https://orcid.org/0000-0002-0648-3555
Abstract
Despite being the leading cause of mortality from bloodstream infections worldwide, little is known about regional variation in treatment practices for Staphylococcus aureus bacteremia (SAB). The aim of this study was to identify global variation in management, diagnostics, and definitions of SAB.
During a 20-day period in 2022, physicians throughout the world were surveyed on SAB treatment practices. The survey was distributed through listservs, e-mails, and social media.
In total, 2031 physicians from 71 different countries on 6 continents (North America [701, 35%], Europe [573, 28%], Asia [409, 20%], Oceania [182, 9%], South America [124, 6%], and Africa [42, 2%]) completed the survey. Management-based responses differed significantly by continent for preferred treatment of methicillin-susceptible S. aureus (MSSA) and methicillin-resistant S. aureus (MRSA) bacteremia, use of adjunctive rifampin for prosthetic material infection, and use of oral antibiotics (P < .01 for all comparisons). The 18F-FDG PET/CT scans were most commonly used in Europe (94%) and least frequently used in Africa (13%) and North America (51%; P < .01). Although most respondents defined persistent SAB as 3–4 days of positive blood cultures, responses ranged from 2 days in 31% of European respondents to 7 days in 38% of Asian respondents (P < .01).
Large practice variations for SAB exist throughout the world, reflecting the paucity of high-quality data and the absence of an international standard of care for the management of SAB.
Staphylococcus aureus is the leading cause of mortality by bloodstream infections worldwide [1], and methicillin resistant S. aureus (MRSA) is the leading cause of mortality attributable to antimicrobial resistance [2]. Despite its global distribution and an incidence of approximately 30 per 100 000 person-years [3, 4], the optimal approach to S. aureus bacteremia (SAB) is poorly understood. Despite the fact that SAB has been a major theme in the medical literature for decades, basic treatment elements such as the optimal antibiotic regimen, the role of adjunct and oral antibiotics, the optimal treatment duration, and the definition of persistent SAB remain fundamentally unknown. Even less is known about global differences in treatment practices for SAB.
The aim of this study was to identify global variation in management, diagnostics, and definitions of SAB. To do this, we used a variety of social media platforms to reach a large number of clinicians throughout the world for a survey on SAB treatment practices.
METHODS
Survey Development and Distribution
We conducted this study on geographic practice variation in SAB by modifying a recently developed survey that was deployed in five European countries [5]. The modified survey was tested among an independent expert panel and adjusted where appropriate. The survey focused on unsettled aspects of the disease in clinical practice: first-choice antimicrobial agents, intravenous to oral switch of antimicrobial therapy, treatment duration, the use of 18F-fluoro-deoxyglucose positron emission tomography/computed tomography (PET/CT) (18F-FDG PET/CT) scan, and the definition of persistent SAB (Supplementary Appendix 1). When relevant, questions were provided separately for both methicillin-susceptible S. aureus (MSSA) and MRSA bacteremia. The survey was anonymous and voluntary. Country of practice was asked to determine geographic region and subsequently respondents were grouped by continent. The survey was developed in English. Target respondents included infectious diseases, clinical microbiology and internal medicine physicians (both adults and pediatrics) treating SAB patients throughout the world. The survey was distributed through a public URL link on listservs, e-mails, Twitter, and WeChat. Respondents were asked to share the survey link with their professional network. The link was accessible between 2 November and 22 November 2022.
Ethical Approval
Given the anonymous and voluntary aspects of the survey, a declaration of exemption was issued by the institutional review board of Duke University.
Definitions
Uncomplicated bacteremia was defined as SAB that was not community-acquired, with <48 hours of positive blood cultures under appropriate antibiotic treatment, and no signs of metastatic infections. Oral switch therapy was defined as prescribing at least part of the treatment course orally. Both definitions were provided with the relevant questions. The estimated percentage of SAB patients in whom oral switch therapy was used was defined as never or uncommonly (<20% of SAB patients), sometimes (20%–60% of SAB patients) or frequently (>60% of SAB patients). All questions concerning antibiotic treatment assumed that the isolate was susceptible to the drug.
Data Collection and Management
Study data were collected and managed using Research Electronic Data Capture (REDCap) tools hosted at Duke University [6]. Respondents who completed 0 or 1 question only were removed from the analysis, as well as respondents that did not enter their country of practice. In order to remove potential non-targeted respondents, records were screened for straight-liners (respondents that failed to differentiate between response alternatives by, for example, answering always only the first answer, or only the “other” option to every multiple-choice question) and for nonsensical answers to open-ended questions. Because the survey was distributed through different listservs and social media, the number of times the public survey URL link was opened was used to provide the best estimation of the response rate.
Statistical Analysis
Descriptive statistics were used to summarize the data. Data were presented as percentages or proportions of the number of respondents that answered the question for categorical variables, and as medians plus interquartile range (IQR) for continuous variables. Pearson χ2 tests were performed to analyze differences between continents. All analyses were carried out using SPSS statistics version 28.0.1.1 (IBM Corporation, Armonk, New York, USA).
