term follow-up study of adult pa- tients with infectious diseases of suspected bacterial origin.

Methods: A prospective observational study of patients admitted to the emergency department during 1.10.2017-31.03.2018. We used Cox regression to estimate adjusted Hazard ratios (aHR) with 95% confidence in- tervals for mortality.

Results: A total of 2110 patients were included (median age 73 years). After a median follow-up of 2.1 years 758 (35.9%, 95% CI 33.9-38.0%) patients had died. Age (aHR1.05; 1.04-1.05), male gender (aHR 1.21; 1.17-1.25), can- cer (aHR 1.80; 1.73-1.87), misuse of alcohol (aHR 1.30; 1.22-1.38), if admitted with sepsis within the last year before index admission (aHR 1.56;1.50-1.61), a Sequential Organ Failure Assessment score >=2 (aHR 1.90; 1.83-1.98), SIRS criteria >=2 (aHR 1.23;1.18-1.28) at admission to the ED, length of stay (aHR 1.05; 1.04-1.05) and devices and implants as sources of infection (aHR 7.0; 5.61-8.73) were independently associated with mortality. skin infections and increasing haemoblobin values reduced the risk of death.

Conclusions: More than one-third of a population of patients admitted to the emergency department with infec- tious diseases of suspected bacterial origin had died during a median follow up of 2.1 years. The study identified several independent predictors for mortality.

(C) 2022 Published by Elsevier Inc.

  1. Introduction

Infectious diseases are common diagnoses presenting in the Emer- gency department (ED) [1,2] and are a leading cause of death world- wide [3]. The ED visit rate due to infectious diseases has increased in the United States by more than a third between 1997 and 2007 [4]. In- fectious disease may lead to potentially lethal conditions such as sepsis and other poor outcomes in certain groups of patients, particularly in patients with a compromised immune system, older people, and in pa- tients with comorbidities [5,6].

A system that aims to quantify and characterize the Epidemiological features of patients suffering from infectious disease is important to im- prove national surveillance, guide triage, and Treatment decisions, opti- mize resource utilization, and facilitate quality assurance measurements ultimately to address preventable morbidity and mortality. Such a system should identify the most at-risk patients and identify a low-risk subgroup that may not require high Levels of care.

* Corresponding author at: Aalborggade 9, 1. tv., 2100 Copenhagen, OE, Denmark.

E-mail addresses: [email protected] (L. Chafranska), [email protected] (R.H. Sorensen), [email protected] (F.E. Nielsen).

Several studies [5-11] have examined predictors for long-term mor- tality following sepsis, septic shock, or bacteraemia, and in patients with specific infectious diseases such as community-acquired pneumonia [12,13]. However, to our knowledge, there are no studies examin- ing predictors for mortality in long-term follow-up studies of adult pa- tients admitted to the ED with bacterial infectious diseases.

The aim of the present long-term follow-up study was, therefore, to examine predictors for mortality among adult ED patients admitted with a broad spectrum of common infectious diseases of suspected bac- terial origin.

  1. Patients & methods

This study was reported according to the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) guidelines [14].

    1. Study design and settings

This is a secondary analysis of a single-centre prospective observa- tional cohort study of the Prognostic ability of qSOFA [15] among 0735-6757/(C) 2022 Published by Elsevier Inc.

adult (>=18 years) patients admitted with suspected bacterial infections to the ED at Slagelse Hospital, Denmark, during the period October 1, 2017, to March 31, 2018. Slagelse Hospital is a tertiary care centre with acute medical, surgical, and trauma care. The ED has an uptake area of 198,000 adult inhabitants with approximately 26,500 visits an- nually. The Danish health care system offers equal access for all resi- dents. Most patients are referred to the ED directly by their general practitioner or arrive at the ED by an emergency ambulance without any preceding contact. Privately funded Danish hospitals have no acute patient intake [16].

    1. Selection of patients with infection

Triage records and electronic patient records for all patients admit- ted to the ED during the study period were every working day during the six-month study period reviewed by the authors to include patients fulfilling the criteria for infections [15]. Initially, all patients with either documented or suspected infection diagnosed by the ED physician and patients who were administered intravenous or peroral antibiotics within 24 h from arrival were registered as patients with infectious dis- eases of suspected bacterial origin [15].

    1. Inclusion and exclusion criteria

In the present study, we included patients with either suspicion of infection or documented infection and treated with antibiotics within 24 h after presenting to the ED and continuing the treatment for at least 48 h. Patients discharged within 48 h after receiving antibiotics in the ED and continuing the treatment at home were also included. Ex- clusion criteria were: Foreign nationality without a Danish Civil Regis- tration number, if antibiotic treatment was discontinued within 48 h, Prophylactic antibiotic treatment in relation to surgery, registration er- rors, transfer to another hospital within 24 h, and patients that had pre- viously been included during the study period [15].

    1. Definitions

To classify patients with chronic diseases, we used the Charlson Co- morbidity Index (CCI) [17]. The classification was based on available in- formation in the electronic patient records. It was graded in three levels of severity based on the weighted sums of 19 diseases included in the CCI classification system: low (CCI score 0), moderate (CCI score 1-2) and high (CCI score >= 3).

The definitions of Quick Sequential Organ Failure Assessment and Systemic Inflammatory Response Syndrome followed the original guidelines [18-20]. The total Sequential Organ Fail- ure (SOFA) score [20,21] was adjusted for chronic diseases with a po- tential impact on the score in our study. The method behind this adjustment has previously been described [22]. Patients without chronic diseases had a baseline SOFA score of zero. Patients with known dementia had a baseline SOFA score of 1. The adjustments for other diseases were based on knowledge of the level of chronicity (mild, moderate, or severe kidney and liver disease) from the CCI classi- fication and the arrival creatinine and bilirubin values. The adjustment for chronic pulmonary disease was based on information on pulmonary disease according to the CCI classification and if different grades of de- creased arrival PaO2 values at the ED were assessed as chronically re- duced [22].

new-onset atrial fibrillation was defined as episodes of atrial fibrilla- tion documented on a 12-lead electrocardiogram on admission and without a history of prior atrial fibrillation.

    1. Data collection

All patients admitted to the ED during the study period were triaged by a nurse at admission according to a standardized procedure. This

included the assessment of Vital parameters, which were registered in triage forms.

