We highlighted the clinical impact of higher adherence to SAB-QIs on better prognosis in patients with SAB, which can be summarized as follows. First, the higher the SAB-QI score, the better the patient outcome. The 30-day prognosis of patients with higher QI scores (9–10) and the highest QI scores (11–13) was clearly better than that of patients with lower QI scores. Second, the SAB-QI scores varied greatly among the enrolled patients, ranging from 3 to 13, suggesting that the management of patients with SAB differs between institutions and physicians. Adherence to the QIs was worse for echocardiography (21.1%), antibiotic treatment (23.2%), and other management (17.0%). Third, the demographic data obtained in this study will greatly help understand the clinical features of SAB as an intractable disease.

In our patient cohort, SAB accounted for 8.7% of all cases of bacteremia (387/4,448), which may be lower than those previously reported; for example, 16.9% of healthcare-onset and 14.9% of community-onset cases in Canada [24], and 27.1% of nosocomial-onset cases in Japan [25]. MRSA infections were detected in 33.9% of patients in the present study, which is comparable to the findings of previous studies reported in Japan (32.8–44.9%) [9, 14, 26]. As reported previously [27], this is corroborated by data from a Japanese national database indicating MRSA isolation rates of 30.7–67.3% in recent years [28]. Potential infectious foci of SAB were reported to be bone and joint (2.4–24.2%), skin and soft tissue (12.5–17.0%), CRBSI (11.6–12.3%), IE (5.5–10.9%), respiratory tract (5.9–15.3%), surgical site (5.3%), and urinary tract infections (3.9%), and primary/unknown cases (19.1–42.7%) [7, 25, 26]. Of the 289 cases finally included, the common infectious foci were primary bacteremia (30.1%) and CRBSI (23.5%), indicating that CRBSI was relatively common in our patients.

The proportion of complicated SAB was high (85.5%) in the present study compared to previously reported data (305/530, 57.5%) [7]. Over one-third (34.3%) of the SAB cases were accompanied by disseminated lesions, especially those involving osteomyelitis and arthritis (49.5%). These results emphasize that systemic examination and management are indispensable for patients with SAB. The lower the SAB-QI score, the higher the primary bacteremia and fewer the disseminated lesions. This could be attributed to the absence of follow-up blood cultures and inadequate systemic assessment, resulting in an insufficient classification of uncomplicated and complicated cases.

The clinical utility of QI-oriented management in ensuring the quality of medical care for patients with SAB [10, 13,14,15,16,17,18,19]. A promising relationship between high adherence to QIs and favorable prognosis has been corroborated worldwide [14, 29, 30]. Our data underscore the clinical significance of adherence to recommended SAB-QIs in improving patient outcomes. In the present study, the median SAB-QI score was 9 points (maximum, 13 points). It is not feasible to compare our data with those of previous studies, and the adequacy of this compliance rate for maintaining medical safety in real-world settings remains unclear. However, SAB-QI scores can be applied to interhospital comparisons or longitudinal evaluations in hospitals.

In our study, follow-up blood cultures after initiating antimicrobial therapy were not performed in one-third of the SAB cases. A recent study reported that follow-up blood cultures were tested in only 18.8% of cases managed in Japanese emergency and critical care departments, indicating inadequate management of patients with SAB [26]. Repeated blood culture testing is indispensable to distinguish between complicated and uncomplicated SAB. Without proper diagnosis, it is impossible to establish an appropriate period of antibiotic treatment [10], possibly increasing the likelihood of therapeutic failure. The fact that most previous studies have included confirmation of negative blood culture as one of the recommended QIs suggests its clinical importance [14,15,16, 18, 19]. In our study, a criterion for defining uncomplicated SAB was negative blood culture results during follow-up. Potentially complicated SAB cases were classified as either uncomplicated or complicated SAB if follow-up blood culture testing had been performed. In this study, the prevalence rates were 14.5–47.4% for uncomplicated SAB and 52.6–85.5% for complicated SAB, which is equivalent to or surpassed complicated SAB prevalence rates reported previously (46.9%) [13].

Interestingly, our data demonstrated a relationship between hospital size and the SAB-QI score. The median SAB-QI score significantly increased as the hospital volume increased. However, paradoxically, the SAB-QI score of the largest hospital group (> 600 beds) was lower than that of the second-largest hospital group (401–600 beds). Generally, the larger the hospital, the better equipped the testing facilities and the more specialists are employed there. Thus, this result is informative as the clinical management of patients with SAB is not necessarily better in high-volume hospitals. We assume that it is difficult for specialized departments (such as Infectious Diseases [ID], Cardiology, and Cardiac Surgery) and laboratory divisions (such as Microbiology and Echocardiography) to fully collaborate in large hospitals. Studies on hospital volume and the quality of medical care have been conducted in various medical fields. For instance, the results of a retrospective multifacility cohort study indicated that the survival rate of patients with ovarian cancer may increase depending on hospital volume [31]. Admission to a high-volume hospital may be associated with lower mortality or higher quality of medical care, although some data suggest that large hospitals do not necessarily provide better medical care to every patient in proportion to hospital size [32, 33]. The diagnostic accuracy and treatment strategies at small-scale hospitals may be suboptimal mainly due to the unavailability of in-hospital facilities for blood culture and a lack of current medical knowledge. Considering that SAB is also a common disease in such small hospitals, it is imperative to evaluate and monitor the quality of care provided to patients with SAB there as well. To draw solid conclusions, the association between SAB management quality and hospital size should be further explored in future studies.

