Introduction
Persistent bacteremia (PB) is defined as the continued or recurrent presence of pathogens in the bloodstream despite appropriate antimicrobial therapy. It is closely associated with increased organ dysfunction, prolonged hospitalization, and significantly elevated mortality, contributing to greater clinical severity and poor prognosis.1–3
Staphylococcus aureus remains the most common pathogen responsible for cases of PB, accounting for 8–39% of all S. aureus bacteremia (SAB) episodes.4 Its virulence factors—including hemolysins and coagulases—facilitate immune evasion and tissue invasion and promote the development of persister phenotypes following infection.5 In addition, S. aureus can form unique biofilms—structured microbial communities encased in extracellular polymeric substances. Biofilms promote bacterial adherence to host tissues and medical devices, impede antibiotic penetration, and significantly reduce clearance by host phagocytes. Within biofilms, S. aureus can enter a metabolically dormant state as persister cells, further enhancing antimicrobial tolerance.6 These complex mechanisms provide a biologically plausible and pathophysiological basis for persistent SAB (pSAB) and underscore the importance of early identification and management.
However, notably there is currently no globally unified standard for the diagnosis and management of pSAB.7 Over the past two decades, the diagnostic criteria for PB have evolved.8 Early definitions, such as that proposed by Fowler et al in the early 2000s, required the presence of bacteremia to persist for at least seven days.9 However, more recent studies have demonstrated that a prolonged duration of SAB is significantly associated with increased mortality and a higher risk of metastatic infections.2 In 2020, Kuehl et al published a multinational cohort study involving 1,588 SAB patients and reported that when the duration of bacteremia increased from 1 day to 2–4 days, the 90-day mortality rate nearly doubled from 22% to 39%, respectively. On the basis of these findings, the authors suggested redefining the cutoff duration for PB as 2 days or more despite active antibiotic therapy.4
Although previous studies have described the clinical characteristics of pSAB and suggested potential predictive factors—such as infective endocarditis, metastatic infection, cardiovascular implantable electronic devices (CIEDs), and delayed source control5—the evolving diagnostic thresholds may alter the interpretation and validity of these predictors, warranting reevaluation.
Moreover, the incidence of pSAB varies across regions and time periods. A 2012 study from South Korea reported that among 79 patients with methicillin-resistant Staphylococcus aureus (MRSA) bacteremia, 31 (39.2%) remained bacteremic for more than seven days.10 In contrast, a 2023 Korean study analyzing 1,300 patients with SAB revealed that only 17% had positive blood cultures persisting beyond two days.11 This shift may reflect changes in diagnostic criteria and advances in treatment strategies. Furthermore, two studies from Taiwan, China, reported lower rates of pSAB than those reported in South Korea,12,13 suggesting notable differences in incidence across countries. However, epidemiological data on pSAB in mainland China remain limited.
To address this knowledge gap, we conducted a retrospective analysis to quantify the incidence of pSAB in a tertiary-care Chinese hospital from 2020 to 2024, and to identify clinical predictors through comparative and multivariate analyses.
Materials and Methods
Research Design and Population
This retrospective, case–control study was conducted at the Second Affiliated Hospital of Zhejiang University School of Medicine, a tertiary-care center in Zhejiang Province, China, with more than 2000 beds. The hospital microbiological database was screened to identify patients who had positive blood cultures for S. aureus. Based on prior studies and with appropriate adjustments for our patient population and research aims,14–16 patients were included in the study if they met all of the following criteria: 1) Hospitalized and aged ≥18 years; 2) at least one positive blood culture yielding S. aureus; 3) clinical signs or symptoms consistent with SAB; 4) received appropriate antibiotic therapy for S. aureus after the index blood culture; and 5) blood cultures performed between 48 hours and 7 days after the initiation of appropriate antibiotic therapy. Patients were excluded if they met any of the following criteria: 1) Polymicrobial bacteremia (ie, S. aureus isolated together with other pathogens in the same episode); 2) absence of clinical manifestations consistent with bacteremia (suggesting contamination rather than true infection); 3) no follow-up blood cultures performed between 48 hours and 7 days after the initiation of appropriate antibiotic treatment; and 4) did not receive appropriate antimicrobial therapy or died or were lost to follow-up within 48 hours of treatment initiation. All eligible patients were categorized into two groups: pSAB and nonpersistent SAB (non-pSAB). pSAB was defined as at least one follow-up blood culture positive for S. aureus obtained between 48 hours and 7 days after the initiation of appropriate antimicrobial therapy.4,14 Non-pSAB was defined as all follow-up blood cultures obtained between 48 hours and 7 days after the initiation of appropriate antimicrobial therapy being negative for S. aureus. As a retrospective case-control study, controls were randomly selected from the non-pSAB group at a 1:1 ratio to the number of pSAB cases.
