Introduction
Sepsis is defined as a life-threatening disturbance of organs function due to regulatory changes of a macro organism response to infection. Nowadays no evident and valid criteria to define a point of child transition from infection to sepsis [1]. The cause of child’s admission to intensive care unit (ICU) is a sepsis in 8 % among all cases. Also, there are annual more than 4.5 million of child’s deaths due to sepsis around the world [2–4]. This is a challenge to recognize a sepsis in children so it is related to a high prevalence of febrile conditions, a low specificity of clinical infectious signs, and an existence of anatomical and physiological features [2].
The definitions of Sepsis-3 emphasizes an importance of organ dysfunction recognizing in adults so it led to increasing and expansion of trials, improving of understanding sepsis pathobiology and patient’s response to invasions [5]. In this light the specific pathophysiology phenotyping has a great importance for evident assessment of patient’s response to treatment [6, 7]. Therefore, it is very important to recognize a sepsis on early stages of organ dysfunctions in children that especially concerning lower- resource healthcare.
However recent published guideline by Surviving sepsis campaign (SSC) did not propose any specific criteria for recognition a sepsis in children [2]. This circumstance was also noted by authors of Russian recommendations on diagnosis and treatment of sepsis in children [8]. Although there were a few valid pediatric organ dysfunction scores before publishing of SSC guideline like pediatric sequential organ failure assessment (pSOFA) and pediatric logistic organ dysfunction 2 (PELOD 2) [9]. Then the international research group (Pediatric Organ Dysfunction Information Update Mandate — PODIUM) was created and began their active work on organ dysfunctions in children [10–12].
In this light the exigency exists to identify clinical data which will be valid criteria for diagnostic sepsis in children. These criteria have to be applicable as globally as for different clinical situations [8]. Ideally sepsis criteria have to be: 1) quite sensitive so medical staff could determine a sepsis on early stages; specific to save a healthcare resources and avoid an inappropriate and excessive treatment, for example antibiotics prescribing; 2) applicable globally and adaptive locally; 3) correlating to biologically significant sepsis phenotypes to provide relevant patient’s selection for effective specific therapy and organs support.
In this connection the Society of critical care medicine in USA (SCCM) had convened a task force of 35 critical care experts, specialists of emergency medicine, infectious diseases, pediatrics, nurses experts, specialists of healthcare and neonatology from Australia, Bangladesh, Brazil, Canada, France, India, Italy, Japan, Switzerland, South Africa, Great Britain and United States of America [13].
The task force group of SCCM had been using data of international survey, systematic review and meta-analysis and also the new assessment of organs dysfunction which was designed and based on a more than 3 million medical cases from 10 sites and 4 continents. Then this group recommends to use the Phoenix sepsis score for sepsis identification in children with infectious process and the value of Phoenix sepsis score 2 points and more is suggested as life threatened organ dysfunction [14]. Originally, values of eight organ-specific scores were calculated in first 24 hours from admission as it showed on organ dysfunction in children with sepsis [14].
Finally, the model was based on only four organ systems criteria (cardiovascular, respiratory, neurological and coagulation systems) as comparable to assessment with using of eight organ systems. Another four systems (such as renal, hepatic, endocrine, immune) were added to above systems (Phoenix 8 score) [15]. The septic children had average Phoenix Sepsis score value of 3 as in countries with higher-resource healthcare as in countries with lower-resource healthcare (interquartile range, IQR, 2–4) [14]. The authors of this conception consider that the new Phoenix sepsis score criteria were derived and validated by SCCM task force using large international database of lower- and higher-resource pediatric hospitals so its use has a potential to improve diagnostic sepsis and septic shock in children compared to existing criteria [16]. In the same time this score has limitations accordingly an opinion of L.N. Sanchez-Pinto et al. One of that significant limitation is a fact that only 3.1 % of database were validated from lower-resource countries what limit the accuracy of this score [11]. Notably the Phoenix Sepsis score had demonstrated very low sensitivity compared to previous criteria (23 vs 77 %) in one of the lower-resource hospitals. It emphasizes a necessary to continue trials before routine implementation of Phoenix sepsis score in lower-resource hospitals [11]. Furthermore, Phoenix sepsis score developers admit that the score has certain advantages over only the organ’s dysfunction system assessment published by International pediatric sepsis consensus conference (IPSCC) in 2005 [16].
Objective
Objective of our study is the comparative assessment of discriminatory ability pSOFA, PELOD 2 and Phoenix Sepsis scores for sepsis in children.
Materials and methods
The design of our trial is retrospective observational multi-center. We used database of Republic Bashkortostan Children’s clinical hospital (37 patients), St.-Petersburgh Children’s clinical hospital named N.F. Filatov (33 patients), Krasnodar Children’s regional clinical hospital (70 patients). The framework time of this study was from June, 1 2022 till June, 26 2023. The inclusion criteria were children from 9 months to 17 years old with sepsis and septic shock. The sepsis and shock were basically identified according Russian recommendation on diagnostic and treatment of sepsis [8]. The exclusion criteria were inborn errors of metabolism, staying in ICU less than 24 hour and renal failure with indication to renal replacement therapy during first 6 hours after admission. A total 140 children had a inclusion criteria. There were 37 children (27.1 %) with septic shock. 29 patients had died (20.7 %). The most frequent cause of sepsis was a pneumonia (117 children, 83.6 %). Also children were included with peritonitis (12 patients, 8.2 %), with infection of skin and soft tissues (6 children, 4.3 %), with ENT infections such as otitis, pansinusitis (3 patients, 2.1 %), with sepsis due to multiple trauma (2 patients, 1.4 %). The endpoint of our trial was a 28-days mortality.
