166
Original article
Mortality and prognostic factors of patients who have blood cultures performed in the emergency department: a cohort study Katrine P. Lindviga, Stig L. Nielsenb, Daniel P. Henriksena, Thøger G. Jensenc, Hans Jørn Kolmosc, Court Pedersenb, Pernille J. Vinholtd and Annmarie T. Lassena Background Early identification and treatment of patients with severe infection improve their prognosis. The aims of this study were to describe the 30-day mortality and to identify prognostic factors among blood-cultured patients in a medical emergency department (MED). Patients and methods This was a hospital-based cohort study including all adult (≥15 years old) blood-cultured patients at the MED at Odense University Hospital between 1 August 2009 and 31 August 2011. Results During the study period, 5499/11 988 (45.9%) patients had blood cultures performed within 72 h of arrival and were included in the study. Of those included, 2631 (47.8%) were men, median age 69 years (range 15–103), and 418 (7.6%) were diagnosed with bacteraemia. The overall 30-day mortality among blood-cultured patients was 11.0% (10.2–11.9). In a multivariate Cox regression model, age of more than 80 years [hazard ratio (HR) 4.6 (95% CI 3.6–6.0)], at least two organ failure [HR 3.6 (2.9–4.5)], bacteraemia [HR 1.4 (1.1–1.8)], Charlson Comorbidity Index of at least 2 h [HR 1.7 (1.3–2.0)], SIRS [HR 1.5 (1.2–1.7)], a history of alcohol dependency [HR 1.7 (1.3–2.3)] and late drawing of blood cultures 24–48 h after arrival [HR 1.7 (1.3–2.2)] were
found to be prognostic factors of mortality among bloodcultured patients in the MED. Conclusion Among blood-cultured patients in the MED, we found an 11.0% overall 30-day mortality. Factors associated with 30-day mortality were age more than 80 years, at least two organ failure, bacteraemia, Charlson Comorbidity Index of at least 2, SIRS, a history of alcohol dependency and late drawing of blood cultures. European Journal of Emergency Medicine 23:166–172 Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved. European Journal of Emergency Medicine 2016, 23:166–172 Keywords: bacteraemia, blood cultures, cohort, emergency medicine, prognostic factors and mortality Departments of aEmergency Medicine, bInfectious Diseases, cClinical Microbiology and dClinical Biochemistry and Pharmacology, Odense University Hospital, Odense C, Denmark Correspondence to Katrine P. Lindvig, Department of Emergency Medicine, Odense University Hospital, Sdr. Boulevard 29, 5000 Odense C, Denmark Tel: + 45 65 41 47 67; fax: + 45 6613 0950/ + 45 654 11571; e-mail: [email protected] Received 12 August 2014 Accepted 14 January 2015
Introduction
Patients and methods
Community-onset bacteraemia is common among patients in the emergency department (ED) [1–3] and is associated with sepsis and septic shock [1]. Early identification and appropriate antibiotic therapy have been shown to reduce the in-hospital mortality of septic shock from 80 to 20–30% [4]. Thus, it could be expected that early diagnosis and treatment of ED patients with suspected bacteraemia would also relate to a better outcome [2].
Study design and participants
It is furthermore essential for the clinician to be able to assess the patient’s risk for a complicated outcome, whether organ failure or death, already at the beginning of the treatment in the ED [1,2,5]. The aims of this study were to describe the 30-day mortality rate and to identify prognostic factors among patients with suspected bacteraemia to accurately identify patients at risk of a complicated outcome in the medical emergency department (MED). 0969-9546 Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
We carried out a population-based prospective cohort study consecutively enrolling all first-time admitted adult patients (age ≥ 15 years) at the MED at Odense University Hospital, who had blood cultures drawn upon arrival or within 72 h of admission, from 1 August 2009 to 31 August 2011. Patients were included at their first MED contact in the observation period. Odense University Hospital is an 1100-bed level 1-trauma centre and a university teaching hospital with all specialities present. It serves both as a tertiary and as a primary hospital, with a primary catchment area of 288 000 individuals. The MED had ∼ 9000 admissions per year during the observation period. Data sources
As part of the standard care at the ED at Odense University Hospital, all patients had their blood pressure, pulse rate, respiratory frequency, oxygen saturation, DOI: 10.1097/MEJ.0000000000000250
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
Blood cultured patients in the ED Lindvig et al. 167
rectal temperature and level of consciousness/Glasgow Coma Scale measured or calculated upon arrival, and had standard blood samples drawn including leucocyte count, platelet count, C-reactive protein, creatinine, international normalized ratio and bilirubin. In addition, a proportion of the patients were subjected to an arterial blood test including PaO2, PaCO2, lactate and pH. The decision to draw blood cultures was made by the attending physicians. All vital values were recorded electronically in the hospital database at the time of measurement, and we did not have to extract these data manually. Information on microbiological and biochemical results as well as comorbid conditions was obtained by linking all included patients at an individual level to five large populationbased registers: (1) The Danish Civil Registration System, which covers data on births, deaths, migration, etc. Since 1968, each Danish resident has been assigned a unique 10-digit civil registration number that enables unambiguous linkage between registries [6]. (2) The Danish National Patient Register, covering all somatic and psychiatric inpatietns and outpatients in all Danish hospitals, enabling true population-based epidemiological research [7]. (3) Odense Pharmaco Epidemiological Database, which contains information on redeemed prescription at pharmacies in the Region of Southern Denmark. The database enables the collection of longitudinal drug histories and allows for linkage of prescription data to other population-based Danish registries [8,9]. (4) The Clinical Microbiological Registration System at Odense University Hospital, containing all blood cultures obtained in the area since 2000. (5) The Clinical Biochemical Registration System at Odense University Hospital containing all laboratory tests obtained in the area since 2000. Microbiological methods
The blood cultures were incubated and screened for the growth of microorganisms for 6 days or until detected positive, using the Bactec 9240 system (Becton Dickinson, Franklin Lakes, New Jersey, USA) until January 2011 and the Bact/Alert system (BioMérieux, Rhône-Alpes, France) thereafter. Routine methods for identification of bacteria were based on conventional characterization [10], the Danish reference programme (http://www.dskm.dk) and automated identification using Vitek 2 (BioMérieux) and MALDI-TOF (SARAMIS; BioMérieux). Definitions
Community-onset bacteraemia was defined as having a positive blood culture drawn within the first 2 days of admission. A blood culture consisted of two blood culture
sets, each comprising one aerobic and one anaerobic bottle, and we defined bacteraemia as either: (a) recognized pathogens detected in at least one blood culture or (b) common skin contaminants (coagulase-negative staphylococci, Bacillus spp., Propionibacterium spp., Corynebacterium spp., viridans group streptococci, Aerococcus spp. or Micrococcus spp.) detected in at least two blood culture sets within 5 days [11–13]. The date of the first positive blood culture set was considered the date of bacteraemia. Polymicrobial bacteraemia was defined as isolation of at least two different microorganisms, deemed to represent bacteraemia within 2 days [14]. The timing of blood cultures was defined as day 0 (0–24 h), day 1 (24–48 h) and day 2 (48–72 h). Late timing of blood cultures was defined from day 1 and onwards (24–72 h). To account for comorbidity, all patients were classified according to the Charlson Comorbidity Score into groups: 0, no comorbidity; 1, light/moderate comorbidity or; ≥ 2, high comorbidity [15]. Systemic Inflammatory Response Syndrome (SIRS) was defined as being present if at least two of the following four criteria were fulfilled: body temperature more than 38.0 or less than 36.0°C, respiratory frequency more than 20 breaths/min or PaCO2 less than 4.3 kPa, pulse rate more than 90 beats/min and leucocyte count more than 12.0 or less than 4.0 × 109/l [16]. Patients were classified as immunocompromised if they had a moderate to high intake of immunosuppressant medication, for example 40 mg prednisone or its equivalent, Primary immunodeficiency diseases, AIDS, a newly diagnosed malignancy registered in the Danish Cancer Registry up to a year before admission or a diagnosis of organ transplantation (Table 1). Patients were classified as having a history of alcohol dependency if they had at least admissions with a discharge diagnosis of an acute alcohol episode, at least one admission with a discharge diagnosis of a chronic alcohol related diagnosis, a redeemed prescription of disulfiram from 2007 and onwards or a registration in The National Register of Alcohol Dependency Treatment (Table 1). Organ system failure was defined on the basis of accessible clinical parameters. Central nervous system failure: Glasgow Coma Scale less than 14 or if the patient is unresponsive or presenting with confusion and disorientation. Respiratory failure: PaO2 less than or equal to 9.75 kPa or PO2 less than 92%. Circulatory failure: systolic blood pressure 90 mmHg or less. Perfusion failure: lactate at least 2.5 mmol/l or pH 7.3 or less. Renal failure: Creatinine at least 177 or a 100 μmol/l increase in p-creatinine among patients with known renal disease. Coagulation failure: international normalized ratio more than 1.6 or platelet count 100 × 109/l or less. Hepatic failure: bilirubin at least 43 μmol/l.