The authors reviewed the electronic records and triage forms on the following working day after the index admission. The triage forms and electronic patient records provided information on demographic data, comorbidities, triage variables, admission laboratory tests, and other ex- aminations. This included age, gender, information on comorbidities (CCI), medical treatment before admission, time of arrival to the ED, mental state (Glasgow Coma Scale (GCS)) or Alert Voice Pain Unrespon- sive (AVPU), systolic blood pressure, respiratory rate (RR), heart rate, body temperature, Leucocyte count, peripheral oxygen saturation, C- reactive protein (CRP), creatinine, bilirubin, platelet count, lactate, glu- cose, results of blood cultures obtained on admission, rhythm on ECG, source of infection, transfer to Intensive Care Unit (ICU), treatment in the ICU, and time of discharge from the hospital. Information on sources of infections was based on a review of all records at discharge with spe- cific details on infectious sources diagnosed and documented in the re- cords by the physicians during the hospital stay. Foci of the infection were specified by bacterial culturing of possibly infected tissues and body fluids.

If patients required hospitalization for more than 48 h, they were transferred from the ED to a medical ward. Critically ill patients were transferred to the ICU.

Information on death within the follow-up period was obtained from The Regional Zealand Patient Registration System, which is linked to the Danish Civil Registration System with daily updated information on the vital status of all Danish citizens [23].

We performed a 3-day pilot study before the initiation of the study to ensure the collection of all data defined in the study protocol and that the information needed was available in a form suitable for abstrac- tion [15]. The data collection and data entry process were randomly controlled by the authors (SMOBA, RHS). Further on, the researchers met regularly to discuss and clarify disputes regarding the collected data and analyses [15].

The data were entered into an electronic database. The data collec- tion and data entry were controlled by the authors regularly.

    1. Statistical analysis

The primary outcome was mortality (time to death) within the long- term follow-up period. Continuous data are presented as medians with interquartile ranges (IQR). Proportions are presented with 95% confi- dence intervals (CI). We followed patients from the date of discharge from the index admission until the end of the follow-up period, emigra- tion or death, whichever came first.

Unadjusted and adjusted HRs (aHR) with 95% confidence intervals

(CI) were generated from univariate and multivariate Cox proportional hazards models, respectively. The candidate predictor variables in the multivariate Cox regression model (age, gender, misuse of alcohol, med- ication, chronic diseases, previous sepsis admission, SOFA, SIRS, atrial fi- brillation, length of stay, laboratory variables, sites of infection) included in the initial model were determined by the increase of Hazard Ratio (HR) in the crude analyses. We chose to include the specific chronic dis- eases in the regression models instead of CCI. However, since chronic kidney, liver, and pulmonary diseases were considered in the calcula- tion of the SOFA scores, we did not include these diseases in the final re- gression model. The validity of the proportional hazard assumption was verified using Schoenfelds and scaled Schoenfelds residuals. The statis- tic minus twice the logarithm of maximized likelihood, -2logL, was used to compare different models fitted to the data. The different models were compared by examining the change in the value of

-2logL on adding terms into a model or deleting terms from the model. Variables deleted from the model during the selection procedure were finally included in the model one by one to examine if they signif- icantly changed the value of -2logL.

Cox-Snell residuals, estimated based on Martingale residuals, were used to check the overall model fit. The final model check was done by using the Cox-Snell residuals as the time variable in a plot along with the Nelson-Aalen cumulative hazard function. The model fits the data if the cumulative hazard plot versus Cox-Snell residuals is a straight line with slope 1 (45-degree line).

We used multiple imputations by the Markov Chain Monte Carlo procedure under missing at random assumptions to impute missing laboratory variables in the regression analyses [24]. The imputation model using ten imputed datasets included the outcome variable (long term death) and the covariates: age, gender, chronic diseases (cancer, heart failure, ischemic heart disease, diabetes, and cerebrovas- cular disease) if admitted with sepsis within the last year before index admission, alcohol consumption, SOFA, SIRS, atrial fibrillation, length of stay, sites of infection and laboratory variables. Stata 15.1 and 17.0 SE (StataCorp, College Station, Texas, USA) was used for all analyses.

  1. Results
    1. Population

A total of 12,092 patients were admitted during the study period and 3176 patients were treated with antibiotics at arrival or within 24 h after arrival [15]. A total of 1066 patients were excluded due to foreign status (n = 6), prophylactic antibiotic treatment in relation to surgery (n = 240), discontinuation of antibiotic treatment within 48 h after ad- mission (n = 174), missing data (n = 56) and registration error (n = 12), transfer to or from other hospitals (n = 224), and previously in- cluded in the study (n = 354) [15] leaving an infection cohort of 2110 ED patients (51.3% female) with a median age of 73.1 years (IQR 60.2-82.7).

    1. Baseline characteristics

Heart failure (10.7%), ischemic heart disease (11.0%), cerebrovascu- lar diseases (13.9%), chronic obstructive pulmonary disease (26.0%), di- abetes (16.8%) malignancies (13.3%), and hypertension (32.5%) were relatively common chronic diseases in the cohort (Table 1). About one-fourth had been admitted due to sepsis within the last year before the index admission. A qSOFA score >=2, SOFA score >=2, and SIRS >=2 at ad- mission to the ED was found in 8.2%, 33.8%, and 47.9%, respectively (Table 1). The median CRP value was 65 umol/L (Table 1).

The most common site of infection was the lungs (52.5%), followed

by urinary tract infections (25.7%) and Abdominal infections (10.7%) (Table 2). A total of 144 (6.8%) had at least two foci of infection (Table 2). A total of 1825 (86.5%) were examined with x-ray, computed tomography, or ultrasound, and 801(43.9%) of the examinations re- vealed a focus of infection (Table 2).

Blood cultures were drawn in 53.0% (11.1% positive). (Table 2). Almost 85% were treated with at least one medication before admis-

sion, and 27.6% were treated with at least five medications (Table 3). A total of 434 (20.6%) patients had antibiotics prescribed in the primary care sector (e.g. family physician/ general practitioner) before admis- sion (Table 3).

    1. Outcomes

After a median follow-up of 753 days (2.1 years) (IQR 340-821 days), a total of 758 (35.9%; 33.9%-38.0%) patients had died. In- hospital and 28-day mortality was 3.7% (2.9%-4.5%) and 7.5% (6.4%-

8.7%), respectively, and 365-day mortality was 25.6% (23.8%-27.5%)

(Table 4). A total of 155 (7.4%; 6.3%-8.5%) were transferred to the ICU (Table 4). The median length of stay was 4.4 days (IQR: 2.0-7.9).

Table 1

Baseline characteristics among patients admitted to the emergency department with suspected bacterial infections.