RESULTS
A total of 2229 individual survey responses were obtained. The URL link was opened 5679 times (response rate 39%). Nine percent (198/2229) of records were removed from the analyses because of completion of ≤1 questions (88/2229, 4%) or not entering the country of practice (110/2229, 5%). No non-targeted responses were identified. The remaining survey records of 2031 respondents from 71 different countries on 6 continents (North America [701, 35%], Europe [573, 28%], Asia [409, 20%], Oceania [182, 9%], South America [124, 6%], and Africa [42, 2%]) were included in the analysis (Figure 1). Respondents stated they were physicians in adult infectious diseases (74%), clinical microbiology (10%), internal medicine (6%), and pediatric infectious diseases (5%). Thirteen percent of respondents were still in training, and 44% had been registered as a consultant for more than 10 years.
Global distribution of survey respondents. Respondents per country: 71 unique countries participated, and participation ranged from 1 to 654 respondents per country.
Antimicrobial Management of SAB
Antibiotic treatment for SAB differed significantly between continents (Figure 2). For MSSA bacteremia, cefazolin was the first-choice antibiotic treatment in North America (78% of respondents), whereas anti-staphylococcal penicillins were preferred in all other continents (51%–82%; P < .01) (Figure 2A). For MRSA bacteremia, vancomycin was the preferred first-choice antibiotic agent in all continents, but with a broad range of 53%–97% of respondents. Daptomycin was identified as the first-choice antibiotic agent for MRSA bacteremia in 23% of European respondents but in <10% of respondents of all other continents (Figure 2B; P < .01 for all comparisons above.)
Antibiotic treatment preferences for S. aureus bacteremia per continent. Percentage of total respondents of the question per continent, and count of respondents per continent. *Listed prosthetic materials: cardiac device, endovascular device, joint prosthesis, heart valve, and spondylodesis material.
Adjunctive Rifampin
The practice of adding adjunctive rifampin in cases of SAB associated with infected prosthetic material was most frequently reported in Europe: 94% of European respondents would add it in at least 1 of the listed prosthetic material infections (cardiac device, endovascular device, joint prosthesis, prosthetic heart valve, and/or spondylodesis material infection). In Oceania and Africa rifampin was least often used in SAB patients with infected prosthetic material: 26% and 38% never added rifampin for this indication, respectively (Figure 2C).
Oral Switch Therapy
The estimated percentage of SAB patients in whom oral therapy was used was lowest in North America, where 76% of physicians indicated that they never or uncommonly used oral switch antibiotic therapy. Acceptance of oral therapy was highest in Europe, where 55% of physicians indicated that they used it frequently in their SAB patients (Figure 3A). The majority of respondents from all continents except Oceania (48%) indicated that they would use oral switch therapy in uncomplicated SAB (57%–71%). Over half (54%–66%) of respondents from every continent identified patients with SAB originating from skin or soft tissue infection as a suitable patient group for safe oral switch therapy. By contrast, respondents differed widely on their views of the acceptability of oral therapy for SAB associated with spondylodiscitis, ranging from 19% in Africa to 60% in Oceania (Table 1). Source control and absence of a central nervous system infection were the only criteria for oral switch therapy for which there was broad agreement among respondents (79% and 69%, respectively) (Supplementary Table 1).
Oral switch therapy in SAB. Percentage (count) of total respondents of the question per continent. A, Estimated percentage of SAB patients per physician that are treated orally for at least part of the treatment course. B, Days of positive blood cultures while receiving adequate treatment to define persistent bacteremia, in S. aureus bacteremia. Abbreviation: SAB, S. aureus bacteremia.
