Abstract
Introduction:
In this study, we investigated whether the Sequential Organ Failure Assessment (SOFA) score performance differs based on the type of infection among patients admitted to the intensive care unit (ICU) with infection.
Materials and Methods:
Single-center, retrospective study of adult ICU patients admitted with infection between January 2008 and April 2018 at an urban tertiary care center. Patients were uniquely classified into different infection types based on International Classification of Diseases, Ninth Revision (ICD-9) and ICD-10 codes. Infection types included were pneumonia, meningitis, bacteremia, cellulitis, cholangitis/cholecystitis, intestinal and diarrheal disease, endocarditis, urinary tract infection (UTI), and peritonitis. The SOFA score performance and mortality in relation to SOFA score were compared across infection types.
Results:
A total of 12 283 patients were included. Of these, 50.6% were female and the median age was 70 years (interquartile range: 57–82). The most common infection types were pneumonia (32.2%) and UTI (31.0%). Overall, 1703 (13.9%) patients died prior to hospital discharge. The median baseline SOFA score (within 24 hours of ICU admission) for the cohort was 5 (3–8). Patients with peritonitis had the highest median SOFA score, 7 (4–9), and patients with cellulitis and UTI had the lowest median SOFA score, 4 (2–7). The SOFA score discrimination to predict mortality was highest among patients with endocarditis (area under the receiver operating characteristic [AUC]: 0.79, 95% CI: 0.69–0.90) and lowest for patients with isolated bacteremia (AUC: 0.59, 95% CI: 0.49–0.70). Observed mortality by quartile of SOFA score differed substantially across infection types.
Conclusions:
Type of infection is an important consideration when interpreting the SOFA score. This is relevant as SOFA emerges as an important tool in the definition and prognostication of sepsis.
Keywords: organ dysfunction scores, sepsis, mortality, area under curve, clinical decision-making, critical care
Introduction
The definition of sepsis has evolved from 1991 to the present.1–3 The Society of Critical Care Medicine and European Society of Intensive Care Medicine Sepsis III Task Force recently redefined sepsis as life-threatening organ dysfunction caused by a dysregulated host response to infection. To provide for the clinical evaluation of “life-threatening organ dysfunction,” an acute change in total Sequential Organ Failure Assessment (SOFA) score of ≥2 was suggested.3,4 The SOFA score has been shown to have good prognostic accuracy for mortality in patients with sepsis.4–7
While the SOFA score can be useful for discriminating between patients who will survive versus those who will not survive,7 SOFA score calculation does not account for the source of infection. The primary site of infection may affect mortality outcomes.8,9 Failing to distinguish between different sources of infection may overpredict or underpredict mortality in certain patient groups based on the primary site of infection.
We hypothesized that SOFA score performance to predict mortality differs based on the type of infection among patients admitted to the intensive care unit (ICU).
Material and Methods
Study Design and Cohort Selection
This was a single-center, retrospective study performed at an urban tertiary center. Adult patients (aged 17 or older) admitted to the ICU between January 2008 and April 2018 with suspected infection were included. Suspected infection was defined by the collection of any microbial cultures and the initiation of antibiotics, both within 24 hours of ICU admission time.4 Based on previously validated International Classification of Diseases, Ninth Revision (ICD-9) and ICD-10 codes (see Supplement Table 1S), patients having at least 1 code for pneumonia, meningitis, bacteremia, cellulitis, cholangitis/cholecystitis, intestinal and diarrheal disease (eg, infectious gastroenteritis and infectious diarrhea), urinary tract infection (UTI), endocarditis, or peritonitis were included. Patients with ICD codes for more than 1 infection type were not included in their respective infection type. Instead they were classified into a separate group “multiple infection types.” Only the index ICU admission was included for each patient, and only patients admitted directly from the emergency department to the ICU were included. The institutional review board of Beth Israel Deaconess Medical Center approved this study.
