The management of patients with infectious processes admitted to Spanish emergency rooms (ER) represented 15% of all daily care provided before the start of the COVID-19 pandemic.1 Also, the severity of infectious processes at their clinical presentation (patients with sepsis, relevant comorbidities, neutropenia, elderly patients with suspected bacteremia, that is, what is known as severe infection), and the 30-day short-term mortality rate have also increased over the last decade.1

We carefully read the article recently published by Caramello et al.,2 and we wish to congratulate the authors over their results and comments they make revealing the difficulties and limitations of the PIRO system (predisposition, infection, response, organ failure)3 used at the ER as a mortality risk stratification tool and to see the need for ICU admission on a routine daily basis. At the ER, both suspicion and diagnosis of severe infection or sepsis is essential. However, it is also necessary to estimate the prognosis of the patient. To this date, this is often done using the quick Sepsis-related Organ Failure Assessment (qSOFA) score.4

We used the database of a recent study conducted at our ER5 with a similar profile of patients to that used by Caramello et al.2 to see the predictive capabilities regarding mortality and ICU admission of the PIRO system that obtained better results compared to those analyzed by Howell et al.3 and compared to the qSOFA score and the 5MPB-Toledo model to predict bacteremia. Therefore, we have reproduced both the inclusion criteria and the methodology published by the authors.2 Our series included 1263 patients aged >18 years from July 1 2018 through August 1 2019 who met the criteria of sepsis from whom hemocultures were obtained. A total of 57% of these patients were men with a mean age of 59 ± 19 years. The overall mortality rate within the first 24 h was 1.5% while the 30-day mortality rate was 9.8%. The hemocultures of 18% of these patients tested positive while 9% had to be admitted to the ICU. The rate of dead patients according to the PIRO categories was scores <5 (4%); scores from 5 to 9 (12%); scores from 10 to 14 (21%); scores from 15 to 20 (43%), and scores >20 (73%). In our sample, the area under the ROC curve of the PIRO system, the qSOFA score, and the 5MPB-Toledo model score regarding the 30-day mortality rate were 0.753 (95%CI, 0.689–0.817), 0.741 (95%CI, 0.678–0.805), and 0.732 (95%CI, 0.668–0.796), respectively. Regarding the ICU admission, the scores were 0.598 (95%CI, 0.546–0.650), 0.612 (95%CI, 0.560–0.664), and 0.587 (95%CI, 0.535–0.639), respectively. The study was evaluated and approved by the Complejo Hospitalario Universitario de Toledo (Spain) clinical research ethics committee (reference No.: 2019/398).

With these data added to the results obtained from the authors, we believe that the limitations of the PIRO system are enough for us to not back up its use compared to the qSOFA score or even the 5MPB-Toledo model that also predicts the presence of bacteremia.