Regional Practice Patterns for S. aureus Bacteremia
| Total N = 2031 | North America N = 701 | Europe N = 573 | Asia N = 409 | Oceania N = 182 | South America N = 124 | Africa N = 42 | P+ | |
|---|---|---|---|---|---|---|---|---|
| Cefazolin versus ASP in MSSAB | ||||||||
| Regarded equally effective | 1470 (81.4) | 557 (85.2) | 416 (78.8) | 235 (76.5) | 148 (85.5) | 86 (81.1) | 28 (75.7) | <.01 |
| Reasons to extend therapy from 2 wks to 4 wks or more | <.01 | |||||||
| Positive BC 48–72 h after start antibiotics | 1248 (84.0) | 516 (89.7) | 363 (83.3) | 169 (76.1) | 132 (86.8) | 54 (65.9) | 14 (77.8) | |
| Immunocompromised status | 873 (59.5) | 411 (72.0) | 187 (43.2) | 134 (62.9) | 94 (63.1) | 36 (44.4) | 11 (55.0) | |
| Unknown portal of entry | 787 (54.3) | 386 (68.4) | 178 (41.7) | 102 (48.1) | 87 (58.0) | 28 (36.4) | 6 (31.6) | |
| Fever 72 h after first positive BC | 746 (51.6) | 331 (58.8) | 203 (47.3) | 94 (45.4) | 78 (52.0) | 32 (41.0) | 8 (40.0) | |
| Community acquisition | 460 (31.3) | 231 (40.7) | 96 (22.1) | 44 (20.4) | 71 (46.4) | 14 (17.9) | 4 (21.1) | |
| Oral switch therapy in different infection focia | ||||||||
| Skin/soft tissue infection | 922 (59.3) | 330 (56.2) | 305 (66.3) | 140 (58.8) | 85 (54.5) | 47 (54.0) | 15 (55.6) | .01 |
| Osteomyelitis | 840 (54.0) | 260 (44.3) | 301 (65.4) | 106 (44.5) | 105 (67.3) | 60 (69.0) | 8 (29.6) | <.01 |
| Spondylodiscitis | 659 (42.4) | 155 (26.4) | 275 (59.8) | 84 (35.3) | 94 (60.3) | 46 (52.9) | 5 (18.5) | <.01 |
| Prosthetic joint septic arthritis | 610 (39.2) | 144 (24.5) | 260 (56.5) | 75 (31.5) | 89 (57.1) | 37 (42.5) | 5 (18.5) | <.01 |
| Prosthetic valve endocarditis | 130 (8.4) | 24 (4.1) | 54 (11.7) | 14 (5.9) | 32 (20.5) | 5 (5.7) | 1 (3.7) | <.01 |
| All of the above | 120 (7.7) | 22 (3.7) | 44 (9.6) | 30 (12.6) | 16 (10.3) | 5 (5.7) | 3 (11.1) | <.01 |
| 18F-FDG PET/CT for SAB | ||||||||
| PET/CT readily available | 829 (55.6) | 278 (48.9) | 341 (78.0) | 106 (46.7) | 78 (51.3) | 24 (28.6) | 2 (9.1) | <.01 |
| PET/CT for SAB covered by insurance | 610 (42.2) | 177 (33.2) | 332 (77.4) | 41 (18.2) | 37 (24.3) | 21 (25.0) | 2 (9.1) | <.01 |
| PET/CT use in some/all SAB patients | 1009 (67.8) | 293 (51.1) | 409 (93.8) | 125 (57.0) | 124 (82.8) | 50 (61.0) | 4 (12.7) | <.01 |
| Never use PET/CT in SAB patients | 479 (32.2) | 281 (49.0) | 27 (6.2) | 96 (43.0) | 26 (17.2) | 32 (39.0) | 17 (77.3) | <.01 |
| Available, but never use PET/CT in SABb | 101 (12.2) | 63 (22.7) | 5 (1.5) | 24 (23.1) | 8 (10.3) | 1 (4.2) | <.01 | |
| Actions following diagnosis of persistent SAB | ||||||||
| Additional diagnostic testing | 1610 (95.8) | 605 (97.0) | 484 (96.6) | 242 (92.0) | 161 (97.6) | 91 (96.8) | 27 (79.4) | <.01 |
| Change antibiotic management | 1246 (74.7) | 523 (83.5) | 316 (63.8) | 207 (80.2) | 106 (64.6) | 72 (76.7) | 22 (68.8) | <.01 |
| Total N = 2031 | North America N = 701 | Europe N = 573 | Asia N = 409 | Oceania N = 182 | South America N = 124 | Africa N = 42 | P+ | |
|---|---|---|---|---|---|---|---|---|
| Cefazolin versus ASP in MSSAB | ||||||||
| Regarded equally effective | 1470 (81.4) | 557 (85.2) | 416 (78.8) | 235 (76.5) | 148 (85.5) | 86 (81.1) | 28 (75.7) | <.01 |
| Reasons to extend therapy from 2 wks to 4 wks or more | <.01 | |||||||
| Positive BC 48–72 h after start antibiotics | 1248 (84.0) | 516 (89.7) | 363 (83.3) | 169 (76.1) | 132 (86.8) | 54 (65.9) | 14 (77.8) | |
| Immunocompromised status | 873 (59.5) | 411 (72.0) | 187 (43.2) | 134 (62.9) | 94 (63.1) | 36 (44.4) | 11 (55.0) | |
| Unknown portal of entry | 787 (54.3) | 386 (68.4) | 178 (41.7) | 102 (48.1) | 87 (58.0) | 28 (36.4) | 6 (31.6) | |
| Fever 72 h after first positive BC | 746 (51.6) | 331 (58.8) | 203 (47.3) | 94 (45.4) | 78 (52.0) | 32 (41.0) | 8 (40.0) | |
