Highlights
- •The aim of this study was to identify risk factors associated with 30-day mortality in patients with MRSA BSI.
- •1,168 patients with confirmed MRSA BSI were identified over a 9-year period in which 30-day all-cause mortality was 16%.
- •There was no significant variability in 30-day mortality over our 9-year study period.
- •Our study showed that age, cancer, heart disease, neurologic disease, nursing home residence and Charlson score >3 are risk factors for 30-day mortality in patients with MRSA BSI.
- •Diabetes, PVD and readmission because of infection have statistically significant protective effects on 30-day mortality
Summary
Objectives
Methicillin-resistant Staphylococcus aureus (MRSA) blood stream infections (BSI) are a major health care problem accounting for a large percentage of nosocomial infections. The aim of this study was to identify risk factors associated with 30-day mortality in patients with MRSA BSI.
Methods
This was a retrospective study performed in Southeast Michigan. Over a 9- year period, a total of 1,168 patients were identified with MRSA BSI. Patient demographics and clinical data were retrieved and evaluated using electronic medical health records.
Results
30-day mortality during the 9-year study period was 16%. Significant risk factors for 30-day mortality were age, cancer, heart disease, neurologic disease, nursing home residence and Charlson score >3 with Odds Ratio (OR) of 1.03 (CI 1.02–1.04), 2.29 (CI 1.40–3.75), 1.78 (CI 1.20–2.63), 1.65 (CI 1.08–2.25), 1.66 (CI 1.02 − 2.70) and 1.86 (CI 1.18 − 2.95) correspondingly. Diabetes mellitus, peripheral vascular disease (PVD), and readmission were protective factors for 30-day mortality with OR of 0.53 (CI 0.36–0.78), 0.46 (CI 0.26–0.84) and 0.13 (CI0.05 − 0.32) respectively.
Conclusions
Our study identified significant risk factors for 30-day mortality in patients with MRSA BSI. Interestingly, diabetes mellitus, PVD and readmission were protective effects on 30-day mortality. There was no statistically significant variability in 30-day mortality over the 9-year study period.
Keywords
Introduction
MRSA infection was first described in 1961 and continues to be a major public health concern, accounting for significant nosocomial and increasing community-acquired infections (
Barber, 1961
, Pastagia et al., 2012
). From 1993 to 2009, septicemia related hospitalizations increased to 153 percent, adding to healthcare cost, up to 15.4 billion dollars (Elixhauser et al., 2011
). Recent data show that the incidence is still increasing (Elixhauser et al., 2011
). In the 2013 National Center for Health Statistics Report showed an overall decrease in mortality but an increase in bloodstream related deaths by 17% (Hall et al., 2013
). In the 2010 National Healthcare Safety Network (NHSN) report on antimicrobial-resistant pathogens associated with healthcare infections, S. aureus was the most common pathogen reported, of which 54.6% were methicillin-resistant (Sievert et al., 2013
). MRSA BSI can cause devastating complications with mortality reported up to 39% (- Sievert Dawn M.
- Ricks Philip
- Edwards Jonathan R.
- Schneider Amy
- Patel Jean
- Srinivasan Arjun
- et al.
Antimicrobial-Resistant Pathogens Associated with Healthcare- Associated Infections: Summary of Data Reported to the National Healthcare Safety Network at the Centers for Disease Control and Prevention, 2009–2010.