Ethical Approval
This retrospective case–control study was approved by the Institutional Review Board of the Second Affiliated Hospital of Zhejiang University School of Medicine (Reference No. 2025–0034). The requirement for informed consent was waived because the study involved only a retrospective review of existing medical records and laboratory data, posed no potential risk to patients, and used de-identified information. All patient data were fully anonymized prior to analysis to ensure confidentiality. The study was conducted in accordance with the ethical standards of the institutional research committee and the principles of the Declaration of Helsinki (as revised in 2013).
Definitions and Data Collection
SAB is defined as patients fulfilling the diagnostic criteria for bloodstream infection, with blood cultures positive for S. aureus.17 The duration of bacteremia was calculated from the day the first positive blood culture was collected to the day of the last positive follow-up blood culture.11 Relapse was defined as the reoccurrence of SAB after two consecutive days of negative blood cultures.18 Appropriate antibiotic therapy was defined as the administration of at least one intravenous agent within 24 hours after the index blood culture, which was subsequently confirmed to be active against the isolated S. aureus strain in vitro and recommended by current clinical guidelines.19,20 Metastatic infection was defined as infection at a distant anatomical site unrelated to the primary source.21 Sources of bacteremia were classified as either hospital-acquired or community-acquired; community-acquired SAB was diagnosed if positive blood cultures were obtained within 48 hours of admission, whereas hospital-acquired SAB was diagnosed when cultures became positive after 48 hours of hospitalization.22 The primary outcome was 30-day mortality.
Clinical data were extracted from electronic medical records. The variables analyzed included demographics (age, sex, comorbidities) and infection severity scores.16 (SOFA score, Pitt score), clinical manifestations (onset of bacteremia, metastatic lesions, fever duration, laboratory parameters, duration of bacteremia), treatments (surgical intervention, antibiotic regimens), and outcomes (hospital stay, 3-month relapse, and 30-day mortality).
Statistical Analysis
Categorical variables are expressed as counts and percentages, and continuous variables are expressed as the means with standard deviations or medians with interquartile ranges (IQRs), depending on the distribution. For univariate analysis, categorical variables were compared using the chi-square test or Fisher’s exact test, and continuous variables were analyzed using Student’s t test or the Mann–Whitney U-test, as appropriate. For multivariate analysis, logistic regression models were used to identify independent predictors, and p values were calculated using the Wald test. A two-tailed p value <0.05 was considered to indicate statistical significance. Statistical analyses were performed using IBM SPSS Statistics for Windows, version 25 (IBM Corp., Armonk, NY, USA).
Results
Epidemiology and Demographics
Over a five-year period, 499 patients with blood culture-confirmed SAB were identified, including 48 patients (9.6%) with PB. The lowest incidence of pSAB was recorded in 2022 (4/80, 4.8%), while the highest was observed in 2024 (17/105, 13.9%). A gradual increase in incidence was noted over the past four years (Figure 1A).
|
Figure 1 (A) Proportion of pSAB among all SAB cases over the past five years. (B) Composition of pSAB by onset of bacteremia. Abbreviations: SA, Staphylococcus aureus; SAB, Staphylococcus aureus bacteremia; pSAB, persistent Staphylococcus aureus bacteremia.