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
168 European Journal of Emergency Medicine 2016, Vol 23 No 3
Table 1
Definition of immunosuppression
Criteria
ICD-10 codes
A moderate to high intake of immunosuppressant medication Primary immunodeficiency diseases AIDS defining diseases A newly diagnosed malignancy registered in the Danish Cancer Registry (up to a year before admission) A diagnosis of organ transplantation Definition of a history of alcohol dependency An acute alcohol episode A chronic alcohol-related diagnosis A redeemed prescription of disulfiram
ATC: H02AB or L00–L04 ≥120 cumulative defined daily dose in the preceding 120 days with a compliance rate of 80% ICD-10: D61; D70–D71; D80–D82; D84; E70.3; G11.3 ICD-10: B20–B24; F02.41–F02.44 ICD-10: C00–C79; C97; D37–D44, D48 ICD-10: Z94 F10.0–F10.1; T519 ICD-10: F10.2–F10.9; E24.4; E52.9A; G31.2; G62.1; G72.1; I42.6; K29.2; K70; K85.2; K86.0; L27.8A; O35.4; P04.3; T50.0A; Z71.4; Z72.1; Z50.2 ATC: N07BB01 from 2007 onwards
Patients registered in The National Register of Alcohol Abuse Treatment.
Statistical analysis
Baseline characteristics were presented in contingency tables as numbers and percentages. Univariate analyses were carried out and tested using the χ2-test, which was two-sided. A P-value of less than 0.05 was considered statistically significant. The 30-day mortality was reported as the proportion of patients who died within 30 days and was shown in a Kaplan–Meier failure plot. In case of missing values, variables were registered as normal. The association with death of predefined variables (bacteraemia, sex, age, Charlson Comorbidity Score, alcohol dependency, SIRS, immunosuppression, organ failure and timing of blood cultures) was evaluated in univariate analyses using the log-rank test. The bacteraemias were grouped as Gram-negative bacteraemia, Gram-positive bacteraemia and polymicrobial bacteraemia and analysed in a univariate Cox regression analysis. All predefined variables with a significance level of less than 0.05 were included in the multivariate Cox regression model, except for female sex [hazard ratio (HR) 1.1 (0.9–1.4)], which was finally included in the final model because of predefined interest in the variable. The proportional hazard assumptions were tested graphically using the Cox–Snell residuals and found appropriate. Statistical analyses were carried out using Stata 12.1 (Stata Corp LP, College Station, Texas, USA).
Table 2 Baseline characteristics of patients with blood cultures in the Medical Emergency Department
Characteristics Sex Male Female Age Mean: 65 years, SD: 20.5, range: 15–102 15–64 65–79 > 80 Charlson comorbidity index 0 1 2 Immune compromised No Yes Alcohol use No Yes SIRS No Yes Organ failure 0 failure 1 failure ≥ 2 failures Timing of blood cultures Day 0 Day 1 Day 2
Blood-cultured patients total (n = 5499) [n (%)]
Missing values (%) 0
2631 (47.8) 2868 (52.2) 0
2357 (42.9) 1527 (27.8) 1615 (29.3) 0 2258 (41.1) 1226 (22.3) 2015 (36.6) 0 4714 (85.7) 785 (14.3) 14 (0.2) 4954 (90.1) 531 (9.7) 2623 (47.7) 2876 (52.3) 2994 (54.4) 1715 (31.2) 790 (14.4) 0 4540 (82.6) 646 (11.7) 313 (5.7)
Baseline characteristics of blood-cultured patients in the Medical Emergency Department.
Ethics
The study was approved by the Danish Data Protection Agency (No. 2008-58-0035) and the Danish National Board of Health (No. 3-3013-35). In observational studies such as in this, review by an Ethics Board is not required according to Danish law.
Results Descriptive
During the study period, we identified 12 027 patients with a first-time admission, of whom 5499 (45.7%) had blood cultures performed within 2 days from admission and were included in the study. Of these, 418 (7.6%) had bacteraemia, corresponding to 3.5% of all MED patients.
A baseline description of patients with blood cultures in the MED is presented in Table 2. In general, patients who had blood cultures drawn were more often female (52.2%), younger (15–64 years old), had no comorbidity (41.1%), were not immunosuppressed (85.7%), had no alcohol dependency (90.3%) and had no organ failure (54.5%). However, they were likely to present with SIRS and had blood cultures drawn on the day of arrival. The most frequent pathogens detected were Escherichia coli (30.4%), Streptococcus pneumoniae (13.0%) and Staphylococcus aureus (10.3%).
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
Blood cultured patients in the ED Lindvig et al. 169
Mortality
In the present study, we found an overall 30-day mortality rate of 11.0% [P < 0.01, 95% confidence interval (CI): 10.2–11.9] among blood-cultured patients in the MED, with a 30-day mortality rate of 19.6% (P < 0.01, 95% CI: 15.6–24.3) for bacteraemic patients compared with 10.3% (P < 0.01, 95% CI: 9.5–11.3) among bloodcultured nonbacteraemic patients (Fig. 1).
11.0% (10.2–11.9). These results are high compared with previous studies of blood-cultured ED patients, which have reported in-hospital mortality rates of 3.6–4.1% [5,17]. This might be because of the different clinical thresholds used to obtain blood cultures in the EDs or because of fundamental differences in the patient population. Positive blood culture
Prognostic factors
In the multiple Cox regression analysis, we found age older than 80 years (HR 4.6, 95% CI: 3.6–6.0) and at least two organ failure (HR 3.6, 95% CI 2.9–4.5) to be the strongest predictors of 30-day mortality among blood-cultured patients in the MED (Table 3). Furthermore, bacteraemia [HR 1.4 (1.1–1.8)], Charlson Comorbidity Index at least 2 [HR 1.7 (1.3–2.0)], SIRS [HR 1.5 (1.2–1.7)], a history of alcohol dependency [HR 1.7 (1.3–2.3)] and drawing of blood cultures 24–48 h after arrival [HR 1.7 (1.3–2.2)] were found to be prognostic factors of mortality. In the group with blood cultures drawn from 48–72 h, 16/106 (15.1%) died within 30 days, representing a HR of 1.6 (0.9–2.5), although not significant. In a univariate Cox regression analysis, we found polymicrobial bacteraemia to be associated with increased mortality [HR 2.9 (1.5–5.7)] compared with patients with Gram-negative bacteraemia (Table 4).