The assessment of discriminatory ability of pSOFA, PELOD 2 and Phoenix Sepsis scores performed on admission day in ICU. It related to determination by SCCM experts an informative significance of Phoenix Sepsis score during first 24 hours from child’s admission [14]. The content of Phoenix Sepsis score is presented in Table 1.
| Variables | 0 points | 1 point | 2 points | 3 points |
|---|---|---|---|---|
| Respiratory, 0–3 points | PaО2:FiO2 > 400 or SpO2:FiO2 > 292b | PaO2:FiO2 < 400 on any respiratory support or SpO2:FiO2 < 292 on any respiratory supportb,c | PaO2:FiO2 100–200 and IMV or SpO2:FiO2 148–220 and IMVb | PaO2:FiO2 < 100 and IMV or SpO2:FiO2 < 148 and IMVb |
| Cardiovascular, 0–6 points | — | 1 point each (up to 3) | 2 points each (up to 6) | — |
| No vasoactive medicationsd | 1 vasoactive medicationd | ≥ 2 vasoactive medicationsd | ||
| Lactate < 5 mmol/le | Lactate 5–10.9 mmol/le | Lactate ≥ 11 mmol/le | ||
| Age basedf mean arterial pressure, mm Hgg | ||||
| < 1 month | > 30 | 17–30 | < 17 | |
| 1 to 11 months | > 38 | 25–38 | < 25 | |
| 1 to < 2 year | > 43 | 31–43 | < 31 | |
| 2 to < 5 years | > 44 | 32–44 | < 32 | |
| 5 to < 12 years | > 48 | 36–48 | < 36 | |
| 12 to 17 years | > 51 | 38–51 | < 38 | |
| Coagulation, 0–2 pointsh | — | 1 point each (maximum 2 points) | — | |
| Platelets ≥ 100 × 103/μl | Platelets < 100 × 103/μl | |||
| International normalized ratio ≤ 1.3 | International normalized ratio > 1.3 | |||
| D-dimer ≤ 2 mg/l FEU | D-dimer > 2 mg/l FEU | |||
| Fibrinogen ≥ 100 mg/dl | Fibrinogen < 100 mg/dl | |||
| Neurological, 0–2 pointsi | Glasgow Coma Scale score > 10; pupils reactivej | Glasgow Coma Scale score ≤ 10j | Fixed pupils bilaterally | |
| PHOENIX SEPSIS Criteria | ||||
| Sepsis | Suspected infection and Phoenix Sepsis Score ≥ 2 points | |||
| Sepsis shock | Sepsis with ≥ 1 cardiovascular point(s) | |||
The demographic and clinical data were presented as median values with interquartile ranges of mean and standard deviation, percentages and frequencies concerning feature characteristics. The continuous variables were compared with Mann-Whitney U-test using. The statistical significance was confirmed if p was less than 0.05. The discriminatory power of the scores was defined with receiver operating characteristic (ROC) analysis and determination of area under ROC curve (AUC).
Results
We analyzed score’s values to compare data between survivors and deceased on a Day 1 (admission day in ICU). This analysis is presented in Table 2.
| Scores | Survivors (n = 111) | Non-survivors (n = 29) | р |
|---|---|---|---|
| pSOFA, points | 4 [3–7] | 9 [6–12] | < 0.05 (0.029) |
| PELOD 2, points | 4 [3–5] | 6 [4–7] | < 0.05 (0.032) |
| Phoenix Sepsis Score, points | 3 [2–4] | 4 [3–5] | < 0.05 (0.048) |
These data allow to claim that all of the three evaluating systems showed an evident difference between survived and deceased children with sepsis. We performed the ROC analysis to receive more accurate information concerning discriminatory ability of scores (Fig. 1).
Fig. 1. Discriminatory ability of the studied scales for sepsis (n = 140)
| Scores | |||||
|---|---|---|---|---|---|
| Scores | Area under curve | Standard deviation | Asymptotic values | Asymptotic confident interval 95 % | |
| Low level | High level | ||||
| pSOFA | 0.668 | 0.080 | 0.068 | 0.510 | 0.826 |
| PELOD 2 | 0.638 | 0.091 | 0.135 | 0.459 | 0.817 |
| Phoenix Sepsis Score | 0.591 | 0.110 | 0.326 | 0.375 | 0.806 |
The AUC’s determination allowed to claim that an informative significance was a similar for all of the three scores (a significance of differences between Phoenix sepsis score and pSOFA was a 0.57, Phoenix sepsis score and PELOD2 was a 0.80, pSOFA and PELOD2 was a 0.74). Then the Phoenix sepsis score has no any advantages over pSOFA and PELOD2 scores in children with sepsis. Therefore, we decided to perform a comparative analysis of scores discriminatory ability for children with septic shock (Fig. 2).
