Best diagnostic accuracy of sepsis combining SIRS criteria or qSOFA score with Procalcitonin and Mid-Regional pro-Adrenomedullin outside ICU

Early diagnosis and treatment significantly reduce sepsis mortality. Currently, no gold standard has been yet established to diagnose sepsis outside the ICU. The aim of the study was to evaluate the diagnostic accuracy of sepsis defined by SIRS Criteria of 1991, Second Consensus Conference Criteria of 2001, modified Second Consensus Conference Criteria of 2001 (obtaining SIRS Criteria and SOFA score), Third Consensus Conference of 2016, in addition to the dosage of Procalcitonin (PCT) and MR-pro-Adrenomedullin (MR-proADM). In this prospective study, 209 consecutive patients with clinical diagnosis of sepsis were enrolled (May 2014–June 2018) outside intensive care unit (ICU) setting. A diagnostic protocol could include SIRS criteria or qSOFA score evaluation, rapid testing of PCT and MR-proADM, and SOFA score calculation for organ failure definition. Using this approach outside the ICU, a rapid diagnostic and prognostic evaluation could be achieved, also in the case of negative SIRS, qSOFA or SOFA scores with high post-test probability to reduce mortality and improve outcomes.


Scientific RepoRtS
| (2020) 10:16605 | https://doi.org/10.1038/s41598-020-73676-y www.nature.com/scientificreports/ gender, prior or current use of antibiotics, immunosuppressive treatments, immune status (active malignancy or other causes of an immunocompromised state), comorbidities and clinical presentation were recorded. For each patients a physical examination including cardiac, abdominal, respiratory and neurological evaluations was performed. The real-world control group included fifty patients admitted to the Diagnostic and Therapeutic Medicine Department of Campus Bio-Medico of Rome for cardiac, kidney, liver, pulmonary and cancer diseases being responsible for a non-infectious related SIRS, qSOFA, or SOFA criteria positivity. clinical and laboratory parameters, blood gas analysis, blood and microbiological cultures. The following clinical and laboratory parameters have been collected: body temperature, blood pressure, heart and respiratory rate, complete blood counts (CBC), PCT, MR-proADM, bilirubin, creatinine, lactate, PaO 2 /FIO 2 , and blood and microbiological cultures at the diagnosis and when clinically necessary.
Blood and microbiological cultures. Blood specimens from patients were collected in BACTEC bottles containing anaerobic or aerobic broth and resins. Blood culture bottles (BC) were incubated in BACTEC FX instrument (Becton Dickinson, Meylan, France) until they resulted positive for bacterial growth or for a maximum of 5 days. Positive BC samples were cultivated in selective agar media. Growing colonies were identified by MALDI-TOF 22 32 . χ 2 for proportions test was used to compare the relative percentage of patients with positivity and/or negativity to SIRS criteria, SOFA score, qSOFA score, PCT and MR-proADM. p value < 0.05 were considered as significant.
Pretest odds, posttest odds, and the consequent posttest probability and χ 2 test for proportions have been computed to investigate whether combination of PCT, MR-proADM, lactate, SOFA, qSOFA scores, SIRS criteria of 1991, Second Consensus Conference Criteria, modified Second Consensus Conference Criteria improves post-test probability 33 .

Results
Patients characteristics. The demographic and clinical characteristics of the study group including 209 patients with sepsis and of the 50 real-world control group patients are reported in Table 1. The control group included patients with cardiac, kidney, liver, pulmonary and cancer diseases being responsible for a non-infectious related SIRS, qSOFA, or SOFA criteria positivity.
Septic patients and control group were similar except for the presence of chronic cardiac failure that was significantly more represented in control population (p = 0.0002), acute kidney injury 34 and solid cancer that were more prevalent in study group (p = 0.0062 and p = 0.0002, respectively). Septic shock was diagnosed in 82 out of 209 (39%) patients (Table 1). In 162/209 (77.5%) patients antimicrobial therapy was administered before sepsis diagnosis (Table 1). 30-day mortality was 8% in patients with sepsis, reaching values as high as 33% in case of septic shock, whereas 90-days mortality was 13% in sepsis and 38% in septic shock (  1B).
PCT, MR-proADM, SIRS criteria, qSOFA and SOFA score values in study population. Median values, interquartile ranges (25th percentile and 75th percentile), and Mann-Whitney's comparison of the different variables are reported in Table 3. In particular, the median number of SIRS criteria registered was two, the median qSOFA score was 1, the median SOFA score was 4, PCT and MR-proADM median levels were 1.16 ng/ mL and 2.55 nmol/L, respectively. All variables resulted significantly higher in septic patients than control group (p < 0.0001) ( Table 3).

ROC curves analysis and areas under the curves (AUCs).
In septic patients, the AUCs values for SIRS criteria, Second Consensus Conference Criteria, modified Second Consensus Conference Criteria, qSOFA and SOFA score are reported in Table 4.
ROC curves comparison between SIRS criteria, Second Consensus Conference Criteria, modified Second Consensus Conference Criteria, qSOFA and SOFA score has been reported in Fig. 2. Any statistically significant difference has been highlighted. Adding PCT and MR-proADM to the ROC curve analysis, PCT AUC was significantly higher (p < 0.05) than all other variables ( Fig. 3; Table 4).
The combination of PCT, SIRS or qSOFA and MR-proADM provide a diagnostic and prognostic evaluation in 99.9% of patients with a turnaround time of about 45 min, whereas the combination of PCT, SOFA score and MR-proADM reaching comparable accuracy (99.9%) requires a turnaround time of about 90 min.