Discussion Mortality Negative blood culture
In this hospital-based cohort study of blood-cultured patients in the MED, we found a 30-day mortality of
Among blood-cultured patients with a positive blood culture, we found a relatively high 30-day mortality of 19.6%; however, the results are comparable to the previous literature, ranging from 14 to 37% [18], and overall mortality of 22%, irrespective of the patient population in a recently published study by Nielsen et al. [19]. In the present study, the clinical threshold for obtaining blood cultures was entirely based on the physician’s clinical decision. In comparison, in a study of blood-cultured patients in the ED, Shapiro et al. [5] reported an overall 28-day mortality rate of 4.1%. They found that SIRS did not add any additional prognostic value, whereas each additional organ dysfunction doubles the 1-year mortality risk. To understand this marked difference in mortality between the two studies, it is necessary to outline the differences in patient populations. Shapiro et al. [5] have included all patients at risk of infection; if they had blood cultures drawn within 3 h of admission; hereof 16% were treated as outpatients. The patients were younger (mean age 59.9 vs. 64.5 years), and a lower proportion had SIRS (29.2 vs. 52.3%) [5]. In the present study, we included only admitted patients, who, in general, are more likely to
Fig. 1
Probability of death (%)
0.30 0.25 0.20 0.15 0.10 0.05 0.00 0
5
10
15
20
25
30
Days Number at risk Blood culture 5081 Bacteraemia 418
4795 360 Bacteraemia = 0
4648 346
4561 338
Bacteraemia = 1
Kaplan–Meier failure estimates for blood-cultured patients in the MED.
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
170
European Journal of Emergency Medicine 2016, Vol 23 No 3
Table 3
Prognostic factors among emergency medical patients with blood cultures
Total
Patients alive within 30 days [n (%)]
Patients dead within 30 days [n (%)]
4556 (89.7) 336 (80.4)
525 (10.3) 82 (19.6)
2.0 (1.6–2.6)
1.0 1.4 (1.1–1.8)
2573 (89.7) 2319 (88.1)
295 (10.3) 312 (11.9)
1.1 (0.9–1.4)
1.0 1.1 (0.9–1.2)
2263 (96.0) 1343 (88.0) 1286 (79.6)
94 (4.0) 184 (12.0) 329 (20.4)
2.2 (2.0–2.5)
1.0 2.4 (1.9–3.2) 4.6 (3.6–6.0)
2120 (93.9) 1098 (89.6) 1674 (83.1)
138 (6.1) 128 (10.4) 341 (16.9)
1.7 (1.5–1.9)
1.0 1.1 (0.9–1.4) 1.7 (1.3–2.0)
4218 (89.5) 674 (85.9)
496 (10.5) 111 (14.4)
1.4 (1.1–1.7)
1.0 1.1 (0.9–1.4)
2254 (92.0) 2638 (86.5)
195 (8.0) 412 (13.5)
1.8 (1.5–2.1)
1.0 1.5 (1.2–1.7)
2834 1512 546
179 (5.6) 210 (10.9) 218 (22.2)
2.4 (2.1–2.6)
1.0 1.6 (1.3–1.9) 3.6 (2.9–4.5)
4423 (89.3) 458 (86.2)
531 (10.7) 73 (13.8)
1.3 (1.0–1.7)
1.0 1.7 (1.3–2.3)
4413 (89.5) 389 (84.0) 90 (84.9)
517 (10.5) 74 (16.0) 16 (15.1)
1.4 (1.1–1.6)
1.0 1.7 (1.3–2.2) 1.6 (0.9–2.5)
Bacteraemia No (5081) Yes (418) Sex Female (2868) Male (2631) Age 15–64 (2357) 65–79 (1527) > 80 (1615) Charlson comorbidity index 0 (2258) 1 (1226) ≥ 2 (2015) Immune compromised No (4714) Yes (785) SIRS No SIRS (2449) SIRS (3050) Organ failure 0 failure (3013) 1 failure (1722) ≥ 2 failures (764) Alcohol use No (4954) Yes (531) Timing of blood cultures Day 0 (0–24 h) (4930) Day 1 (24–48 h) (463) Day 2 (48–72 h) (106)
Crude hazard ratio (95% CI)
Multivariate COX hazard ratio (95% CI)
CI, confidence interval. Prognostic factors among blood-cultured patients in the Medical Emergency Department, total n = 5499. Patients alive within 30 days, n = 4892. Patients dead within 30 days, n = 607.
Table 4
Bacteraemic patients – difference in mortality divided by Gram colour Patients at start (n = 418) [n (%)]
Crude 30-day mortality (n = 82) [n (%)]
Univariate log-rank P-value
Univariate COX analysis HR
211 (50.5) 177 (42.3) 30 (7.2)
33 (15.6) 37 (20.9) 12 (40.0)
< 0.001
1.0 1.4 (0.9–2.2) 2.9 (1.5–5.7)
Gram-negative Gram-positive Polymicrobial
Univariate Cox regression analysis. HR, hazard ratio.
have a poorer prognosis compared with the outpatient population. These fundamental differences in patient population may explain the difference in short-term mortality and may explain why the two studies do not agree that SIRS is a prognostic factor among bloodcultured patients. However, differences in the quality of care provided to patients at risk of bacteraemia, such as time to initiation of treatment with antibiotics and fluid, may induce fundamental variations in mortality among different EDs. The present study could not access time to antibiotics and fluid administration.
disease, chronic liver disease, residence in a nursing home and malignancy (with and without metastasis), infection (pneumonia and skin/soft tissue infection), response (tachypnoea, tachycardia and bandemia) and organ dysfunction (renal, respiratory, cardiac, metabolic and haematologic) as prognostic factors for bloodcultured patients in the ED [20]. The present study thereby confirms the identified prognostic factors as remaining relevant even after 8 years, and despite a different healthcare setting. Organ failure
Prognostic factors
Previous studies have identified prognostic factors among blood-cultured patients in the ED. A study by Howell et al. [20] identified age, chronic obstructive pulmonary
It has been shown previously that the combination of sepsis and organ dysfunction is associated with increased morbidity and mortality [4,5,21–23]. In agreement with these studies, we found organ failure to be an important
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
Blood cultured patients in the ED Lindvig et al. 171
prognostic factor among blood-cultured patients in the MED.
Timing of blood cultures in the emergency department
We found that the timing of blood cultures in the MED was related to an increased risk of mortality if the blood culture was performed the day after arrival, which was especially evident among bacteraemic patients (the 30-day mortality rate for patients’ blood cultured day 0, 1 and 2 is 10.5, 16.0 and 15.1%, respectively). Perhaps the observed increased risk is because of the selection of patients, who, several hours after arrival, experience deterioration in their clinical condition, followed by medical re-evaluation and blood cultures, or perhaps, it is an indicator of suboptimal treatment and evaluation of the patient at arrival. In the present observational study, we could not distinguish between these explanations, but a manual review of all patients with blood cultures performed 2 days after arrival showed a high proportion of patients in whom the blood cultures had been ordered when the patient arrived, however, by mistake first performed the following day. To our knowledge, no other studies have evaluated what impact the timing blood cultures have as a prognostic factor for mortality in the ED. However, studies of severe sepsis have found that early diagnosis and treatment of patients with severe sepsis improve their diagnosis [21], thereby pinpointing the need for early diagnosis and treatment. Thus, the prognostic factor of late drawing of blood cultures supports the theory of early goal-directed therapy because it indicates that late diagnosis and deterioration in clinical condition might worsen the patients’ outcome. In a univariate analysis, we found that patients with polymicrobial bacteraemia had a high mortality rate and an increased HR of dying within 30 days compared with patients with Gram-negative bacteraemia, which is in accordance with other studies [24]. The strengths of our study are the consecutive inclusion of all adult first-time admission patients with blood cultures drawn upon arrival to the MED within the study period and the ability to follow up on all the patients included because of the unique personal identification number assigned to all Danish citizens. Furthermore, blood-cultured patients had two sets of blood cultures drawn, representing four bottles for each blood-cultured patient, which increases the credibility of a positive blood culture [25,26]. However, a source of error for blood cultures is the presence of contaminants. This study defined that for a common commensal bacterium to be considered a positive bacteraemia, the bacterium must be isolated from at least two blood culture sets, within a 5-day period; otherwise, the isolated bacterium is considered to be a contaminant, and thereby excluded from the study, which minimizes – but does not rule out – the risk of inclusion of false-positive bacteraemic episodes.
