Abstract
-
Objective
Sepsis, a life-threatening organ dysfunction, is a major global health concern. Early detection remains challenging due to nonspecific symptoms. Noninvasive tools such as the shock index, diastolic shock index, capillary refill time (CRT), and mottling score (MS) could help clinicians assess hemodynamic status and predict mortality, but a comprehensive comparison of their prognostic value is lacking. This study compares the performance of those four tools in predicting mortality in septic patients at 24 hours, 7 days, and 28 days.
-
Methods
A single-center, prospective observational study was conducted from January to September 2024. Adult patients (≥18 years) with suspected infection and a National Early Warning Score-2 of ≥5 were enrolled. Demographic data, vital signs, and CRT and MS results were collected at presentation, and mortality outcomes were recorded at 24 hours, 7 days, and 28 days.
-
Results
In total, 135 patients were included (median age, 85 years [interquartile range, 79–90 years]; 44.4% female). The mortality rates were 15.6% at 24 hours, 25.2% at 7 days, and 35.6% at 28 days. CRT showed the highest predictive value for 24-hour mortality (area under the curve [AUC], 0.829; 95% confidence interval [CI], 0.755–0.889), and MS had the best performance at 7 days (AUC, 0.732; 95% CI, 0.646–0.806) and at 28 days (AUC, 0.749; 95% CI, 0.662–0.823). No significant differences emerged in pairwise comparisons.
-
Conclusion
Although no one tool significantly outperformed the others, all four tools may provide useful, noninvasive mortality prediction in sepsis. CRT may be most effective for early risk stratification, and MS correlates with mid-term outcomes, supporting their integration into clinical assessments.
-
Keywords: Capillary refill time; Diastolic shock index; Mottling score; Sepsis; Septic shock
Capsule Summary
What is already known
Noninvasive tools such as capillary refill time (CRT) and the mottling score (MS) are used to assess hemodynamic status in septic patients, but a comprehensive comparison of their predictive efficacy is lacking.
What is new in the current study
This study finds that CRT is the strongest predictor of early mortality (24 hours), and MS is more effective for predicting mortality from sepsis at 7 and 28 days. The findings support integrating CRT and MS into clinical protocols to enhance early and ongoing risk assessment for septic patients, potentially improving management strategies.
INTRODUCTION
Sepsis is a life-threatening organ dysfunction caused by a dysregulated host response to infection [
1]. According to the Sepsis-3 consensus definition, sepsis occurs when an infection leads to acute changes in the Sequential Organ Failure Assessment (SOFA) score, indicating organ dysfunction [
1]. Septic shock, a subset of sepsis, is characterized by persistent hypotension despite adequate fluid resuscitation and significantly increases the risk of in-hospital mortality [
1,
2]. The global incidence of sepsis is alarming, with estimates suggesting that it affects more than 30 million individuals annually, resulting in approximately 6 million deaths worldwide. In high-income countries, the overall mortality rate for sepsis ranges from 15% to 30%, whereas in low- and middle-income countries, it exceeds 50% [
3,
4]. This underscores the critical need for effective early detection and management strategies [
5]. Early diagnosis and prognosis of sepsis remain significant challenges in clinical practice. The initial presentation of sepsis can be insidious, with nonspecific symptoms that can mimic less severe conditions. This often results in delayed recognition and treatment, so establishing protocols for early recognition and treatment is crucial for improving patient outcomes. Current guidelines emphasize the importance of early identification, with recommendations to assess patients for risk factors and clinical signs of sepsis within the first hours of presentation [
2]. However, despite advances in diagnostic criteria and the introduction of early warning scores (e.g., National Early Warning Score-2 [NEWS-2]), timely identification is still insufficient, particularly in emergency settings. As a result, many patients continue to present with advanced disease, and sepsis morbidity and mortality rates remain high [
2,
5].
Given the importance of early assessment, noninvasive prognostic tools have emerged as valuable adjuncts in the management of septic patients. These tools, including the shock index (SI), diastolic SI (DSI), capillary refill time (CRT), and mottling score (MS), allow clinicians to rapidly evaluate a patient’s hemodynamic status and perfusion [
6–
9]. The SI, defined as the ratio of the heart rate to systolic blood pressure, provides insight into a patient’s cardiovascular status [
6]. The DSI, an adaptation focusing on diastolic blood pressure, is intended to enhance the prognostic accuracy of the SI [
7]. Both the SI and DSI have been proposed as potential early markers of circulatory impairment before overt hypotension occurs [
6,
7]; however, a comprehensive comparison of their prognostic efficacy is lacking. CRT assesses peripheral perfusion and has been shown to correlate with in-hospital mortality in critically ill patients [
8]. The MS evaluates skin perfusion and can identify microcirculatory failure [
9]. Each of these tools contributes to the early identification of patients at high risk for deterioration, allowing for timely intervention.