|
Among the 48 patients included in the pSAB group, 26 were male (54.2%), with a median age of 61.5 years (IQR 54.0–71.0). These demographic characteristics were not significantly different from those of the non-pSAB group. However, the Charlson comorbidity index (CCI) was significantly greater in the pSAB group (median 4.0 [IQR 2.0–7.0] vs 3.0 [IQR 2.0–4.5], p = 0.030). Although the proportion of patients with a CCI ≥3 did not differ significantly between the two groups (70.8% vs 58.3%, OR=1.74; p=0.200), a significantly greater proportion of patients with a CCI ≥7 were in the pSAB group (27.1% vs 10.4%; OR = 3.19; p = 0.036). These findings indicate that patients with a greater burden of comorbidities were more prevalent in the pSAB group, suggesting that severe comorbidity may be a potential predictor of the development of pSAB (Table 1).
 |
Table 1 Baseline Characteristics and Clinical Features in Patients with pSAB vs Non-pSAB
|
Clinical Characteristics of pSAB
In the pSAB group, 18 patients (37.5%) were infected with MRSA, and 16 patients (33.3%) had community-acquired infections. In comparison, the corresponding proportions in the non-pSAB group were 47.9% and 31.3%, respectively. No significant differences were found between the pSAB and non-pSAB groups in these aspects.
With respect to the onset of bacteremia, the most common primary site of infection in the pSAB group was infective endocarditis (IE) (13 cases, 27.1%), followed by skin and soft tissue infections (10 cases, 20.8%), vascular and device-related infections (9 cases, 18.8%), bone and joint infections (9 cases, 18.8%), and pulmonary infections (6 cases, 12.5%). In one case (2.1%), the primary source of infection was unknown (Figure 1B). With the exception of IE, which was significantly more common in the pSAB group (27.1% vs 4.2%, OR = 8.54; p = 0.002), there were no statistically significant differences between the two groups in terms of other primary infection sites.
With respect to clinical manifestations, patients in the pSAB group had a higher incidence of metastatic infections (22.9% vs 4.2%, OR=6.84; p=0.007), and a greater proportion had experienced persistent fever for ≥7 days prior to admission (31.3% vs 12.5%; OR=3.18; p=0.026). However, there were no significant differences between the pSAB and non-pSAB groups in terms of the levels of markers of infection severity, such as white blood cell count, C-reactive protein level, procalcitonin level, interleukin-6 level, platelet count, D-dimer level, blood lactate level, SOFA score, and Pitt score.
In terms of management, surgical intervention was more common in the pSAB group (37.5% vs 18.8%, OR=2.06; p=0.041). Combination therapy with two anti-MRSA agents was also more frequently utilized in the pSAB group (22.9% vs 0.00%, p<0.001), whereas β-lactam monotherapy was less common (14.6% vs 45.8%, OR=0.20, p=0.001). Other antibiotic agents, including vancomycin, daptomycin, and linezolid, as well as combination therapy with these agents and β-lactams, did not significantly differ between the two groups.
The clinical outcomes for pSAB patients were notably worse. The proportion of patients whose length of hospitalization was ≥7 days after SAB was significantly greater in the pSAB group than in the non-pSAB group (77.1% vs 45.8%, OR = 3.98; p = 0.002). Although the overall duration of postbacteremia hospital stay was not significantly different between the two groups (median 22.5 [IQR 16.5–40.8] vs 18.0 [IQR 11.8–33.3], p=0.140), patients in the pSAB group tended to have longer hospitalizations. Furthermore, the duration of fever was significantly longer in the pSAB group (median 9.5 [IQR 7.0–15.0] vs 5.0 [IQR 1.0–12.0], p=0.014). In addition, the pSAB group had higher 3-month relapse rates (20.8% vs 0.0%, p=0.001) and higher 30-day mortality rates (16.7% vs 2.1%, OR=9.40, p=0.036) (Figure 2), indicating significantly worse clinical outcomes in patients with pSAB.
 |
Figure 2 Thirty-day Kaplan-Meier survival curves for patients with pSAB. Abbreviation: pSAB, persistent Staphylococcus aureus bacteremia.
