1 Introduction

Dyspnea is one of the most common reasons for admission to the emergency department [1]. Oxygen saturation has great importance in determining the triage status of patients admitted to the hospital with dyspnea and planning the emergency treatment [2].

Peripheral perfusion index (PI), which shows tissue oxygenation is a noninvasive way of demonstrating tissue perfusion in critically ill patients. Studies have shown that PI is an accurate, fast and reliable pulse oximetry-based indicator of tissue perfusion [3,4,5]. PI shows the perfusion status of the tissue in the applied area for an instant and a certain time interval. The PI value ranges from 0.02% (very weak) to 20% (strong) [6].

Triage scales are used to distinguish emergency and non-emergency patients. The emergency triage system is used to quickly determine the care priorities of patients during admission to the emergency department [7, 8].

It is important to make the triage classification for dyspnea in emergency services quickly and accurately to start the treatment protocols early as possible. Perfusion index has an important role in the determination of tissue hypoxia and peripheral perfusion [9]. In our study, we aimed to determine the relationship between perfusion index and the emergency triage classification in patients admitted to the emergency department with dyspnea.

2 Materials & methods

This study was conducted prospectively on 2471 patients who were admitted to the tertiary University Hospital Emergency Medicine Clinic between January 1, 2019 and January 1, 2021 with dyspnea. Ethics committee approval and clinical trial number were obtained for the study. (Accept Number:10.10.2018 and 20.478.486- ClinicalTrials.gov Identifier: NCT05742113)

Adult patients over the age of 18 who presented with dyspnea and whose perfusion index values ​​were measured at the time of first admission (arrival time), at the first hour and the second hour of admission were included in the study. Patients who applied due to the mechanism of trauma, patients with known vascular disease (Buerger’s disease, peripheral artery disease, etc.), patients with COVID-19 PCR positivity and whose perfusion index measurement could not be completed due to hospitalization or discharge were excluded from the study. The triage classification ( red, yellow, green codes) was made according to the regulation on the Implementation Procedures and Principles of Emergency Services in the Inpatient Health Facilities of the Ministry of Health [10,11,12]. Masimo Radical 7 pulse oximetry device (Masimo Corporation, Irvine, CA) measuring non-invasive carboxyhemoglobin was used for perfusion index measurement. The patient’s vital signs(heart rate, blood pressure, respiratory rate, and temperature), perfusion index and oxygen saturation, clinical status information were recorded at the first hour and second hour of his arrival to the emergency department, while the patient’s treatment, consultation, and examination processes were being processed. Oxygen saturations measured with the probe were noted into the file. Perfusion measurements were made from the middle finger of the same side, with the patient’s hand at heart level and room temperature.

The diagnosis of the patients was recorded in the study form. The results were evaluated statistically. The perfusion index and oxygen saturation measured by finger probes were compared and the accuracy in detecting the emergency triage classification was compared.

2.1 Statistical analysis

Study data were analyzed using the Number Cruncher Statistical System 2007 Statistical Software package program (NCSS, LLC, Kaysville, Utah). Descriptive statistical methods (mean, standard deviation, median, frequency, percentage, minimum, maximum) were used. The conformity of the quantitative data to the normal distribution was evaluated with the Shapiro-Wilk test and graphical examinations. Mann-Whitney U test was used for comparisons between two groups of quantitative variables that did not show normal distribution. Kruskal-Wallis test and Dunn-Bonferroni test were used for comparisons between groups of more than two quantitative variables that did not show normal distribution. The Friedman Test was used for in-group comparisons of quantitative variables that did not show normal distribution, and the Wilcoxon signed ranks test with Bonferroni correction was used for the evaluation of pairwise comparisons. Pearson chi-square test and Fisher- Freeman- Halton exact test were used to compare qualitative data. Spearman correlation analysis was used to evaluate the relationships between quantitative variables. ROC Curve analysis was used to determine the sensitivity of Red Code Triage. Statistical significance was accepted as p < 0.05.

3 Results

We excluded 981 patients from the study and it was completed on 1490 patients. Of the cases participating in the study, 34.2% (n = 510) were female and 65.8% (n = 980) were male. It was determined that the triage code was yellow in 64.4% (n = 960), and red in 35.6% (n = 530) of the cases.

Arrival perfusion index, 1st -hour and 2nd -hour perfusion indexes of cases with yellow triage code were found to be statistically significantly higher than those with red triage code. (p = 0.001; p < 0.05). A statistically significant difference was found between the perfusion index measurements at the time of arrival, 1st- hour, and 2nd -hour of the cases whose triage code was red (p = 0.001; p < 0.05).