Pastagia et al., 2012
, Keynan and Rubinstein, 2013
, Mansur et al., 2012
, Wi et al., 2012
, Lee et al., 2013
, Gasch et al., 2013a
). Identifying risk factors associated with mortality is essential to improving patient outcomes. Studies have shown that the most important factors determining patient outcomes in MRSA BSI are age, presence of comorbidities, and appropriate initial antibiotic treatment (Pastagia et al., 2012
, Keynan and Rubinstein, 2013
, Ok et al., 2013
). This study aims to identify possible predictors of 30-day mortality in MRSA BSI patients and to evaluate changes in mortality rate over a 9 year period.Methods
Study Design and Patient Identification
This was a retrospective study performed at an integrated 4-hospital health system in southeast Michigan. The primary outcome of this study was 30-day all-cause mortality in patients with MRSA BSI. All patients ≥ 18 years of age with confirmed MRSA BSI were identified over a 9-year period, from July 2005 to June 2014, via review of microbiology laboratory records. Electronic medical health records were utilized to obtain clinical information, patient demographics and 30-day mortality. Thirty-day all cause mortality was defined as death within 30 days of index blood culture. Patient demographics and comorbidities included: age, gender, presence of immunosuppressive disease, cancer, heart disease, gastrointestinal (GI) disease, neurological disease, renal disease, diabetes mellitus, human immunodeficiency virus (HIV), neutropenia, solid organ transplant recipient, intravenous drug use (IVDA), peripheral vascular disease (PVD) and connective tissue disease. The following healthcare exposures were analyzed: prior hospitalization including admission to intensive care unit (ICU), recent surgery within the past 1 year, nursing home residence (NHR), and presence of indwelling central venous catheter (CVC). Source of acquisition was categorized as follows: CVC related, osteomyelitis, endocarditis, graft infection, skin or wound infections, intra-abdominal (IA), respiratory, urinary tract (UTI) or undetermined source. Duration of bacteremia and recurrence were evaluated. Recurrence was defined as positive blood culture within 30 days of completing treatment; duration of bacteremia was divided in 3 groups: ≤3 days, 4–6 days and ≥ 7 days. Patients with recurring positive blood cultures during one admission were included once. Readmission was defined as any infection-related readmissions within 30 days of completing therapy. The Charlson Comorbidity Index was calculated to compare patients’ comorbidities and 30-day mortality (
Charlson et al., 1987
).Statistical Analysis
Continuous data were described using means and standard deviations, while all categorical data were presented as counts and percentages. Five patients were excluded from the analysis because of incomplete data. Univariate two-group tests were performed to compare clinical and demographic data between patients who expired within 30 days from index blood culture with alive patients. All predictors of mortality with univariate P-value <0.2 were included in a multivariable logistic regression model. A final model was produced using manual backwards selection, meaning non-significant predictors were removed one by one in descending P-value order and the model re-run each time until each predictor remaining was significant. Variability in 30-day all-cause mortality over the 9-year study period was determined using a chi-square test. All statistical analyses were completed using SAS 9.4 (SAS Institute Inc., Cary, NC, USA).
Results
Over the 9-year period there were 1,168 individual patients with MRSA BSI, of which 193 (16%) expired within 30 days of index blood culture. Over the 9-year period there was no significant variability in 30-day mortality (p = 0.1934).
Results from univariate and multivariate logistic regression models evaluating the association between risk factors and 30-day mortality are shown in Table 1. The following risk factors were associated with increased risk of 30-day mortality: cancer OR 2.29 (CI 1.40-3.75 p value 0.001), heart disease OR 1.78 (CI 1.20-2.63 p-value 0.0039), neurologic disease OR 1.65 (CI 1.08-2.52 p-value 0.02), nursing home residence OR 1.66 (CI 1.02-1.70 p value 0.0421), and Charlson score >3 OR 1.88 (CI 1.19-2.97 p value 0.0068). Age was a statistically significant risk factor (p-value <0.001). The odds of 30-day mortality increased by 2.9% for every year increase in age when controlling for other variables. Diabetes, PVD, and readmission were protective factors for 30-day mortality with OR of 0.60 (CI 0.42-0.86 p-value 0.0056), 0.53 (CI 0.30 −0.93 p-value 0.0277), 0.12 (CI 0.05–0.30 p- value <0.001) respectively. There was a relationship between lower mortality and infection sources including CVC OR 0.39 (CI 0.24-0.63 p value 0.0001), skin/wound OR 0.48 (CI 0.31-0.71 p value 0.001) and GU OR 0.40 (CI 0.18-0.88 p value 0.022).