|
Clinical Predictors of pSAB
Multivariate analysis revealed a Charlson comorbidity index ≥7 (OR=4.09; 95% CI: 1.19–14.00; p=0.025), infective endocarditis (OR=8.66; 95% CI: 1.52–49.03; p=0.015), and fever duration ≥7 days before admission (OR=5.06; 95% CI: 1.61–15.90; p=0.006) as clinical features that can predict pSAB presence (Table 2).
 |
Table 2 Multivariate Analysis of Predictors for pSAB
|
Discussion
The incidence of pSAB has recently increased,20 a trend that was also observed in our study. As a highly virulent gram-positive pathogen, S. aureus accounts for 30%–40% of PB cases.23 Despite active treatment, pSAB is significantly associated with poor clinical outcomes.2 Similarly, in our study, patients in the pSAB group experienced a prolonged duration of fever and an extended hospital stay, and both the 30-day mortality rate and the 3-month recurrence rate were higher, highlighting the significantly worse clinical outcomes in patients with pSAB.
We identified infective endocarditis, fever duration exceeding 7 days before admission, and a CCI ≥7 as clinical features that can predict the presence of pSAB, which is partially consistent with the findings of previous literature. For example, Fowler et al established infective endocarditis as a critical predictor of pSAB.24 The CCI reflects the burden of comorbidities in patients. Previous reports have identified a CCI ≥6 as an independent risk factor for 30-day mortality in pSAB patients.11 In our study, the proportion of patients with a CCI ≥7 was significantly greater in the pSAB group than in the non-pSAB group, suggesting that a high comorbidity burden may have predictive value for the development of pSAB. This may be related to impaired immune function and reduced infection control capacity in these patients. Notably, we identified prolonged fever prior to hospitalization as a novel predictive factor, suggesting that it may serve as a marker of refractory infections or delayed therapeutic intervention. Clinically, close attention should be given to patients with pSAB who present with prolonged fever at admission, with early evaluation and source control prioritized.
We observed a greater proportion of surgical interventions and increased use of combination therapy with two anti-MRSA antibiotics (such as vancomycin combined with linezolid or tigecycline) in the pSAB group, reflecting greater complexity in infection control. Nevertheless, patients with pSAB experienced longer fever durations and significantly higher rates of recurrence and short-term mortality, which is consistent with the findings of previous studies.25,26 For example, Chang et al reported that even under optimized treatment conditions, patients with pSAB faced an elevated risk of treatment failure because of difficulty in lesion clearance or biofilm formation.27 Our findings support this perspective and suggest that intensified antibacterial therapy alone may not compensate for inadequate source control. Therefore, in pSAB management, particular emphasis should be placed on early and effective lesion control, especially in patients with implant-related infections, abscesses, or endocarditis.
Unexpectedly, no significant differences were observed between the pSAB and non-pSAB groups regarding early inflammatory indicators (such as white blood cell count and procalcitonin) or severity scores (SOFA and Pitt scores) within 48 hours of bacteremia onset. These findings contrast with those of studies by Giannella et al, who reported stronger early inflammatory responses in patients with pSAB.28 Our results suggest that the early inflammatory response to bacteremia may be similar regardless of subsequent persistence. Additionally, preadmission antibiotic use in some patients may have masked the true intensity of early infection responses.
This study had several limitations. First, it was a single-center, retrospective study with a relatively small sample size, potentially introducing selection bias. Second, variability in the timing and frequency of follow-up blood cultures during bacteremia may have influenced pSAB recognition. Furthermore, antibiotic regimens were not fully standardized, and individualized treatment decisions may have affected outcomes. Nonetheless, our study provides valuable clinical evidence for the identification and management of pSAB, particularly highlighting the importance of early detection and source control in high-risk populations.
Conclusion
This study revealed a rising incidence of pSAB in a tertiary-care hospital in China, with significantly higher 30-day mortality and 3-month relapse rates in patients with pSAB than in those without pSAB. We identified infective endocarditis, a CCI ≥7, and fever lasting ≥7 days before admission as independent clinical predictors of pSAB. Notably, our findings revealed prolonged fever before admission as a previously overlooked but significant predictor of pSAB. These findings underscore the importance of early identification and prompt source control, particularly in patients with complex comorbidities or prolonged febrile symptoms. The implementation of standardized screening protocols may help clinicians detect pSAB earlier and improve patient outcomes. Future multicenter and prospective studies are warranted to validate these findings and refine clinical decision-making strategies.