Age, median years (IQR) 73.1(60.2-82.7)

Female gender, n (%) 1083 (51.3)

Comorbidities, n (%)

Charlson Comorbidity Index


662 (31.4)


992 (47.0)


456 (21.6)

Congestive Heart failure

226 (10.7)

Ischemic heart disease1

231 (11.0)

Cerebrovascular disease2

293 (13.9)

Chronic pulmonary disease

548 (26.0)

Diabetes mellitus3

354 (16.8)


281 (13.3)

Chronic kidney disease5

118 (5.6)

Chronic mild or severe liver disease

35 (1.7)


686 (32.5)


74 (3.5)

Admitted with sepsis within last year, n (%)

572 (27.1)

History of alcohol misuse, n (%)

Smoking, n (%)6

119 (5.6)

No smoking

627 (39.8)

Current smoker

395 (25.1)

A history with smoking

552 (35.1)

Risk scores upon admission, n (%)

q-SOFA 0

1211 (57.4)


725 (34.4)


174 (8.2)



882 (41.8)


514 (24.4)


714 (33.8)



497 (23.6)


603 (28.6)


1010 (47.9)

Antibiotic treatment, n (%)

Intravenous 1614 (76.5)

Peroral 496 (23.5)

Atrial fibrillation, n (%)


1695 (80.3)

A history of atrial fibrillation

334 (15.8)

New onset atrial fibrillation

83 (3.9)

Length of stay; days, median (IQR)

Laboratory results on admission, median (IQR)

Haemoglobin (mmol/L)7

4.4 (2.0-7.9)

8.1 (7.2-8.8)

White blood cell count (x10/L)8

11.1 (8.4-14.7)

C-reactive protein (mg/L)9

65 (20-135)

Platelets (x 109/L)10

242 (189-316)

Potassium (mmol/L)11

3.9 (3.6-4.2)

Creatinine (umol/L)12

82 (63-113)

Bilirubin (mmol/L)13

9 (6-13)

Lactate (mmol/L)14

1.2 (0.8-1.9)

Glucose mmol/L)15

6.6 (5.8-8.1)

ICD, Implantable cardioverter defibrillator; IQR, interquartile range; qSOFA, quick sequen- tial organ failure assessment; SIRS, systemic inflammatory response syndrome; SOFA, se- quential organ failure assessment.

1 Myocardial infarction or other ischemic heart disease prior to admission.

2 A history of cerebrovascular disease including transient ischemic attacks.

3 Uncomplicated or with end-organ damage.

4 Included cancer with and without metastasis and hematologic malignancies.

5 Moderate or severe chronic kidney disease.

6 Smoking unknown for 536 patients.

7 11 missing.

8 5 missing.

9 11 missing.

10 26 missing.

11 20 missing.

12 10 missing.

13 66 missing.

14 1,336 missing (measurement of lactate only a routine test in patients with suspicion of sepsis).

15 97 missing.

Microbiological findings, results of imaging examinations and sites of infection.

Outcomes among 2110 patients admitted to the emergency department with suspected bacterial infections.

Blood cultures

n (%)

n (%; 95% CI)

Number blood cultures obtained 1119 (53.0)

positive blood cultures 124 (11.1)

Other positive cultures/total cultures obtained1

Expectorate 94/285

Urine 360/971

Faeces 5/64)

Skin 33/77

Cerebrospinal fluid 3/11


Primary outcome:

Mortality within long-term follow-up period1 758 (35.9; 33.9-38.0)

In-hospital 77 (3.7; 2.9-4.5)

28-day 158 (7.5;6.4-8.7)

365-day 541 (25.6;23.8-27.5)

Transfer to Intensive Care Unit; n (%; 95% CI) 155 (7.4;6.3-8.5)

CI, confidence interval.


Foci of infection2


1115 (52.5)


545 (25.7)


228 (10.7)


196 (9.2)


9 (0.4)

Central nervous system

9 (0.4)


4 (0.2)


11 (0.5)


9 (0.4)


145 (6.8)

2 infection foci

139 (6.6)

3+ infection foci

Infiltrates/abscesses on imaging exams1

5 (0.2)

X-ray (thorax)


Computed tomography (CT) scan





1 Median 2.1 years.

1 Number of abnormal findings/total tests or examinations performed.

2 Some patients had more than one site of infection.

    1. Regression models
      1. Univariate analysis

The crude regression analysis (Table 5) showed that age, male gen- der, CCI 1-2 and CCI 3+, different chronic diseases, number of medica- tions, tobacco consumption, and if admitted with sepsis within the last

Table 3

Medical treatment among patients admitted to the emergency department with suspected bacterial infections.

n (%)


Antibiotics 434 (20.6)



467 (22.1)

Calcium antagonists

370 (17.5)


85 (4.0)


619 (29.3)


879 (41.7)


726 (34.4)

Lipid lowering drugs

536 (25.4)


Inhalation medications 483 (22.9)


Insulin/per oral antidiabetics 294 (13.9)

Psychopharmacology medication3 342 (16.2)


Opioids 357 (16.9)

Other analgesics4 652 (30.9)

Number of medications

0 357 (16.9)

1-2 530 (25.1)

3-4 642 (30.4)

>=5 581 (27.6)

1 acetylsalicylic acid, Non-vitamin K antagonist oral anticoagulants, Warfarine/ Dicoumarol.

2 Loop-acting diuretics, Potassium-sparing diuretics, Thiazide diuretics.

3 Antipsychotics, Lithium, anti-depressives, benzodiazepines.

4 Paracetamol, caffeine-phenazon, codeine, Non-steroidal anti-inflammatory drugs, others.

year before the index admission were factors associated with increased risk of mortality. Misuse of alcohol increased the risk of death insignifi- cantly. Increasing scores of qSOFA, SOFA, and SIRS at admission to the ED, atrial fibrillation, length of stay, elevated white blood cell count, po- tassium, and the lactate level at admission were also associated with mortality in unadjusted analyses (Table 5). The lungs and unknown foci of infections were associated with increased risk of mortality, while abdominal and skin infections and increasing haemoglobin values reduced the risk (Table 5).

Urinary tract infections and a positive blood culture increased the risk of death insignificantly (Table 5). Devices and implants as infectious foci were also associated with mortality (Table 5).

