Abstract
Oral switch to linezolid is a promising alternative to standard parenteral therapy (SPT) in Staphylococcus aureus bacteremia (SAB).
We conducted a prospective cohort study of all adult cases of SAB between 2013 and 2017 in a Spanish university hospital. We compared the efficacy, safety, and length of hospital stay of patients receiving SPT and those where SPT was switched to oral linezolid between days 3 and 9 of treatment until completion. We excluded complicated SAB and osteoarticular infections. A k-nearest neighbor algorithm was used for propensity score matching with a 2:1 ratio.
After propensity score matching, we included 45 patients from the linezolid group and 90 patients from the SPT group. Leading SAB sources were catheter related (49.6%), unknown origin (20.0%), and skin and soft tissue (17.0%). We observed no difference in 90-day relapse between the linezolid group and the SPT group (2.2% vs 4.4% respectively; P = .87). No statistically significant difference was observed in 30-day all-cause mortality between the linezolid group and the SPT group (2.2% vs 13.3%; P = .08). The median length of hospital stay after onset was 8 days in the linezolid group and 19 days in the SPT group (P < .01). No drug-related events leading to discontinuation were noted in the linezolid group.
Treatment of SAB in selected low-risk patients with an oral switch to linezolid between days 3 and 9 of treatment until completion yielded similar clinical outcomes as SPT, allowing earlier discharge from the hospital.
Staphylococcus aureus is one of the leading causes of both nosocomial and community-onset bloodstream infections worldwide [1]. Staphylococcus aureus bacteremia (SAB) is characterized by a high incidence of bacterial complications, including distant metastatic foci, local extension, and relapses, with a crude mortality that ranges from 20% to 40% [2–4].
Current evidence indicates that complications, especially relapses, can be minimized by an adequate length of antimicrobial therapy, supporting longer treatment schedules than for other bloodstream infections [5–7]. Unlike other infectious diseases and despite the existence of effective oral agents with high bioavailability, current guidelines do not contemplate oral switch due to the lack of sound clinical evidence. Previous studies have identified low-risk patients with a low incidence of SAB-related complications [8, 9] who could potentially benefit from an early oral switch. An early oral switch may reduce the risk of complications and increased costs due to prolonged hospital stay and the use of intravenous catheters, while contributing to patients’ quality of life.
Linezolid is an oxazolidinone with a bioavailability of virtually 100% active against staphylococci and other gram-positive bacteria, including those resistant to methicillin. Previous studies have demonstrated its efficacy and safety in S. aureus catheter-related bloodstream infections and other serious infections, such as complicated skin and soft tissue infections, methicillin-resistant S. aureus (MRSA) infections, and nosocomial pneumonia, including the subset with bacteremia [10–16].
Our aim was to study the efficacy and safety of completing SAB treatment with oral linezolid in low-risk patients by comparing the clinical outcomes of low-risk adult patients with SAB who switched to oral linezolid between days 3 and 9 of treatment until completion vs those who received standard parenteral therapy (SPT).
METHODS
Setting and Study Design
This prospective, observational study was performed at the University Hospital Vall d’Hebron (Barcelona, Spain), a 1000-bed acute-care teaching hospital. From January 2013 to January 2017, all adult patients (age ≥18 years) diagnosed with monomicrobial SAB were included. In each patient, only the first episode of SAB during the study period was included.
Clinical management was left to the discretion of the attending physician. After completing SAB treatment, patients were followed up for 90 days or until death, whichever occurred first. If clinical monitoring concluded before 90 days after treatment, follow-up was completed by telephone interview. Information was corroborated by checking primary care records and other regional hospital registries.
For additional information on methods, see the Supplementary Materials.
Study Population
The present report is a subanalysis of the entire SAB cohort. Our hypothesis was that patients with low risk for complications, including those clinically stable, with appropriate source control and negative follow-up blood cultures, would be suitable for an early oral switch to linezolid.
Patients were divided into 2 groups according to the antimicrobial SAB treatment: (1) SPT if antibiotics were administered intravenously during the entire antimicrobial course, and (2) oral linezolid if parenteral treatment was switched to oral linezolid (600 mg every 12 hours) 3 to 9 days after another active SAB treatment was started, until completion.
Patients meeting the following criteria were excluded: (1) death ≤7 days after the index culture; (2) complicated SAB, defined as persistence of positive blood cultures after ≥3 days of appropriate treatment (including source control), development of septic thrombophlebitis, infective endocarditis, infected arterial aneurysm, endovascular graft infection, or other metastatic distant foci before concluding parenteral therapy, and/or any device-related infection where the device could not be removed in the first 3 days [4, 17–22]; (3) osteoarticular infections; (4) oral switch to a different antibiotic than linezolid; (5) oral switch to linezolid outside of the 3- to 9-day range from treatment start; and (6) incomplete follow-up information.
The decision to switch to oral linezolid was made on an individual basis by the attending physician considering personal experience, confidence in previously published literature [10–16], and the possibility of early hospital discharge.
Outcomes
The primary outcome was 90-day relapse, defined as a new positive blood culture for S. aureus, or a microbiologically confirmed S. aureus infection of any location, within 90 days after completing antibiotic treatment.
The secondary outcomes were (1) 30-day all-cause mortality, defined as death from any cause within 30 days after the index culture was obtained; and (2) length of hospital stay after the index culture.
Definitions
The Charlson comorbidity index, validated as predictive of mortality among patients with SAB, was used to measure the severity of chronic underlying conditions [23, 24]. Type of acquisition was classified as community-acquired, healthcare-associated, or nosocomial following Friedman criteria [25].
To evaluate severity of illness, vasopressor requirement in the first 24 hours after onset was recorded. Septic shock was defined as vasopressor requirement during at least 1 hour to maintain a mean arterial pressure of ≥65 mm Hg [26].
Source of bacteremia was defined according to Centers for Disease Control and Prevention criteria [27]. Catheter-related SAB was considered if the Infectious Diseases Society of America guidelines’ criteria for a definitive diagnosis of catheter-related bloodstream infection were fulfilled [28] or if purulence around the insertion site or clinical signs of phlebitis with no other plausible primary source were described by the treating physician.