We found statistically significant difference between the arrival to hospital, 1st hour and 2nd hour perfusion index measurements of the cases according to the outcomes of patients. According to the results of the pairwise comparison made to determine the difference; the arrival, 1st -hour and 2nd- hour perfusion index of the cases who were sent to the intensive care unit were found to be significantly lower than the cases who were discharged and admitted to the service. Likewise, the 1st hour and 2nd hour perfusion indexes of the patients who were admitted to the ward were found to be significantly lower than those who were discharged (p = 0.033, p = 0.011; p < 0.05) (Table 1).

Table 1 Evaluation of Perfusion Indices by Triage Code and Outcome of Patients
Full size table

For the 0.9 cut- off value of the arrival PI level according to the triage status; sensitivity 79.25%; specificity 78.12%; positive predictive value is 66.7 and negative predictive value is 87.2. A statistically significant correlation was found between the triage status and the 0.9 cut- off value of the admission PI level (p = 0.001; p < 0.05). We can say that the ODDS rate of red triage is 13.63 times (95% CI: 5.99–31.01) times higher in cases with a PI level of 0.9 and below (Fig. 1).

Fig. 1
figure 1

ROC Curve of Arrival perfusion index by Triage Status

Full size image

In the ROC analysis, the cut-off value of 1.1 and above the admission PI level was determined as the most appropriate point for discharge. Sensitivity at this cut-off value is 83.5% (95% CI:71-91.6); specificity 73.3% (95%CI:63-82.1) ; positive predictive value is 67.1% ( 95%CI:55.1–77.7) and negative predictive value is 86.8% (95%CI:77.1–93.5). The area under the ROC curve obtained was found 82.0% ( 95%CI:0.748–0.878) with a standard error of 3.5%. We obtained a statistically significant correlation between discharge status and ≥ 1.1 cut-off value of admission PI level (p = 0.001; p < 0.01). We can say that the rate of being discharged from the hospital is 14.11 times (95% CI: 5.92–33.65) more in cases with a PI level of 1.1 and above than under the level 1.1 (Table 2).

Table 2 Diagnostic Screening Tests for Arrival PI according to discharge status
Full size table

A statistically significant difference was found between admission PI measurements according to outcome (p < 0.01); arrival PI measurements of the patients admitted to the intensive care unit were found to be significantly lower than those who were discharged and hospitalized (Table 3) (Fig. 2).

Table 3 Evaluation of Arrival PI Measurements According To The Outcome Status
Full size table
Fig. 2
figure 2

Distribution of Arrival PI Measurements by Outcome

Full size image

Arrival SpO2, 1st hour and 2nd hour values ​​were found to be statistically significantly higher in cases with yellow triage code than in cases with red triage code (p = 0.001, p = 0.002, p = 0.012, p < 0.05, respectively).

We found a statistically significant difference between arrival, 1st hour and 2nd hour SpO2 measurements in cases with red triage code (p = 0.001; p < 0.05).

We found a statistically significant difference between the arrival, 1st hour and 2nd hour SpO2 measurements according to the outcome type (p = 0.001; p < 0.01). According to the results of the pairwise comparison made to determine the difference; the finger probe value of the discharged cases was found to be significantly higher than the cases sent to the service and taken to the intensive care unit (p = 0.001; p = 0.001; p < 0.05) (Table 4).

Table 4 Evaluation of Finger Probe Measurements by Triage Code, Diagnosis and Outcome
Full size table

When we look at the areas under the ROC curve as to which of the PI and Pulse measurements related to defining the triage code as red, the PI on arrival is 82%; the pulse device at the time of admission was 79.8% and no significant difference was found between the two.

When we look at the areas under the ROC curve to predict discharge, which of the PI and pulse measurements is better, the PI on arrival is 82.4%; On the other hand, the rate of pulse device at admission was 75.2%, and no significant difference was found between the two (p = 0.154; p > 0.05). However, in both cases, the predictive level of PI measures for definition was higher.(Fig. 3).

Fig. 3
figure 3

ROC curve of PI-SpO2 predicting discharge

Full size image

A weak positive correlation (perfusion index increases as the pH/pO2/SpO2 value increases) was statistically significant between the arrival pH, pO2 and SpO2 measurements and the arrival perfusion index measurements of the subjects participating in the study (r = 0.258,p = 0.002 / r = 0.301, p = 0.001 / r = 0.369, p = 0.001 inrespectively).

There was a weak negative correlation (the perfusion index decreases as the lactate/pCO2 value increases) between the arrival lactate, pCO2 measurements and the admission perfusion index measurements of the subjects participating in the study (r=-0.316, p = 0.001 / r = 0.217, p = 0.008 inrespectively).

No statistically significant correlation was found between the admission HCO3 measurements and the admission perfusion index measurements of the cases participating in the study(p > 0.05).

4 Discussion

Dyspnea is a serious symptom seen in half of the patients admitted to tertiary health institutions and in a quarter of the patients receiving outpatient treatment [13]. Most of the patients with dyspnea apply to the emergency services and their treatment is initially started in the emergency services. Patients hospitalized directly from the emergency room to the intensive care unit have a better prognosis than the patients hospitalized from the ward to the intensive care unit [14]. This depends on effective and prompt triage and intervention.