Despite the individual merits of these noninvasive assessments, no comprehensive studies have compared their efficacy in predicting sepsis mortality at various time frames. Therefore, our aim in this study was to evaluate and compare the prognostic performance of the SI, DSI, CRT, and MS in predicting 24-hour, 7-day, and 28-day mortality in septic patients. By conducting pairwise comparisons among these tools, we sought to identify which assessment method offers the greatest predictive value in the early stages of sepsis management. Ultimately, this research aims to enhance clinical decision-making and improve outcomes for patients experiencing sepsis and septic shock.
METHODS
Ethics statement
This study was approved by the Ethics Committee of the Area Vasta Emilia Centro (No. 301/2023/Oss/AOUFe). Written informed consent was obtained from all patients or their legally authorized representatives. The study was conducted in accordance with the ethical standards of the responsible committee on human experimentation and the 1975 Declaration of Helsinki.
Study design and setting
This single-center observational, prospective, and noninterventional study was performed between January and September 2024. The study was conducted in a tertiary care center with approximately 600 beds and an emergency department that manages more than 70,000 visits annually. The inclusion criteria were as follows: (1) age ≥18 years; (2) suspicion of infectious disease; (3) NEWS-2 of ≥5; and (4) a signed informed consent form from each involved patient (or their relatives in cases with overall severe clinical conditions). Sepsis was confirmed using the Sepsis-3 diagnostic criteria [
1]. NEWS-2 was selected as an inclusion criterion due to its widespread clinical use as an early indicator of sepsis risk [
2,
5]. The collected data were age, sex, vital parameters (i.e., systolic and diastolic blood pressure, heart rate, respiratory rate, peripheral oxygen saturation, oxygen requirement, body temperature, and neurological status), CRT, MS, and 24-hour, 7-day, and 28-day mortality from sepsis that were confirmed using medical records. All noninvasive parameters (SI, DSI, CRT, MS) were measured at the time of initial clinical evaluation in the emergency department, prior to any therapeutic intervention. CRT was measured by applying pressure to the distal phalanx for 5 seconds and assessing the time for color return [
8]. The MS was evaluated using a validated five-grade scale (
Fig. 1) [
9], though interobserver reliability was not assessed in this study. It should be noted that the MS was assessed exclusively at the anterior aspect of the knee, according to the original validation study [
9]. Although that site offers a standardized and practical approach, mottling is a systemic phenomenon that can manifest in other regions, such as the hands, feet, or abdomen. This limitation might reduce the sensitivity of the score in patients with atypical perfusion patterns or in those with dark skin, which can make mottling less visible. Future studies should consider broader anatomical assessments to improve diagnostic accuracy.
Categorical data are presented as absolute frequencies and percentages, and continuous variables are expressed as medians and interquartile ranges (IQRs). The Mann-Whitney test was used to compare continuous variables between two independent groups, and the chi-square test was used to compare the distribution of categorical data between two independent groups. Associations between two continuous variables were assessed using Spearman correlation. A multivariable logistic regression on 24-hour mortality was performed to adjust for potential confounding variables (age, Charlson Comorbidity Index [CCI], and altered mental status). A second mutlivariable logistic regression on the same outcomes was performed to adjust for SOFA scores and lactates. The discriminatory ability of each predictive score was evaluated using the area under the curve (AUC) of the receiver operating characteristic (ROC) curve. Pairwise comparisons of ROC curves were conducted using the DeLong test.
To further explore the prognostic performance of the noninvasive tools under investigation, we conducted a series of additional statistical analyses. First, for each parameter (SI, DSI, CRT, and MS), we used the Youden index to determine the optimal cutoff values for predicting 24-hour, 7-day, and 28-day mortality. Based on those thresholds, we calculated sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and overall accuracy. In addition, we performed a Kendall tau-b correlation analysis to assess the relationships among the four tools and evaluate whether they reflect overlapping or distinct physiological domains. Lastly, we conducted an age-stratified analysis by dividing the study population into four predefined age groups: <65, 65–75, 75–85, and >85 years. Notably, no deaths occurred in the <65 years group, preventing an outcome analysis in that category.
Statistical analyses were conducted using jamovi ver. 2.5 (the jamovi project) and MedCalc ver. 19.8 (MedCalc Software Ltd).
RESULTS
We prospectively enrolled 135 patients in this study, and their median age was 85 years (IQR, 79–90 years). Among them 60 (44.4%) were female. The mortality rate was 15.6% at 24 hours, 25.2% at 7 days, and 35.6% at 28 days. Further details are summarized in
Table 1.