Appropriate source control was defined as the removal of all vascular lines (confirmed or suspected as a source of SAB) present at least 24 hours before the first positive blood culture, removal of infected hardware, or drainage of infected fluid collections. In the absence of a vascular line, infected hardware, and a drainable collection, we also considered the source appropriately controlled.
Antibiotic treatment was considered appropriate if the strain was susceptible in vitro to at least one of the administered antibiotics, except for aminoglycosides, which were considered inappropriate regardless of the sensitivity tests.
Definitive therapy was defined as the main antimicrobial administered from day 4 after the index culture. The cut-point of 4 days assumed that susceptibility results are usually available in this period [29].
Microbiological Studies
Blood cultures were performed with the BacT/ALERT 3D system (bioMérieux, Marcy l’Étoile, France) and isolate identification was performed using the VITEK 2 or VITEK MS systems (bioMérieux) or by commercial molecular test (Cepheid Xpert MRSA/SA BC, Sunnyvale, California). Antimicrobial susceptibility testing was done in accordance with the European Committee on Antimicrobial Susceptibility Testing guidelines by use of disk diffusion techniques.
In case a new S. aureus isolate was found in a clinical sample during follow-up, whole genome sequencing (WGS) was used to differentiate relapse from reinfection. WGS was performed on a MiSeq Instrument (Illumina, San Diego, California) using the Nextera DNA Flex Library Prep Kit (Illumina) and paired-end sequenced with a MiSeq Reagent Kit version 2, 250 bp (Illumina), with an average insertion size of 300 bp. The sequence files were analyzed by Ridom SeqSphere+ software version 5 (Ridom GmbH, Muenster, Germany), with default parameters for quality trimming and assembly. Bacterial relatedness was established by comparing the 1861 target genes of the S. aureus core genome multilocus sequence typing scheme [30].
Statistical Analysis
Continuous variables were described as median and interquartile range (IQR), whereas qualitative variables were described by absolute count and relative percentage. Comparisons between groups were assessed by Mann-Whitney U test for quantitative variables and χ2 test–based analyses for qualitative variables.
We compared the 2 treatment groups: SPT and oral linezolid. To avoid potential treatment selection bias, a propensity score matching analysis was performed. As potential confounders with treatment group as outcome, we included for propensity score matching (1) all nonredundant variables with a P value ≤.20: age, hematologic and solid malignancy, cirrhosis, chronic renal disease, hemodialysis, predisposing risk factor for endocarditis and endovascular graft infection, type of acquisition, septic shock, intensive care unit hospitalization, time from onset to appropriate treatment, and infectious disease (ID) consultation; and (2) those considered clinically relevant with P > .20: gender, methicillin resistance, Charlson comorbidity index, immunosuppression, neutropenia, and source of bacteremia. Using the MatchIt package, a k-nearest neighbor algorithm was used for propensity score matching with a 2:1 ratio, matching every patient of the linezolid group with 2 patients treated with SPT. Competitive risk survival analysis was used to confirm the results of the primary outcome.
Logistic regression was used to estimate the odds ratios (ORs) and associated 95% confidence intervals (CIs) for potential risk factors associated with 90-day relapse and with 30-day all-cause mortality. First, we performed a univariate analysis and evaluated the individual influence on the outcome. Nonredundant variables with a P value ≤.20 on the univariate analysis, or those considered clinically relevant including treatment group, were included in the multivariate analysis. A backward stepwise elimination algorithm was used for final variable selection of explanatory variables.
Data analysis was performed using R software (version 3.4.2) under RStudio IDE (version 1.0.153). A 2-sided P value <.05 was considered statistically significant for all tests.
RESULTS
During the study period, 381 patients with SAB were identified, and 152 patients met eligibility criteria (Figure 1). Of these, 45 patients switched to oral linezolid (linezolid group). After propensity score matching with a 2:1 ratio, the 45 patients of the linezolid group and 90 patients of the SPT group were included for final analysis. Leading sources of infection were catheter-related bacteremia (49.6%), bacteremia of unknown origin (20.0%), and skin and soft tissue infection (17.0%).
Study flowchart. aComplicated Staphylococcus aureus bacteremia includes episodes with persistence of positive blood cultures after ≥3 days of appropriate treatment (including source control), development of septic thrombophlebitis, infective endocarditis, infected arterial aneurysm, endovascular graft infection, or other metastatic distant foci before concluding parenteral therapy, and/or any device-related infection where the device could not be removed in the first 3 days.
Baseline characteristics of the 2 matched groups were generally balanced, except that patients in the linezolid group were less likely to have chronic renal disease (4.4% vs 22.2%). Although statistically not significant, the SPT group presented a higher proportion of patients on hemodialysis and with endocarditis-predisposing conditions, endovascular grafts, orthopedic devices, septic shock, unknown source of SAB, and absence of ID consultation. A comparison of baseline characteristics between linezolid and SPT groups before and after propensity score matching is shown in Table 1.