Table 1Univariate and Multivariate Logistic Regressionn Analysis.
| Predictor | Demographics and Clinical Characteristics | Univariate Logistic Regression | Multivariate Logistic Regression | ||||
|---|---|---|---|---|---|---|---|
| Unadjusted OR | 95% CI | P-Value | Adjusted OR | 95% CI | P-Value | ||
| Age, years, mean(SD) | 60.3 (17.6) | 1.04 | 1.03-1.05 | <0.001 | 1.03 | 1.02-1.04 | <0.001 |
| Gender, male | 689 (59%) | 1.13 | 0.83-1.55 | 0.4417 | |||
| Race | |||||||
| African American | 694 (60%) | ||||||
| Caucasian | 407 (35%) | ||||||
| Other | 56 (5%) | ||||||
| Cancer | 140 (12%) | 2.73 | 1.84-4.07 | <0.0001 | 2.29 | 1.40-3.75 | 0.001 |
| Connective tissue | 23 (2%) | 1.12 | 0.38-3.36 | 0.8341 | |||
| CVC | 127 (11%) | 0.61 | 0.34-1.08 | 0.0913 | |||
| Diabetes | 467 (40%) | 0.75 | 0.54-1.04 | 0.0800 | 0.53 | 0.36-0.78 | 0.0015 |
| Duration of bacteremia, days (SD) | 3.9 | 1.32 | 0.78-2.23 | 0.3099 | |||
| GI disease | 238 (20%) | 1.06 | 0.73-1.55 | 0.7486 | |||
| Heart disease | 329 (28%) | 2.25 | 1.64-3.10 | 0.0001 | 1.78 | 1.20-2.63 | 0.0039 |
| HIV | 35 (3%) | 0.30 | 0.07-1.26 | 0.1009 | |||
| Immunosuppressive disease | 75 (6%) | 1.83 | 1.06-3.15 | 0.0306 | |||
| IV drug use | 206 (18%) | 0.40 | 0.24-0.68 | 0.0006 | |||
| Neurologic disease | 236 (20%) | 2.44 | 1.73-3.43 | 0.0001 | 1.65 | 1.08-2.52 | 0.02 |
| Neutropenia | 13 (1%) | 3.20 | 1.04-9.89 | 0.0432 | |||
| Nursing home resident | 129 (11%) | 2.67 | 1.77-4.02 | <0.0001 | 1.66 | 1.02-2.70 | 0.0421 |
| Prior hospitalization | 597 (51%) | 1.25 | 0.92-1.70 | 0.1599 | |||
| Prior ICU stay | 183 (16%) | 1.31 | 0.88-1.96 | 0.1876 | |||
| Prior surgery | 102 (9%) | 1.10 | 0.65-1.89 | 0.7165 | |||
| PVD | 149 (13%) | 0.66 | 0.39-1.11 | 0.1191 | 0.46 | 0.26-0.84 | 0.0106 |
| Readmit infection | 169 (14%) | 0.13 | 0.05-0.33 | <0.0001 | 0.13 | 0.05-0.32 | <0.001 |
| Recurrence 30d | 84 (7%) | 0.24 | 0.09-0.65 | 0.0055 | |||
| Renal disease | 550 (47%) | 1.29 | 0.95-1.76 | 0.1077 | |||
| Respiratory disease | 208 (18%) | 2.10 | 1.47-3.01 | <0.0001 | |||
| Solid organ transplant | 31 (3%) | 1.79 | 0.79-4.05 | 0.1655 | |||
| Source: Graft | 31 (3%) | 1.22 | 0.49-3.01 | 0.6675 | |||
| Source: CVC | 244 (21%) | 0.63 | 0.41-0.96 | 0.0316 | 0.39 | 0.24-0.63 | 0.0001 |
| Source: IE | 0 | N/A | N/A | ||||
| Source: IA | 10 (1%) | 2.18 | 0.56-8.52 | 0.2609 | |||
| Source: None | 247 (21%) | 1.91 | 1.36-2.70 | 0.0002 | |||
| Source: Respiratory | 97 (8%) | 2.26 | 1.41-3.62 | 0.0007 | |||
| Source: Skin/Wound | 370 (32%) | 0.51 | 0.35-0.75 | 0.0005 | 0.48 | 0.31-0.71 | 0.001 |
| Source: GU | 61 (5%) | 1.14 | 0.58-2.24 | 0.6986 | 0.40 | 0.18-0.88 | 0.022 |
| Source: Osteomyelitis | 7 (1%) | 0.00 | 0.001- > 999 | 0.9838 | |||
| Source: Endocarditis | 106 (9%) | 0.89 | 0.51-1.55 | 0.6779 | |||
| Charlson Score, N, mean (SD) | 1154, 3.2 (2.4) | 1.77 | 1.11-1.25 | <0.001 | |||
| Charlson Score Category (> = 3 vs <3) | 2.40 | 1.69-3.40 | <0.001 | 1.86 | 1.18-2.95 | 0.0079 | |
Discussion
Consequences of MRSA bacteremia are devastating with increase in hospital stay, morbidity, and mortality. Therefore, early identification of MRSA BSI is important to optimize patient outcomes (
Lee et al., 2013
, Gasch et al., 2013a
, Bassetti et al., 2011
, Hernandez et al., 2015
, Gasch et al., 2013b
). Prior studies have shown two modifiable risk factors for 30-day mortality: 1) appropriate empiric antimicrobial therapy, 2) prompt removal of focus (Keynan and Rubinstein, 2013
, Lee et al., 2013
, Gasch et al., 2013a
, Ok et al., 2013
). Notably, with time we found no improvement in overall mortality over time despite advancement in antimicrobial treatment. Thus far, the role of different antimicrobial agents against MRSA infection in clinical setting is uncertain.The strength of this study is a large sample size with comprehensive clinical data over an extended period of time. Despite advances in medical therapy, use of newer antibiotics, and changes in patient characteristics and strains, we did not find any statistically significant change in the rate of 30-day mortality over the 9-year study period. Our study found 16.5% all cause 30-day mortality in MRSA BSI; this is lower than rates reported in other studies, which vary between 21–39% (
