Our scoring system is the first tool to predict IE in patients with BHS bacteraemia. Heart murmur and elevated C-reactive protein levels are independent predictors of IE, and by adding GBS infection, low platelet count, and low blood pressure to the prediction model, the likelihood of IE occurrence can be estimated.
Several studies have scored the risk of IE in gram-positive cocci bacteraemia, such as the PREDICT study [2] and VIRSTA study [11] for S. aureus bacteraemia, HANDOC score for non-BHS bacteraemia [3], and DENOVA scores for Enterococcus faecalis [4]. However, to the best of our knowledge, there is no scoring system for BHS IE. The mortality rate of BHS IE is high, ranging from 18.4–29.3% [1, 12, 13], and is comparable to that of IE caused by S. aureus [13, 14]. Therefore, early therapeutic intervention after hospitalization is necessary.
This scoring tool used vital signs, physical examination results, and blood test results to determine the risk of IE at a relatively early clinical stage. As hypotension was reported to be associated with a higher rate of septic shock in an article comparing BHS IE with other streptococci and enterococci [13], it was appropriate to include it in this score. Platelet count, which is also included in the SOFA score [15], is considered a valid indicator of IE. As shown by the NOVA [16] and DENOVA scores [4], which evaluate the necessity of transoesophageal echocardiography (TEE) in enterococcal bacteraemia, a heart murmur is an essential physical finding in IE. CRP is an inflammatory protein often used for scoring, such as in the LRINEC score [17]. Although CRP levels have no correlation with the likelihood of IE occurrence as per previous studies [18], our study found that they could be a potential marker of IE, especially in BHS bacteraemia. In addition, CRP can be used to determine the response to treatment and prognosis by measuring its levels over time; therefore, CRP can be measured at the beginning of treatment [19–21]. The ability of bacteria to bind fibrinogen is thought to be a factor that causes IE [22]. Seo [22] have shown the ability of GBS to bind to fibrinogen via the glycoprotein Srrl, which explains that GBS is more likely to cause IE among BHS [1].
In our study, 74.5% of the follow-up blood cultures were collected from the IE group, but none showed persistent bacteraemia. This result was the same as that reported in previous studies [7]. Some physicians may recommend TEE since IE cannot be completely ruled out even if the BETA-GBS score is less than two in BHS bacteraemia; however, the low rate of persistent bacteraemia makes it challenging to determine the indication for invasive testing such as TEE in BHS bacteraemia. Therefore, our scoring is based on minimally invasive testing, utilizing physical examination and laboratory findings, and we believe that the combination of these findings reduces the probability of IE to a very low level. Some papers [7, 23] have indicated that unnecessary follow-up blood cultures increase the cost of medical care, and there is a need for more appropriate testing. The scoring system that was developed in this study was highly sensitive and can help rule out IE in BHS bacteraemia, and has potential reduce the unnecessary burden on patients and medical staff.
Our study had some limitations. First, this was a relatively small, single-centre, observational study. External validation for the implementation of this score by comparison with patients with BHS in other countries or facilities and prospective validation of the results of this study are warranted. Second, the prevalence of IE may be overestimated because the definition of IE included ‘Possible IE’ in Duke’s criteria. However, since our purpose was to create a scoring system that can exclude IE, it may be helpful as a screening tool with high sensitivity. Third, in this retrospective study, it is possible that some information regarding the physical examination and medical history was missing. In addition, echocardiography was not performed in all patients in the BHS-non-IE group; therefore, IE evaluation may not have been sufficient. Finally, because follow-up blood cultures were not collected from all patients, persistent bacteraemia in BHS IE may have been underestimated. However, because persistent bacteraemia tended not to occur in a previous study [7], follow-up blood cultures may not be necessary.
In summary, our BETA-GBS score is a tool for clinicians and may enable them to assess reduced likelihood of IE with a score of less than two in BHS bacteraemia.