Baseline Characteristics of Adult Patients With Staphylococcus aureus Bacteremia Comparing Treatment With Early Oral Switch to Linezolid and Standard Parenteral Treatment
| Characteristic | Whole Cohort | Propensity Score-matched Cohort | ||||
|---|---|---|---|---|---|---|
| Oral Linezolid (n = 45) | Standard Treatment (n = 107) | P Value | Oral Linezolid (n = 45) | Standard Treatment (n = 90) | P Value | |
| Male sex | 30 (66.7) | 68 (63.5) | .86 | 30 (66.7) | 59 (65.6) | 1.00 |
| Age, y, median (IQR) | 61 (45–69) | 63 (49–76) | .16 | 61 (45–69) | 64 (47–77) | .14 |
| MRSA strain | 5 (11.1) | 19 (17.8) | .43 | 5 (11.1) | 12 (13.3) | .93 |
| Preexisting condition | ||||||
| Charlson index, median (IQR) | 2 (1–5) | 2 (1–5) | .22 | 2 (1–5) | 2 (1–5) | .48 |
| Diabetes mellitus | 10 (22.2) | 29 (27.1) | .67 | 10 (22.2) | 22 (24.4) | .94 |
| Malignancy | 22 (48.9) | 37 (34.6) | .14 | 22 (48.9) | 33 (36.7) | .24 |
| Liver cirrhosis | 0 (0.0) | 7 (6.5) | .18 | 0 (0.0) | 1 (1.1) | 1.00 |
| Chronic renal disease | 2 (4.4) | 24 (22.4) | .01 | 2 (4.4) | 20 (22.2) | .02 |
| Dialysis | 1 (2.2) | 15 (14.0) | .06 | 1 (2.2) | 12 (13.3) | .08 |
| Immunosuppressiona | 19 (42.2) | 35 (32.7) | .35 | 19 (42.2) | 27 (30.0) | .22 |
| HIV infection | 0 (0.0) | 3 (2.8) | .62 | 0 (0.0) | 2 (2.2) | .80 |
| Neutropenia (<0.5 neutrophils/μL) | 4 (8.9) | 3 (2.8) | .23 | 4 (8.9) | 3 (3.3) | .34 |
| Solid organ transplant | 2 (4.4) | 10 (9.3) | .49 | 2 (4.4) | 5 (5.6) | 1.00 |
| Endocarditis-predisposing condition or endovascular graftb | 8 (17.8) | 33 (30.8) | .15 | 8 (17.8) | 26 (28.9) | .23 |
| Orthopedic device | 3 (6.7) | 11 (10.3) | .69 | 3 (6.7) | 10 (11.1) | .61 |
| Acquisition | .13 | .40 | ||||
| Hospital-acquired | 22 (48.9) | 67 (62.6) | 22 (48.9) | 55 (61.1) | ||
| Healthcare-related | 4 (8.9) | 7 (6.5) | 4 (8.9) | 6 (6.7) | ||
| Community-acquired | 19 (42.2) | 33 (30.8) | 19 (42.2) | 29 (32.2) | ||
| Clinical data | ||||||
| Septic shock | 2 (4.4) | 16 (15.0) | .12 | 2 (4.4) | 9 (10.0) | .44 |
| ICU stay | 4 (8.9) | 27 (25.2) | .04 | 4 (8.9) | 16 (17.8) | .27 |
| Source of bacteremia | .69 | .45 | ||||
| Catheter-related | 25 (55.6) | 50 (46.7) | 25 (55.6) | 42 (46.7) | ||
| Skin and soft tissue | 7 (15.6) | 20 (18.7) | 7 (15.6) | 16 (17.8) | ||
| Pneumonia or pleural empyema | 4 (8.9) | 13 (12.1) | 4 (8.9) | 9 (10.0) | ||
| Genitourinary | 3 (6.7) | 2 (1.9) | 3 (6.7) | 2 (2.2) | ||
| Unknown | 6 (13.3) | 22 (20.6) | 6 (13.3) | 21 (23.3) | ||
| Antibiotic treatment | ||||||
| Time from onset to appropriate treatment, d, median (IQR) | 0 (0–2) | 0 (0–1) | .18 | 0 (0–2) | 0 (0–1) | .10 |
| Duration of appropriate treatment, d, median (IQR) | 15 (14–16) | 15 (13–19) | .92 | 15 (14–16) | 15 (14–19) | .49 |
| Duration of intravenous treatment, d, median (IQR) | 7 (6–8) | 15 (13–19) | <.01 | 7 (6–8) | 15 (14–19) | <.01 |
| Source control <72 h | 45 (92.1) | 103 (96.3) | .45 | 45 (92.1) | 86 (95.6) | .37 |
| ID consultation | 45 (100.0) | 91 (85.0) | .01 | 45 (100.0) | 85 (94.4) | .26 |
| Time from index culture to visit, d, median (IQR) | 2 (1–2) | 2 (1–3) | .61 | 2 (1–2) | 2 (1–3) | .68 |
| Characteristic | Whole Cohort | Propensity Score-matched Cohort | ||||
|---|---|---|---|---|---|---|
| Oral Linezolid (n = 45) | Standard Treatment (n = 107) | P Value | Oral Linezolid (n = 45) | Standard Treatment (n = 90) | P Value | |
| Male sex | 30 (66.7) | 68 (63.5) | .86 | 30 (66.7) | 59 (65.6) | 1.00 |
| Age, y, median (IQR) | 61 (45–69) | 63 (49–76) | .16 | 61 (45–69) | 64 (47–77) | .14 |
| MRSA strain | 5 (11.1) | 19 (17.8) | .43 | 5 (11.1) | 12 (13.3) | .93 |
| Preexisting condition | ||||||
| Charlson index, median (IQR) | 2 (1–5) | 2 (1–5) | .22 | 2 (1–5) | 2 (1–5) | .48 |
| Diabetes mellitus | 10 (22.2) | 29 (27.1) | .67 | 10 (22.2) | 22 (24.4) | .94 |
| Malignancy | 22 (48.9) | 37 (34.6) | .14 | 22 (48.9) | 33 (36.7) | .24 |
| Liver cirrhosis | 0 (0.0) | 7 (6.5) | .18 | 0 (0.0) | 1 (1.1) | 1.00 |
| Chronic renal disease | 2 (4.4) | 24 (22.4) | .01 | 2 (4.4) | 20 (22.2) | .02 |
| Dialysis | 1 (2.2) | 15 (14.0) | .06 | 1 (2.2) | 12 (13.3) | .08 |
| Immunosuppressiona | 19 (42.2) | 35 (32.7) | .35 | 19 (42.2) | 27 (30.0) | .22 |
| HIV infection | 0 (0.0) | 3 (2.8) | .62 | 0 (0.0) | 2 (2.2) | .80 |