Among the tested tools, CRT was the best predictor of mortality at 24 hours, with an AUC of 0.829 (MS, 0.827; SI, 0.754; DSI, 0.755; P<0.001), indicating strong prognostic value in the early phase of sepsis (
Fig. 2A). For longer-term mortality prediction, the MS emerged as the most reliable tool, with the best AUCs at both 7 days (MS, 0.732; SI, 0.701; DSI, 0.701; CRT, 0.667; P<0.001) (
Fig. 2B) and 28 days (MS, 0.749; SI, 0.691; CRT, 0.687; DSI, 0.673; P<0.001) (
Fig. 2C). No statistically significant differences were observed among the four tools in the pairwise comparisons (
Table 2). In the univariable analysis, all four tools were found to be independent risk factors for mortality at 24 hours, 7 days, and 28 days (
Table 3). These results were further explored through a corrected multivariable logistic regression analysis (
Table 4), which evaluated each noninvasive tool (SI, DSI, CRT, and MS) independently while adjusting for age, CCI, and altered mental status. All four tools were found to be independently associated with mortality across all time points. Specifically, SI showed an odds ratio (OR) of 10.949 (95% confidence interval [CI], 4.065–29.480) for 24-hour mortality, 7.299 (95% CI, 3.012–17.689) for 7-day mortality, and 10.379 (95% CI, 4.070–26.467) for 28-day mortality. DSI had ORs of 4.492 (95% CI, 2.402–8.638), 3.960 (95% CI, 2.225–7.045), and 4.116 (95% CI, 2.317–7.312), respectively, at the same time points. CRT also retained statistical significance in all models with ORs of 1.810 (95% CI, 1.451–2.258), 1.422 (95% CI, 1.195–1.692), and 1.358 (95% CI, 1.149–1.605), respectively. Similarly, MS was associated with increased mortality with ORs of 3.007 (95% CI, 2.117–4.270), 2.203 (95% CI, 1.679–2.890), and 2.103 (95% CI, 1.609–2.748) at 24 hours, 7 days, and 28 days, respectively. In addition, we used the Youden index to identify optimal cutoff values for each tool and calculated corresponding sensitivity, specificity, PPV, NPV, and accuracy metrics. Those results are presented in
Suppl. 1 and provide clinically useful thresholds to guide bedside interpretation of results from these tools.
To assess whether the tools reflect overlapping physiological domains or offer complementary information, we performed a Kendall tau-b correlation analysis. SI and DSI were closely associated with parameters such as PaO
2/FiO
2 (fraction of inspired oxygen), platelet count, and hepatic dysfunction. In contrast, CRT and MS showed strong correlations with mean arterial pressure and the Glasgow Coma Scale. Full results are shown in
Suppl. 2.
In addition to the primary analyses, a supplementary multivariable analysis was conducted for each noninvasive parameter independently, controlling for total SOFA scores and lactate levels. CRT and MS emerged as independent predictors of 24-hour mortality, even when adjusted for established prognostic markers. Conversely, SI and DSI did not demonstrate independent predictive value within these adjusted models, suggesting that their prognostic utility might be contingent on concurrent hemodynamic data (
Suppl. 3).
Finally, we conducted an age-stratified ROC analysis to evaluate prognostic performance across different age groups. Patients were divided into four predefined age groups (<65, 65–75, 75–85, and >85 years); however, due to the absence of deaths in the <65 years group, the ROC analysis was limited to the remaining three groups. The curves presented in
Suppl. 4 confirm consistent trends across strata, with CRT performing better in early mortality prediction and MS showing stronger performance at the later time points.
DISCUSSION
The results of this study show that CRT had the highest AUC for predicting mortality at 24 hours, whereas MS had the highest AUC for predicting mortality at 7 and 28 days. Furthermore, pairwise comparisons among the tools did not reveal significant differences, suggesting that each tool has merit, and they can be used interchangeably in practice without one clearly outperforming the others.
However, it is important to acknowledge that the tools evaluated in this study reflect distinct physiological domains: SI and DSI are primarily indicators of macrocirculatory function [
6,
7], whereas CRT and MS assess microcirculatory status [
8,
9]. Our intention was to compare these tools side by side to evaluate their relative prognostic performance, despite their differing physiological targets. This approach allows clinicians to understand whether one domain offers superior predictive value in practice. Nevertheless, we recognize that future studies could benefit from grouping these tools by physiological domain or exploring composite models that integrate both macro- and microcirculatory indicators to enhance risk stratification.