In the emergency department, there is a need for effective methods that will help to unterstand the severity of the patient’s condition and the decision of triage. Oxygen saturation is a very important parameter in emergency conditions and the medical evaluation of critically ill patients [4]. Fingertip probes (pulse oximetry) and blood gas measurement are used in emergency services to display oxygen saturation.

Arterial blood gas measurement is an important invasive method used to evaluate patients presenting with dyspnea. Arterial blood gas measurement helps to understand the pathophysiology and mechanism of respiratory failure, the degree of compensation, the definition and monitoring of the acid-base state, the depth of dyspnea and the amount of oxygen that should be given to the patient [15]. Arterial blood gas measurement, which is another marker used to show oxygen saturation, helps to show the depth of dyspnea. However, since it is an invasive method, it has little role in determining the triage category in the first application.

Fingertip probes (pulse oximetry) are effective clinical tools used to display patients’ oxygen status. The advantages of pulse oximetry are safe, painless, easy to use, non-invasive, and have quick result. These features make the pulse oximeter valuable in evaluating oxygen saturation, planning the patient’s treatment, and triage classification [16]. However, it may be affected by various conditions such as movement, tremor, vasoconstriction, septic shock, and hypothermia [3]. Due to such factors, the degree of desaturation of the patient cannot be determined and the intervention may be delayed or not performed. Even with the continuous alarm signal of the pulse oximeter, emergency service or intensive care workers may lose their sensitivity to this sound over time, which may cause delays in understanding the seriousness of the situation. In a study conducted in the intensive care unit, an accuracy of 66% oxygen saturation was observed with pulse oximetry; and 89% with the PI measuring device [3]. In the studies, it is stated that the PI measuring device system is a more accurate, faster and reliable saturation indicator compared to the pulse oximeters in routine use [3, 4].

In our study, it was found that fingertip probe values ​​were powerful in showing the triage code of the patients. Fingertip saturation values ​​at admission, 1st- hour and 2nd- hour were found to be significantly higher in patients with yellow triage code compared to patients with red triage code. The increase in the 1st hour and 2nd hour values ​​is statistically significant according to the arrival value in the patients with the triage code yellow. We think that the reason for this is closely related to the treatments applied to the patient. The increase in the 1st- hour and 2nd- hour values ​​compared to the arrival fingertip saturation values ​​in patients with a red triage code is statistically significant. We think that this may be related to the rapid respiratory airway management and treatment of patients in the red triage category.

In recent years, the use of peripheral perfusion index has gained popularity as a noninvasive and easily measurable parameter, as it indicates impaired peripheral perfusion in critically ill patients. In a study conducted with newborns, it was stated that oxygen saturation of patients could be measured accurately and quickly without using any invasive method within 27 s, faster than other methods with PI measurement device [4].

Studies on the predictor of peripheral PI on hospitalization and mortality in emergency room patients are very limited in the literature. Oskay et al. found in their study that PI was an insignificant parameter in terms of hospitalization and mortality estimation in emergency room patients in the critical area [17]. Lima et al. examined whether decreased peripheral perfusion in critically ill patients reflects the clinical signs and showed that the PI value of 1.4 and below in critically ill patients is a strong indicator of impaired perfusion [5]. In our study, we found that the impaired peripheral PI value was statistically significant in determining critically ill and intensive care admissions similar to Lima et al. However, the cut-off value was found to be 0.9, and a different value from the literature was determined. We think that the cut-off value is different because the number of our patients is higher than the publications in the literature.

It is very important to determine the triage category at the first examination, especially in patients who present to the emergency department with dyspnea. It is seen that there are still disruptions and researches on this subject. We could find one study about the use of the perfusion index in the classification of triage in patients presenting to the emergency department. While this publication generally makes a general triage classification; no specific publication could be found on patients presenting with dyspnea [17]. The main purpose of our study is to evaluate the role of PI measurement in triage classification.

When PI measurement in patients presenting with dyspnea is compared with other screening tests, the sensitivity for the 0.9 cut-off of the admission perfusion index level is 79.25%, the specificity is 78.12%, the positive predictive value is 66.7, and the negative predictive value is 87.2. The arrival perfusion index level was found to be statistically significant according to triage status. The incidence of red triage was found to be 13.63 times higher in cases with perfusion index level of 0.9 and below. This result shows the importance of PI in determining the triage category.

4.1 Study limitations

The most important limitation of our study is that the relationship between the perfusion index and the need for intubation was not examined. It is thought that more comprehensive studies are needed to evaluate the effect of peripheral perfusion index measurement in intensive care patients and the relationship between endotracheal intubation and PI.

In Conclusion, the perfusion index can help to determine the triage classification in emergency departments for dyspnea.