This study has some potential limitations. The definition of organ failure is not validated; however, to a large extent, our definitions of severe sepsis are similar to the recommendations by the Surviving Sepsis Campaign from 2012 [21] and the definitions proposed by Shapiro et al. [2], but we were restricted by an inadequate systematic registration of treatment with oxygen therapy, making it impossible to estimate the PaO2/FiO2 ratio. The MED at Odense University Hospital does not receive obvious cardiological, oncological, haematological, nephrological or patients with upper gastrointestinal bleeding, and paediatric patients. Consequently, the results do not apply to all acute medical patients. Furthermore, this is a single-centre study, reflecting the standard care at Odense University Hospital within this period, and therefore, the results may not be entirely generalizable to other wards and hospitals. Conclusion
As it is essential for the clinician to accurately assess the patient’s risk of mortality and assumed prognosis, for those with suspected bacteraemia in the MED, this study provides further information. We found an 11.0% overall 30-day mortality rate. Among blood-cultured patients in the MED older than 80 years of age, at least two organ failure, bacteraemia, Charlson Comorbidity Index of at least 2, SIRS, a history of alcohol dependency and drawing of blood cultures 24–48 h after arrival were associated with an increased 30-day mortality among blood-cultured patients in the MED.
Acknowledgements Katrine Prier Lindvig is supported by an unrestricted grant from Fonden af 17121981. Annmarie Lassen is supported by an unrestricted grant from the private philanthropic foundation TrygFoundation to University of Southern Denmark. Court Pedersen has received payment for educational activities from Gilead, Abbott and MSD. Conflicts of interest
There are no conflicts of interest.
References 1
2
3
4 5
Shapiro NI, Wolfe RE, Wright SB, Moore R, Bates DW. Who needs a blood culture? A prospectively derived and validated prediction rule. J Emerg Med 2008; 35:255–264. Shapiro NI, Wolfe RE, Moore RB, Smith E, Burdick E, Bates DW. Mortality in Emergency Department Sepsis (MEDS) score: a prospectively derived and validated clinical prediction rule. Crit Care Med 2003; 31:670–675. Lindvig KP, Henriksen DP, Nielsen SL, Jensen TG, Kolmos HJ, Pedersen C, et al. How do bacteraemic patients present to the emergency department and what is the diagnostic validity of the clinical parameters; temperature, C-reactive protein and systemic inflammatory response syndrome? Scand J Trauma Resusc Emerg Med 2014; 22:39. Angus DC, van der Poll T. Severe sepsis and septic shock. N Engl J Med 2013; 369:840–851. Shapiro N, Howell MD, Bates DW, Angus DC, Ngo L, Talmor D. The association of sepsis syndrome and organ dysfunction with mortality in emergency department patients with suspected infection. Ann Emerg Med 2006; 48:583–590.
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
172
6 7 8 9 10 11
12
13
14 15
16 17
European Journal of Emergency Medicine 2016, Vol 23 No 3
Pedersen CB. The Danish Civil Registration System. Scand J Public Health 2011; 39 (7 Suppl):22–25. Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient Register. Scand J Public Health 2011; 39 (Suppl):30–33. Gaist D, Sorensen HT, Hallas J. The Danish prescription registries. Dan Med Bull 1997; 44:445–448. Kildemoes HW, Sorensen HT, Hallas J. The Danish National Prescription Registry. Scand J Public Health 2011; 39 (7 Suppl):38–41. Murray PR BE, Pfaller MA, Tenover FC, Yolken RH. Manual of clinical microbiology, 7th ed. Washington, DC: Americal Society for Microbiology; 1999. Gradel KO, Knudsen JD, Arpi M, Ostergaard C, Schonheyder HC, Sogaard M. Classification of positive blood cultures: computer algorithms versus physicians’ assessment - development of tools for surveillance of bloodstream infection prognosis using population-based laboratory databases. BMC Med Res Methodol 2012; 12:139. Leal J, Gregson DB, Ross T, Flemons WW, Church DL, Laupland KB. Development of a novel electronic surveillance system for monitoring of bloodstream infections. Infect Control Hosp Epidemiol 2010; 31:740–747. Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control 2008; 36:309–332. Roberts FJ. Definition of polymicrobial bacteremia. Rev Infect Dis 1989; 11:1029–1030. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40:373–383. Nystrom PO. The systemic inflammatory response syndrome: definitions and aetiology. J Antimicrob Chemother 1998; 41 (Suppl A):1–7. Chase M, Klasco RS, Joyce NR, Donnino MW, Wolfe RE, Shapiro NI. Predictors of bacteremia in emergency department patients with suspected infection. Am J Emerg Med 2012; 30:1691–1697.
18
19
20
21
22
23
24
25
26
Coburn B, Morris AM, Tomlinson G, Detsky AS. Does this adult patient with suspected bacteremia require blood cultures? Jama 2012; 308:502–511. Nielsen SL, Lassen AT, Gradel KO, Jensen TG, Kolmos HJ, Hallas J, et al. Bacteremia is associated with excess long-term mortality: A 12-year population-based cohort study. J Infect 2015; 70:111–126. Howell MD, Talmor D, Schuetz P, Hunziker S, Jones AE, Shapiro NI. Proof of principle: the predisposition, infection, response, organ failure sepsis staging system. Crit Care Med 2011; 39:322–327. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, et al. Surviving Sepsis Campaign: international guidelines for management of severe sepsis and septic shock, 2012. Intensive Care Med 2013; 39:165–228. Arnold RC, Sherwin R, Shapiro NI, O’Connor JL, Glaspey L, Singh S, et al. Multicenter observational study of the development of progressive organ dysfunction and therapeutic interventions in normotensive sepsis patients in the emergency department. Acad Emerg Med 2013; 20:433–440. Jones AE, Trzeciak S, Kline JA. The Sequential Organ Failure Assessment score for predicting outcome in patients with severe sepsis and evidence of hypoperfusion at the time of emergency department presentation. Crit Care Med 2009; 37:1649–1654. Lin JN, Lai CH, Chen YH, Chang LL, Lu PL, Tsai SS, et al. Characteristics and outcomes of polymicrobial bloodstream infections in the emergency department: a matched case–control study. Acad Emerg Med 2010; 17:1072–1079. Arpi M, Bentzon MW, Jensen J, Frederiksen W. Importance of blood volume cultured in the detection of bacteremia. Eur J Clin Microbiol Infect Dis 1989; 8:838–842. Mermel LA, Maki DG. Detection of bacteremia in adults: consequences of culturing an inadequate volume of blood. Ann Intern Med 1993; 119:270–272.