| Community acquisition | 460 (31.3) | 231 (40.7) | 96 (22.1) | 44 (20.4) | 71 (46.4) | 14 (17.9) | 4 (21.1) | |
| Oral switch therapy in different infection focia | ||||||||
| Skin/soft tissue infection | 922 (59.3) | 330 (56.2) | 305 (66.3) | 140 (58.8) | 85 (54.5) | 47 (54.0) | 15 (55.6) | .01 |
| Osteomyelitis | 840 (54.0) | 260 (44.3) | 301 (65.4) | 106 (44.5) | 105 (67.3) | 60 (69.0) | 8 (29.6) | <.01 |
| Spondylodiscitis | 659 (42.4) | 155 (26.4) | 275 (59.8) | 84 (35.3) | 94 (60.3) | 46 (52.9) | 5 (18.5) | <.01 |
| Prosthetic joint septic arthritis | 610 (39.2) | 144 (24.5) | 260 (56.5) | 75 (31.5) | 89 (57.1) | 37 (42.5) | 5 (18.5) | <.01 |
| Prosthetic valve endocarditis | 130 (8.4) | 24 (4.1) | 54 (11.7) | 14 (5.9) | 32 (20.5) | 5 (5.7) | 1 (3.7) | <.01 |
| All of the above | 120 (7.7) | 22 (3.7) | 44 (9.6) | 30 (12.6) | 16 (10.3) | 5 (5.7) | 3 (11.1) | <.01 |
| 18F-FDG PET/CT for SAB | ||||||||
| PET/CT readily available | 829 (55.6) | 278 (48.9) | 341 (78.0) | 106 (46.7) | 78 (51.3) | 24 (28.6) | 2 (9.1) | <.01 |
| PET/CT for SAB covered by insurance | 610 (42.2) | 177 (33.2) | 332 (77.4) | 41 (18.2) | 37 (24.3) | 21 (25.0) | 2 (9.1) | <.01 |
| PET/CT use in some/all SAB patients | 1009 (67.8) | 293 (51.1) | 409 (93.8) | 125 (57.0) | 124 (82.8) | 50 (61.0) | 4 (12.7) | <.01 |
| Never use PET/CT in SAB patients | 479 (32.2) | 281 (49.0) | 27 (6.2) | 96 (43.0) | 26 (17.2) | 32 (39.0) | 17 (77.3) | <.01 |
| Available, but never use PET/CT in SABb | 101 (12.2) | 63 (22.7) | 5 (1.5) | 24 (23.1) | 8 (10.3) | 1 (4.2) | <.01 | |
| Actions following diagnosis of persistent SAB | ||||||||
| Additional diagnostic testing | 1610 (95.8) | 605 (97.0) | 484 (96.6) | 242 (92.0) | 161 (97.6) | 91 (96.8) | 27 (79.4) | <.01 |
| Change antibiotic management | 1246 (74.7) | 523 (83.5) | 316 (63.8) | 207 (80.2) | 106 (64.6) | 72 (76.7) | 22 (68.8) | <.01 |
N = the total number of respondents of the survey. Values are counts (%). Not all respondents answered every question, therefore the percentages represent the percentage of the total respondents of the continent who answered this question.
Abbreviations: ASP, anti-staphylococcal penicillin; BC, blood culture; h, hours; MSSAB, methicillin-susceptible Staphylococcus aureus bacteremia; PET/CT, positron emission tomography/computed tomography; SAB, Staphylococcus aureus bacteremia.
Results are still significant with P < .01 when continents with n ≤ 5 were excluded from analysis.
This represents the number and percent of respondents that indicated that PET/CT is readily available but still never use PET/CT in SAB patients.
Regional Practice Patterns for S. aureus Bacteremia
| Total N = 2031 | North America N = 701 | Europe N = 573 | Asia N = 409 | Oceania N = 182 | South America N = 124 | Africa N = 42 | P+ | |
|---|---|---|---|---|---|---|---|---|
| Cefazolin versus ASP in MSSAB | ||||||||
| Regarded equally effective | 1470 (81.4) | 557 (85.2) | 416 (78.8) | 235 (76.5) | 148 (85.5) | 86 (81.1) | 28 (75.7) | <.01 |
| Reasons to extend therapy from 2 wks to 4 wks or more | <.01 | |||||||
| Positive BC 48–72 h after start antibiotics | 1248 (84.0) | 516 (89.7) | 363 (83.3) | 169 (76.1) | 132 (86.8) | 54 (65.9) | 14 (77.8) | |
| Immunocompromised status | 873 (59.5) | 411 (72.0) | 187 (43.2) | 134 (62.9) | 94 (63.1) | 36 (44.4) | 11 (55.0) | |
| Unknown portal of entry | 787 (54.3) | 386 (68.4) | 178 (41.7) | 102 (48.1) | 87 (58.0) | 28 (36.4) | 6 (31.6) | |
| Fever 72 h after first positive BC | 746 (51.6) | 331 (58.8) | 203 (47.3) | 94 (45.4) | 78 (52.0) | 32 (41.0) | 8 (40.0) | |
| Community acquisition | 460 (31.3) | 231 (40.7) | 96 (22.1) | 44 (20.4) | 71 (46.4) | 14 (17.9) | 4 (21.1) | |
| Oral switch therapy in different infection focia | ||||||||
| Skin/soft tissue infection | 922 (59.3) | 330 (56.2) | 305 (66.3) | 140 (58.8) | 85 (54.5) | 47 (54.0) | 15 (55.6) | .01 |