Data Collection and SOFA Score Calculation
Demographic data, vital signs, laboratory values, variables for SOFA score calculation, and outcomes were extracted from the electronic medical record. Comorbidities were categorized based on Elixhauser guidance.10
The SOFA score11 was calculated using the worst available measure for each variable recorded within the 24 hours of ICU admission time. Accuracy of extracted data was confirmed through manual chart review of randomly selected charts. A modified SOFA score, utilizing oxygen saturation/fraction of inspired oxygen (SaO2/FIO2) ratio as opposed to partial pressure of oxygen/fraction of inspired oxygen (PaO2/FIO2) ratio, was used given the limited number of arterial blood gas samples available in our cohort.12 Missing SOFA variables were assumed to be normal and assigned a score of 0 as performed in prior high impact studies.4,13 For complete details of SOFA score calculation, please see Supplement Table 2S.
Statistical Analysis
Descriptive statistics were reported as means with SDs, medians with interquartile ranges, or counts with frequencies, depending on the type and distribution of data. Between-group comparisons were made with Fisher exact test or χ2 test for categorical data and T test or Wilcoxon rank-sum, analysis of variance, or Kruskal-Wallis test for continuous data as appropriate. For all analyses, patients with a single infectious type were analyzed separately from those with coding for multiple infections.
The discrimination and calibration of SOFA to predict in-hospital mortality was calculated for each infection type using the area under the receiver operating characteristic (AUC) and Hosmer-Lemeshow test, respectively. Change in in-hospital mortality with rising SOFA score in different diseases was calculated and represented visually. For this representation, the total SOFA score was divided into quartiles based on the distribution of SOFA score for the complete cohort (SOFA score: 0–3, 4–5, 6–8, >8). To assess whether the addition of infection type to SOFA score improved model discrimination for the outcome of in-hospital mortality, 2 ROC curves were constructed and their AUCs compared using a test of equality of ROC areas (Stata roccomp command). The above models were adjusted for measures of baseline risk including age, gender, and select comorbidities using logistic regression (as in Table 1).
Table 1.
Baseline Cohort Characteristics.
| Whole cohort | Cellulitis | Cholangitis/cholecystitis | Endocarditis | Meningitis | Peritonitis | Pneumonia | UTI | Bacteremia | |
|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||
| n (%) | 12 283 (100%) | 597 (4.86) | 478 (3.89) | 98 (0.80) | 49 (0.40) | 396 (3.22) | 3950 (32.16) | 3811 (31.03) | 349 (2.84) |
| Age, median (IQR) | 70 (57–82) | 61 (48–71) | 71 (62–81) | 55.5 (38–69) | 53 (38–64) | 64 (53–75) | 68.5 (56–81) | 74 (61–84) | 63 (52–77) |
| Female gender, n (%) | 6217 (50.61) | 236 (39.53) | 186 (38.91) | 32 (32.65) | 22 (44.90) | 175 (44.19) | 1694 (42.89) | 2421 (63.53) | 145 (41.55) |
| Caucasian ethnicity, n | 8549 | 455 | 370 | 70 | 35 | 291 | 2637 | 2687 | 218 |
| Comorbidities, n (%) | |||||||||