| Neutropenia (<0.5 neutrophils/μL) | 4 (8.9) | 3 (2.8) | .23 | 4 (8.9) | 3 (3.3) | .34 |
| Solid organ transplant | 2 (4.4) | 10 (9.3) | .49 | 2 (4.4) | 5 (5.6) | 1.00 |
| Endocarditis-predisposing condition or endovascular graftb | 8 (17.8) | 33 (30.8) | .15 | 8 (17.8) | 26 (28.9) | .23 |
| Orthopedic device | 3 (6.7) | 11 (10.3) | .69 | 3 (6.7) | 10 (11.1) | .61 |
| Acquisition | .13 | .40 | ||||
| Hospital-acquired | 22 (48.9) | 67 (62.6) | 22 (48.9) | 55 (61.1) | ||
| Healthcare-related | 4 (8.9) | 7 (6.5) | 4 (8.9) | 6 (6.7) | ||
| Community-acquired | 19 (42.2) | 33 (30.8) | 19 (42.2) | 29 (32.2) | ||
| Clinical data | ||||||
| Septic shock | 2 (4.4) | 16 (15.0) | .12 | 2 (4.4) | 9 (10.0) | .44 |
| ICU stay | 4 (8.9) | 27 (25.2) | .04 | 4 (8.9) | 16 (17.8) | .27 |
| Source of bacteremia | .69 | .45 | ||||
| Catheter-related | 25 (55.6) | 50 (46.7) | 25 (55.6) | 42 (46.7) | ||
| Skin and soft tissue | 7 (15.6) | 20 (18.7) | 7 (15.6) | 16 (17.8) | ||
| Pneumonia or pleural empyema | 4 (8.9) | 13 (12.1) | 4 (8.9) | 9 (10.0) | ||
| Genitourinary | 3 (6.7) | 2 (1.9) | 3 (6.7) | 2 (2.2) | ||
| Unknown | 6 (13.3) | 22 (20.6) | 6 (13.3) | 21 (23.3) | ||
| Antibiotic treatment | ||||||
| Time from onset to appropriate treatment, d, median (IQR) | 0 (0–2) | 0 (0–1) | .18 | 0 (0–2) | 0 (0–1) | .10 |
| Duration of appropriate treatment, d, median (IQR) | 15 (14–16) | 15 (13–19) | .92 | 15 (14–16) | 15 (14–19) | .49 |
| Duration of intravenous treatment, d, median (IQR) | 7 (6–8) | 15 (13–19) | <.01 | 7 (6–8) | 15 (14–19) | <.01 |
| Source control <72 h | 45 (92.1) | 103 (96.3) | .45 | 45 (92.1) | 86 (95.6) | .37 |
| ID consultation | 45 (100.0) | 91 (85.0) | .01 | 45 (100.0) | 85 (94.4) | .26 |
| Time from index culture to visit, d, median (IQR) | 2 (1–2) | 2 (1–3) | .61 | 2 (1–2) | 2 (1–3) | .68 |
Data are presented as No. (%) unless otherwise indicated. Patients may have >1 risk factor.
Abbreviations: HIV, human immunodeficiency virus; ICU, intensive care unit; ID, infectious disease; IQR, interquartile range; MIC, minimum inhibitory concentration; MRSA, methicillin-resistant Staphylococcus aureus.
aImmunosuppression includes systemic corticosteroid therapy for >2 weeks (>40 mg prednisolone or equivalent), antineoplastic chemotherapy, and any other immunosuppressive therapy.
bIncludes 9 endocavitary cardiac devices, 4 prosthetic valves, 37 other predisposing valve conditions, and 3 endovascular grafts.
Baseline Characteristics of Adult Patients With Staphylococcus aureus Bacteremia Comparing Treatment With Early Oral Switch to Linezolid and Standard Parenteral Treatment
| Characteristic | Whole Cohort | Propensity Score-matched Cohort | ||||
|---|---|---|---|---|---|---|
| Oral Linezolid (n = 45) | Standard Treatment (n = 107) | P Value | Oral Linezolid (n = 45) | Standard Treatment (n = 90) | P Value | |
| Male sex | 30 (66.7) | 68 (63.5) | .86 | 30 (66.7) | 59 (65.6) | 1.00 |
| Age, y, median (IQR) | 61 (45–69) | 63 (49–76) | .16 | 61 (45–69) | 64 (47–77) | .14 |
| MRSA strain | 5 (11.1) | 19 (17.8) | .43 | 5 (11.1) | 12 (13.3) | .93 |
| Preexisting condition | ||||||
| Charlson index, median (IQR) | 2 (1–5) | 2 (1–5) | .22 | 2 (1–5) | 2 (1–5) | .48 |
| Diabetes mellitus | 10 (22.2) | 29 (27.1) | .67 | 10 (22.2) | 22 (24.4) | .94 |
| Malignancy | 22 (48.9) | 37 (34.6) | .14 | 22 (48.9) | 33 (36.7) | .24 |
| Liver cirrhosis | 0 (0.0) | 7 (6.5) | .18 | 0 (0.0) | 1 (1.1) | 1.00 |
| Chronic renal disease | 2 (4.4) | 24 (22.4) | .01 | 2 (4.4) | 20 (22.2) | .02 |
| Dialysis | 1 (2.2) | 15 (14.0) | .06 | 1 (2.2) | 12 (13.3) | .08 |
| Immunosuppressiona | 19 (42.2) | 35 (32.7) | .35 | 19 (42.2) | 27 (30.0) | .22 |
| HIV infection | 0 (0.0) | 3 (2.8) | .62 | 0 (0.0) | 2 (2.2) | .80 |
| Neutropenia (<0.5 neutrophils/μL) | 4 (8.9) | 3 (2.8) | .23 | 4 (8.9) | 3 (3.3) | .34 |
| Solid organ transplant | 2 (4.4) | 10 (9.3) | .49 | 2 (4.4) | 5 (5.6) | 1.00 |
| Endocarditis-predisposing condition or endovascular graftb | 8 (17.8) | 33 (30.8) | .15 | 8 (17.8) | 26 (28.9) | .23 |
| Orthopedic device | 3 (6.7) | 11 (10.3) | .69 | 3 (6.7) | 10 (11.1) | .61 |
| Acquisition | .13 | .40 | ||||
| Hospital-acquired | 22 (48.9) | 67 (62.6) | 22 (48.9) | 55 (61.1) | ||
| Healthcare-related | 4 (8.9) | 7 (6.5) | 4 (8.9) | 6 (6.7) | ||
| Community-acquired | 19 (42.2) | 33 (30.8) | 19 (42.2) | 29 (32.2) | ||