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
Original article
Mortality and prognostic factors of patients who have blood cultures performed in the emergency department: a cohort study Katrine P. Lindviga, Stig L. Nielsenb, Daniel P. Henriksena, Thøger G. Jensenc, Hans Jørn Kolmosc, Court Pedersenb, Pernille J. Vinholtd and Annmarie T. Lassena Background Early identification and treatment of patients with severe infection improve their prognosis. The aims of this study were to describe the 30-day mortality and to identify prognostic factors among blood-cultured patients in a medical emergency department (MED). Patients and methods This was a hospital-based cohort study including all adult (≥15 years old) blood-cultured patients at the MED at Odense University Hospital between 1 August 2009 and 31 August 2011. Results During the study period, 5499/11 988 (45.9%) patients had blood cultures performed within 72 h of arrival and were included in the study. Of those included, 2631 (47.8%) were men, median age 69 years (range 15–103), and 418 (7.6%) were diagnosed with bacteraemia. The overall 30-day mortality among blood-cultured patients was 11.0% (10.2–11.9). In a multivariate Cox regression model, age of more than 80 years [hazard ratio (HR) 4.6 (95% CI 3.6–6.0)], at least two organ failure [HR 3.6 (2.9–4.5)], bacteraemia [HR 1.4 (1.1–1.8)], Charlson Comorbidity Index of at least 2 h [HR 1.7 (1.3–2.0)], SIRS [HR 1.5 (1.2–1.7)], a history of alcohol dependency [HR 1.7 (1.3–2.3)] and late drawing of blood cultures 24–48 h after arrival [HR 1.7 (1.3–2.2)] were
found to be prognostic factors of mortality among bloodcultured patients in the MED. Conclusion Among blood-cultured patients in the MED, we found an 11.0% overall 30-day mortality. Factors associated with 30-day mortality were age more than 80 years, at least two organ failure, bacteraemia, Charlson Comorbidity Index of at least 2, SIRS, a history of alcohol dependency and late drawing of blood cultures. European Journal of Emergency Medicine 23:166–172 Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved. European Journal of Emergency Medicine 2016, 23:166–172 Keywords: bacteraemia, blood cultures, cohort, emergency medicine, prognostic factors and mortality Departments of aEmergency Medicine, bInfectious Diseases, cClinical Microbiology and dClinical Biochemistry and Pharmacology, Odense University Hospital, Odense C, Denmark Correspondence to Katrine P. Lindvig, Department of Emergency Medicine, Odense University Hospital, Sdr. Boulevard 29, 5000 Odense C, Denmark Tel: + 45 65 41 47 67; fax: + 45 6613 0950/ + 45 654 11571; e-mail: [email protected] Received 12 August 2014 Accepted 14 January 2015
Introduction
Patients and methods
Community-onset bacteraemia is common among patients in the emergency department (ED) [1–3] and is associated with sepsis and septic shock [1]. Early identification and appropriate antibiotic therapy have been shown to reduce the in-hospital mortality of septic shock from 80 to 20–30% [4]. Thus, it could be expected that early diagnosis and treatment of ED patients with suspected bacteraemia would also relate to a better outcome [2].
Study design and participants
It is furthermore essential for the clinician to be able to assess the patient’s risk for a complicated outcome, whether organ failure or death, already at the beginning of the treatment in the ED [1,2,5]. The aims of this study were to describe the 30-day mortality rate and to identify prognostic factors among patients with suspected bacteraemia to accurately identify patients at risk of a complicated outcome in the medical emergency department (MED). 0969-9546 Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
We carried out a population-based prospective cohort study consecutively enrolling all first-time admitted adult patients (age ≥ 15 years) at the MED at Odense University Hospital, who had blood cultures drawn upon arrival or within 72 h of admission, from 1 August 2009 to 31 August 2011. Patients were included at their first MED contact in the observation period. Odense University Hospital is an 1100-bed level 1-trauma centre and a university teaching hospital with all specialities present. It serves both as a tertiary and as a primary hospital, with a primary catchment area of 288 000 individuals. The MED had ∼ 9000 admissions per year during the observation period. Data sources
As part of the standard care at the ED at Odense University Hospital, all patients had their blood pressure, pulse rate, respiratory frequency, oxygen saturation, DOI: 10.1097/MEJ.0000000000000250
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
Blood cultured patients in the ED Lindvig et al. 167
rectal temperature and level of consciousness/Glasgow Coma Scale measured or calculated upon arrival, and had standard blood samples drawn including leucocyte count, platelet count, C-reactive protein, creatinine, international normalized ratio and bilirubin. In addition, a proportion of the patients were subjected to an arterial blood test including PaO2, PaCO2, lactate and pH. The decision to draw blood cultures was made by the attending physicians. All vital values were recorded electronically in the hospital database at the time of measurement, and we did not have to extract these data manually. Information on microbiological and biochemical results as well as comorbid conditions was obtained by linking all included patients at an individual level to five large populationbased registers: (1) The Danish Civil Registration System, which covers data on births, deaths, migration, etc. Since 1968, each Danish resident has been assigned a unique 10-digit civil registration number that enables unambiguous linkage between registries [6]. (2) The Danish National Patient Register, covering all somatic and psychiatric inpatietns and outpatients in all Danish hospitals, enabling true population-based epidemiological research [7]. (3) Odense Pharmaco Epidemiological Database, which contains information on redeemed prescription at pharmacies in the Region of Southern Denmark. The database enables the collection of longitudinal drug histories and allows for linkage of prescription data to other population-based Danish registries [8,9]. (4) The Clinical Microbiological Registration System at Odense University Hospital, containing all blood cultures obtained in the area since 2000. (5) The Clinical Biochemical Registration System at Odense University Hospital containing all laboratory tests obtained in the area since 2000. Microbiological methods
The blood cultures were incubated and screened for the growth of microorganisms for 6 days or until detected positive, using the Bactec 9240 system (Becton Dickinson, Franklin Lakes, New Jersey, USA) until January 2011 and the Bact/Alert system (BioMérieux, Rhône-Alpes, France) thereafter. Routine methods for identification of bacteria were based on conventional characterization [10], the Danish reference programme (http://www.dskm.dk) and automated identification using Vitek 2 (BioMérieux) and MALDI-TOF (SARAMIS; BioMérieux). Definitions
Community-onset bacteraemia was defined as having a positive blood culture drawn within the first 2 days of admission. A blood culture consisted of two blood culture
sets, each comprising one aerobic and one anaerobic bottle, and we defined bacteraemia as either: (a) recognized pathogens detected in at least one blood culture or (b) common skin contaminants (coagulase-negative staphylococci, Bacillus spp., Propionibacterium spp., Corynebacterium spp., viridans group streptococci, Aerococcus spp. or Micrococcus spp.) detected in at least two blood culture sets within 5 days [11–13]. The date of the first positive blood culture set was considered the date of bacteraemia. Polymicrobial bacteraemia was defined as isolation of at least two different microorganisms, deemed to represent bacteraemia within 2 days [14]. The timing of blood cultures was defined as day 0 (0–24 h), day 1 (24–48 h) and day 2 (48–72 h). Late timing of blood cultures was defined from day 1 and onwards (24–72 h). To account for comorbidity, all patients were classified according to the Charlson Comorbidity Score into groups: 0, no comorbidity; 1, light/moderate comorbidity or; ≥ 2, high comorbidity [15]. Systemic Inflammatory Response Syndrome (SIRS) was defined as being present if at least two of the following four criteria were fulfilled: body temperature more than 38.0 or less than 36.0°C, respiratory frequency more than 20 breaths/min or PaCO2 less than 4.3 kPa, pulse rate more than 90 beats/min and leucocyte count more than 12.0 or less than 4.0 × 109/l [16]. Patients were classified as immunocompromised if they had a moderate to high intake of immunosuppressant medication, for example 40 mg prednisone or its equivalent, Primary immunodeficiency diseases, AIDS, a newly diagnosed malignancy registered in the Danish Cancer Registry up to a year before admission or a diagnosis of organ transplantation (Table 1). Patients were classified as having a history of alcohol dependency if they had at least admissions with a discharge diagnosis of an acute alcohol episode, at least one admission with a discharge diagnosis of a chronic alcohol related diagnosis, a redeemed prescription of disulfiram from 2007 and onwards or a registration in The National Register of Alcohol Dependency Treatment (Table 1). Organ system failure was defined on the basis of accessible clinical parameters. Central nervous system failure: Glasgow Coma Scale less than 14 or if the patient is unresponsive or presenting with confusion and disorientation. Respiratory failure: PaO2 less than or equal to 9.75 kPa or PO2 less than 92%. Circulatory failure: systolic blood pressure 90 mmHg or less. Perfusion failure: lactate at least 2.5 mmol/l or pH 7.3 or less. Renal failure: Creatinine at least 177 or a 100 μmol/l increase in p-creatinine among patients with known renal disease. Coagulation failure: international normalized ratio more than 1.6 or platelet count 100 × 109/l or less. Hepatic failure: bilirubin at least 43 μmol/l.