      1. Multivariate analysis

Age (aHR 1.05; 1.04-1.05), male gender (aHR 1.21; 1.17-1.25), a his- tory of cancer (aHR 1.80;1.73-1.87), admission with sepsis within the last year before index admission (aHR 1.56; 1.50-1.61) and a history of alcohol misuse (aHR 1.30; 1.22-1.38) were independently associated with mortality. A SOFA score >=2 (aHR 1.90; 1.83-1.98), SIRS criteria >=2 (aHR 1.23; 1.18-1.28) and length of stay 1.05(1.04-1.05) were associ- ated with death (Table 6). Devices and implants increased the risk of mortality substantially (aHR 7.00; 5.61-8.73). Skin infections (aHR 0.62;0.58-0.66) and increasing values of haemoglobin (aHR 0.86; 0.85-0.87) reduced the risk of death (Table 6).

The cumulative hazard plot of the Cox-Snell residuals is shown in Fig. 1. The hazard function is a reasonably straight line that approxi- mates the 45-degree line closely except for large values of time. Based on this finding we conclude that the final regression model (Table 6) fits the data well.

  1. Discussion

After a median follow-up of 2.1 years, we found that 35.9% of unse- lected ED patients with suspected bacterial infectious diseases had died. High age, male gender, history with cancer, previous admission with sepsis, alcohol misuse, patients fulfilling at least two SOFA or SIRS criteria at admission to the ED, devices, and implants as the source of infection, and length of hospital stay were identified as predictors of mortality. Skin infections and increasing values of haemoglobin were associated with a reduced risk of death.

The literature on ED patients with infectious disease, long-term mor- tality, and predictors for mortality is sparse. A study by Murray et al. [25] found a one-year mortality rate of 22% in 3102 ED patients (median age 61 years) admitted with suspected infections. Ittisanyakorn et al. [26] assessed the long-term mortality among 463 infected elderly (median age 78 years) ED patients. They found a one-year mortality rate of 39.1%, a considerably higher one-year mortality than our study with a lower median age. Goto et al. [27] examined a nationwide registry- based sample of more than three million infectious disease-related ED visits of elderly (> 65 years) patients. They found in-hospital mortality of 4%, similar to the findings in our study.

Table 5

Crude hazard ratios for mortality within the long-term follow-up period.

Table 6

predictive variables in a Cox regression model of mortality within a long-term follow-up period of patients admitted to the ED with suspected bacterial infections.

Hazard ratio (95% CI)

Age 1.05(1.04-1.06)

Male gender 1.22 (1.06-1.41)


Charlson Comorbidity Index

0 Reference

1-2 2.55 (2.05-3.16)

3+ 5.60 (4.49-6.99)

Congestive Heart failure 1.69 (1.39-2.07)

Ischemic heart disease1 1.66 (1.36-2.02)

Cerebrovascular disease2 1.71 (1.43-2.04)

Chronic pulmonary disease 1.22 (1.05-1.43)

Diabetes mellitus3 1.33 (1.11-1.60)

Malignancy4 2.53 (1.14-3.00)

Chronic kidney disease5 1.76 (1.40-2.28)

Chronic mild or severe liver disease 1.90 (1.23-2.92)

Hypertension 1.11 (0.96-1.29)

Pacemaker/ICD 1.46 (1.04-2.06)

Admitted with sepsis within last year 1.86 (1.56-2.09)

History of alcohol misuse 1.26 (0.95-1.68)


No smoking Reference

Current smoking 1.02 (0.82-1.28)

A history with smoking 1.29 (1.06-1.56)

Risk scores upon admission

q-SOFA score

0 Reference

1 1.84 (1.58-2.15)

>= 2 3.25 (2.61-4.03)

SOFA score

0 Reference

1 1.48 (1.21-1.80)

>=2 2.51 (2.12-2.97)

SIRS criteria

0 Reference

1 1.27 (1.03-1.56)

>= 2 1.29 (1.07-1.56)

Atrial fibrillation

None Reference

A history of atrial fibrillation 1.73 (1.46-2.06)

New onset atrial fibrillation 1.88 (1.38-2.58)

Length of stay 1.04 (1.03-1.04)

Laboratory results

Haemoglobin7 0.74 (0.70-0.78)

White blood cell count8 1.02 (1.01-1.04)

C-reactive protein9 0.99 (0.99-1.00)

Platelets10 1.00 (1.00-1.00)

Potassium11 1.05 (1.02-1.07)

Creatinine12 1.00 (1.00-1.00)

Bilirubin13 1.00 (1.00-1.01)

Lactate 14 1.06 (1.03-1.10)

Glucose 15 1.02 (1.00-1.03)


Positive blood cultures 1.26 (0.94-1.69)

Sites of infection

Lungs 1.23 (1.07-1.43)

Urinary 1.14 (0.98-1.34)

Abdominal 0.63 (0.48-0.83)

Skin 0.50 (0.37-0.69)

Endocarditis 0.91 (0.29-2.82)

Central nervous system 0.28 (0.04-1.98)

Devices/implants 3.73 (1.20-11.60)

Facial/Teeth 0.21 (0.03-1.51)

Others 1.78 (0.74-4.28)

Unknown 1.33 (1.03-1.72)

>1 foci of infection 1.17 (0.90-1.53)

Number of medications

0 Reference

1-2 2,63 (1.94-3.57)

3-4 3.18 (2.36-4.27)

>=5 3.67 (2.74-4.94)

CI, confidence interval; qSOFA, quick sequential organ failure assessment; SIRS, systemic inflammatory response syndrome;

SOFA, sequential organ failure assessment.

1-15 See footnotes in Table 1.

Hazard ratio (95% CI)

Age1 1.05 (1.04-1.05)

Male gender 1.21 (1.17-1.25)

A history of cancer 1.80 (1.73-1.87)

Admitted with sepsis within last year 1.56 (1.50-1.61)

History of alcohol misuse 1.30 (1.22-1.38)

Risk scores upon admission:

SOFA score2

0 Reference

1 1.09 (1.04-1.14)

>=2 1.90 (1.83-1.98)

SIRS criteria

0 Reference

1 1.17 (1.12-1.22)

>=2 1.23 (1.18-1.28)

Length of stay (days)1 1.05 (1.04-1.05)

Devices/implants3 7.00 (5.61-8.73)

Skin infection 0.62 (0.58-0.66)

Haemoglobin value on arrival1 0.86 (0.85-0.87)

CI, confidence interval; SOFA, sequential organ failure assessment; SIRS, systemic inflam- matory response syndrome.

1 Change of hazard ratio per unit increase of variable.

2 Adjusted for chronic diseases (kidney, respiratory, liver, dementia) with impact on SOFA calculations.

3 Devices/implants (n = 4) were considered to be the cause of infection and three out of the four patients had died during long-term follow up.