The strong predictive value of CRT at 24 hours underscores its utility as a rapid assessment tool in the emergency setting. CRT is a simple bedside test that evaluates peripheral perfusion and can be performed quickly, making it an accessible option for frontline clinicians [
8,
10,
11]. In septic patients, abnormal CRT, which reflects impaired peripheral perfusion, has been associated with higher in-hospital mortality and worse outcomes [
8,
10]. Research shows that a prolonged CRT after fluid resuscitation correlates with a high risk of adverse events, including longer intensive care unit stays, need for mechanical ventilation, and renal replacement therapy [
10]. One meta-analysis of multiple studies demonstrated that prolonged CRT is a reliable predictor of in-hospital mortality in septic patients, particularly those with acute circulatory failure [
11]. The significant correlation between CRT and in-hospital mortality in our study highlights the importance of hemodynamic status early in the course of sepsis. Delayed or inadequate perfusion is a critical indicator of worsening clinical condition [
10], and CRT serves as a timely alert for healthcare providers.
In contrast, the superiority of MS in predicting 7- and 28-day mortality suggests its effectiveness in assessing mid-term outcomes in septic patients. This indicates that although immediate interventions based on CRT are crucial, ongoing assessment with the MS can provide valuable insights into a patient's evolving clinical picture. Mottling, a manifestation of microcirculatory failure, reflects the body's systemic response to sepsis and can be an important indicator of severe disease progression [
9,
12–
14]. The ability of MS to remain a reliable predictor over time can be attributed to its assessment of skin perfusion and overall hemodynamic stability. Studies have demonstrated that a higher MS is associated with worse outcomes in septic shock [
12]. For instance, one study showed that patients with a higher MS had significantly lower skin oxygen saturation and higher mortality at 28 days [
12]. The MS, which ranges from 0 (no mottling) to 5 (severe mottling), can be a quick bedside indicator of the severity of septic shock and help to predict in-hospital mortality. Patients with mottling scores of 3 or more typically have a poor prognosis [
12,
13].
Moreover, the duration of mottling is also a critical factor. Persistent mottling for more than 6 hours is linked with increased in-hospital mortality, emphasizing the importance of early detection and management of circulatory failure in septic shock patients [
14]. However, the MS has limitations, particularly in darker-skinned individuals, and requires training for consistent interpretation [
12–
14].
Both SI and DSI generally correlate with the development of hypotension. Nonetheless, although hypotension is indeed a late feature of shock states, both SI and DSI have been used as potential markers of early sepsis due to their ability to detect circulatory impairment before overt hypotension occurs [
6,
7]. These indexes provide valuable insights into a patient's hemodynamic status, allowing for earlier identification and intervention in septic patients, which is crucial for improving outcomes.
The median age of our study population (85 years) can affect the interpretation of our results. Older patients typically exhibit a more complex clinical picture, including multiple comorbidities that can influence both sepsis presentation and outcomes [
5,
15,
16]. The frailty and diminished physiological performance common in elderly people can also contribute to an increased risk of in-hospital mortality [
15]. Other factors that might affect elderly patients’ prognosis are age-related organ decline and delayed diagnosis due to atypical symptom presentation [
17]. Our findings highlight the need for tailored assessment strategies in these patients, in whom traditional metrics might not fully capture the nuances of sepsis presentation and progression. Additionally, the small sample size raises concerns about the generalizability of our findings. Although our results provide valuable insights, larger studies are needed to validate these findings and enhance their statistical power. The lack of significant differences in the pairwise comparisons indicates that both CRT and the MS are promising tools, but further research with a more substantial cohort could reveal nuanced differences in their predictive capabilities that were not evident in our study.
The clinical implications of our findings are significant. The early identification of patients at high risk of in-hospital mortality is critical to optimize treatment strategies in sepsis. Although we did not find substantial differences in the predictive capabilities of the assessed tools, the strong performance of CRT and the MS suggests that they could serve as complementary assessments. For instance, CRT could be used as an initial rapid assessment, followed by the MS to monitor ongoing risk over time. Incorporating these noninvasive tools into routine clinical practice could help clinicians recognize deteriorating patients and prompt timely interventions. Moreover, their ease of use makes them particularly valuable in resource-limited settings, especially where access to advanced monitoring technologies is restricted.
In the multivariable analysis, each tool demonstrated a significant association with mortality at all three tested time points, indicating that they are valuable predictors of patient outcomes in sepsis. Despite their differences in focus (from cardiovascular function to peripheral perfusion), all four tools consistently identified patients at a high risk of death at specific time points. This result confirms that, in clinical practice, any of these scores can be used effectively to stratify risk in septic patients, supporting timely and targeted interventions to improve survival rates.
Our findings are further supported by additional analyses. The multivariable logistic regression models confirm that each tool independently predicts mortality at all time points, even after adjusting for key confounders such as age, comorbidity burden, and mental status. This reinforces the robustness of their prognostic value in the clinical setting. We also identified optimal cutoff values for each parameter to provide practical thresholds for frontline application and decision-making. The correlation analysis showed moderate interdependence between tools, particularly within the same physiological domain, but no complete overlap, suggesting that combining macrocirculatory (SI, DSI) and microcirculatory (CRT, MS) markers could offer additive prognostic value. A subgroup analysis across elderly age groups confirmed the reproducibility of these trends in even very old populations, reinforcing the potential utility of these tools in geriatric sepsis assessment.