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
168 European Journal of Emergency Medicine 2016, Vol 23 No 3
Table 1
Definition of immunosuppression
Criteria
ICD-10 codes
A moderate to high intake of immunosuppressant medication Primary immunodeficiency diseases AIDS defining diseases A newly diagnosed malignancy registered in the Danish Cancer Registry (up to a year before admission) A diagnosis of organ transplantation Definition of a history of alcohol dependency An acute alcohol episode A chronic alcohol-related diagnosis A redeemed prescription of disulfiram
ATC: H02AB or L00–L04 ≥120 cumulative defined daily dose in the preceding 120 days with a compliance rate of 80% ICD-10: D61; D70–D71; D80–D82; D84; E70.3; G11.3 ICD-10: B20–B24; F02.41–F02.44 ICD-10: C00–C79; C97; D37–D44, D48 ICD-10: Z94 F10.0–F10.1; T519 ICD-10: F10.2–F10.9; E24.4; E52.9A; G31.2; G62.1; G72.1; I42.6; K29.2; K70; K85.2; K86.0; L27.8A; O35.4; P04.3; T50.0A; Z71.4; Z72.1; Z50.2 ATC: N07BB01 from 2007 onwards
Patients registered in The National Register of Alcohol Abuse Treatment.
Statistical analysis
Baseline characteristics were presented in contingency tables as numbers and percentages. Univariate analyses were carried out and tested using the χ2-test, which was two-sided. A P-value of less than 0.05 was considered statistically significant. The 30-day mortality was reported as the proportion of patients who died within 30 days and was shown in a Kaplan–Meier failure plot. In case of missing values, variables were registered as normal. The association with death of predefined variables (bacteraemia, sex, age, Charlson Comorbidity Score, alcohol dependency, SIRS, immunosuppression, organ failure and timing of blood cultures) was evaluated in univariate analyses using the log-rank test. The bacteraemias were grouped as Gram-negative bacteraemia, Gram-positive bacteraemia and polymicrobial bacteraemia and analysed in a univariate Cox regression analysis. All predefined variables with a significance level of less than 0.05 were included in the multivariate Cox regression model, except for female sex [hazard ratio (HR) 1.1 (0.9–1.4)], which was finally included in the final model because of predefined interest in the variable. The proportional hazard assumptions were tested graphically using the Cox–Snell residuals and found appropriate. Statistical analyses were carried out using Stata 12.1 (Stata Corp LP, College Station, Texas, USA).
Table 2 Baseline characteristics of patients with blood cultures in the Medical Emergency Department
Characteristics Sex Male Female Age Mean: 65 years, SD: 20.5, range: 15–102 15–64 65–79 > 80 Charlson comorbidity index 0 1 2 Immune compromised No Yes Alcohol use No Yes SIRS No Yes Organ failure 0 failure 1 failure ≥ 2 failures Timing of blood cultures Day 0 Day 1 Day 2
Blood-cultured patients total (n = 5499) [n (%)]
Missing values (%) 0
2631 (47.8) 2868 (52.2) 0
2357 (42.9) 1527 (27.8) 1615 (29.3) 0 2258 (41.1) 1226 (22.3) 2015 (36.6) 0 4714 (85.7) 785 (14.3) 14 (0.2) 4954 (90.1) 531 (9.7) 2623 (47.7) 2876 (52.3) 2994 (54.4) 1715 (31.2) 790 (14.4) 0 4540 (82.6) 646 (11.7) 313 (5.7)
Baseline characteristics of blood-cultured patients in the Medical Emergency Department.
Ethics
The study was approved by the Danish Data Protection Agency (No. 2008-58-0035) and the Danish National Board of Health (No. 3-3013-35). In observational studies such as in this, review by an Ethics Board is not required according to Danish law.
Results Descriptive
During the study period, we identified 12 027 patients with a first-time admission, of whom 5499 (45.7%) had blood cultures performed within 2 days from admission and were included in the study. Of these, 418 (7.6%) had bacteraemia, corresponding to 3.5% of all MED patients.
A baseline description of patients with blood cultures in the MED is presented in Table 2. In general, patients who had blood cultures drawn were more often female (52.2%), younger (15–64 years old), had no comorbidity (41.1%), were not immunosuppressed (85.7%), had no alcohol dependency (90.3%) and had no organ failure (54.5%). However, they were likely to present with SIRS and had blood cultures drawn on the day of arrival. The most frequent pathogens detected were Escherichia coli (30.4%), Streptococcus pneumoniae (13.0%) and Staphylococcus aureus (10.3%).
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
Blood cultured patients in the ED Lindvig et al. 169
Mortality
In the present study, we found an overall 30-day mortality rate of 11.0% [P < 0.01, 95% confidence interval (CI): 10.2–11.9] among blood-cultured patients in the MED, with a 30-day mortality rate of 19.6% (P < 0.01, 95% CI: 15.6–24.3) for bacteraemic patients compared with 10.3% (P < 0.01, 95% CI: 9.5–11.3) among bloodcultured nonbacteraemic patients (Fig. 1).
11.0% (10.2–11.9). These results are high compared with previous studies of blood-cultured ED patients, which have reported in-hospital mortality rates of 3.6–4.1% [5,17]. This might be because of the different clinical thresholds used to obtain blood cultures in the EDs or because of fundamental differences in the patient population. Positive blood culture
Prognostic factors
In the multiple Cox regression analysis, we found age older than 80 years (HR 4.6, 95% CI: 3.6–6.0) and at least two organ failure (HR 3.6, 95% CI 2.9–4.5) to be the strongest predictors of 30-day mortality among blood-cultured patients in the MED (Table 3). Furthermore, bacteraemia [HR 1.4 (1.1–1.8)], Charlson Comorbidity Index at least 2 [HR 1.7 (1.3–2.0)], SIRS [HR 1.5 (1.2–1.7)], a history of alcohol dependency [HR 1.7 (1.3–2.3)] and drawing of blood cultures 24–48 h after arrival [HR 1.7 (1.3–2.2)] were found to be prognostic factors of mortality. In the group with blood cultures drawn from 48–72 h, 16/106 (15.1%) died within 30 days, representing a HR of 1.6 (0.9–2.5), although not significant. In a univariate Cox regression analysis, we found polymicrobial bacteraemia to be associated with increased mortality [HR 2.9 (1.5–5.7)] compared with patients with Gram-negative bacteraemia (Table 4).