| Osteomyelitis | 840 (54.0) | 260 (44.3) | 301 (65.4) | 106 (44.5) | 105 (67.3) | 60 (69.0) | 8 (29.6) | <.01 |
| Spondylodiscitis | 659 (42.4) | 155 (26.4) | 275 (59.8) | 84 (35.3) | 94 (60.3) | 46 (52.9) | 5 (18.5) | <.01 |
| Prosthetic joint septic arthritis | 610 (39.2) | 144 (24.5) | 260 (56.5) | 75 (31.5) | 89 (57.1) | 37 (42.5) | 5 (18.5) | <.01 |
| Prosthetic valve endocarditis | 130 (8.4) | 24 (4.1) | 54 (11.7) | 14 (5.9) | 32 (20.5) | 5 (5.7) | 1 (3.7) | <.01 |
| All of the above | 120 (7.7) | 22 (3.7) | 44 (9.6) | 30 (12.6) | 16 (10.3) | 5 (5.7) | 3 (11.1) | <.01 |
| 18F-FDG PET/CT for SAB | ||||||||
| PET/CT readily available | 829 (55.6) | 278 (48.9) | 341 (78.0) | 106 (46.7) | 78 (51.3) | 24 (28.6) | 2 (9.1) | <.01 |
| PET/CT for SAB covered by insurance | 610 (42.2) | 177 (33.2) | 332 (77.4) | 41 (18.2) | 37 (24.3) | 21 (25.0) | 2 (9.1) | <.01 |
| PET/CT use in some/all SAB patients | 1009 (67.8) | 293 (51.1) | 409 (93.8) | 125 (57.0) | 124 (82.8) | 50 (61.0) | 4 (12.7) | <.01 |
| Never use PET/CT in SAB patients | 479 (32.2) | 281 (49.0) | 27 (6.2) | 96 (43.0) | 26 (17.2) | 32 (39.0) | 17 (77.3) | <.01 |
| Available, but never use PET/CT in SABb | 101 (12.2) | 63 (22.7) | 5 (1.5) | 24 (23.1) | 8 (10.3) | 1 (4.2) | <.01 | |
| Actions following diagnosis of persistent SAB | ||||||||
| Additional diagnostic testing | 1610 (95.8) | 605 (97.0) | 484 (96.6) | 242 (92.0) | 161 (97.6) | 91 (96.8) | 27 (79.4) | <.01 |
| Change antibiotic management | 1246 (74.7) | 523 (83.5) | 316 (63.8) | 207 (80.2) | 106 (64.6) | 72 (76.7) | 22 (68.8) | <.01 |
| Total N = 2031 | North America N = 701 | Europe N = 573 | Asia N = 409 | Oceania N = 182 | South America N = 124 | Africa N = 42 | P+ | |
|---|---|---|---|---|---|---|---|---|
| Cefazolin versus ASP in MSSAB | ||||||||
| Regarded equally effective | 1470 (81.4) | 557 (85.2) | 416 (78.8) | 235 (76.5) | 148 (85.5) | 86 (81.1) | 28 (75.7) | <.01 |
| Reasons to extend therapy from 2 wks to 4 wks or more | <.01 | |||||||
| Positive BC 48–72 h after start antibiotics | 1248 (84.0) | 516 (89.7) | 363 (83.3) | 169 (76.1) | 132 (86.8) | 54 (65.9) | 14 (77.8) | |
| Immunocompromised status | 873 (59.5) | 411 (72.0) | 187 (43.2) | 134 (62.9) | 94 (63.1) | 36 (44.4) | 11 (55.0) | |
| Unknown portal of entry | 787 (54.3) | 386 (68.4) | 178 (41.7) | 102 (48.1) | 87 (58.0) | 28 (36.4) | 6 (31.6) | |
| Fever 72 h after first positive BC | 746 (51.6) | 331 (58.8) | 203 (47.3) | 94 (45.4) | 78 (52.0) | 32 (41.0) | 8 (40.0) | |
| Community acquisition | 460 (31.3) | 231 (40.7) | 96 (22.1) | 44 (20.4) | 71 (46.4) | 14 (17.9) | 4 (21.1) | |
| Oral switch therapy in different infection focia | ||||||||
| Skin/soft tissue infection | 922 (59.3) | 330 (56.2) | 305 (66.3) | 140 (58.8) | 85 (54.5) | 47 (54.0) | 15 (55.6) | .01 |
| Osteomyelitis | 840 (54.0) | 260 (44.3) | 301 (65.4) | 106 (44.5) | 105 (67.3) | 60 (69.0) | 8 (29.6) | <.01 |
| Spondylodiscitis | 659 (42.4) | 155 (26.4) | 275 (59.8) | 84 (35.3) | 94 (60.3) | 46 (52.9) | 5 (18.5) | <.01 |
| Prosthetic joint septic arthritis | 610 (39.2) | 144 (24.5) | 260 (56.5) | 75 (31.5) | 89 (57.1) | 37 (42.5) | 5 (18.5) | <.01 |
| Prosthetic valve endocarditis | 130 (8.4) | 24 (4.1) | 54 (11.7) | 14 (5.9) | 32 (20.5) | 5 (5.7) | 1 (3.7) | <.01 |
| All of the above | 120 (7.7) | 22 (3.7) | 44 (9.6) | 30 (12.6) | 16 (10.3) | 5 (5.7) | 3 (11.1) | <.01 |
| 18F-FDG PET/CT for SAB | ||||||||
| PET/CT readily available | 829 (55.6) | 278 (48.9) | 341 (78.0) | 106 (46.7) | 78 (51.3) | 24 (28.6) | 2 (9.1) | <.01 |
| PET/CT for SAB covered by insurance | 610 (42.2) | 177 (33.2) | 332 (77.4) | 41 (18.2) | 37 (24.3) | 21 (25.0) | 2 (9.1) | <.01 |
| PET/CT use in some/all SAB patients | 1009 (67.8) | 293 (51.1) | 409 (93.8) | 125 (57.0) | 124 (82.8) | 50 (61.0) | 4 (12.7) | <.01 |