A strength of this study is its exclusive focus on noninvasive tools that require no laboratory testing or specialized equipment. Unlike biomarkers such as lactate, which require blood sampling and laboratory infrastructure, the tools evaluated here can be applied in any clinical setting (including prehospital care) at no cost. This makes them particularly valuable in resource-limited environments, where rapid risk stratification is essential but access to advanced diagnostics might be restricted.
Several limitations of this study warrant discussion. First, the study was conducted in a single center, limiting the generalizability of the findings. Patient populations, healthcare resources, and clinical practices vary significantly across hospitals, especially between countries with differing healthcare systems. Second, our small sample size could affect the statistical power of our results and our ability to detect subtle differences among the prognostic tools. Furthermore, a formal power calculation was not performed. A larger cohort would provide more robust and generalizable results. Finally, the median age of our study population was 85 years. Although that reflects the higher vulnerability of elderly patients to sepsis, it limits the applicability of the findings to younger or middle-aged patients, who can exhibit different physiological responses and outcomes. External validation in more diverse and age-balanced cohorts is warranted to confirm the applicability of our results across broader populations.
Future studies should aim to include larger and more diverse patient populations to further validate our findings. Additionally, our study did not include a control group, which limits some comparative analyses; however, our primary objective was to assess the predictive capacity of these tools in septic patients. Exploring the integration of these tools into clinical decision-making frameworks could provide insights into how they might enhance patient outcomes. Investigating potential combinations of these tools or their use in conjunction with traditional scoring systems (such as the SOFA score) could also yield promising results.
In conclusion, this study highlights the prognostic value of noninvasive tools in the assessment of sepsis mortality. CRT was the parameter with the highest AUC at 24 hours, whereas MS had the highest AUC at 7 and 28 days. Although our findings did not reveal significant differences among the tools, the complementary nature of these scores provides an opportunity to improve clinical practice. Furthermore, because each parameter independently predicted mortality at all three of the tested time points, future studies should explore whether combining the tools (potentially integrating both macro- and microcirculatory indicators) could enhance prognostic accuracy. These findings further support the hypothesis that microcirculatory markers, such as CRT and MS, retain independent prognostic relevance in the early phase of sepsis, even when accounting for comprehensive tools such as the SOFA score and serum lactate. The lack of predictive independence for SI and DSI in this context could reflect their stronger overlap with hemodynamic parameters already embedded in the SOFA score, raising the possibility of multicollinearity in multivariable models. Future studies should explore hybrid assessment strategies that integrate both macrocirculatory and microcirculatory indices to optimize early risk stratification. Further research is needed to enhance our understanding of the roles of these tools in the management of sepsis, particularly in vulnerable populations such as elderly people.
NOTES
-
Author contributions
Conceptualization: MG; Data curation: MG, GM, PB, CG, AZ, SE, MB, MDS; Formal analysis: MG, BP, MDS; Validation: MG, BP, CP, MM, CC, RdeG; Writing–original draft: MG; Writing–review & editing: all authors. All authors read and approved the final manuscript.
-
Conflicts of interest
The authors have no conflicts of interest to declare.
-
Funding
The authors received no financial support for this study.
-
Data availability
Data analyzed in this study are not publicly available due to privacy policy, but they are available from the corresponding author upon reasonable request.
Supplementary materials
Suppl. 1.
Optimal cutoff values and diagnostic performance metrics for each noninvasive tool (SI, DSI, CRT, MS) in predicting 24-hour, 7-day, and 28-day mortality.
ceem-25-075-Suppl-1.pdf
Suppl. 2.
Spearman correlation coefficients between each noninvasive prognostic tool (SI, DSI, CRT, and MS) and the SOFA score.
ceem-25-075-Suppl-2.pdf
Suppl. 3.
Multivariable logistic regression models evaluating the associations between SI, DSI, CRT, and MS and mortality at 24 hours, 7 days, and 28 days. Each model was stratified by SOFA scores and lactate levels.
ceem-25-075-Suppl-3.pdf
Suppl. 4.
Receiver operating characteristic (ROC) curves of SI, DSI, CRT, and MS for predicting 24-hour, 7-day, and 28-day mortality across four predefined age groups: <65, 65–75, 75–85, and >85 years.
ceem-25-075-Suppl-4.pdf
Fig. 1.A graphical representation of the mottling score. As the severity increases, in-hospital mortality rises accordingly.
Fig. 2.Receiver operating characteristic curves for diastolic shock index (DSI), SI, capillary refill time (CRT), and mottling score (MS) in assessing (A) 24-hour mortality, (B) 7-day mortality, and (C) 28-day mortality from sepsis and septic shock. AUC, area under the curve.