Sievert et al., 2013
, - Sievert Dawn M.
- Ricks Philip
- Edwards Jonathan R.
- Schneider Amy
- Patel Jean
- Srinivasan Arjun
- et al.
Antimicrobial-Resistant Pathogens Associated with Healthcare- Associated Infections: Summary of Data Reported to the National Healthcare Safety Network at the Centers for Disease Control and Prevention, 2009–2010.
Wi et al., 2012
, Gasch et al., 2013a
, Ok et al., 2013
, Gasch et al., 2013b
).Our study showed age and cardiac disease as risk factors for 30-day mortality which is consistent with findings in other studies (
Pastagia et al., 2012
, Mansur et al., 2012
, Ok et al., 2013
, Gasch et al., 2013b
). A study evaluating outcomes in patients over 65 year of age with community onset bacteremia found 52 out of 2,065 (2.51%) patients with MRSA BSI. Although the most common pathogen isolated was Escherichia coli, MRSA had the highest mortality (28%). Mortality was significantly higher in patients over 75 year of age, and administration of inappropriate antibiotic treatment (Gasch et al., 2013b
). This stresses the importance of identifying these patients early in order to improve their outcomes. Our analysis found that nursing home residence increased the risk of 30 day mortality, which is likely due to advanced age and increased number of comorbidities.Contrary to other studies, we found that cancer was a significant risk factor for 30-day mortality with an OR of 2.84 (1.82–4.44 p-value <0.001). Various studies have reported that the presence of a previous rapid fatal underlying disease is an important risk factor, but only one study analyzed cancer as an independent risk factor which was not statistically significant (
Lee et al., 2013
, Ok et al., 2013
, Gasch et al., 2013b
). Cancer, as well as other conditions, is included in the Charlson weighted index of comorbidity score. Although it can be a good tool that can help clinicians identify patients who are at increased risk for early mortality due to MRSA BSI, other factors such as source of acquisition are not included (Retamar et al., 2012
, Lesens et al., 2003
, Charlson et al., 1987
).Another study found that persistent bacteremia (defined as positive blood cultures >7 days) had greater mortality than non-persistent bacteremia 58.1% vs. 16.2% (p-value <0.001) (
Hernandez et al., 2015
). However, our study did not find duration of bacteremia to be a statistically significant risk factor for 30-day mortality. All other risk factors that were evaluated including presence of immunosuppressive disease, GI disease, renal disease, diabetes, HIV status, neutropenia, solid organ transplant recipient, IVDA, connective tissue disease, ICU stay, and prior surgery were not significantly associated with 30-day mortality. Continuous observation of patients with MRSA BSI is essential to improve our understanding of risk factors associated with 30-day mortality.In our study diabetes and PVD were protective factors for 30-day mortality in patients with MRSA BSI. Other studies included diabetes in their evaluation of risk factors, finding them to be protective factors, though only one found it to be statistically significant (
Pastagia et al., 2012
, Lee et al., 2013
, Gasch et al., 2013a
). This is an unexpected finding, but we believe that it can be associated with the source of infection. Patients with PVD and diabetes are usually less acutely ill and present with skin/wound infections which are more easily managed, and started earlier on appropriate antibiotic treatment. There has been recent research on hypoxia inducible factor 1 (HIF-1) playing an important role in the innate response to infections (Bento and Pereira, 2011
, Shalova et al., 2015