Discussion Mortality Negative blood culture
In this hospital-based cohort study of blood-cultured patients in the MED, we found a 30-day mortality of
Among blood-cultured patients with a positive blood culture, we found a relatively high 30-day mortality of 19.6%; however, the results are comparable to the previous literature, ranging from 14 to 37% [18], and overall mortality of 22%, irrespective of the patient population in a recently published study by Nielsen et al. [19]. In the present study, the clinical threshold for obtaining blood cultures was entirely based on the physician’s clinical decision. In comparison, in a study of blood-cultured patients in the ED, Shapiro et al. [5] reported an overall 28-day mortality rate of 4.1%. They found that SIRS did not add any additional prognostic value, whereas each additional organ dysfunction doubles the 1-year mortality risk. To understand this marked difference in mortality between the two studies, it is necessary to outline the differences in patient populations. Shapiro et al. [5] have included all patients at risk of infection; if they had blood cultures drawn within 3 h of admission; hereof 16% were treated as outpatients. The patients were younger (mean age 59.9 vs. 64.5 years), and a lower proportion had SIRS (29.2 vs. 52.3%) [5]. In the present study, we included only admitted patients, who, in general, are more likely to
Fig. 1
Probability of death (%)
0.30 0.25 0.20 0.15 0.10 0.05 0.00 0
5
10
15
20
25
30
Days Number at risk Blood culture 5081 Bacteraemia 418
4795 360 Bacteraemia = 0
4648 346
4561 338
Bacteraemia = 1
Kaplan–Meier failure estimates for blood-cultured patients in the MED.
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
170
European Journal of Emergency Medicine 2016, Vol 23 No 3
Table 3
Prognostic factors among emergency medical patients with blood cultures
Total
Patients alive within 30 days [n (%)]
Patients dead within 30 days [n (%)]
4556 (89.7) 336 (80.4)
525 (10.3) 82 (19.6)
2.0 (1.6–2.6)
1.0 1.4 (1.1–1.8)
2573 (89.7) 2319 (88.1)
295 (10.3) 312 (11.9)
1.1 (0.9–1.4)
1.0 1.1 (0.9–1.2)
2263 (96.0) 1343 (88.0) 1286 (79.6)
94 (4.0) 184 (12.0) 329 (20.4)
2.2 (2.0–2.5)
1.0 2.4 (1.9–3.2) 4.6 (3.6–6.0)
2120 (93.9) 1098 (89.6) 1674 (83.1)
138 (6.1) 128 (10.4) 341 (16.9)
1.7 (1.5–1.9)
1.0 1.1 (0.9–1.4) 1.7 (1.3–2.0)
4218 (89.5) 674 (85.9)
496 (10.5) 111 (14.4)
1.4 (1.1–1.7)
1.0 1.1 (0.9–1.4)
2254 (92.0) 2638 (86.5)
195 (8.0) 412 (13.5)
1.8 (1.5–2.1)
1.0 1.5 (1.2–1.7)
2834 1512 546
179 (5.6) 210 (10.9) 218 (22.2)
2.4 (2.1–2.6)
1.0 1.6 (1.3–1.9) 3.6 (2.9–4.5)
4423 (89.3) 458 (86.2)
531 (10.7) 73 (13.8)
1.3 (1.0–1.7)
1.0 1.7 (1.3–2.3)
4413 (89.5) 389 (84.0) 90 (84.9)
517 (10.5) 74 (16.0) 16 (15.1)
1.4 (1.1–1.6)
1.0 1.7 (1.3–2.2) 1.6 (0.9–2.5)
Bacteraemia No (5081) Yes (418) Sex Female (2868) Male (2631) Age 15–64 (2357) 65–79 (1527) > 80 (1615) Charlson comorbidity index 0 (2258) 1 (1226) ≥ 2 (2015) Immune compromised No (4714) Yes (785) SIRS No SIRS (2449) SIRS (3050) Organ failure 0 failure (3013) 1 failure (1722) ≥ 2 failures (764) Alcohol use No (4954) Yes (531) Timing of blood cultures Day 0 (0–24 h) (4930) Day 1 (24–48 h) (463) Day 2 (48–72 h) (106)
Crude hazard ratio (95% CI)
Multivariate COX hazard ratio (95% CI)
CI, confidence interval. Prognostic factors among blood-cultured patients in the Medical Emergency Department, total n = 5499. Patients alive within 30 days, n = 4892. Patients dead within 30 days, n = 607.
Table 4
Bacteraemic patients – difference in mortality divided by Gram colour Patients at start (n = 418) [n (%)]
Crude 30-day mortality (n = 82) [n (%)]
Univariate log-rank P-value
Univariate COX analysis HR
211 (50.5) 177 (42.3) 30 (7.2)
33 (15.6) 37 (20.9) 12 (40.0)
< 0.001
1.0 1.4 (0.9–2.2) 2.9 (1.5–5.7)
Gram-negative Gram-positive Polymicrobial
Univariate Cox regression analysis. HR, hazard ratio.
have a poorer prognosis compared with the outpatient population. These fundamental differences in patient population may explain the difference in short-term mortality and may explain why the two studies do not agree that SIRS is a prognostic factor among bloodcultured patients. However, differences in the quality of care provided to patients at risk of bacteraemia, such as time to initiation of treatment with antibiotics and fluid, may induce fundamental variations in mortality among different EDs. The present study could not access time to antibiotics and fluid administration.
disease, chronic liver disease, residence in a nursing home and malignancy (with and without metastasis), infection (pneumonia and skin/soft tissue infection), response (tachypnoea, tachycardia and bandemia) and organ dysfunction (renal, respiratory, cardiac, metabolic and haematologic) as prognostic factors for bloodcultured patients in the ED [20]. The present study thereby confirms the identified prognostic factors as remaining relevant even after 8 years, and despite a different healthcare setting. Organ failure
Prognostic factors
Previous studies have identified prognostic factors among blood-cultured patients in the ED. A study by Howell et al. [20] identified age, chronic obstructive pulmonary
It has been shown previously that the combination of sepsis and organ dysfunction is associated with increased morbidity and mortality [4,5,21–23]. In agreement with these studies, we found organ failure to be an important
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
Blood cultured patients in the ED Lindvig et al. 171
prognostic factor among blood-cultured patients in the MED.
Timing of blood cultures in the emergency department
We found that the timing of blood cultures in the MED was related to an increased risk of mortality if the blood culture was performed the day after arrival, which was especially evident among bacteraemic patients (the 30-day mortality rate for patients’ blood cultured day 0, 1 and 2 is 10.5, 16.0 and 15.1%, respectively). Perhaps the observed increased risk is because of the selection of patients, who, several hours after arrival, experience deterioration in their clinical condition, followed by medical re-evaluation and blood cultures, or perhaps, it is an indicator of suboptimal treatment and evaluation of the patient at arrival. In the present observational study, we could not distinguish between these explanations, but a manual review of all patients with blood cultures performed 2 days after arrival showed a high proportion of patients in whom the blood cultures had been ordered when the patient arrived, however, by mistake first performed the following day. To our knowledge, no other studies have evaluated what impact the timing blood cultures have as a prognostic factor for mortality in the ED. However, studies of severe sepsis have found that early diagnosis and treatment of patients with severe sepsis improve their diagnosis [21], thereby pinpointing the need for early diagnosis and treatment. Thus, the prognostic factor of late drawing of blood cultures supports the theory of early goal-directed therapy because it indicates that late diagnosis and deterioration in clinical condition might worsen the patients’ outcome. In a univariate analysis, we found that patients with polymicrobial bacteraemia had a high mortality rate and an increased HR of dying within 30 days compared with patients with Gram-negative bacteraemia, which is in accordance with other studies [24]. The strengths of our study are the consecutive inclusion of all adult first-time admission patients with blood cultures drawn upon arrival to the MED within the study period and the ability to follow up on all the patients included because of the unique personal identification number assigned to all Danish citizens. Furthermore, blood-cultured patients had two sets of blood cultures drawn, representing four bottles for each blood-cultured patient, which increases the credibility of a positive blood culture [25,26]. However, a source of error for blood cultures is the presence of contaminants. This study defined that for a common commensal bacterium to be considered a positive bacteraemia, the bacterium must be isolated from at least two blood culture sets, within a 5-day period; otherwise, the isolated bacterium is considered to be a contaminant, and thereby excluded from the study, which minimizes – but does not rule out – the risk of inclusion of false-positive bacteraemic episodes.