Acknowledgments
The authors thank Mr. Jinliang Liu for his expert guidance on statistical methods.
Disclosure
The authors report no conflicts of interest in this work.
References
1. Kitaya S, Kanamori H, Ishii R, et al. Clinical characteristics and outcomes of persistent bacteremia in patients with head and neck cancer in a tertiary care hospital. Front Med. 2024;11:1406983. doi:10.3389/fmed.2024.1406983
2. Sathya Kumar AM, George MM, Bhanuprasad K, et al. Persistent bacteremia predicts poor outcomes among neutropenic patients with carbapenem-resistant gram-negative bloodstream infections receiving appropriate therapy. Ann Clin Microbiol Antimicrob. 2023;22(1):12. doi:10.1186/s12941-023-00561-7
3. Guimaraes AO, Cao Y, Hong K, et al. A prognostic model of persistent bacteremia and mortality in complicated staphylococcus aureus bloodstream infection. Clin Infect Dis. 2019;68(9):1502–1511. doi:10.1093/cid/ciy739
4. Kuehl R, Morata L, Boeing C, et al. Defining persistent Staphylococcus aureus bacteraemia: secondary analysis of a prospective cohort study. Lancet Infect Dis. 2020;20(12):1409–1417. doi:10.1016/S1473-3099(20)30447-3
5. Parsons JB, Westgeest AC, Conlon BP, Fowler VG. Persistent methicillin-resistant staphylococcus aureus bacteremia: host, pathogen, and treatment. Antibiotics. 2023;12(3):455. doi:10.3390/antibiotics12030455
6. Kaushik A, Kest H, Sood M, Steussy BW, Thieman C, Gupta S. Biofilm producing methicillin-resistant staphylococcus aureus (MRSA) infections in humans: clinical implications and management. Pathogens. 2024;13(1):76. doi:10.3390/pathogens13010076
7. Westgeest AC, Buis DTP, Sigaloff KCE, et al. Global differences in the management of staphylococcus aureus bacteremia: no international standard of care. Clin Infect Dis. 2023;77(8):1092–1101. doi:10.1093/cid/ciad363
8. Holland TL, Bayer AS, Fowler VG. Persistent methicilin-resistant staphylococcus aureus bacteremia: resetting the clock for optimal management. Clin Infect Dis. 2022;75(9):1668–1674. doi:10.1093/cid/ciac364
9. Fowler VG Jr, Sakoulas G, McIntyre LM, et al. Persistent bacteremia due to methicillin-resistant Staphylococcus aureus infection is associated with agr dysfunction and low-level in vitro resistance to thrombin-induced platelet microbicidal protein. J Infect Dis. 2004;190(6):1140–1149. doi:10.1086/423145
10. Ok HS, Lee HS, Park MJ, et al. Predictors and clinical outcomes of persistent methicillin-resistant Staphylococcus aureus bacteremia: a prospective observational study. Korean J Intern Med. 2013;28(6):678–686. doi:10.3904/kjim.2013.28.6.678
11. Kim SH, Jeon M, Jang S, Mun SJ. Factors for mortality in patients with persistent Staphylococcus aureus bacteremia: the importance of treatment response rather than bacteremia duration. J Microbiol Immunol Infect. 2023;56(5):1007–1015. doi:10.1016/j.jmii.2023.07.010
12. Hsu MS, Huang YT, Hsu HS, Liao CH. Sequential time to positivity of blood cultures can be a predictor of prognosis of patients with persistent Staphylococcus aureus bacteraemia. Clin Microbiol Infect. 2014;20(9):892–898. doi:10.1111/1469-0691.12608
13. Lin SH, Liao WH, Lai CC, et al. Risk factors for mortality in patients with persistent methicillin-resistant Staphylococcus aureus bacteraemia in a tertiary care hospital in Taiwan. J Antimicrob Chemother. 2010;65(8):1792–1798. doi:10.1093/jac/dkq188
14. Jung J, Song KH, Jun KI, et al. Predictive scoring models for persistent gram-negative bacteremia that reduce the need for follow-up blood cultures: a retrospective observational cohort study. BMC Infect Dis. 2020;20(1):680. doi:10.1186/s12879-020-05395-8
15. Bai AD, Daneman N, Brown KA, Boyd JG, Gill SS. Long-term morbidity and mortality of patients who survived past 30 days from bloodstream infection: a population-based retrospective cohort study. J Infect. 2024;89(5):106283. doi:10.1016/j.jinf.2024.106283
16. Kang CK, Song KH, Kim SE, et al. A prospective cohort study of durations of staphylococcus aureus bacteremia according to different phenotypes and a new concept of persistent bacteremia. Antimicrob Agents Chemother. 2019;64(1):e01656–19. doi:10.1128/AAC.01656-19.