| Other neurological condition | 1953 (15.9) | 63 (10.55) | 32 (6.69) | 12 (12.24) | 21 (42.86) | 22 (5.56) | 493 (12.48) | 810 (21.25) | 53 (15.19) |
| Coagulopathy | 2492 (20.29) | 96 (16.08) | 123 (25.73) | 42 (42.86) | 10 (20.41) | 98 (24.75) | 772 (19.54) | 652 (17.11) | 74 (21.20) |
| Congestive heart failure | 4001 (32.5) | 168 (28.14) | 103 (21.55) | 43 (43.88) | 4 (8.16) | 57 (14.39) | 1444 (36.56) | 1160 (30.44) | 106 (30.37) |
| Pulmonary disease | 3359 (27.35) | 141 (23.62) | 95 (19.87) | 20 (20.41) | 2 (4.08) | 68 (17.17) | 1494 (37.82) | 754 (19.78) | 72 (20.63) |
| Liver disease | 1283 (10.45) | 52 (8.71) | 73 (15.27) | 11 (11.22) | 3 (6.12) | 128 (32.32) | 317 (8.03) | 308 (8.08) | 63 (18.05) |
| Renal disease | 3109 (25.31) | 151 (25.29) | 93 (19.46) | 25 (25.51) | 4 (8.16) | 76 (19.19) | 978 (24.76) | 910 (23.88) | 126 (36.1) |
| Cancer | 904 (7.36) | 17 (2.85) | 76 (15.9) | 5 (5.1) | 2 (4.08) | 36 (9.09) | 382 (9.67) | 199 (5.22) | 22 (6.3) |
| Lymphoma | 216 (1.76) | 6 (1.01) | 12 (2.51) | 4 (4.08) | 0 | 4 (1.01) | 73 (1.85) | 50 (1.31) | 10 (2.87) |
| Mortality (rate in %) | 1703 (13.87) | 29 (4.86) | 49 (10.25) | 10 (10.20) | 6 (12.24) | 79 (19.95) | 714 (18.08) | 350 (9.18) | 39 (11.17) |
Abbreviations: IQR, interquartile range; UTI, urinary tract infection.
A sensitivity analysis was performed in the cohort of patients with an explicit ICD-9 or ICD-10 code for sepsis or septic shock (ICD9: 995.91, 785.52; ICD10: A41.9, 65.21). The selection criteria for this analysis are more specific for sepsis.14
All analyses were performed using Stata version 16 (Stata-Corp 2019). Graphical representations were performed using R (R core team 2019). Two-sided P value <.05 was considered to be statistically significant.
Results
Cohort Characteristic
A total of 12 283 patients were included (see Figure 1 for details of cohort selection). The median age of the population sample was 70 (IQR: 57–82), with 50.6% female. Of the total cohort, 8549 (69.6%) patients were white. The most common infection types were pneumonia (32.2%) and UTI (31.0%). A total of 2555 (20.8%) patients were coded as having more than 1 infection type. No patient was coded as having only an intestinal infection (ie, patients with intestinal infection also had concurrent other infection types). A total of 1703 (13.9%) patients died prior to hospital discharge. Complete cohort characteristics can be found in Table 1.
Figure 1.
Cohort selection flow diagram.
The SOFA Score Calculations
The overall median SOFA score for the cohort was 5 (3–8). The median SOFA score was highest in peritonitis, 7 (4–9). The respiratory SOFA subscore was significantly higher in patients with pneumonia as compared to the rest of the cohort excluding patients with pneumonia, 1 (0–2) versus 0 (0–2), P < .001. Similarly, the central nervous system SOFA subscore was higher among patients with meningitis, 3 (1–4) versus 1 (0–3), P < .001, and the hepatobiliary SOFA subscore was higher among patients with cholangitis/cholecystitis, 2 (1–2) versus 0 (0–0), P < .001. See Table 2 for additional details.
Table 2.
Sequential Organ Failure Assessment (SOFA) Scores for Each Infection Type and Outcomes.