Table 1.Main clinical features and outcomes of enrolled patients
Table 1.
|
Characteristic |
Total (n=135) |
24-hr mortality
|
7-day mortality
|
28-day mortality
|
|
Survivors (n=114, 84.4%) |
Nonsurvivors (n=21, 15.6%) |
Survivors (n=101, 74.8%) |
Nonsurvivors (n=34, 25.2%) |
Survivors (n=87, 64.4%) |
Nonsurvivors (n=48, 35.6%) |
|
Age (yr) |
85 (79–90) |
86 (79–91) |
83 (80–89) |
87 (79–91) |
85 (79–90) |
86 (79–90) |
86 (78–90) |
|
Female sex |
60 (44.4) |
49 (43.0) |
11 (52.4) |
39 (38.6) |
21 (61.4) |
31 (35.6) |
29 (64.4) |
|
CCI |
3 (1–4) |
3 (2–5) |
2 (1–3) |
3 (2–5) |
3 (1–4) |
3 (1–4) |
3 (1–5) |
|
SBP (mmHg) |
100 (90–130) |
105 (95–130) |
85 (70–100) |
110 (95–135) |
90 (70–110) |
110 (95–135) |
90 (75–110) |
|
DBP (mmHg) |
60 (50–80) |
60 (60–80) |
50 (40–60) |
60 (60–80) |
50 (40–60) |
60 (60–80) |
60 (40–70) |
|
HR (bpm) |
102 (90–115) |
100 (90–113) |
105 (90–125) |
101 (90–112) |
105 (90–120) |
100 (90–110) |
105 (90–120) |
|
RR (breaths/min) |
24 (20–28) |
24 (20–28) |
28 (22–32) |
24 (20–28) |
24 (20–28) |
24 (18–28) |
27 (22–30) |
|
SpO2 (%) |
93 (90–96) |
94 (91–96) |
92 (90–96) |
94 (91–97) |
93 (90–96) |
94 (91–97) |
93 (90–96) |
|
FiO2 (%) |
0.24 (0.21–0.40) |
0.21 (0.21–0.35) |
0.40 (0.21–0.80) |
0.21 (0.21–0.36) |
0.29 (0.21–0.60) |
0.21 (0.21–0.35) |
0.28 (0.21–0.60) |
|
GCS score |
14 (11–15) |
14 (12–15) |
9 (8–12) |
14 (12–15) |
10 (8–13) |
14 (13–15) |
11 (9–14) |
|
Body temperature (°C) |
38.0 (36.9–38.5) |
38.0 (36.9–38.5) |
37.8 (37.0–38.2) |
37.9 (36.9–38.4) |
38.0 (37.0–38.4) |
38.0 (36.9–38.3) |
37.8 (36.7–38.3) |
|
DSI |
1.5 (1.3–1.9) |
1.5 (1.3–1.8) |
2.0 (1.8–2.9) |
1.5 (1.3–1.8) |
2.0 (1.4–2.6) |
1.5 (1.3–1.8) |
1.8 (1.4–2.5) |
|
SI |
0.9 (0.7–1.1) |
0.9 (0.7–1.1) |
1.2 (1.0–1.9) |
0.9 (0.7–1.1) |
1.1 (0.9–1.6) |
0.9 (0.7–1.1) |
1.1 (0.9–1.5) |
|
CRT (sec) |
3.5 (3.0–5.0) |
3.0 (2.5–5.0) |
6.0 (5.0–8.0) |
3.3 (2.5–5.0) |
5.0 (3.0–7.0) |
3.0 (2.0–4.5) |
5.0 (3.0–7.0) |
|
MS |
1 (1–2) |
1 (0–2) |
3 (2–4) |
1 (0–2) |
3 (1–4) |
1 (0–2) |
2 (1–3) |
|
SOFA score |
5 (4–7) |
5 (3–6) |
8 (6–9) |
4 (3–6) |
8 (6–9) |
4 (3–6) |
7 (5–9) |
Table 2.AUC comparison among DSI, SI, CRT, and MS in predicting 24-hour, 7-day, and 28-day mortality and pairwise comparisons of ROC curves
Table 2.