Age is a predictor for long-term mortality following ED admission for an infection, supported by the results from the study by Ittisanyakorn et al. [26] and is in accordance with the existing literature [28-30]. El- derly have an increased risk of invasion by pathogenic organisms due to alterations of the barriers posed by skin, lungs, and the gastrointesti- nal tract (and other mucosal linings). Furthermore, the elderly suffers from more comorbidities (e.g., diabetes, chronic obstructive pulmonary disease, or heart failure), resulting in greater impairment in immunity and insufficient vaccine response [31,32]. All these factors may contrib- ute to a more life-threatening infection in the elderly and poorer long- term outcomes.

We found higher mortality among men compared to females. Mor- tality is higher for men than for women of all ages [29]. This is partly due to behavioural differences and men having an earlier onset of chronic diseases. Men are also more likely to suffer from lethal

Image of Fig. 1

Fig. 1. Cumulative hazard plot of the Cox-Snell residuals of the final proportional hazards Cox regression model.

conditions like heart disease, stroke, and diabetes [33]. Furthermore, differences in health-seeking behaviours and compliance between men and women may, to some extent, explain why men have higher mortality rates than women [34].

Multimorbidity generally leads to a decreased quality of life and an increased risk of hospitalisations and mortality [35]. Similar to Murray et al. [25], we found that a higher burden of comorbidities (CCI >=1) was associated with long-term mortality in the crude regression analy- sis. This is comparable to the study by Ittisanyakorn et al. [26], who found that a CCI score of >=5 predicted one-year mortality among elderly ED patients admitted with infection.

Among the chronic diseases included in the multivariate regression model, we found that history of cancer was independently associated with mortality. A history of cancer was relatively common in our study population. Our results indicate that cancer was the most impor- tant risk factor of death compared to other chronic diseases included in the CCI.

Camou et al. [36] similarly found that cancer patients admitted to the ICU with septic shock had an increased risk of 180-day mortality com- pared to non-cancer patients.

Previous sepsis admission within one year before index admission was an independent factor for mortality. Studies have shown that at prior admission for sepsis was associated with an increased risk of read- mission [37,38]. Some studies have found an increased risk of mortality in the post-discharge period partly due to compromised homeostasis, pulmonary and renal complications and secondary infection [39-41].

Misuse of alcohol was independently associated with mortality. Nu- merous studies have pinpointed that alcohol misuse leads to higher long-term mortality among patients hospitalized for different condi- tions; sepsis and septic shock in the ICU [42], severe bloodstream infec- tions [43], and CAP [44,45]. This patient group is more susceptible to getting infected, have impaired host response to infections and a worse prognosis following infections [46].

Although the number of patients with devices or implants was small and the estimates, therefore, are imprecise, it is remarkable that three out of four patients died within the follow-up period and that the con- dition was independently associated with death. Sohail et al. [47] like- wise discovered that patients with cardiovascular implantable electronic device infections had increased device-dependent long- term mortality even after successful infection treatment. De Bie et al.

[48] found that cardiac device infections were responsible for more than twofold increased mortality than patients who remained free from infection. Another study by Iordanou et al. [49] found a crude ICU mortality rate of 40% for the patients who acquired device- associated infections and 17.9% for patients who did not. Catheter- associated urinary tract infections have also been associated with higher mortality [50].

A SOFA score >=2 and SIRS >=2 was independently associated with mortality. SOFA determines the severity of organ dysfunction and is a key criterion for sepsis [51]. A SOFA score of >=2 identifies infectious pa- tients with a significantly increased risk of poor outcomes [20,22,51-53]. Although SIRS is obsolete as a tool to identify patients with sepsis [22,51-55], SIRS may still be used as an important prognosticator in re-

lation to mortality.

Length of stay was also independently associated with long-term mortality, which is consistent with previous findings among sepsis sur- vivors [30]. Unfortunately, we cannot conclude on causes of prolonged length of stay among patients in our study. However, we assume that the patients with longer lengths of stay are more severely ill, and that may explain the increased risk of death.

Increasing values of haemoglobin were associated with a reduced risk of mortality. Anaemia is common among the Geriatric population and may indicate serious chronic diseases such as cancer or a specific treatable condition. A cohort study by Culleton et al. [56] concluded that anaemia was associated with increased risk for hospitalization and mortality in older adults. A study by Dunne et al. [57] found that

perioperative anaemia was associated with increased mortality, in- creased postoperative pneumonia, and increased hospital length of stay. Interestingly, a study by Purtle et al. [58] showed that an elevated red cell distribution width, which can be correlated to anaemia, at the time of discharge was a robust predictor of subsequent all-cause patient mortality in patients treated with critical care who survive hospitaliza- tion. Investigations of the associations between different types of anae- mia and mortality among infected patients is warranted to identify potentially modifiable factors associated with death.

Skin infections were associated with a reduced risk of death. Sub

analyses showed that the proportion of patients with a skin infection and signs of organ failure (SOFA >=2) was significantly lower (24.5% vs. 34.8%, 95% CI on difference 3.9-10.7) compared to all other infections with organ failure. A better outcome seen in this patient group may be due to the visibility of skin infections making early identification and treatment of the infection focus possible.

  1. Implications

We have identified several independent predictor variables for mor- tality within a long-term follow-up period. Our results are important for both clinicians and patients. We are aware that we have not presented a Predictive performance model with estimations of how accurately our prediction model can predict mortality. It may therefore be challenging to apply the findings in our study clinically. However, our results re- garding determinants for death should be used in new and bigger test data sets using more advanced statistical and machine learning algo- rithms for prediction, e.g. Random forest analyses, and with estimations of absolute risk for individual patients. Such data may be important to guide clinical decision-making and risk stratification of the patients. In this perspective, our data are very useful.

Our study has also identified several interesting determinants of death, including male gender, alcohol misuse, and devices. The causality and the way in which these individual factors contribute to death should be further explored in future research for the benefit of both cli- nicians and patients. If modifiable factors with impact on mortality could be defined, it may be possible for clinicians to identify high-risk patients prior to discharge to ensure optimal management and follow- up of these patients aiming to reduce the risk of death. However, until such data are available, we recommend clinicians to be aware of the high-risk patients we have identified and to offer the patients timely follow-up.

  1. Strengths and limitations

The main strength of this study was its prospective design, which minimizes the risk of bias and missing data. The study was designed to identify all infected patients consecutively admitted to our ED, and a complete follow-up was ensured by linkage of the patient’s unique personal identification number and the Danish Civil Registration System with information on the vital status of all Danish citizens [23].