| Clinical data | ||||||
| Septic shock | 2 (4.4) | 16 (15.0) | .12 | 2 (4.4) | 9 (10.0) | .44 |
| ICU stay | 4 (8.9) | 27 (25.2) | .04 | 4 (8.9) | 16 (17.8) | .27 |
| Source of bacteremia | .69 | .45 | ||||
| Catheter-related | 25 (55.6) | 50 (46.7) | 25 (55.6) | 42 (46.7) | ||
| Skin and soft tissue | 7 (15.6) | 20 (18.7) | 7 (15.6) | 16 (17.8) | ||
| Pneumonia or pleural empyema | 4 (8.9) | 13 (12.1) | 4 (8.9) | 9 (10.0) | ||
| Genitourinary | 3 (6.7) | 2 (1.9) | 3 (6.7) | 2 (2.2) | ||
| Unknown | 6 (13.3) | 22 (20.6) | 6 (13.3) | 21 (23.3) | ||
| Antibiotic treatment | ||||||
| Time from onset to appropriate treatment, d, median (IQR) | 0 (0–2) | 0 (0–1) | .18 | 0 (0–2) | 0 (0–1) | .10 |
| Duration of appropriate treatment, d, median (IQR) | 15 (14–16) | 15 (13–19) | .92 | 15 (14–16) | 15 (14–19) | .49 |
| Duration of intravenous treatment, d, median (IQR) | 7 (6–8) | 15 (13–19) | <.01 | 7 (6–8) | 15 (14–19) | <.01 |
| Source control <72 h | 45 (92.1) | 103 (96.3) | .45 | 45 (92.1) | 86 (95.6) | .37 |
| ID consultation | 45 (100.0) | 91 (85.0) | .01 | 45 (100.0) | 85 (94.4) | .26 |
| Time from index culture to visit, d, median (IQR) | 2 (1–2) | 2 (1–3) | .61 | 2 (1–2) | 2 (1–3) | .68 |
| Characteristic | Whole Cohort | Propensity Score-matched Cohort | ||||
|---|---|---|---|---|---|---|
| Oral Linezolid (n = 45) | Standard Treatment (n = 107) | P Value | Oral Linezolid (n = 45) | Standard Treatment (n = 90) | P Value | |
| Male sex | 30 (66.7) | 68 (63.5) | .86 | 30 (66.7) | 59 (65.6) | 1.00 |
| Age, y, median (IQR) | 61 (45–69) | 63 (49–76) | .16 | 61 (45–69) | 64 (47–77) | .14 |
| MRSA strain | 5 (11.1) | 19 (17.8) | .43 | 5 (11.1) | 12 (13.3) | .93 |
| Preexisting condition | ||||||
| Charlson index, median (IQR) | 2 (1–5) | 2 (1–5) | .22 | 2 (1–5) | 2 (1–5) | .48 |
| Diabetes mellitus | 10 (22.2) | 29 (27.1) | .67 | 10 (22.2) | 22 (24.4) | .94 |
| Malignancy | 22 (48.9) | 37 (34.6) | .14 | 22 (48.9) | 33 (36.7) | .24 |
| Liver cirrhosis | 0 (0.0) | 7 (6.5) | .18 | 0 (0.0) | 1 (1.1) | 1.00 |
| Chronic renal disease | 2 (4.4) | 24 (22.4) | .01 | 2 (4.4) | 20 (22.2) | .02 |
| Dialysis | 1 (2.2) | 15 (14.0) | .06 | 1 (2.2) | 12 (13.3) | .08 |
| Immunosuppressiona | 19 (42.2) | 35 (32.7) | .35 | 19 (42.2) | 27 (30.0) | .22 |
| HIV infection | 0 (0.0) | 3 (2.8) | .62 | 0 (0.0) | 2 (2.2) | .80 |
| Neutropenia (<0.5 neutrophils/μL) | 4 (8.9) | 3 (2.8) | .23 | 4 (8.9) | 3 (3.3) | .34 |
| Solid organ transplant | 2 (4.4) | 10 (9.3) | .49 | 2 (4.4) | 5 (5.6) | 1.00 |
| Endocarditis-predisposing condition or endovascular graftb | 8 (17.8) | 33 (30.8) | .15 | 8 (17.8) | 26 (28.9) | .23 |
| Orthopedic device | 3 (6.7) | 11 (10.3) | .69 | 3 (6.7) | 10 (11.1) | .61 |
| Acquisition | .13 | .40 | ||||
| Hospital-acquired | 22 (48.9) | 67 (62.6) | 22 (48.9) | 55 (61.1) | ||
| Healthcare-related | 4 (8.9) | 7 (6.5) | 4 (8.9) | 6 (6.7) | ||
| Community-acquired | 19 (42.2) | 33 (30.8) | 19 (42.2) | 29 (32.2) | ||
| Clinical data | ||||||
| Septic shock | 2 (4.4) | 16 (15.0) | .12 | 2 (4.4) | 9 (10.0) | .44 |
| ICU stay | 4 (8.9) | 27 (25.2) | .04 | 4 (8.9) | 16 (17.8) | .27 |
| Source of bacteremia | .69 | .45 | ||||
| Catheter-related | 25 (55.6) | 50 (46.7) | 25 (55.6) | 42 (46.7) | ||
| Skin and soft tissue | 7 (15.6) | 20 (18.7) | 7 (15.6) | 16 (17.8) | ||
| Pneumonia or pleural empyema | 4 (8.9) | 13 (12.1) | 4 (8.9) | 9 (10.0) | ||
| Genitourinary | 3 (6.7) | 2 (1.9) | 3 (6.7) | 2 (2.2) | ||
| Unknown | 6 (13.3) | 22 (20.6) | 6 (13.3) | 21 (23.3) | ||
| Antibiotic treatment | ||||||
| Time from onset to appropriate treatment, d, median (IQR) | 0 (0–2) | 0 (0–1) | .18 | 0 (0–2) | 0 (0–1) | .10 |
| Duration of appropriate treatment, d, median (IQR) | 15 (14–16) | 15 (13–19) | .92 | 15 (14–16) | 15 (14–19) | .49 |
| Duration of intravenous treatment, d, median (IQR) | 7 (6–8) | 15 (13–19) | <.01 | 7 (6–8) | 15 (14–19) | <.01 |
| Source control <72 h | 45 (92.1) | 103 (96.3) | .45 | 45 (92.1) | 86 (95.6) | .37 |
| ID consultation | 45 (100.0) | 91 (85.0) | .01 | 45 (100.0) | 85 (94.4) | .26 |
| Time from index culture to visit, d, median (IQR) | 2 (1–2) | 2 (1–3) | .61 | 2 (1–2) | 2 (1–3) | .68 |
Data are presented as No. (%) unless otherwise indicated. Patients may have >1 risk factor.