This study has some potential limitations. The definition of organ failure is not validated; however, to a large extent, our definitions of severe sepsis are similar to the recommendations by the Surviving Sepsis Campaign from 2012 [21] and the definitions proposed by Shapiro et al. [2], but we were restricted by an inadequate systematic registration of treatment with oxygen therapy, making it impossible to estimate the PaO2/FiO2 ratio. The MED at Odense University Hospital does not receive obvious cardiological, oncological, haematological, nephrological or patients with upper gastrointestinal bleeding, and paediatric patients. Consequently, the results do not apply to all acute medical patients. Furthermore, this is a single-centre study, reflecting the standard care at Odense University Hospital within this period, and therefore, the results may not be entirely generalizable to other wards and hospitals. Conclusion
As it is essential for the clinician to accurately assess the patient’s risk of mortality and assumed prognosis, for those with suspected bacteraemia in the MED, this study provides further information. We found an 11.0% overall 30-day mortality rate. Among blood-cultured patients in the MED older than 80 years of age, at least two organ failure, bacteraemia, Charlson Comorbidity Index of at least 2, SIRS, a history of alcohol dependency and drawing of blood cultures 24–48 h after arrival were associated with an increased 30-day mortality among blood-cultured patients in the MED.
Acknowledgements Katrine Prier Lindvig is supported by an unrestricted grant from Fonden af 17121981. Annmarie Lassen is supported by an unrestricted grant from the private philanthropic foundation TrygFoundation to University of Southern Denmark. Court Pedersen has received payment for educational activities from Gilead, Abbott and MSD. Conflicts of interest
There are no conflicts of interest.
References 1
2
3
4 5
Shapiro NI, Wolfe RE, Wright SB, Moore R, Bates DW. Who needs a blood culture? A prospectively derived and validated prediction rule. J Emerg Med 2008; 35:255–264. Shapiro NI, Wolfe RE, Moore RB, Smith E, Burdick E, Bates DW. Mortality in Emergency Department Sepsis (MEDS) score: a prospectively derived and validated clinical prediction rule. Crit Care Med 2003; 31:670–675. Lindvig KP, Henriksen DP, Nielsen SL, Jensen TG, Kolmos HJ, Pedersen C, et al. How do bacteraemic patients present to the emergency department and what is the diagnostic validity of the clinical parameters; temperature, C-reactive protein and systemic inflammatory response syndrome? Scand J Trauma Resusc Emerg Med 2014; 22:39. Angus DC, van der Poll T. Severe sepsis and septic shock. N Engl J Med 2013; 369:840–851. Shapiro N, Howell MD, Bates DW, Angus DC, Ngo L, Talmor D. The association of sepsis syndrome and organ dysfunction with mortality in emergency department patients with suspected infection. Ann Emerg Med 2006; 48:583–590.
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
172
6 7 8 9 10 11
12
13
14 15
16 17
European Journal of Emergency Medicine 2016, Vol 23 No 3
Pedersen CB. The Danish Civil Registration System. Scand J Public Health 2011; 39 (7 Suppl):22–25. Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient Register. Scand J Public Health 2011; 39 (Suppl):30–33. Gaist D, Sorensen HT, Hallas J. The Danish prescription registries. Dan Med Bull 1997; 44:445–448. Kildemoes HW, Sorensen HT, Hallas J. The Danish National Prescription Registry. Scand J Public Health 2011; 39 (7 Suppl):38–41. Murray PR BE, Pfaller MA, Tenover FC, Yolken RH. Manual of clinical microbiology, 7th ed. Washington, DC: Americal Society for Microbiology; 1999. Gradel KO, Knudsen JD, Arpi M, Ostergaard C, Schonheyder HC, Sogaard M. Classification of positive blood cultures: computer algorithms versus physicians’ assessment - development of tools for surveillance of bloodstream infection prognosis using population-based laboratory databases. BMC Med Res Methodol 2012; 12:139. Leal J, Gregson DB, Ross T, Flemons WW, Church DL, Laupland KB. Development of a novel electronic surveillance system for monitoring of bloodstream infections. Infect Control Hosp Epidemiol 2010; 31:740–747. Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control 2008; 36:309–332. Roberts FJ. Definition of polymicrobial bacteremia. Rev Infect Dis 1989; 11:1029–1030. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40:373–383. Nystrom PO. The systemic inflammatory response syndrome: definitions and aetiology. J Antimicrob Chemother 1998; 41 (Suppl A):1–7. Chase M, Klasco RS, Joyce NR, Donnino MW, Wolfe RE, Shapiro NI. Predictors of bacteremia in emergency department patients with suspected infection. Am J Emerg Med 2012; 30:1691–1697.
18
19
20
21
22
23
24
25
26
Coburn B, Morris AM, Tomlinson G, Detsky AS. Does this adult patient with suspected bacteremia require blood cultures? Jama 2012; 308:502–511. Nielsen SL, Lassen AT, Gradel KO, Jensen TG, Kolmos HJ, Hallas J, et al. Bacteremia is associated with excess long-term mortality: A 12-year population-based cohort study. J Infect 2015; 70:111–126. Howell MD, Talmor D, Schuetz P, Hunziker S, Jones AE, Shapiro NI. Proof of principle: the predisposition, infection, response, organ failure sepsis staging system. Crit Care Med 2011; 39:322–327. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, et al. Surviving Sepsis Campaign: international guidelines for management of severe sepsis and septic shock, 2012. Intensive Care Med 2013; 39:165–228. Arnold RC, Sherwin R, Shapiro NI, O’Connor JL, Glaspey L, Singh S, et al. Multicenter observational study of the development of progressive organ dysfunction and therapeutic interventions in normotensive sepsis patients in the emergency department. Acad Emerg Med 2013; 20:433–440. Jones AE, Trzeciak S, Kline JA. The Sequential Organ Failure Assessment score for predicting outcome in patients with severe sepsis and evidence of hypoperfusion at the time of emergency department presentation. Crit Care Med 2009; 37:1649–1654. Lin JN, Lai CH, Chen YH, Chang LL, Lu PL, Tsai SS, et al. Characteristics and outcomes of polymicrobial bloodstream infections in the emergency department: a matched case–control study. Acad Emerg Med 2010; 17:1072–1079. Arpi M, Bentzon MW, Jensen J, Frederiksen W. Importance of blood volume cultured in the detection of bacteremia. Eur J Clin Microbiol Infect Dis 1989; 8:838–842. Mermel LA, Maki DG. Detection of bacteremia in adults: consequences of culturing an inadequate volume of blood. Ann Intern Med 1993; 119:270–272.
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