The study has some limitations. The method used to adjust the SOFA score for chronic diseases with potential impact on the baseline SOFA value was not protocolized before the study start. All the SOFA correc- tions were done retrospectively, and the information on chronic dis- eases were based on diseases that were registered in the CCI. We did not have precise information on the severity of chronic diseases before admission to ED, and our method used to adjust the SOFA values has not been validated. Therefore, we cannot exclude the risk of misclassifi- cation of patients based on the SOFA score. Furthermore, we only used admission variables to calculate the SOFA score. Serial SOFA measure- ments in the first hours of admission may have detected some patients with clinical deterioration and fulfilling the Sepsis criteria (SOFA >=2) during the initial ED stay. A total of 123 (5.6%) patients of all infected pa- tients were excluded due to missing values that could influence the cal- culation of SOFA and SIRS scores. However, among the excluded

patients, only five patients had died within 28-day, and over 50% of the exclusions were due to missing bilirubin values or platelets counts, which were not obtained at the admission. We cannot exclude the risk of misclassification of patients fulfilling either SOFA >=2, qSOFA >=2 or SIRS >=2 criteria due to missing values to calculate the scores. Other potential predictors, e.g. tobacco smoking [59], drug misuse [60], phys- ical activity [61], mental health [62], newly major surgery [63], social support [64], and frailty [65], were not examined in our study. They should also be examined and included in future prediction models. In- formation on tobacco consumption was missing in many patients, and we considered the risk of incorrect classification based on the informa- tion in the records to be too high to include data on tobacco consump- tion in the analyses. The number of patients with infections due to devices or implants was small, which was reflected in the wide confi- dence intervals and imprecise estimates in the regression models. The study period preceded the COVID-19 pandemic making it impossible to study the impact of COVID-19 compared to other prognostic factors. Future research of long-term outcomes among patients admitted with infections should also include COVID-19. Two authors were involved in the abstraction of data from the records. Despite the lack of analyses of the interrater reliability, our opinion is that the methods used ensure a low risk of interobserver disagreement between the data abstractors. Finally, the study was undertaken at a single centre.

  1. Conclusion

More than one-third of a population of patients admitted to an ED with infectious diseases of suspected bacterial origin on admission had died during a median follow-up of 2.1 years. Age, male gender, a history of cancer, misuse of alcohol, sepsis admission within one year before index admission, signs of organ dysfunction, fulfilling more than two SIRS criteria upon admission, devices, and implants as sources of infec- tions, and length of stay were independent predictors of long-term mor- tality. Skin infections and higher haemoglobin concentrations reduced the risk of death.


Ethics approval and consent to participate: The study was reported to The Danish Data Protection Agency (REG-

105-2017). The study was the 16th of May 2017 defined as a quality project by the Secretariat of The Committee on Health Research Ethics of Region Zealand, which acts as an Institutional Review Board. The pro- ject was therefore not covered by Committee Act and was not obligated to report for the ethic committee system. Administrative permission to access the data was acquired 25 September 2018 from the Head of Slagelse Hospital Administration.

Availability of data and materials

The datasets used are available by reasonable request to the corre- sponding author.


This work was supported by Region Zealand Health Research Foun- dation, Naestved, Slagelse and Ringsted Hospitals Research Fund, and the Department of Emergency Medicine, Copenhagen University Hospi- tal, Bispebjerg and Frederiksberg, Denmark.

CRediT authorship contribution statement

Lana Chafranska: Writing – review & editing, Writing – original draft, Visualization, Validation, Conceptualization. Oscar Overgaard Stenholt: Conceptualization, Writing – review & editing. Rune Husas Sorensen: Project administration, Investigation, Formal analysis, Data

curation. S.M. Osama Bin Abdullah: Data curation, Formal analysis, Investigation. Finn Erland Nielsen: Visualization, Validation, Supervi- sion, Software, Resources, Project administration, Funding acquisition, Formal analysis, Conceptualization.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influ- ence the work reported in this paper.


Results from this study have previously been presented as an e-poster at the 14th European Congress on Emergency Medicine (EUSEM), 19-22 September 2020 (Virtual Congress).


  1. Vest-Hansen B, Riis AH, Sorensen HT, Christiansen CF. Acute admissions to medical departments in Denmark: diagnoses and patient characteristics. Eur J Intern Med. 2014;25:639-45.
  2. Kennedy JL, Haberling DL, Huang CC, Lessa FC, Lucero DE, Daskalakis DC, et al. Infec- tious disease hospitalizations: United States, 2001 to 2014. Chest. 2019;156:255-68.
  3. Armstrong GL, Conn LA, Pinner RW. Trends in infectious disease mortality in the United States during the 20th century. JAMA. 1999;281:61-6. 1001/jama.281.1.61.
  4. Schuur JD, Venkatesh AK. The growing role of emergency departments in hospital admissions. New England Journal of Medicine. 2012;367:391-3. 1056/NEJMp1204431.
  5. Shankar-Hari M, Harrison DA, Ferrando-Vivas P, Rubenfeld GD, Rowan K. Risk fac- tors at index hospitalization associated with longer-term mortality in adult Sepsis sur- vivors. JAMA Netw Open. 2019;2:e194900. 2019.4900.
  6. Schmidt K, Gensichen J, Fleischmann-Struzek C, Bahr V, Pausch C, Sakr Y, et al. M E D I C I N E Long-term survival following Sepsis results of a single-center registry study with 4-year follow-up. Dtsch Arztebl int; 2020..
  7. Nasir N, Jamil B, Siddiqui S, Talat N, Khan FA, Hussain R. Mortality in sepsis and its relationship with gender. Pakistan Journal of Medical Sciences. 2015;31:1201-6.
  8. Sasse KC, Nauenberg E, Long A, Anton B, Tucker HJ, Hu TW. Long-term survival after intensive care unit admission with sepsis. Crit Care Med. 1995;23:1040-7. https://
  9. Laupland KB, Zygun DA, Doig CJ, Bagshaw SM, Svenson LW, Fick GH. One-year mor- tality of bloodstream infection-associated sepsis and septic shock among patients presenting to a regional critical care system. Intensive Care Med. 2005;31:213-9.
  10. Brink A, Alsma J, Verdonschot RJCG, Rood PPM, Zietse R, Lingsma HF, et al. Predicting mortality in patients with suspected sepsis at the emergency department; a retro- spective cohort study comparing qSOFA, SIRS and National Early Warning Score. PLoS ONE. 2019;14.
  11. Nesseler N, Defontaine A, Launey Y, Morcet J, Malledant Y, Seguin P. Long-term mor- tality and quality of life after septic shock: a follow-up observational study. Intensive Care Med. 2013;39:881-8.
  12. Kothe H, Bauer T, Marre R, Suttorp N, Welte T, Dalhoff K, et al. Outcome of community-acquired pneumonia: influence of age, residence status and antimicro- bial treatment. Eur Respir J. 2008;32:139-46. 00092507.
  13. Fine MJ, Smith MA, Carson CA, Mutha SS, Steadman SS, Weissfeld LA, et al. Prognosis and outcomes or patients with communityacquired pneumonia. Pneumologie. 1997;51:88.
  14. Vandenbroucke JP, Von Elm E, Altman DG, Gotzsche PC, Mulrow CD, Pocock SJ, et al. Strengthening the reporting of observational studies in epidemiology (STROBE): ex- planation and elaboration. PLoS Med. 2007;4:1628-54. journal.pmed.0040297.
  15. Bin Abdullah SMO, Sorensen RH, Dessau RBC, Sattar SMRU, Wiese L, Nielsen FE. Prognostic accuracy of qSOFA in predicting 28-day mortality among infected pa- tients in an emergency department: a prospective validation study. Emerg Med J. 2019;36:722-8.
  16. Schmidt M, Schmidt SAJ, Adelborg K, Sundboll J, Laugesen K, Ehrenstein V, et al. The Danish health care system and epidemiological research: from health care contacts to database records. Clin Epidemiol. 2019;11:563-91. CLEP.S179083.
  17. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prog- nostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373-83.
  18. Levy MM, Fink MP, Marshall JC, Abraham E, Angus D, Cook D, et al. SCCM/ESICM/ ACCP/ATS/SIS international Sepsis definitions conference. Critical care medicine, vol.