Discussion
The physiopathology of sepsis highlights the need of unambiguous diagnostic criteria for a rapid patients identification and adequate therapy administration, within one hour from symptoms presentation 35 . Sepsis definition and diagnostic criteria proposed from 1991 until now still lack of specificity [36][37][38][39] .
Confounding factors influencing body temperature, heart and respiratory rates and white blood cell count included in SIRS comprehended beta-blockers, calcium-antagonists and other antiarrhythmic drugs or pacemaker DDD (heart rate), paracetamol, anti-inflammatory drugs and antimicrobials (body temperature); benzodiazepine, sedative and chronic oxygen administration (respiratory rate); immunosuppressive drugs and Ideally, the best criteria should be as rapid as practically reliable for an early diagnosis and treatment of sepsis. In this prospective study, sepsis was diagnosed according to SIRS Criteria of 1991, Second Consensus Conference Criteria, modified Second Consensus Conference Criteria, Third Consensus Conference Criteria, in comparison with PCT and MR-proADM measurement. ROC curve analysis used to evaluate the diagnostic accuracy of the different criteria showed complete overlapping of the curves. On this basis, it should be convenient to prefer using bedside SIRS criteria or qSOFA in non-ICU setting rather than SOFA score requiring laboratory screening, Glasgow coma scale determination and knowledge of patients' comorbidities or previous organ failures. Second Consensus Criteria of 2001 require the measurement of multiple clinical as well as bioumoral parameters needing long determination time. In this study, SIRS criteria and qSOFA allowed a diagnosis in 97% and 96% of patients, respectively, in case of suspicion of sepsis outside ICU. In the last years, plasma biomarkers have been proposed as tools for a rapid diagnosis and good indicator of prognosis. Among these, PCT and MR-proADM showed the best diagnostic and prognostic accuracy for the complementary nature of given information. PCT was optimal for etiological diagnosis and antimicrobial therapy management 12 , whereas MR-proADM was significantly correlated with organ failure and worse prognosis. In the present study, ROC analysis showed that besides clinical scores, PCT measurement represent the best diagnostic accuracy in sepsis, as previously described 12,[21][22][23][24]41,42 allowing early tailored antimicrobial therapy administration and daily follow-up. It should be reliable to combine bedside SIRS criteria or qSOFA with PCT laboratory determination for early identification of sepsis, followed by SOFA score calculation for severity and prognosis evaluation. In this study, about 35% of patients were negative for SIRS criteria or qSOFA, and SOFA score or for all, despite evidence of positive blood culture and documented microbiological isolate or clinical diagnosis of infection. In these patients, the use of MR-proADM was essential to provide early diagnosis and confirm the suspicion of sepsis.
These results suggest that in case of suspected sepsis, SIRS criteria or qSOFA should be bedside evaluated together with PCT measurement. These combinations reach a post-test probability of 99.9%. Besides PCT and MR-proADM, a marker of organ failure, even if comparable to SOFA score in sepsis severity prediction, showed the ability to anticipate SOFA and qSOFA score and the advantage to be more objective and fasten measured, as previously described outside ICU 24 . Exactly, in case of clinical suspicious of infection the presence of SIRS criteria ≥ 2, qSOFA ≥ 2, PCT ≥ 0.5 and MR-proADM ≥ 1.5 nmol/L identifies sepsis in 99.9% of cases. This approach, reliable in about 45 min, could allow an early diagnosis of sepsis within the first hour, even outside the intensive care contest to reduce the need for ICU transfer and mortality, as previously reported 24,43 .
Data from the prospective study highlighted comparable diagnostic accuracy between SIRS criteria from the First Consensus Conference of 1991, Criteria from the Second Consensus Conference of 2001 and from the Third Consensus Conference of 2016. Moreover, the use of the modified Second Consensus Conference Criteria of 2001, based on SIRS criteria plus SOFA score for sepsis diagnosis, did not improve diagnostic accuracy more than PCT and MR-proADM. In this study, SIRS criteria allowed a diagnosis in 97% of patient and, when combined with PCT measurement, identified 99.9% of septic patients. Moreover, MR-proADM values > 1 nmol/L showed the ability to identify septic patients when SIRS, SOFA and PCT were still negative. These results confirm those reported by other authors where MR-proADM anticipates by 24 h the organ failure development 44 .
Through SIRS criteria, qSOFA, PCT and MR-proADM determination, sepsis diagnosis can be achieved within the first hour from suspicion as recommended 35 to improve outcome and decrease mortality. Furthermore, MR-proADM ≥ 1 nmol/L, even in case of negative PCT, qSOFA, SOFA, absence of Second Consensus Conference Criteria, identified septic patients with positive blood culture.
In conclusion, data from this study could suggest a diagnostic protocol for sepsis management outside ICU setting including, within 30 min from sepsis suspicion, bedside SIRS criteria or qSOFA score evaluation; within 1 h, PCT and MR-proADM measurement, microbiological culture collection, empiric sepsis therapy set up Table 7. Post-test probability analysis used to define the diagnostic value derived from the combined use of PCT, MR-proADM, SOFA score and SIRS criteria in patients with sepsis or septic shock. www.nature.com/scientificreports/ and SOFA score calculation. From these prompt actions, rapid diagnostic and prognostic evaluation of sepsis could be achieved also in case of negative SIRS, qSOFA or SOFA score with high post-test probability to reduce mortality and improve outcome.