| Median (IQR) | SOFA | Respiratory SOFA | Cardiovascular SOFA | CNSSOFA | Renal SOFA | Coagulation SOFA | Liver SOFA |
|---|---|---|---|---|---|---|---|
|
| |||||||
| Whole cohort | 5 (3–8) | 1 (0–2) | 1 (0–1) | 1 (0–3) | 1 (0–1) | 0 (0–1) | 0 (0–0) |
| Cellulitis | 4 (2–7) | 0 (0–2) | 1 (0–1) | 1 (0–2) | 1 (0–1) | 0 (0–1) | 0 (0–0) |
| Cholangitis/cholecystitis | 6 (4–8) | 0 (0–1) | 1 (1–3) | 0 (0–1) | 1 (0–1) | 0 (0–1) | 2 (1–2) |
| Endocarditis | 5 (3–8) | 1 (0–2) | 1 (1–3) | 0.5 (0–3) | 1 (0–2) | 0 (0–1) | 0 (0–0) |
| Meningitis | 5 (2–7) | 2 (0–2) | 1 (0–1) | 3 (1–4) | 0 (0–1) | 0 (0–1) | 0 (0–0) |
| Peritonitis | 7 (4–9) | 0 (0–2) | 1 (1–4) | 1 (0–3) | 1 (0–2) | 0 (0–1) | 0 (0–2) |
| Pneumonia | 5 (3–8) | 1 (0–2) | 1 (0–1) | 1 (0–3) | 0 (0–1) | 0 (0–1) | 0 (0–0) |
| UTI | 4 (2–7) | 0 (0–2) | 1 (0–1) | 1 (0–3) | 0 (0–1) | 0 (0–1) | 0 (0–0) |
| Bacteremia | 5 (3–8) | 1 (0–2) | 1 (0–1) | 1 (0–3) | 1 (0–2) | 0 (0–1) | 0 (0–0) |
Abbreviations: IQR, interquartile range; UTI, urinary tract infection; SOFA, Sequential Organ Failure Assessment.
The SOFA Discrimination by Infection Type
The overall discrimination of the SOFA score to predict hospital mortality for the complete cohort was 0.69 (95% CI: 0.68–0.70). When divided by infection type, the SOFA score discrimination was highest for endocarditis, meningitis, and peritonitis (AUC: 0.79 [95% CI: 0.69–0.90], 0.77 [95% CI: 0.58–0.95], and 0.77 [95% CI: 0.71–0.82], respectively]. The SOFA score discrimination was lowest in patients with bacteremia (AUC: 0.60, 95% CI: 0.49–0.70). Both AUC and adjusted AUC for all infection types and graphical representation of AUC are provided in Table 3 and Figure 2, respectively.
Table 3.
Area under the Receiver Operating Characteristic (AUC) for each Infection Type.
| Infection type | n (%) | Mortality (rate in %) | AUC (95% CI) | Adjusted AUC (95% CI) | H-L test, P value |
|---|---|---|---|---|---|
|
| |||||
| Whole cohort | 12 283 (100) | 1703 (13.87) | 0.69 (0.68–0.70) | 0.75 (0.74–0.76) | .13 |
| Cellulitis | 597 (4.86) | 29 (4.86) | 0.74 (0.66–0.83) | 0.88 (0.83–0.94) | .83 |
| Cholangitis/cholecystitis | 478 (3.89) | 49 (10.25) | 0.73 (0.65–0.8l) | 0.78 (0.71–0.85) | .73 |
| Endocarditis | 98 (0.80) | 10 (10.20) | 0.79 (0.69–0.90) | 0.92 (0.85–0.98) | .99 |
| Meningitis | 49 (0.40) | 6 (12.24) | 0.77 (0.58–0.95) | 0.84 (0.69–0.98) | >.99 |
| Peritonitis | 396 (3.22) | 79 (19.95) | 0.77 (0.71–0.82) | 0.82 (0.77–0.87) | .42 |
| Pneumonia | 3950 (32.16) | 714 (18.08) | 0.67 (0.65–0.69) | 0.75 (0.74–0.77) | .29 |
| UTI | 3811 (31.03) | 350 (9.18) | 0.71 (0.68–0.74) | 0.77 (0.74–0.79) | .93 |
| Bacteremia | 349 (2.84) | 39 (11.17) | 0.60 (0.49–0.70) | 0.79 (0.72–0.86) | .43 |
Abbreviations: AUC, area under the receiver operating characteristic; H-L, Hosmer-Lemeshow; UTI, urinary tract infection.
Figure 2.