Abbreviations: HIV, human immunodeficiency virus; ICU, intensive care unit; ID, infectious disease; IQR, interquartile range; MIC, minimum inhibitory concentration; MRSA, methicillin-resistant Staphylococcus aureus.
aImmunosuppression includes systemic corticosteroid therapy for >2 weeks (>40 mg prednisolone or equivalent), antineoplastic chemotherapy, and any other immunosuppressive therapy.
bIncludes 9 endocavitary cardiac devices, 4 prosthetic valves, 37 other predisposing valve conditions, and 3 endovascular grafts.
Follow-up blood cultures were obtained from the 45 patients (100%) in the linezolid group and the 89 patients (98.9%) in the SPT group. Transthoracic or transesophageal echocardiography was performed in 16 patients (35.6%) in the linezolid group and in 50 patients (55.6%) in the SPT group.
In the SPT group, the most frequently prescribed definitive therapy for the 78 methicillin-susceptible S. aureus infections was cloxacillin (41 patients [52.6%]), followed by cefazolin (21 patients [26.9%]), β-lactam/β-lactamase inhibitor combinations (8 patients [10.3%]), carbapenems (4 patients [5.1%]), daptomycin (2 patients [2.6%]), cefepime (1 patient [1.3%]), and teicoplanin (1 patient [1.3%]). For the 12 MRSA infections in the SPT group, the most frequently prescribed definitive therapies were daptomycin (8 patients [66.7%]), vancomycin (2 patients [16.7%]), and intravenous linezolid (2 patients [16.7%]).
The median duration of appropriate antibiotic treatment was 15 days in both groups (IQR, 14–16 days in the linezolid group and 14–19 days in the SPT group). The median interval between onset and the start of an appropriate antibiotic treatment (generally an antistaphylococcal β-lactam, vancomycin, or daptomycin) was 0 days in both groups (IQR, 0–2 days in the linezolid group and 0–1 day in the SPT group). The median interval between appropriate antibiotic treatment start and oral switch to linezolid was 7 days (IQR, 6–8 days). None of the patients in the linezolid group presented with adverse events that forced them to interrupt oral linezolid.
Primary Outcome
After propensity score matching, there were 1 (2.2%) and 4 (4.4%) relapses within the 90-day follow-up period in the linezolid group and the SPT group, respectively (Table 2). We observed no difference in 90-day relapse between the groups after competitive risk survival analysis (P = .83). None of the patients with chronic renal disease relapsed within the 90-day follow-up period. The same bacterial clone was confirmed to be responsible for the relapse through WGS in 4 of 5 relapses. The remaining relapse, which belonged to the SPT group, could not be confirmed due to loss of one of the bacterial isolates.
Outcomes in Adult Patients With Staphylococcus aureus Bacteremia Comparing Treatment With Early Oral Switch to Linezolid and Standard Parenteral Treatment
| Outcome | Whole Cohort | Propensity Score–matched Cohort | ||||
|---|---|---|---|---|---|---|
| Oral Linezolid (n = 45) | Standard Treatment (n = 107) | P Value | Oral Linezolid (n = 45) | Standard Treatment (n = 90) | P Value | |
| 90-d relapse in survivors | 1 (2.2) | 4 (3.7) | 1.00 | 1 (2.2) | 4 (4.4) | .87 |
| 14-d mortality | 0 (0.0) | 10 (9.3) | .08 | 0 (0.0) | 6 (6.7) | .18 |
| 30-d mortality | 1 (2.2) | 17 (15.9) | .04 | 1 (2.2) | 12 (13.3) | .08 |
| Length of hospital stay after index culture, d, median (IQR)a | 8 (7–10) | 19 (15–32) | <.01 | 8 (7–10) | 19 (15–30) | <.01 |
| Outcome | Whole Cohort | Propensity Score–matched Cohort | ||||
|---|---|---|---|---|---|---|
| Oral Linezolid (n = 45) | Standard Treatment (n = 107) | P Value | Oral Linezolid (n = 45) | Standard Treatment (n = 90) | P Value | |
| 90-d relapse in survivors | 1 (2.2) | 4 (3.7) | 1.00 | 1 (2.2) | 4 (4.4) | .87 |
| 14-d mortality | 0 (0.0) | 10 (9.3) | .08 | 0 (0.0) | 6 (6.7) | .18 |
| 30-d mortality | 1 (2.2) | 17 (15.9) | .04 | 1 (2.2) | 12 (13.3) | .08 |
| Length of hospital stay after index culture, d, median (IQR)a | 8 (7–10) | 19 (15–32) | <.01 | 8 (7–10) | 19 (15–30) | <.01 |
Data are presented as No. (%) unless otherwise indicated.
Abbreviation: IQR, interquartile range.
aExcluding those who died during hospitalization.
Outcomes in Adult Patients With Staphylococcus aureus Bacteremia Comparing Treatment With Early Oral Switch to Linezolid and Standard Parenteral Treatment
| Outcome | Whole Cohort | Propensity Score–matched Cohort | ||||
|---|---|---|---|---|---|---|
| Oral Linezolid (n = 45) | Standard Treatment (n = 107) | P Value | Oral Linezolid (n = 45) | Standard Treatment (n = 90) | P Value | |
| 90-d relapse in survivors | 1 (2.2) | 4 (3.7) | 1.00 | 1 (2.2) | 4 (4.4) | .87 |
| 14-d mortality | 0 (0.0) | 10 (9.3) | .08 | 0 (0.0) | 6 (6.7) | .18 |
| 30-d mortality | 1 (2.2) | 17 (15.9) | .04 | 1 (2.2) | 12 (13.3) | .08 |
| Length of hospital stay after index culture, d, median (IQR)a | 8 (7–10) | 19 (15–32) | <.01 | 8 (7–10) | 19 (15–30) | <.01 |
| Outcome | Whole Cohort | Propensity Score–matched Cohort | ||||
|---|---|---|---|---|---|---|
| Oral Linezolid (n = 45) | Standard Treatment (n = 107) | P Value | Oral Linezolid (n = 45) | Standard Treatment (n = 90) | P Value | |
| 90-d relapse in survivors | 1 (2.2) | 4 (3.7) | 1.00 | 1 (2.2) | 4 (4.4) | .87 |
| 14-d mortality | 0 (0.0) | 10 (9.3) | .08 | 0 (0.0) | 6 (6.7) | .18 |
| 30-d mortality | 1 (2.2) | 17 (15.9) | .04 | 1 (2.2) | 12 (13.3) | .08 |
| Length of hospital stay after index culture, d, median (IQR)a | 8 (7–10) | 19 (15–32) | <.01 | 8 (7–10) | 19 (15–30) | <.01 |
Data are presented as No. (%) unless otherwise indicated.