17. Minter DJ, Appa A, Chambers HF, Doernberg SB. Contemporary management of staphylococcus aureus bacteremia-controversies in clinical practice. Clin Infect Dis. 2023;77(11):e57–e68. doi:10.1093/cid/ciad500
18. Thornley EM, Alexander R, Popelka J, Montgomery CP. Risk factors for persistent staphylococcus aureus bacteremia in children. Pediatr Infect Dis J. 2024;43(11):1040–1045. doi:10.1097/INF.0000000000004439
19. Bussini L, Rosselli Del Turco E, Pasquini Z, et al. Risk factors for persistent enterococcal bacteraemia: a multicentre retrospective study. J Glob Antimicrob Resist. 2022;29:386–389. doi:10.1016/j.jgar.2022.05.003
20. Yang E, Cho YG, Kim E, et al. Clinical and microbiological characteristics of persistent Staphylococcus aureus bacteremia, risk factors for mortality, and the role of CD4(+) T cells. Sci Rep. 2024;14(1):15472. doi:10.1038/s41598-024-66520-0
21. Peri AM, Rafiei N, O’Callaghan K, et al. Host response signature trends in persistent bacteraemia and metastatic infection due to Staphylococcus aureus and Gram-negative bacilli: a prospective multicentre observational study. Infect Dis. 2024;56(4):268–276. doi:10.1080/23744235.2023.2294122
22. Chung H, Kim E, Yang E, et al. C-reactive protein predicts persistent bacteremia caused by community-acquired methicillin-resistant Staphylococcus aureus strain. Eur J Clin Microbiol Infect Dis. 2021;40(12):2497–2504. doi:10.1007/s10096-021-04303-5
23. Tong SYC, Fowler VG Jr, Skalla L, Holland TL. Management of Staphylococcus aureus Bacteremia. JAMA. 2025;334:798. doi:10.1001/jama.2025.4288
24. Fowler VG Jr, Miro JM, Hoen B, et al. Staphylococcus aureus endocarditis: a consequence of medical progress. JAMA. 2005;293(24):3012–3021. doi:10.1001/jama.293.24.3012
25. Choi SH, Dagher M, Ruffin F, et al. Risk factors for recurrent staphylococcus aureus bacteremia. Clin Infect Dis. 2021;72(11):1891–1899. doi:10.1093/cid/ciaa801
26. Russell CD, Berry K, Cooper G, et al. Distinct clinical endpoints of staphylococcus aureus bacteraemia complicate assessment of outcome. Clin Infect Dis. 2024;79(3):604–611. doi:10.1093/cid/ciae281
27. Chang TH, Tang HJ, Chen CC, Chen CJ. Clinical characteristics and genomic changes of recurrent methicillin-resistant staphylococcus aureus bacteremia. J Microbiol Immunol Infect. 2025;58(2):251–257. doi:10.1016/j.jmii.2024.11.008
28. Giannella M, Pascale R, Pancaldi L, et al. Follow-up blood cultures are associated with improved outcome of patients with gram-negative bloodstream infections: retrospective observational cohort study. Clin Microbiol Infect. 2020;26(7):897–903. doi:10.1016/j.cmi.2020.01.023