). HIF-1 is a transcription factor that is involved in energy metabolism, cellular adaptation to hypoxia, and immune response. Recent studies have shown that hyperglycemia destabilizes HIF-1, which could contribute to diabetic complications such as infections (Bento and Pereira, 2011
, Howangyin and Silvestre, 2014
). T. Bhandari et al and other groups are conducting research into the possibility of having HIF-1 as a therapeutic target, which would provide increased treatment options to patients with BSI (Bhandari and Nizet, 2014
, Okumura et al., 2012
, Zeitouni et al., 2016
, Li et al., 2014
).We also found that readmission due to infection is a protective factor. The reason of this was not determined, however, patients who were readmitted because of MRSA infection would have probably received an earlier and better management, which may have prevented progression of disease, resulting in a decreased mortality. Source of infection is an important mortality risk factor. Our study found that skin/wound, GU, and CVC source of infection are protective factors for 30-day mortality. Further studies are needed in order to understand the causality of these relationships.
A limitation of this study was that we did not have access to information regarding ID consultation. All patients with MRSA bacteremias were treated with vancomycin or daptomycin, but we do not have information regarding timing of initial treatment. Further studies are needed to evaluate the effects of appropriate initial antibiotic treatment in high risk patients such as patients with cancer, heart and neurologic disease. We did not include MRSA strain types and vancomycin MICs as factors in our analysis.
Conclusions
During the 9-year study period we did not find any statistically significant variability in 30-day mortality. Our study showed that age, cancer, heart disease, neurologic disease, nursing home residence and Charlson score greater than 3 are risk factors for 30-day mortality in patients with MRSA BSI. These findings can help clinicians correlate patient risk factors with poor outcomes and provide early optimal management, including removal of devices, focus or other modifiable risk factors. Further prospective studies are needed to examine this subset of patients, in order to design interventions that can improve outcomes. We found that diabetes, PVD and readmission were statistically significant protective factors for 30-day mortality, though we believe that this is due to the source and severity of infection. Continuous surveillance of patients with multidrug resistant organism infections is essential for our health care system.
Conflicts of interest
M. Z. has received research grants from Pfizer, Cerexa, Cubist, Merck, Tetraphase, Melinta, Paratek and Rempex and Cempra. K.R. has received research grant from Cubist, Merck and Theravance. All other authors have nothing to declare.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Acknowledgments
None.
References
- Methicillin-resistant staphylococci.J Clin Path. 1961; : 385-393
- Risk factors and mortality of healthcare-associated and community- acquired Staphylococcus aureus bacteraemia.Clin Microbiol Infect. 2011; 18: 862-869
- Regulation of hypoxia-inducible factor 1 and the loss of the cellular response to hypoxia in diabetes.Diabetologia. 2011; 54: 1946-1956https://doi.org/10.1007/s00125-011-2191-8
- Hypoxia-Inducible Factor (HIF) as a Pharmacological Target for Prevention and Treatment of Infectious Diseases.Infect Dis Ther. 2014; 3: 159-174https://doi.org/10.1007/s40121-014-0030-1
- A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation.J Chronic Dis. 1987; 40: 373-383https://doi.org/10.1016/0021-9681(87)90171-8
- STATISTICAL BRIEF # 1 22. vol. 348. 2011
- Predictive factors for early mortality among patients with methicillin-resistant staphylococcus aureus bacteraemia.J Antimicrob Chemother. 2013; 68: 1423-1430https://doi.org/10.1093/jac/dkt016
- Predictive factors for mortality in patients with methicillin-resistant Staphylococcus aureus bloodstream infection: Impact on outcome of host, microorganism and therapy.Clin Microbiol Infect. 2013; 19: 1049-1057https://doi.org/10.1111/1469-0691.12108