Abbreviation: IQR, interquartile range.
aExcluding those who died during hospitalization.
Of these 5 relapses, 1 patient (20%) died within the 90-day follow-up period of the relapse, due to an early prosthetic valve endocarditis with aortic graft infection (previous Bentall procedure). The initial episode of SAB in this patient who died was diagnosed as catheter-related and treated for 28 days with SPT. The only relapse in the linezolid group was due to a septic thrombophlebitis with the initial episode of SAB diagnosed as catheter-related and treated for 15 days. Two of the 5 relapses presented a skin and soft tissue infection as the suspected source of the initial SAB with the sources of the relapses being identified as osteoarticular. The remaining relapse had a urinary catheter–associated source in the initial SAB episode and in the relapse.
In the analysis of risk factors associated with 90-day relapse, oral switch to linezolid was not associated with an increased risk of relapse (OR, 0.6 [95% CI, .1–5.4]; P = 0.64; Supplementary Table 1). This result was confirmed on adjusted multivariate analysis (Table 3). Additional factors that remained associated with 90-day relapse in the multivariable model were neutropenia at onset, the presence of a prosthetic valve, time from index culture to appropriate treatment, and time from index culture until ID consultation.
Multivariable Analysis of Potential Risk Factors Associated With 90-Day Relapse of Staphylococcus aureus Bacteremia
| Characteristic | Univariable Analysis | Multivariable Analysis | ||||
|---|---|---|---|---|---|---|
| OR | (95% CI) | P Value | OR | (95% CI) | P Value | |
| Early oral switch to linezolid | 0.6 | (.1–5.4) | .64 | … | … | … |
| Age, per year | 1.0 | (1.0–1.1) | .90 | … | … | … |
| MRSA strain | 3.8 | (.6–24.0) | .16 | … | … | … |
| Vancomycin E-test MIC ≥1.5 | 4.6 | (.5–42.1) | .18 | … | … | … |
| Neutropenia (<0.5 neutrophils/μL) | 5.9 | (.6–60.9) | .14 | 69.9 | (1.4–3589.3) | .04 |
| Prosthetic valve | 12.0 | (1.0–142.1) | .05 | 445.4 | (5.4–36753.3) | <.01 |
| Genitourinary source | 8.9 | (.8–99.1) | .07 | … | … | … |
| Time from index culture to appropriate treatment, per day | 2.9 | (1.4–6.4) | <.01 | 5.8 | (1.7–20.3) | <.01 |
| Time from index culture to ID physician visit, per day | 1.4 | (1.0–1.9) | .07 | 1.6 | (1.0–2.5) | .05 |
| Characteristic | Univariable Analysis | Multivariable Analysis | ||||
|---|---|---|---|---|---|---|
| OR | (95% CI) | P Value | OR | (95% CI) | P Value | |
| Early oral switch to linezolid | 0.6 | (.1–5.4) | .64 | … | … | … |
| Age, per year | 1.0 | (1.0–1.1) | .90 | … | … | … |
| MRSA strain | 3.8 | (.6–24.0) | .16 | … | … | … |
| Vancomycin E-test MIC ≥1.5 | 4.6 | (.5–42.1) | .18 | … | … | … |
| Neutropenia (<0.5 neutrophils/μL) | 5.9 | (.6–60.9) | .14 | 69.9 | (1.4–3589.3) | .04 |
| Prosthetic valve | 12.0 | (1.0–142.1) | .05 | 445.4 | (5.4–36753.3) | <.01 |
| Genitourinary source | 8.9 | (.8–99.1) | .07 | … | … | … |
| Time from index culture to appropriate treatment, per day | 2.9 | (1.4–6.4) | <.01 | 5.8 | (1.7–20.3) | <.01 |
| Time from index culture to ID physician visit, per day | 1.4 | (1.0–1.9) | .07 | 1.6 | (1.0–2.5) | .05 |
Abbreviations: CI, confidence interval; ID, infectious disease; MIC, minimum inhibitory concentration; MRSA, methicillin-resistant Staphylococcus aureus; OR, odds ratio.
Multivariable Analysis of Potential Risk Factors Associated With 90-Day Relapse of Staphylococcus aureus Bacteremia
| Characteristic | Univariable Analysis | Multivariable Analysis | ||||
|---|---|---|---|---|---|---|
| OR | (95% CI) | P Value | OR | (95% CI) | P Value | |
| Early oral switch to linezolid | 0.6 | (.1–5.4) | .64 | … | … | … |
| Age, per year | 1.0 | (1.0–1.1) | .90 | … | … | … |
| MRSA strain | 3.8 | (.6–24.0) | .16 | … | … | … |
| Vancomycin E-test MIC ≥1.5 | 4.6 | (.5–42.1) | .18 | … | … | … |
| Neutropenia (<0.5 neutrophils/μL) | 5.9 | (.6–60.9) | .14 | 69.9 | (1.4–3589.3) | .04 |
| Prosthetic valve | 12.0 | (1.0–142.1) | .05 | 445.4 | (5.4–36753.3) | <.01 |
| Genitourinary source | 8.9 | (.8–99.1) | .07 | … | … | … |
| Time from index culture to appropriate treatment, per day | 2.9 | (1.4–6.4) | <.01 | 5.8 | (1.7–20.3) | <.01 |
| Time from index culture to ID physician visit, per day | 1.4 | (1.0–1.9) | .07 | 1.6 | (1.0–2.5) | .05 |
| Characteristic | Univariable Analysis | Multivariable Analysis | ||||
|---|---|---|---|---|---|---|
| OR | (95% CI) | P Value | OR | (95% CI) | P Value | |
| Early oral switch to linezolid | 0.6 | (.1–5.4) | .64 | … | … | … |
| Age, per year | 1.0 | (1.0–1.1) | .90 | … | … | … |
| MRSA strain | 3.8 | (.6–24.0) | .16 | … | … | … |
| Vancomycin E-test MIC ≥1.5 | 4.6 | (.5–42.1) | .18 | … | … | … |
| Neutropenia (<0.5 neutrophils/μL) | 5.9 | (.6–60.9) | .14 | 69.9 | (1.4–3589.3) | .04 |
| Prosthetic valve | 12.0 | (1.0–142.1) | .05 | 445.4 | (5.4–36753.3) | <.01 |
| Genitourinary source | 8.9 | (.8–99.1) | .07 | … | … | … |
| Time from index culture to appropriate treatment, per day | 2.9 | (1.4–6.4) | <.01 | 5.8 | (1.7–20.3) | <.01 |
| Time from index culture to ID physician visit, per day | 1.4 | (1.0–1.9) | .07 | 1.6 | (1.0–2.5) | .05 |
Abbreviations: CI, confidence interval; ID, infectious disease; MIC, minimum inhibitory concentration; MRSA, methicillin-resistant Staphylococcus aureus; OR, odds ratio.