The area under the receiver operating characteristic for disease-specific Sequential Organ Failure Assessment (SOFA) to predict mortality. The figure does not include curves for meningitis and endocarditis to allow for easier visual interpretation. Area under the curve for each infectious type are: whole cohort: 0.69, cellulitis: 0.74, cholecystitis/cholangitis: 0.73, peritonitis: 0.77, pneumonia: 0.67, UTI: 0.71, and bacteremia: 0.60.
When the variable for infection type was added to SOFA score in the prediction of in-hospital mortality, the discrimination of the model improved (0.69 [95% CI: 0.68–0.70] vs 0.71 [95% CI: 0.70–0.73], P < .001). Adjusting for baseline risk, the AUC of SOFA in the whole cohort was 0.75 (95% CI: 0.74–0.76) and improved to 0.77 (95% CI: 0.75–0.78) after the inclusion of infection type in the model (P < .001).
The SOFA Score and Mortality by Infection Type
Within each quartile of SOFA score, there was a significant difference in mortality across infection types (P < .001 for each quartile). After excluding patients with multiple infection type, patients with pneumonia had a higher mortality in each SOFA quartile as compared to the remaining patients with other infection types (P < .001 for comparison in each quartile). Within each quartile of SOFA score, patients with cellulitis and pneumonia had substantially different mortality rates (quartile 1: 1.2% vs 9.5%, quartile 2: 4.2% vs 15.6%, quartile 3: 6.1% vs 20.8%, quartile 4: 14.8% vs 33.2%). Patients with cellulitis had lower mortality in each SOFA quartile (P < .01 for comparison in each quartile). Mortality for patients with cellulitis was <10%, except for those in the highest SOFA quartile. See Figure 3 for a visual representation of mortality by infection type.
Figure 3.
Mortality rate for disease-specific Sequential Organ Failure Assessment (SOFA). For ease of visual interpretation, endocarditis and meningitis were not included in this figure. Mortality rates by quartile can be found in the data supplement.
Sensitivity Analysis
A total of 2294 (18.7%) patients carried an explicit sepsis code. In this subgroup, the median overall SOFA score was 7 (5–10) and 587 (25.6%) patients died. Overall, SOFA score discrimination was 0.65 (95% CI: 0.62–0.68), which improved to 0.68 (95% CI: 0.65–0.70) when site of infection was included in the model (P < .001). Variation in model discrimination and mortality by infection type was similar to that in the complete cohort. See Supplement Figures 1S and 2S for complete results.
Discussion
Among patients admitted to the ICU with infection in this study, there was substantial variability in both discrimination and observed mortality by SOFA score when patients were categorized by type of infection. Additionally, discrimination of SOFA score to predict mortality improved when infection type was incorporated in the model.
The definition of sepsis and septic shock is now reliant upon the SOFA score as a determinant of life-threatening organ dysfunction. Specifically, the Sepsis III Task Force determined that an acute infection coupled with a SOFA score of at least 2 (or an acute rise of at least 2 points from baseline) is reflective of life-threatening organ dysfunction and makes the definition of sepsis.3,4 Of note, patients with cellulitis, UTI, and cholecystitis/cholangitis had mortality rates substantially less than 10% when SOFA scores were in the first and second quartile (4–5). In fact, patients with cellulitis did not reach a mortality rate of 10% until the fourth SOFA quartile (SOFA score >8). Varying mortality rates among different infections with similar SOFA scores raises the question of whether SOFA score-based sepsis definitions should be considered within the context of the disease site.