| Never use PET/CT in SAB patients | 479 (32.2) | 281 (49.0) | 27 (6.2) | 96 (43.0) | 26 (17.2) | 32 (39.0) | 17 (77.3) | <.01 |
| Available, but never use PET/CT in SABb | 101 (12.2) | 63 (22.7) | 5 (1.5) | 24 (23.1) | 8 (10.3) | 1 (4.2) | <.01 | |
| Actions following diagnosis of persistent SAB | ||||||||
| Additional diagnostic testing | 1610 (95.8) | 605 (97.0) | 484 (96.6) | 242 (92.0) | 161 (97.6) | 91 (96.8) | 27 (79.4) | <.01 |
| Change antibiotic management | 1246 (74.7) | 523 (83.5) | 316 (63.8) | 207 (80.2) | 106 (64.6) | 72 (76.7) | 22 (68.8) | <.01 |
N = the total number of respondents of the survey. Values are counts (%). Not all respondents answered every question, therefore the percentages represent the percentage of the total respondents of the continent who answered this question.
Abbreviations: ASP, anti-staphylococcal penicillin; BC, blood culture; h, hours; MSSAB, methicillin-susceptible Staphylococcus aureus bacteremia; PET/CT, positron emission tomography/computed tomography; SAB, Staphylococcus aureus bacteremia.
Results are still significant with P < .01 when continents with n ≤ 5 were excluded from analysis.
This represents the number and percent of respondents that indicated that PET/CT is readily available but still never use PET/CT in SAB patients.
Treatment Duration
The most commonly identified durations of therapy between geographic regions for SAB-associated syndromes were similar. Thus, the majority of respondents from each continent indicated the same duration of treatment for native valve endocarditis (6 weeks), septic arthritis (4 weeks), and spondylodiscitis (6 weeks). Despite these similarities in practice amongst the majority of practitioners across geographic regions, substantial “within-region” variation existed for these syndromes. For each infectious complication of SAB, individual respondents within each continent indicated longer and shorter durations of therapy (Figure 4).
Treatment duration for S. aureus bacteremia. Percentage (count) of total respondents of the question per continent. P < .01 for comparison between continents for all categories (χ2 test).
The finding of blood cultures positive for S. aureus after 48–72 hours of appropriate therapy was identified as the most important reason to extend therapy duration in SAB patients beyond 2 weeks in all continents (range: 66% in South America to 90% in North America). Immunocompromised status was identified as an indication to extend antibiotic treatment beyond 2 weeks for most North American physicians (72%) but less than half (43%) of European physicians. By contrast, community acquisition of SAB was considered a reason to extend antibiotic treatment in only 20%–41% of physicians (Table 1; P < .01 for all above mentioned comparisons between continents).
18F-FDG PET/CT Scan use
The availability, insurance coverage, and use of 18F-FDG PET/CT scans differed significantly between geographic regions (Table 1). All were highest in Europe and lowest in Africa. The direct availability of 18F-FDG PET/CT scans for SAB patients ranged from 9% in Africa and 29% in South America, to 78% in Europe. 18F-FDG PET/CT scans were used for SAB patients by 94% of European, 83% of Oceanian, 61% of South American, 57% of Asian, and 51% of North American physicians (P < .01 for both above mentioned comparisons between continents). Survey respondents indicating that they ordered 18F-FDG PET/CT scans in patients with SAB were asked to specify for which indications they did so. Globally, the most important and most agreed upon indication for 18F-FDG PET/CT scan in SAB was persistent bacteremia: 62%–70% of physicians in every continent ordered 18F-FDG PET/CT scans for this indication (Supplementary Table 2; P = .66).
