Biomarkers: Another help tool to predict bacteriemia in the emergency department

Dear Editor,

We read with interest the review entitled “Microbiological diagnosis of bacteraemia and fungaemia: blood cultures and molecular methods” by Guna Serrano et al.¹ An excellent article in which it is underlined that the detection of bacteraemia and fungaemia is a priority for the clinical microbiology department due to its significant diagnostic and prognostic relevance. The importance of adequate blood culture collection prior to antibiotic treatment for patients with sepsis² is also highlighted, which may determine their progress and mortality.³ Furthermore, the authors state “nowadays, we cannot discuss the diagnosis of sepsis without taking into consideration the detection of biomarkers”.¹

In this context, bearing in mind that around 60% of sepsis cases are diagnosed in hospital emergency departments (HED)² and that the vast majority of blood cultures, which are sent to the microbiology department for processing, are obtained in these departments,⁴ we would like to highlight the importance of correctly suspecting and predicting the existence of bacteraemia in the HED for the patient. The new sepsis definitions (as well as the traditional ones) have limitations as they are not very specific,⁴ hence why there are disputes and different proposals regarding the inclusion of other criteria for diagnosing infection, sepsis and/or bacteraemia.⁵ In recent years, different reviews have been published which propose different criteria to optimise the indications for blood culture collection, as well as to improve their efficacy (increasing the number of positive blood cultures), effectiveness (reducing the number of contaminated blood cultures) and efficiency (cost of collection and processing, improving the adequacy and timeliness of antibiotic treatment, decision to discharge or admit).⁶ Thus, finding an applicable and genuinely useful predictive model for bacteraemia has become the objective of many authors combining different clinical, epidemiological and analytical variables, among which biomarkers such as procalcitonin (PCT), which increases the predictive performance of these models and is available in the vast majority of HEDs throughout Spain, currently stand out.^4,7

At our centre, since the implementation years ago of “code sepsis” (CS), we use directed triage (where from the initial patient assessment, laboratory tests for lactate and PCT are carried out, and, in case of suspicious findings, blood cultures are performed).⁸ We also follow the recommendations of Julián-Jiménez et al.,⁷ in which blood culture collection is indicated if the PCT concentration is >1 ng/ml. In this regard, we created a retrospective study in relation to the two conditions that result in the highest number of cases of sepsis and bacteraemia in HEDs: urinary tract infections (UTIs) and pneumonia. In this study, we have proven, in the second half of 2017 on adult patients aged ≥18 years, that PCT, which is usually available in HEDs, is the biomarker with the best prognostic performance for bacteraemia (better than C-reactive protein). A total of 346 patients were included, from whom blood cultures were taken (125 cases of pneumonia of which 18 [14.4%] had bacteraemia, and 221 cases of UTIs of which 25 [11.31%] had bacteraemia), with a mean age of 56 ± 24 and 54% of whom were female. To predict the existence of bacteraemia, PCT obtained the greatest area under the Receiver Operating Characteristic curve (AUC-ROC) of 0.94 (95% CI: 0.91–0.98; p < 0.001) and, with a cut-off point of ≥0.95 ng/ml, a sensitivity of 98%, a specificity of 94%, a positive predictive value of 84% and a negative predictive value of 98% were obtained. Meanwhile, CRP (mg/l) obtained a predictive performance with an AUC-ROC of 0.652 (p = 0.28). The mean values when comparing PCT in patients with pneumonia and UTIs with/without bacteraemia were 9.26 ± 16.42 vs. 0.36 ± 0.38 ng/ml; p < 0.001. As a conclusion to our study and experience, and in line with the recommendations of other authors, we can safely say that PCT is of use and provides great diagnostic performance for the prediction of bacteraemia in HEDs, improving both blood culture indications and results. It also helps us to get urgent decisions, such as the timely and adequate administration of antibiotics, and the decision to admit the patient and in the most appropriate place, right.

Therefore, as Guna Serrano et al. state in their article, we consider it relevant and necessary to include biomarkers as support tools in the assessment and confirmation of patients with sepsis, as well as to confirm suspected bacteraemia in the HED.^4,9