The mortality rate of SAB is approximately 20–30% in developed countries despite effective antibacterial therapy and source control [2]. Despite the implementation of active antimicrobial stewardship and infectious disease consultation (11–24%) [34] or evidence-based bundle intervention (17–22%) [13], favorable reductions in mortality have yet to be achieved in patients with SAB. The overall 30-day mortality rate in the present study was 18.0%, with 37 patients (71.2%) dying within 14 days after SAB diagnosis. Among the patients who survived for > 14 days, the 30-day mortality rate was 6.0%. Thus, the prognoses in our patient cohort were favorable as reported in a previous Japanese study (3.4–10.0%) [14]. The cause of the disparity in prognoses reported in Japan and those reported in other developed countries remains unclear. Variations in the prevalence of circulating pathogenic strains across countries may be a contributing factor [35]. Given that SAB may precipitate multiple complications that require prolonged treatment (> 30 days), the assessment of long-term prognosis is imperative. However, prognostic evaluation following inter-hospital transfer posed a challenge in this retrospective study. Older age, the presence of one or more comorbidities, and methicillin resistance are reported potential predictors for mortality in patients with SAB [2, 36]. Our study also corroborated that age, methicillin resistance, multiple comorbidities, and lower SAB-QI scores are associated with 30-day mortality in patients with SAB.

The landscape of therapeutic strategies for patients with SAB is evolving. In standard practice, patients with SAB undergo at least 2 weeks of intravenous antibiotic therapy, and the treatment period is extended to 4 to 6 weeks in complicated cases [5, 37]. The optimal duration remains a subject of debate and two non-inferiority RCTs are currently underway; the SAB7 trial aims to evaluate the efficacy of 7- and 14-day antibiotic treatment in low-risk patients [38], and the SAFE trial aims to compare 4- and 6-week intravenous antibiotic therapy in patients with complicated SAB including native valve infective endocarditis [39]. Another RCT has indicated that selected low-risk patients with uncomplicated SAB can be safely and effectively managed with early oral switch therapy [40]. Although increased evidence on a shorter antibiotic strategy is favorable, patients with the poor prognosticators warrant particular attention to improve their prognosis.

Our data clearly demonstrated that the SAB-QI score was significantly lower in fatal cases (median: 7 vs. 9, p < 0.01) (Fig. 3). Moreover, the lower the SAB-QI score, the lower the 30-day survival rate (Fig. 4), as previously reported [13, 14, 34]. Additionally, multivariate analysis indicated that lower SAB-QI scores are associated with poor prognoses in patients with SAB. Our SAB-QI was developed from the original 25 QIs [10], but its clinical validity remains unclear. However, based on our data, we suggest that the SAB-QI has the potential to play an important role in improving the prognosis of patients with SAB by ensuring the quality of medical care.

Here, we discuss ways to increase SAB-QI scores and enhance patient management. Similar to QI-oriented management, five core interventions for SAB have recently been proposed in a review article: (1) appropriate anti-staphylococcal therapy, (2) screening echocardiography, (3) assessment of metastatic phenomena and source control, (4) decision on antimicrobial therapy duration, and (5) ID consultation [41]. Consultation with ID physicians is quite difficult in Japan, where specialized doctors are not readily available [42]. Systematic education and training curricula for undergraduate students and young doctors are required to increase the number of ID physicians. Another approach, the development of antimicrobial stewardship teams or programs, and bundle management, has been reported to improve the prognosis of patients with SAB [13, 14, 34, 43]. In the absence of ID physicians, especially in small hospitals, such collaborative work may contribute to an increase in SAB-QI scores, subsequently leading to a better patient prognosis.

Our study has two strengths. First, the SAB-QI score, which was established by modifying the original 25 QIs [10], was used for clinical evaluation. Our score was simplified and thus could be made more easily available in any healthcare setting. Second, 289 cases from 14 hospitals were included; thus, generalizability of the analyzed data is warranted. However, this study had several limitations. First, we retrospectively collected clinical and microbiological data from the medical records. Thus, there may be errors or misunderstandings in the past data. In addition, considering the difficulty of evaluating past data, the timeframes to achieve each QI score were not fully defined. Second, although a description of the clinical course of SAB in the medical discharge summary is included in the SAB-QIs [16], its relationship with prognosis is unclear because the patient may have died before completing the summary. Third, several cases of SAB have been treated with ceftriaxone, particularly in cases complicated by central nervous system infections; however, the clinical validity of ceftriaxone treatment in cases of SAB remains uncertain. An additional QI proposal for SAB complicated with central nervous lesions is expected in the future. Fourth, SAB cases from small-scale hospitals (fewer than 200 beds) represented a minor proportion of the cases (4.5%), potentially leading to a selection bias in the population. Fifth, of the 98 excluded cases, nearly half (42 cases, 42.9%) were reported at a single acute care hospital, which may suggest a selection bias. The primary reasons for exclusion were clinical diagnosis of contamination (18 cases, 42.9%); and early transfer and hospitalization duration less than 4 days after diagnosis (13 cases, 31.0%). Given that S. aureus infrequently causes contamination during blood culture testing [44], this observation should be further examined. However, validation of diagnoses via retrospective analysis posed a challenge. In addition, the high patient transfer rate in acute care hospitals is unavoidable. Sixth, the assessment of mortality might have necessitated quantification of the severity of chronic underlying conditions using the Charlson Comorbidity Index. However, because of the retrospective nature of this study, the accurate collection of essential data was not possible. Instead, we compared the frequencies of multiple comorbidities in each QI group. Seventh, the study lacked a clear definition of the appropriate timing for source control measures when collecting the data. This may have caused an overestimation of QI values. Finally, adherence to SAB-QIs may be worse in patients with poor prognoses related to underlying clinical conditions. In such cases, the clinician had potentially adopted a more conservative approach, which may have reduced QI compliance. However, the effects of such confounding factors were not fully evaluated in this retrospective study.