Secondary Outcomes
Without reaching statistical significance, 30-day all-cause mortality was lower in the linezolid group than in the SPT group (1 [2.2%] vs 12 [13.3%] deaths; P = .08; Table 2). In the analysis of risk factors associated with 30-day all-cause mortality, after adjustment for potential confounders, early oral switch to linezolid was associated with a decrease in 30-day mortality (OR, 0.1 [95% CI, .0–.9]; P = .04; Supplementary Table 2). Additional factors that remained associated with 30-day mortality in the multivariable model were Charlson comorbidity index and liver cirrhosis.
The median length of hospital stay after the index culture was 8 days (IQR, 7–10 days) in the linezolid group and 19 days (IQR, 15–30) in the SPT group (P < .01; Table 2).
DISCUSSION
Our findings indicate that in selected low-risk patients with SAB (ie, clinically stable without proven or suspected complicated endovascular infection or distant metastatic foci, with negative follow-up blood cultures and appropriate source control <72 hours), oral switch to linezolid between days 3 and 9 of treatment until completion may be a safe alternative to SPT, while reducing significantly hospital stay.
Relapses within the 90-day follow-up period were similar in the linezolid group than in the SPT group. Although not reaching statistical significance (P = .08), 30-day all-cause mortality was lower in the linezolid group. No drug-related events leading to discontinuation were noted in the linezolid group, probably due to the relatively short duration of most treatments.
Despite the frequency of SAB and the potential advantages of an early oral switch treatment strategy, little evidence exists to support alternative SAB treatment strategies. A major cause of this lack of evidence is the difficulty in designing and executing high-quality trials to answer these therapeutic questions as they require often impractically large sample sizes [31, 32]. Recently the Gram-Positive Committee of the Antibacterial Resistance Leadership Group highlighted oral treatment strategies for invasive S. aureus infections as one of the priorities for future investigations [31].
To our knowledge, only one randomized controlled trial assessed the effectiveness of oral antimicrobial therapy in SAB. This trial included 104 patients who were treated either with an oral combination of a fluoroquinolone (fleroxacin) plus rifampicin or with SPT (flucloxacillin or vancomycin). Even though the cure rate in both groups was similar (82% vs 80%) and the median length of hospital stay was lower in the fleroxacin-rifampicin group (12 days vs 23 days), the study was limited due to a higher incidence of drug toxicity in the fleroxacin-rifampicin group [33]. Despite this lack of evidence, Thwaites et al found that 25% of patients with SAB in 8 centers in the United Kingdom received oral antimicrobials alone for >50% of the treatment duration, indicating that this strategy is common in some hospitals [34].
Other previous studies have shown a reduction in hospital stay and potential economic savings with an early oral treatment strategy in selected patients with gram-positive infections including S. aureus complicated skin and soft tissue infections [35, 36].
The main limitations of this study are its observational design, with the decision to switch to oral linezolid at the discretion of the treating physician, and the relatively small sample size. We tried to overcome confounding by indication, restricting our study population only to those where an early oral switch to linezolid was considered clinically feasible and using propensity score matching. Due to the limited number of controls to match, even though the propensity-matched groups were generally well balanced, patients in the linezolid group were still less likely to have chronic renal disease. Furthermore, although statistically not significant, the SPT group presented a higher proportion of patients with several risk factors associated with clinical failure and relapse such as hemodialysis, endocarditis-predisposing conditions, endovascular grafts, orthopedic devices, septic shock, unknown source of SAB, and absence of ID consultation [4, 22, 37, 38]. Considering the relatively small sample size that hindered statistical significance in the various comparisons and the potential impact of the aggregate of these risk factors, we recommend interpreting these results with caution, especially in the presence of these risk factors. Additionally, we cannot exclude the possibility of unmeasured confounding factors not included in our propensity scores.
One of the strengths of this study is that it includes patients in a real-life scenario. It must be noted that only a minority of patients with SAB were considered suitable for an early oral switch. All patients who were switched to linezolid had a close follow-up evaluation allowing us to detect complications, if any, in an early stage. To avoid an improper use that could lead to the emergence of resistance and put patients at risk of a suboptimal treatment, we believe that this treatment strategy should be carefully planned to secure close follow-up, and preferably by ID physicians with broad experience managing SAB.
In summary, this study suggests that treating SAB in selected low-risk patients with an early oral switch to linezolid between days 3 and 9 of treatment until completion is not associated with an increased risk of mortality or relapse compared to SPT, in the absence of complicated endovascular infection, metastatic distant foci, persistent bacteremia, or inappropriate source control. Furthermore, early oral switch to linezolid may reduce hospital stay compared with SPT. These results should encourage a well-designed randomized controlled trial to strengthen the evidence and enable this early oral treatment strategy with linezolid in clinical practice.
Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Notes
Financial support. This work was supported by Plan Nacional de I+D+I 2013–2016 and Instituto de Salud Carlos III, Subdirección General de Redes y Centros de Investigación Cooperativa, Ministerio de Economía, Industria y Competitividad, Spanish Network for Research in Infectious Diseases (REIPI RD16/0016/0003)—co financed by European Development Regional Fund “A Way to Achieve Europe,” Operative Program Intelligent Growth 2014–2020.
Potential conflicts of interest. All authors: No reported conflict. All authors have submitted the ICMJE form for disclosure of potential conflicts of interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