Persistent S. aureus Bacteremia
The clinical definition of persistent SAB varied widely between continents. The most frequent definition of persistent SAB was a duration of at least 3-4 days of positive blood cultures despite appropriate treatment, identified by >33% of physicians in every continent. However, in Europe (31%) and South America (24%), a significant minority of survey respondents indicated that persistent SAB was present after only 2 or more days of positive blood cultures. By contrast, 38% of Asian physicians indicated that seven or more days of positive blood cultures were required to constitute persistent SAB (Figure 3B). Almost all physicians indicated that they would order additional diagnostic testing in the setting of persistent SAB (79% in Africa, > 90% in all other continents), and a majority of physicians would also change their medical management (range 64% in Europe to 84% in North America; P < .01 for all above mentioned comparisons between continents) (Table 1).
DISCUSSION
This study shows that even the most basic aspects of treating patients with SAB differ profoundly between geographic regions. This variation was most marked in fundamental aspects of decision-making for SAB treatment, including antibiotic choice for MSSA bacteremia, addition of rifampin for prosthetic device infections, and route of administration. An anti-staphylococcal penicillin was treatment of choice for MSSA bacteremia in Europe and Oceania but a distant second to cefazolin in North America. The evidence for superiority of either of the 2 is limited to cohort studies with conflicting results and with underrepresentation of complicated disease [7, 8], emphasizing the need for randomized trials.
The role of adjunctive rifampin in patients with prosthetic material infections also differed by continent. This controversy persists despite the availability of published society guidelines that recommend the use of rifampin in S. aureus infections involving prosthetic valves and arthroplasties [9, 10]. However, the recommendation to use rifampin for prosthetic valve infective endocarditis has a very limited evidence base [11]. Thus, well-designed randomized trials are needed to define any potential role of adjunctive rifampin in prosthesis-associated SAB. Importantly, the wide range of practices regarding the use of rifampin in this survey demonstrates the presence of the global equipoise necessary to ethically conduct such a trial.
The practice of prescribing part of the treatment course for SAB with oral antibiotics was well accepted in all continents except North America, where only a minority of physicians would consider its routine use. This infrequent use of oral therapy in the United States may be due in part to a high prevalence of MRSA, the presence of a well-organized outpatient parenteral antimicrobial therapy system, or concerns related to medical malpractice. This lack of global consensus on the role of oral switch therapy is also reflected by the lack of consensus on which setting in which it should be considered. In fact, only the criteria of “source control” and “absence of central nervous system infection” were considered essential for oral switch by a clear majority. By contrast, all other listed criteria were regarded as essential by approximately half of the respondents—which implies that these were considered non-essential by the other half. Because oral switch therapy has potential to decrease the number of adverse drug events, catheter-associated problems and costs, and the fact that the survey respondents are in equipoise on the question, the need for a well-designed randomized trial seems clear. Current studies such as SABATO and SNAP might provide answers in the future [12, 13].
Broader global consensus existed for treatment duration of SAB. The worldwide similarity of respondents’ views on treatment duration for SAB is noteworthy given that the data for this aspect of SAB is at least as limited as that for the treatment-related aspects outlined above for which there is significant controversy [14, 15]. Although the majority of surveyed physicians throughout the world treat complications of SAB such as endocarditis and osteoarticular infection for a similar duration, a portion of physicians in each continent will treat substantially longer or shorter. This finding suggests that factors influencing treatment duration decision-making may be provider-based and situational rather than simply geographical in nature.
18F-FDG PET/CT use
Our findings also indicate significant geographic variability in the use of 18F-FDG PET/CT as diagnostic tool in SAB, with broad use in Europe and Oceania being balanced by infrequency in other continents. Observational studies have reported that 18F-FDG PET/CT may impact management and reduce mortality in patients with SAB because of higher detection of metastatic foci [16, 17], although the reduced mortality may have been confounded by immortal time bias related to including patients dying before undergoing 18F-FDG PET/CT [18]. Obviously, the associated costs could be a reason to refrain from using 18F-FDG PET/CT in low- and middle-income regions, but this does not explain its highly variable use in high-income regions. Recently, a call to action was published in the United States, advocating for insurance coverage of 18F-FDG PET/CT use in SAB patients [19]. In order to reach that goal, high-quality studies including randomized trials of 18F-FDG PET/CT are warranted.