|
Variable |
AUC |
95% CI |
P-value |
Pairwise comparison of ROC curve P-value
|
|
DSI |
SI |
CRT |
MS |
|
24-hr mortality |
|
|
<0.001 |
|
|
|
|
|
DSI |
0.754 |
0.672–0.824 |
|
- |
0.948 |
0.168 |
0.259 |
|
SI |
0.755 |
0.674–0.825 |
|
0.948 |
- |
0.132 |
0.244 |
|
CRT (sec) |
0.829 |
0.755–0.889 |
|
0.168 |
0.132 |
- |
0.970 |
|
MS |
0.827 |
0.753–0.887 |
|
0.259 |
0.244 |
0.970 |
- |
|
7-day mortality |
|
|
<0.001 |
|
|
|
|
|
DSI |
0.701 |
0.614–0.779 |
|
- |
>0.999 |
0.622 |
0.592 |
|
SI |
0.701 |
0.614–0.779 |
|
>0.999 |
- |
0.606 |
0.579 |
|
CRT (sec) |
0.667 |
0.579–0.748 |
|
0.622 |
0.606 |
- |
0.174 |
|
MS |
0.732 |
0.646–0.806 |
|
0.592 |
0.579 |
0.174 |
- |
|
28-day mortality |
|
|
<0.001 |
|
|
|
|
|
DSI |
0.673 |
0.582–0.756 |
|
- |
0.488 |
0.818 |
0.180 |
|
SI |
0.691 |
0.601–0.772 |
|
0.488 |
- |
0.955 |
0.317 |
|
CRT (sec) |
0.687 |
0.597–0.769 |
|
0.818 |
0.955 |
- |
0.140 |
|
MS |
0.749 |
0.662–0.823 |
|
0.180 |
0.317 |
0.140 |
- |
Table 3.Univariable analysis correlating each of the tested tools with 24-hour, 7-day, and 28-day mortality
Table 3.
|
Variable |
OR (95% CI) |
P-value |
|
24-hr mortality |
|
|
|
DSI |
4.911 (2.182–11.056) |
<0.001 |
|
SI |
13.657 (3.869–48.211) |
<0.001 |
|
CRT (sec) |
1.896 (1.435–2.504) |
<0.001 |
|
MS |
2.371 (1.643–3.422) |
<0.001 |
|
7-day mortality |
|
|
|
DSI |
3.949 (1.907–8.177) |
<0.001 |
|
SI |
8.549 (2.766–26.428) |
<0.001 |
|
CRT (sec) |
1.375 (1.120–1.688) |
0.002 |
|
MS |
1.817 (1.353–2.440) |
<0.001 |
|
28-day mortality |
|
|
|
DSI |
3.638 (1.769–7.481) |
<0.001 |
|
SI |
8.911 (2.790–28.458) |
<0.001 |
|
CRT (sec) |
1.438 (1.171–1.764) |
0.001 |
|
MS |
1.889 (1.399–2.553) |
<0.001 |
Table 4.Multivariable logistic regression models evaluating the associations between each noninvasive tool and mortality at 24 hours, 7 days, and 28 days
Table 4.
|
Model |
24-hr mortality
|
7-day mortality
|
28-day mortality
|
|
OR (95% CI) |
P-value |
OR (95% CI) |
P-value |
OR (95% CI) |
P-value |
|
Shock index |
10.949 (4.065–29.480) |
<0.001 |
7.299 (3.012–17.689) |
<0.001 |
10.379 (4.070–26.467) |
<0.001 |
|
Age (yr) |
1.005 (0.970–1.042) |
0.776 |
1.003 (0.972–1.036) |
0.848 |
1.012 (0.981–1.043) |
0.449 |
|
CCI |
0.990 (0.853–1.148) |
0.890 |
0.992 (0.868–1.113) |
0.901 |
1.135 (0.990–1.301) |
0.070 |
|
Altered mentation |
17.914 (2.283–140.564) |
0.006 |
6.881 (2.240–21.139) |
0.001 |
3.662 (1.596–8.400) |
0.002 |
|
Diastolic shock index |
4.492 (2.402–8.638) |
<0.001 |
3.960 (2.225–7.045) |
<0.001 |
4.116 (2.317–7.312) |
<0.001 |
|
Age (yr) |
1.005 (0.969–1.042) |
0.799 |
1.002 (0.970–1.035) |
0.901 |
1.010 (0.980–1.042) |
0.509 |
|
CCI |
0.993 (0.856–1.152) |
0.928 |
0.992 (0.868–1.113) |
0.901 |
1.122 (0.982–1.283) |
0.092 |
|
Altered mentation |
18.214 (2.341–141.698) |
0.006 |
7.248 (2.342–22.425) |
0.001 |
3.820 (1.659–8.795) |
0.002 |
|
Capillary refill time (sec) |
1.810 (1.451–2.258) |
<0.001 |
1.422 (1.195–1.692) |
<0.001 |
1.358 (1.149–1.605) |
<0.001 |
|
Age (yr) |
1.002 (0.964–1.042) |
0.904 |
0.999 (0.968–1.032) |
0.966 |
1.006 (0.977–1.036) |
0.669 |
|
CCI |
0.923 (0.787–1.080) |
0.321 |
0.957 (0.836–1.095) |
0.523 |
1.096 (0.961–1.249) |
0.172 |
|
Altered mentation |
9.979 (1.277–77.962) |
0.028 |
4.899 (1.613–14.882) |
0.005 |
2.720 (1.214–6.095) |
0.015 |
|
Mottling score |
3.007 (2.117–4.270) |
<0.001 |
2.203 (1.679–2.890) |
<0.001 |
2.103 (1.609–2.748) |
<0.001 |
|
Age (yr) |
0.998 (0.958–1.040) |
0.925 |
0.996 (0.963–1.029) |
0.805 |
1.003 (0.973–1.034) |
0.856 |
|
CCI |
0.936 (0.789–1.111) |
0.450 |
0.964 (0.835–1.114) |
0.622 |
1.122 (0.976–1.291) |
0.106 |
|
Altered mentation |
21.039 (2.199–201.259) |
0.008 |
6.040 (1.847–19.752) |
0.003 |
2.960 (1.259–6.957) |
0.013 |
REFERENCES
- 1. Singer M, Deutschman CS, Seymour CW, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA 2016;315:801-10.