The SOFA score was originally introduced to objectively quantify organ dysfunction during sepsis using easily available measures in ICU.11The SOFA score incorporates 6 measures of organ function but does not include adjustment for the source of infection. To date, studies exploring whether the source of infection is associated with outcome after adjusting for various confounders have presented mixed results.8,9,15 In one prospective study performed by Mansur et al,8mortality outcomes were explored in patients with pneumonia, bloodstream infection of unknown origin, and abdominal infection whereby SOFA score was used to assess organ dysfunction. Their findings showed that primary bacteremia is associated with higher mortality risk compared to pulmonary and intra-abdominal infections in patients with sepsis. However, patients with primary bacteremia had worse SOFA scores compared to the rest of the cohort. Notably, only Caucasians were enrolled in the study by Mansur et al, which limits generalizability.8
In our study, observed mortality varied substantially when categorized by type of infection. For example, the median SOFA score in cholangitis/cholecystitis, 6 (4–8), was similar to peritonitis, 7 (4–9), though their in-hospital mortality rates were substantially different (10.3% vs 20.0%, respectively). Within each quartile of SOFA score, patients with cellulitis and pneumonia had markedly different mortality rates. In a review of the existing literature, few studies have explored SOFA score performance to predict mortality in different infection types. Among those which have, the focus is typically on pneumonia and/or bloodstream infection,5,8,16–20 but not the wider range of infection types described in the present article.
Patients with biliary infection in this study had high overall mean SOFA scores at enrollment but modest relative mortality. Total bilirubin, which is a component of SOFA score, is often elevated in patients with biliary sepsis as a result of the primary disease process as opposed to multiple organ dysfunction from sepsis or septic shock. This is an important consideration when interpreting the SOFA score for diagnostic and prognostic purposes in patients with biliary infection.
Our study has a number of important limitations. First, the study uses retrospective data from a single tertiary hospital, which limits generalizability. Second, although infection types were determined using previously validated ICD-9 and ICD-10 codes, we were not able to independently verify each classification by chart review. Third, only single time point SOFA variables could be obtained because of the limitations of our retrospective data, and hence, effects of change in SOFA score over mortality could not be evaluated. Finally, while our approach to missing data in the SOFA score matched that of previous high-impact studies, the optimal handling of SOFA score missingness has not been defined. This is a common challenge in many SOFA-related studies.16,21
Conclusion
Our study found that SOFA score performance, including discrimination and observed mortality by SOFA quartile, differs substantially based on infection type. Clinicians should consider the infection type when applying SOFA scores to patients with infection in the ICU.
Supplementary Material
Acknowledgments
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Supplemental Material
Supplemental material for this article is available online.
References
- 1.Levy MM, Fink MP, Marshall JC, et al. 2001 SCCM/ESICM/ ACCP/ATS/SIS International Sepsis Definitions Conference. Crit Care Med. 2003;31(4):1250–1256. 10.1097/01.CCM.0000050454.01978.3B [DOI] [PubMed] [Google Scholar]
- 2.Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest. 1992;101(6):1644–1655. [DOI] [PubMed] [Google Scholar]
- 3.Singer M, Deutschman CS, Seymour CW, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA. 2016;315(8):801–810. 10.1001/jama.2016.0287 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Seymour CW, Liu VX, Iwashyna TJ, et al. Assessment of clinical criteria for sepsis: for the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA. 2016;315(8): 762–774. 10.1001/jama.2016.0288 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Raith EP, Udy AA, Bailey M, 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. 2017;317(3):290–300. 10.1001/jama.2016.20328 [DOI] [PubMed] [Google Scholar]