Definition of Persistent SAB
The results of our survey suggest that the identification of persistent SAB may be therapeutically important, as it triggers additional diagnostic testing and changes in medical management for the majority of respondents. However, respondents generally disagreed on how to define it. Although 3–4 days was the most common identified definition of persistent SAB overall, all options in the range of 2–7 or more days were selected by respondents from each continent. Roughly one third of European respondents defined persistent SAB as only 2 days of bacteremia, although a similar portion of Asian respondents indicated that it occurred after 7 or more days. The prognostic significance of persistent SAB has been previously demonstrated [20–22]. Identifying a broadly accepted definition of persistent SAB would thus be helpful to optimize clinical decision-making, as well as to harmonize the terminology used in clinical research.
Perspective
The current study suggests that there is no global standard of practice for SAB. Striking differences were noted, both between and within continents, in what antibiotics were prescribed, and by what route. The lack of a global standard in the management of SAB stands in stark contrast to treatment of other syndromes of comparable lethality. For coronary artery disease, management has been largely standardized by guidelines based on data from randomized controlled trials [23–25]. Coincident with establishing these best treatment practices, the annual US mortality rate from coronary artery disease declined by 17.7% from 2005 to 2015 [26]. By contrast, the 1-month mortality for patients with SAB only decreased by 2.8% over the same time period [27]. The results of this study underscore one key fact: a global standard of care for SAB will be difficult to develop pending more definitive clinical trials data. Indeed, fewer than 3500 patients have been enrolled in published SAB randomized trials over the past 20 years (Supplementary Table 3). Factors other than robust clinical data, such as cultural differences, costs and availability of resources also influence management choices. However, without consensus on best practice, normative and cultural factors gain influence on for example antibiotic prescription behavior [28]. Multinational clinical trials such as the Staphylococcus aureus Network Adaptive Platform (SNAP) [13] are thus essential to standardize clinical definitions, identify treatment strategies, and improve patient outcomes of this common and frequently lethal infection.
Strengths and Limitations
The current study illustrates the potential of using social media to understand global treatment practices and decision making. Although previous studies on physicians’ management of SAB have been conducted [5, 29, 30], none were as extensive and on a global scale as this current study. Our study has several limitations. There were relatively low participation rates from South America and Africa. The respondents were not questioned about their local guideline and adherence to it, and for many countries no national guidelines were available. This made it impossible for us to consider the role of national guidelines in the present study. Given the fact that 71 countries were included in the survey, comparing differences between each of these countries was methodologically infeasible. Therefore, we limited the analyses to continents. We were unable to evaluate spatial clustering of infections. The survey was only available in English, which might have dissuaded non-English speaking physicians. Because the survey was distributed through listservs and social media, the exact number of recipients or proportion of physicians per country is unknown. Therefore, the true response rate is uncertain and could only be estimated by the ratio of the reported surveys and the number of times the URL link was opened. Finally, because it was impossible to control who filled out the survey, respondents theoretically could have been non-physicians or could have completed the survey multiple times. However, because there was no incentive in responding, and we did not discover any nonsensical answers, this seems unlikely. Overall, the advantage of receiving feedback from over 2000 specialists from all over the world outweighs the potential disadvantages of the use of social media platforms.
CONCLUSION
Large practice variations for SAB exist throughout the world, reflecting the absence of an international standard of care for the management of patients with SAB. This article sets the stage and the agenda for multinational or global clinical trials and networks, to address the unresolved aspects of this devastating disease.
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
Author contributions. A. W. and V. F. led the study. D. B., K. S., and M. B. developed the European survey that was the basis for the current study. A. W., V. F., M. B., E. S., M. L., and S. T. modified and designed the final survey. All authors contributed to the study design and the survey distribution. A. W. and V. F. did the analysis and wrote the first draft of the manuscript. All authors revised the manuscript and gave final approval of the manuscript.
Acknowledgments. The authors thank #IDTwitter for their invaluable help in distributing the survey. They also thank all other global respondents and colleagues who helped in distributing and completing the survey.
Data sharing statement. Requests for access to the data generated during this study will be considered on a case-by-case basis by the authors.
Financial support. Work contained in this manuscript was made possible by grant number 1R01-AI165671 (V. G. F.) from the National Institutes of Health (NIH).
References
Author notes
Potential conflicts of interest. V. G. F. reports personal fees from Novartis, Debiopharm, Genentech, Achaogen, Affinium, Medicines Co., MedImmune, Bayer, Basilea, Affinergy, Janssen, Contrafect, Regeneron, Destiny, Amphliphi Biosciences, Integrated Biotherapeutics; C3J, Armata, Valanbio, Akagera, Aridis, Roche; grants from NIH, MedImmune, Allergan, Pfizer, Advanced Liquid Logics, Theravance, Novartis, Merck; Medical Biosurfaces, Locus, Affinergy, Contrafect, Karius, Genentech, Regeneron, Deep Blue, Basilea, Janssen; royalties from UpToDate; stock options from Valanbio and ArcBio, honoraria from Infectious Diseases of America for his service as Associate Editor of Clinical Infectious Diseases, and a patent sepsis diagnostics pending. All other authors report no potential conflicts. 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.