Fig. 2. Discriminatory ability of the analyzed scales for septic shock in children (n = 38)
Scores |
|||||
|---|---|---|---|---|---|
| Scores | Area under curve | Standard deviation | Asymptotic values | Asymptotic confident interval 95% | |
| Low level | High level | ||||
| Phoenix Sepsis Score | 0.494 | 0.119 | 0.950 | 0.260 | 0.727 |
| pSOFA | 0.500 | 0.102 | 1.000 | 0.300 | 0.700 |
| PELOD 2 | 0.535 | 0.106 | 0.730 | 0.328 | 0.743 |
The analysis had demonstrated no one of all scores showed predictable ability for outcomes of septic shock in children. The AUC was evaluated around 0.5 for all scores and a significance of differences between Phoenix sepsis score and pSOFA was a 0.81, Phoenix sepsis score and PELOD 2 was a 0.97, pSOFA and PELOD 2 was a 0.78.
Discussion
Our study is a retrospective clinical trial concerning a comparative assessment of PELOD 2, pSOFA and Phoenix sepsis scores possibilities to predict sepsis outcomes in children.
The SCCM’s task force group declared that Phoenix sepsis criteria have a high discriminatory ability to identify a sepsis in children accurately as for a low-resource hospital as for a high-resource setting. It will facilitate an international dissemination and data collection around the world. Our outside validation of this evaluating system showed that it has a moderate discriminatory ability. Furthermore, PELOD 2, pSOFA and Phoenix Sepsis scores have a similar informative value according to our data and to data of SCCM task force group [16].
Our sepsis identifying based on Project of Russian clinical recommendation for diagnostic and treatment of sepsis in children. All children of this trial had a ≥ 2 points of Phoenix Sepsis score on admission day in ICU. It gives an evidence concerning similar approaches to identifying of infectious generalization [14]. Moreover, the new criteria developing was performing with using of pSOFA, PELOD 2 subscores and PODIUM group achievements [15]. For example, the most effective criteria of cardiovascular and neurological dysfunctions appropriate to a subscores of PELOD 2, hematological/coagulation, respiratory and renal dysfunctions appropriate to pSOFA score and immune and endocrine dysfunctions to PODIUM [14]. Meanwhile, we were surprised by that the SCCM task force group compared the new Phoenix sepsis score criteria to IPSCC recommendations (2005) concerning sepsis identification. In the same time, SCCM task force group had developed a conceptual definition of pediatric sepsis as a suspected infection in presence of life-threatening organ dysfunctions. Also, the task force group had used organ dysfunctions criteria according to “Sepsis-3” conception which related to a higher risk of mortality [14].
We estimated no one analyzed score could predict a survival of children with sepsis at early time (first 24 hours) staying in ICU. It confirms by results of recent published report by A.V. Trembach et al [17]. It is clear that the Phoenix sepsis score’s variables threshold was used for identifying of life-threatening organ dysfunctions in children with infection. These thresholds do not provide a screening of children with high risk sepsis development or an early recognition of children with suspected sepsis. The Phoenix sepsis score criteria is presented as simplification of complicated biological processes which lead to a sepsis in children in presence of heterogeneity of original child’s state as macro-organism and microorganism in context Sepsis-3 conception [16]. The SCCM task force group did not try to describe a certain signs of macro-organism dysregulation and to confirm results by collection higher level biological resolution data, e.g. included genomics and proteomics.
The limitations of this study are a retrospective design and a small number of patients. Despite this our results confirm the SCCM group opinion that sometimes Phoenix Sepsis score system has no advantage compared to pSOFA and PELOD 2 scores as a system of organ dysfunctions assessment [15]. Furthermore, an availability of resources and a local medical practice may affect the identification of infection, e.g. microbiological testing, sensitivity to antibiotics.
Conclusion
Our trials showed that Phoenix sepsis score, pSOFA score and PELOD 2 score were similar in Russian settings. Should we currently recognize a sepsis in children based on only Phoenix sepsis score? At this time pSOFA score has a high dissemination in pediatric ICU in Russia with using not only for septic children. Furthermore, Phoenix sepsis score requires to determine a lactate level, D-dimer, international normalized ratio, fibrinogen. These tests are available in large hospitals and ICU of third level of Russian Federation healthcare. Thereby, we conclude that the well-known pSOFA and PELOD 2 scores have to be currently used along with the Phoenix sepsis score. Although, the Phoenix sepsis score will have to be undergone by additional external international validation in countries with low-resourced healthcare.
Disclosure. The authors declare no competing interests.
Author contribution. All authors according to the ICMJE criteria participated in the development of the concept of the article, obtaining and analyzing factual data, writing and editing the text of the article, checking and approving the text of the article.
Ethics approval. This study was approved by the local Ethical Committee of Kuban State Medical University (reference number 19 — 16.04.2013).
Funding source. This study was not supported by any external sources of funding.
Data Availability Statement. The data that support the findings of this study are available from the corresponding author upon reasonable request.