- Trends in Inpatient Hospital Deaths: National Hospital Discharge Survey, 2000–2010.2013
- Clinical characteristics and outcome of elderly patients with community-onset bacteremia.J Infect. 2015; 70: 135-143https://doi.org/10.1016/j.jinf.2014.09.002
- Diabetes Mellitus and Ischemic Diseases: Molecular Mechanisms of Vascular Repair Dysfunction.Arterioscler Thromb Vasc Biol. 2014; 34: 1126-1135https://doi.org/10.1161/ATVBAHA.114.303090
- Staphylococcus aureus Bacteremia, Risk Factors, Complications, and Management.Crit Care Clin. 2013; 29: 547-562https://doi.org/10.1016/j.ccc.2013.03.008
- Clinical features and risk factors of mortality for bacteremia due to community-onset healthcare-associated methicillin-resistant S. aureus.Diagn Microbiol Infect Dis. 2013; 76: 86-92https://doi.org/10.1016/j.diagmicrobio.2013.01.017
- Role of Comorbidity in Mortality Related to Staphylococcus aureus Bacteremia: A Prospective Study Using the Charlson Weighted Index of Comorbidity.Infect Control Hosp Epidemiol. 2003; 24: 890-896https://doi.org/10.1086/502156
- The relationship between serum hypoxia-inducible factor 1α and coronary artery calcification in asymptomatic type 2 diabetic patients.Cardiovasc Diabetol. 2014; 13: 52https://doi.org/10.1186/1475-2840-13-52
- Does sex affect 30-day mortality in staphylococcus aureus bacteremia?.Gend Med. 2012; 9: 463-470https://doi.org/10.1016/j.genm.2012.10.009
- Predictors and clinical outcomes of persistent methicillin-resistant Staphylococcus aureus bacteremia: a prospective observational study.Korean J Intern Med. 2013; 28: 678-686
- A new pharmacological agent (AKB-4924) stabilizes hypoxia inducible factor-1 (HIF-1) and increases skin innate defenses against bacterial infection.J Mol Med. 2012; 90: 1079-1089https://doi.org/10.1007/s00109-012-0882-3
- Predicting Risk for Death from MRSA Bacteremia.Emerg Infect Dis. 2012; 18: 1072-1080https://doi.org/10.3201/eid1807.101371
- Impact of inadequate empirical therapy on the mortality of patients with bloodstream infections: A propensity score-based analysis.Antimicrob Agents Chemother. 2012; 56: 472-478https://doi.org/10.1128/AAC.00462-11
- Human Monocytes Undergo Functional Re-programming during Sepsis Mediated by Hypoxia-Inducible Factor-1α.Immunity. 2015; 42: 484-498https://doi.org/10.1016/j.immuni.2015.02.001
- Antimicrobial-Resistant Pathogens Associated with Healthcare- Associated Infections: Summary of Data Reported to the National Healthcare Safety Network at the Centers for Disease Control and Prevention, 2009–2010.Infect Control Hosp Epidemiol. 2013; 34: 1-14https://doi.org/10.1086/668770
- High vancomycin minimum inhibitory concentration is a predictor of mortality in meticillin-resistant Staphylococcus aureus bacteraemia.Int J Antimicrob Agents. 2012; 40: 108-113https://doi.org/10.1016/j.ijantimicag.2012.04.003
- Hypoxia Decreases Invasin-Mediated Yersinia enterocolitica Internalization into Caco-2Cells.PLoS One. 2016; 11: e0146103https://doi.org/10.1371/journal.pone.0146103
Article info
Publication history
Published online: May 19, 2017
Accepted:
May 11,
2017
Received in revised form:
May 9,
2017
Received:
February 21,
2017
Corresponding Editor: Eskild Petersen, Aarhus, DenmarkIdentification
Copyright
© 2017 The Authors. Published by Elsevier Ltd on behalf of International Society for Infectious Diseases.
User license
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0) | How you can reuse 
Elsevier's open access license policy
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0)
Permitted
For non-commercial purposes:
- Read, print & download
- Redistribute or republish the final article
- Text & data mine
- Translate the article (private use only, not for distribution)
- Reuse portions or extracts from the article in other works
Not Permitted
- Sell or re-use for commercial purposes
- Distribute translations or adaptations of the article
Elsevier's open access license policy