31. Crit Care Med. 2001;2003:1250-6.


  1. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Chest. 1992;101:1644-55.
  2. Singer M, Deutschman CS, Seymour C, Shankar-Hari M, Annane D, Bauer M, et al. The third international consensus definitions for sepsis and septic shock (sepsis- 3). JAMA – Journal of the American Medical Association. 2016;315:801-10. https://
  3. Vincent JL, Moreno R, Takala J, Willatts S, De Mendonca A, Bruining H, et al. The SOFA (Sepsis-related organ failure assessment) score to describe organ dysfunction/fail- ure. Intensive Care Med. 1996;22:707-10.
  4. Abdullah SO, bin, Sorensen RH, Nielsen FE.. Prognostic accuracy of SOFA, qSOFA, and SIRS for mortality among emergency department patients with infections. Infection and drug resistance. 2021;14:2763-75.
  5. Schmidt M, Pedersen L, Sorensen HT. The Danish civil registration system as a tool in epidemiology. Eur J Epidemiol. 2014;29:541-9. 014-9930-3.
  6. Pedersen AB, Mikkelsen EM, Cronin-Fenton D, Kristensen NR, Pham TM, Pedersen L, et al. Missing data and multiple imputation in clinical epidemiological research. Clin Epidemiol. 2017;9:157-66.
  7. Murray SB, Bates DW, Ngo L, Ufberg JW, Shapiro NI. Charlson index is associated with one-year mortality in emergency department patients with suspected infec- tion. Acad Emerg Med. 2006;13:530-6.
  8. Ittisanyakorn M, Ruchichanantakul S, Vanichkulbodee A, Sri-On J. Prevalence and factors associated with one-year mortality of infectious diseases among elderly emergency department patients in a middle-income country. BMC Infect Dis. 2019;19:662.
  9. Goto T, Yoshida K, Tsugawa Y, Camargo CA, Hasegawa K. Infectious disease-related emergency department visits of elderly adults in the United States, 2011-2012. J Am Geriatr Soc. 2016;64:31-6.
  10. Reilev M, Lundby C, Jensen J, Larsen SP, Hoffmann H, Pottegard A. Morbidity and mortality among older people admitted to nursing home. Age Ageing. 2020;49: 67-73.
  11. Christensen K, Doblhammer G, Rau R, Vaupel JW. Ageing populations: the challenges ahead. The Lancet. 2009;374:1196-208. 61460-4.
  12. Koch AM, Nilsen RM, Eriksen HM, Cox RJ, Harthug S. Mortality related to hospital- associated infections in a tertiary hospital; repeated cross-sectional studies between 2004-2011. Antimicrob Resist Infect Control. 2015;4. s13756-015-0097-9.
  13. Castle SC, Uyemura K, Fulop T, Makinodan T. Host resistance and immune responses in advanced age. Clin Geriatr Med. 2007;23:463-79. 2007.03.005.
  14. Gomez CR, Boehmer ED, Kovacs EJ. The aging innate immune system. Curr Opin Immunol. 2005;17:457-62.
  15. Crimmins EM, Shim H, Zhang YS, Kim JK. Differences between men and women in mortality and the health dimensions of the morbidity process. Clin Chem. 2019; 65:135.
  16. Klinge M, Aasbrenn M, Aasbrenn M, Ozturk B, Christiansen CF, Suetta C, et al. Read- mission of older acutely admitted Medical patients after short-term admissions in Denmark: a nationwide cohort study. BMC Geriatr. 2020;20:203. 10.1186/s12877-020-01599-4.
  17. Willadsen T, Siersma V, Nicolaisdottir D, Koster-Rasmussen R, Jarbol D, Reventlow S, et al. Multimorbidity and mortality: A 15-year longitudinal registry-based nation- wide Danish population study. Journal of Comorbidity. 2018;8. 1177/2235042X18804063. 2235042X18804063.
  18. Camou F, Didier M, Leguay T, Milpied N, Daste A, Ravaud A, et al. Long-term progno- sis of septic shock in cancer patients. Support Care Cancer. 2020;28:1325-33.
  19. Jones TK, Fuchs BD, Small DS, Halpern SD, Hanish A, Umscheid CA, et al. Post-acute care use and Hospital readmission after sepsis. Annals of the American Thoracic So- ciety. 2015;12:904-13.
  20. Stenholt POO, Abdullah SMO, Sorensen RH, Nielsen FE. Independent predictors for 90-day readmission of emergency department patients admitted with sepsis: a pro- spective cohort study. BMC Infect Dis. 2021;21:1-10. s12879-021-06007-9.
  21. Jiang Z, Bo L, Xu Z, Song Y, Wang J, Wen P, et al. An explainable machine learning al- gorithm for risk factor analysis of in-hospital mortality in sepsis survivors with ICU readmission. Comput Methods Programs Biomed. 2021;204:106040. https://doi. org/10.1016/j.cmpb.2021.106040.
  22. Winters BD, Eberlein M, Leung J, Needham DM, Pronovost PJ, Sevransky JE. Long- term mortality and quality of life in sepsis: a systematic review. Crit Care Med. 2010;38:1276-83.
  23. Riche F, Chousterman BG, Valleur P, Mebazaa A, Launay JM, Gayat E. Protracted im- mune disorders at one year after ICU discharge in patients with septic shock. Crit Care. 2018;22:1-10.
  24. O’Brien JM, Lu B, Ali NA, Martin GS, Aberegg SK, Marsh CB, et al. Alcohol dependence is independently associated with sepsis, septic shock, and hospital mortality among adult intensive care unit patients. Crit Care Med. 2007;35:345-50. 10.1097/01.CCM.0000254340.91644.B2.