- 6.Moreno R, Vincent JL, Matos R, et al. The use of maximum SOFA score to quantify organ dysfunction/failure in intensive care. Results of a prospective, multicentre study. Working group on Sepsis Related Problems of the ESICM. Intensive Care Med. 1999;25(7):686–696. 10.1007/s001340050931 [DOI] [PubMed] [Google Scholar]
- 7.Ferreira FL, Bota DP, Bross A, Mélot C, Vincent JL. Serial evaluation of the SOFA score to predict outcome in critically ill patients. JAMA. 2001;286(14):1754–1758. 10.1001/jama.286.14.1754 [DOI] [PubMed] [Google Scholar]
- 8.Mansur A, Klee Y, Popov AF, et al. Primary bacteraemia is associated with a higher mortality risk compared with pulmonary and intra-abdominal infections in patients with sepsis: a prospective observational cohort study. BMJ Open. 2015;5(1): e006616. 10.1136/bmjopen-2014-006616 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Cohen J, Cristofaro P, Carlet J, Opal S. New method of classifying infections in critically ill patients. Crit Care Med. 2004;32(7): 1510–1526. 10.1097/01.ccm.0000129973.13104.2d [DOI] [PubMed] [Google Scholar]
- 10.Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998; 36(1):8–27. 10.1097/00005650-199801000-00004 [DOI] [PubMed] [Google Scholar]
- 11.Vincent JL, Moreno R, Takala J, et al. The SOFA (Sepsis-Related Organ Failure Assessment) score to describe organ dysfunction/ failure. Intensive Care Med. 1996;22(7):707–710. 10.1007/s001340050156 [DOI] [PubMed] [Google Scholar]
- 12.Pandharipande PP, Shintani AK, Hagerman HE, et al. Derivation and validation of Spo2/Fio2 ratio to impute for Pao2/Fio2 ratio in the respiratory component of the Sequential Organ Failure Assessment score. Crit Care Med. 2009;37(4):1317–1321. 10.1097/CCM.0b013e31819cefa9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Knaus WA, Zimmerman JE, Wagner DP, Draper EA, Lawrence DE. APACHE-acute physiology and chronic health evaluation: a physiologically based classification system. Crit Care Med. 1981; 9(8):591–597. 10.1097/00003246-198108000-00008 [DOI] [PubMed] [Google Scholar]
- 14.Iwashyna TJ, Odden A, Rohde J, et al. Identifying patients with severe sepsis using administrative claims: patient-level validation of the angus implementation of the International Consensus Conference Definition of Severe Sepsis. Med Care. 2014;52(6): e39–e43. 10.1097/MLR.0b013e318268ac86 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Zahar JR, Timsit JF, Orgeas MG, et al. Outcomes in severe sepsis and patients with septic shock: pathogen species and infection sites are not associated with mortality. Crit Care Med. 2011; 39(8):1886–1895. 10.1097/CCM.0b013e31821b827c [DOI] [PubMed] [Google Scholar]
- 16.Hattemer A, Hauser A, Diaz M, et al. Bacterial and clinical characteristics of health care- and community-acquired bloodstream infections due to Pseudomonas aeruginosa. Antimicrob Agents Chemother. 2013;57(8):3969–3975. 10.1128/AAC.02467-12 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Asai N, Watanabe H, Shiota A, et al. Could qSOFA and SOFA score be correctly estimating the severity of healthcare-associated pneumonia? J Infect Chemother. 2018;24(3): 228–231. 10.1016/j.jiac.2017.10.004 [DOI] [PubMed] [Google Scholar]
- 18.Namikawa H, Yamada K, Fujimoto H, et al. Clinical characteristics of bacteremia caused by extended-spectrum beta-lactamase-producing Escherichia coli at a tertiary hospital. Intern Med. 2017;56(14):1807–1815. 10.2169/internalmedicine.56.7702 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Dat VQ, Long NT, Hieu VN, et al. Clinical characteristics, organ failure, inflammatory markers and prediction of mortality in patients with community acquired bloodstream infection. BMC Infect. Dis. 2018;18(1):535. 10.1186/s12879-018-3448-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Routsi C, Pratikaki M, Platsouka E, et al. Risk factors for carbapenem-resistant gram-negative bacteremia in intensive care unit patients. Int Care Med. 2013;39(7):1253–1261. 10.1007/s00134-013-2914-z [DOI] [PubMed] [Google Scholar]
- 21.Wang JY, Chen YX, Guo SB, Mei X, Yang P. Predictive performance of quick sepsis-related organ failure assessment for mortality and ICU admission in patients with infection at the ED. Am J Emerg Med. 2016;34(9):1788–1793. 10.1016/j.ajem.2016.06.015 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.