- 2. Evans L, Rhodes A, Alhazzani W, et al. Surviving Sepsis Campaign: international guidelines for management of sepsis and septic shock 2021. Intensive Care Med 2021;47:1181-247.
- 3. Chiu C, Legrand M. Epidemiology of sepsis and septic shock. Curr Opin Anaesthesiol 2021;34:71-6.
- 4. World Health Organization (WHO). Global report on the epidemiology and burden of sepsis: current evidence, identifying gaps and future directions. WHO; 2020.
- 5. Guarino M, Perna B, Cesaro AE, et al. 2023 Update on sepsis and septic shock in adult patients: management in the emergency department. J Clin Med 2023;12:3188.
- 6. Diaztagle Fernández JJ, Castañeda-González JP, Trujillo Zambrano JI, Duarte Martínez FE, Saavedra Ortiz MÁ. Assessment of the shock index in septic shock: a systematic review. Med Intensiva (Engl Ed) 2024;48:e10-9.
- 7. Jeon Y, Kim S, Ahn S, et al. Predicting septic shock in patients with sepsis at emergency department triage using systolic and diastolic shock index. Am J Emerg Med 2024;78:196-201.
- 8. Kattan E, Bakker J, Estenssoro E, et al. Hemodynamic phenotype-based, capillary refill time-targeted resuscitation in early septic shock: the ANDROMEDA-SHOCK-2 randomized clinical trial study protocol. Rev Bras Ter Intensiva 2022;34:96-106.
- 9. Preda G, Bourcier S, Joffre J, et al. Mottling score is associated with 28-day mortality in critically ill patients with sepsis. Minerva Anestesiol 2017;83:664-6.
- 10. Castro R, Kattan E, Ferri G, et al. Effects of capillary refill time- vs. lactate-targeted fluid resuscitation on regional, microcirculatory and hypoxia-related perfusion parameters in septic shock: a randomized controlled trial. Ann Intensive Care 2020;10:150.
- 11. Jacquet-Lagrèze M, Pernollet A, Kattan E, et al. Prognostic value of capillary refill time in adult patients: a systematic review with meta-analysis. Crit Care 2023;27:473.
- 12. Kazune S, Caica A, Volceka K, Suba O, Rubins U, Grabovskis A. Relationship of mottling score, skin microcirculatory perfusion indices and biomarkers of endothelial dysfunction in patients with septic shock: an observational study. Crit Care 2019;23:311.
- 13. Ferraris A, Bouisse C, Mottard N, et al. Mottling score and skin temperature in septic shock: relation and impact on prognosis in ICU. PLoS One 2018;13:e0202329.
- 14. de Moura EB, Amorim FF, da Cruz Santana AN, et al. Skin mottling score as a predictor of 28-day mortality in patients with septic shock. Intensive Care Med 2016;42:479-80.
- 15. Ibarz M, Haas LE, Ceccato A, Artigas A. The critically ill older patient with sepsis: a narrative review. Ann Intensive Care 2024;14:6.
- 16. Stenberg H, Li X, Pello-Esso W, et al. The effects of sociodemographic factors and comorbidities on sepsis: a nationwide Swedish cohort study. Prev Med Rep 2023;35:102326.
- 17. Bouza C, Martínez-Alés G, López-Cuadrado T. The impact of dementia on hospital outcomes for elderly patients with sepsis: a population-based study. PLoS One 2019;14:e0212196.
, Giacomo Maroncelli1, Benedetta Perna1, Paolo Baldin1, Caterina Ghirardi1,2, Alex Zanotto1, Silvia Eichner1, Matteo Bolognesi1, Chiara Pesci2, Martina Maritati3, Carlo Contini3, Roberto De Giorgio1