  25. Laupland KB, Gregson DB, Zygun DA, Doig CJ, Mortis G, Church DL. Severe blood- stream infections: a population-based assessment. Crit Care Med. 2004;32:992-7.
  26. Fernandez Sola J, Junque A, Estruch R, Monforte R, Torres A, Urbano Marquez A. High Alcohol intake as a risk and prognostic factor for community-acquired pneumonia. Arch Intern Med. 1995;155:1649-54. 00430150137014.
  27. Ruiz M, Ewig S, Torres A, Arancibia F, Marco F, Mensa J, et al. Severe community- acquired pneumonia: risk factors and follow-up epidemiology. Am J Respir Crit Care Med. 1999;160:923-9.
  28. Szabo G. Consequences of alcohol consumption on host defence. Alcohol Alcohol. 1999;34:830-41.
  29. Rizwan Sohail M, Henrikson C, Jo Braid-Forbes M, Forbes K, Lerner D. Increased long- term mortality in patients with cardiovascular implantable electronic device infections. Pacing and Clinical Electrophysiology : PACE. 2015;38:231-9. PACE.12518.
  30. De Bie MK, Van Rees JB, Thijssen J, Borleffs CJW, Trines SA, Cannegieter SC, et al. Car- diac device infections are associated with a significant mortality risk. Heart Rhythm. 2012;9:494-8.
  31. Iordanou S, Middleton N, Papathanassoglou E, Raftopoulos V. Surveillance of device associated infections and mortality in a major intensive care unit in the Republic of Cyprus. BMC Infect Dis. 2017;17.
  32. Vincitorio D, Barbadoro P, Pennacchietti L, Pellegrini I, David S, Ponzio E, et al. Risk factors for catheter-associated urinary tract infection in Italian elderly. Am J Infect Control; 2014..
  33. Seymour CW, Liu VX, Iwashyna TJ, Brunkhorst FM, Rea TD, Scherag A, et al. Assess- ment of clinical criteria for Sepsis: for the third international consensus definitions for Sepsis and septic shock (Sepsis-3). JAMA. 2016;315:762-74. 1001/JAMA.2016.0288.
  34. Kovach CP, Fletcher GS, Rudd KE, Grant RM, Carlbom DJ. Comparative prognostic ac- curacy of sepsis scores for hospital mortality in adults with suspected infection in non-ICU and ICU at an academic public hospital. PLoS ONE. 2019;14:e0222563.
  35. Raith EP, Udy AA, Bailey M, McGloughlin S, MacIsaac C, Bellomo R, et al. Prognostic accuracy of the SOFA score, SIRS criteria, and qSOFA score for in-hospital mortality among adults with suspected infection admitted to the intensive care unit. JAMA – Journal of the American Medical Association. 2017;317:290-300. 10.1001/jama.2016.20328.
  36. Liao MM, Lezotte D, Lowenstein SR, Howard K, Finley Z, Feng Z, et al. Sensitivity of systemic inflammatory response syndrome for critical illness among ED patients. American Journal of Emergency Medicine. 2014;32:1319-25. 1016/j.ajem.2014.07.035.
  37. Kaukonen K-M, Bailey M, Pilcher D, Cooper DJ, Bellomo R. Systemic inflammatory re- sponse syndrome criteria in defining severe Sepsis. New England Journal of Medi- cine. 2015;372:1629-38.
  38. Culleton BF, Manns BJ, Zhang J, Tonelli M, Klarenbach S, Hemmelgarn BR. Impact of anemia on hospitalization and mortality in older adults. Blood. 2006;107:3841-6.
  39. Dunne JR, Malone D, Tracy JK, Gannon C, Napolitano LM. Perioperative anemia: an independent risk factor for infection, mortality, and resource utilization in surgery. Journal of Surgical Research. 2002;102:237-44. 6330.
  40. Purtle SW, Moromizato T, McKane CK, Gibbons FK, Christopher KB. The association of red cell distribution width at hospital discharge and out-of-hospital mortality fol- lowing critical illness. Crit Care Med. 2014;42:918-29.


  1. Arcavi L, Benowitz NL. cigarette smoking and infection. Arch Intern Med. 2004;164: 2206-16.
  2. Ronan MV, Herzig SJ. Hospitalizations Related To Opioid abuse/Dependence And As- sociated serious infections Increased Sharply, 2002-12, 35; 2017; 832-7. https://doi. org/10.1377/HLTHAFF.2015.1424.
  3. Koolhaas CM, Dhana K, Schoufour JD, Lahousse L, van Rooij FJA, Ikram MA, et al. physical activity and cause-specific mortality: the Rotterdam study. Int J Epidemiol. 2018;47:1705-13.
  4. Davydow D, Ribe A, Pedersen H, Vestergaard M, Fenger-Gron M. The association of unipolar depression with thirty-day mortality after hospitalization for infection: a population-based cohort study in Denmark. J Psychosom Res. 2016;89:32-8.
  5. Jiang W, Hu X, Hu Z, Tang Z, Wu H, Chen L, et al. Morbidity and mortality of nosoco- mial infection after cardiovascular surgery: a report of 1606 cases. Current. Med Sci. 2018;38(2):329-35.
  6. Laugesen K, Baggesen LM, Schmidt SAJ, Glymour MM, Lasgaard M, Milstein A, et al. social isolation and all-cause mortality: a population-based cohort study in Denmark. Sci Rep. 2018;8(1):1-8.
  7. Owen R, Conroy S, Taub N, Jones W, Bryden D, Pareek M, et al. Comparing associa- tions between frailty and mortality in hospitalised older adults with or without COVID-19 infection: a retrospective observational study using electronic health re- cords. Age Ageing. 2021;50:307-16.