Absolute Eosinophil Counts as a Prognostic Marker in Patients with Sepsis

Abstract

Background:

Sepsis is a leading cause of mortality and morbidity in critically ill patients. It is necessary to have markers of severity that are easily accessible and useful to guide treatment in a timely manner. Eosinophil count could be a potential biomarker in sepsis, and it is routinely checked in clinical practice.

Aims and Objectives:

To assess absolute eosinophil count (AEC) as prognostic marker in patients with sepsis.

Methodology:

This observational study was conducted in a tertiary care hospital in South India. A total of 100 patients admitted with sepsis were included. AEC and Sequential Organ Failure Assessment (SOFA) score were calculated at admission and after 72 h. AEC was correlated with the SOFA score. These patients were clinically followed up during their hospital stay. A receiver operating characteristic curve was developed to determine the optimum AEC cutoff point for predicting mortality.

Results:

Decreasing trend of AEC during the course of hospital stay (after 72 h) of admission was found to have a strong negative correlation with SOFA score. AEC cutoff <50 cells/mm³ after 72 h of admission was associated with increased mortality. Low AEC after 72 h of admission and decreasing trend of AEC were associated with increased risk of requirement of ionotropic support, dialysis, ventilator, and mortality.

Conclusion:

A decline in AEC after 72 h of admission was linked to increased mortality. Therefore, eosinophil count can be used as a cost-effective marker for assessing severity and prognosis in patients with sepsis.

INTRODUCTION

Sepsis is a clinical syndrome that has physiological, biological, and biochemical abnormalities caused due to dysregulated host response to the infection.[1,2] Sepsis and the associated inflammatory response can lead to multiple organ dysfunction syndrome and death.[2,3] As per the Third International Consensus Definitions of Sepsis and Septic Shock (Sepsis 3) by European Society of Intensive Care Medicine (ESICM) - SCCM, sepsis is defined as life-threatening organ dysfunction due to dysregulated host response to infection.[4,5] The diagnostic tool for sepsis is the Sequential Organ Failure Assessment (SOFA) score. Clinical criteria of sepsis are suspected infection with acute organ dysfunction defined as an increase of 2 points or more in the SOFA score. Sepsis and septic shock are increasing in incidence and contributing significantly to morbidity and mortality. Prediction of the outcome for patients with sepsis will facilitate more aggressive interventions. Sepsis is the foremost cause of admission into the Indian Intensive Treatment Units (ITUs) and is the major cause of mortality.[6,7] Hence, it is necessary to have prognostic markers of severity that are increasingly accessible and useful to guide the treatment in a timely manner.

The biomarkers such as C-reactive protein and procalcitonin have been most widely studied and used in patients with sepsis. These biomarkers have limited sensitivity and specificity with high cost and may be raised in other inflammatory diseases.[8] Evidence demonstrates the usefulness of eosinophil count as the prognostic marker of severity. Eosinopenia in acute infection can be due to acute stress mediated by adrenal glucocorticoids and epinephrine. Sequestration could be due to the migration of eosinophils into the inflammatory site by chemotactic mediators released during acute inflammation. The major chemotactic mediator is C5a and fibrin fragments.[8] It has also shown that eosinopenia may be associated with a higher mortality rate. This study was undertaken to assess absolute eosinophil count (AEC) as prognostic marker in patients with sepsis.

METHODOLOGY

Study design

This study was a prospective observational study.

Study setting

This study was conducted at a tertiary care center in South India from February 2021 to October 2022.

Source of data

A total of 100 patients with sepsis who fulfilled both inclusion and exclusion criteria.

Inclusion criteria

Adult patients (>18 years) diagnosed with sepsis according to Sepsis 3 guidelines.

Exclusion criteria

• Pregnant females

• Patients who died or discharged within 72 h of admission

• Patients currently on chemotherapy

• Patients with HIV, on glucocorticoid medications

• Patients with bronchial asthma, hay fever, atopic dermatitis, and allergic conjunctivitis.

Data collection

A total of 100 patients admitted with sepsis were included in the study according to Sepsis 3 guidelines.

Detailed clinical history and examination were done. Cultures of blood/urine/craniospinal fluid/respiratory secretion/appropriate specimen according to clinical circumstances, AEC and SOFA score were calculated at admission and after 72 h. These patients were clinically followed up till their stay in the hospital. The eosinophil count was correlated with the SOFA score. Sensitivity, specificity, and positive and negative likelihood ratio with 95% confidence interval were carried out at various cutoff levels. Receiver operating characteristic (ROC) curve was developed to find out the optimum cutoff points among survivors and nonsurvivors. Patients were divided into two groups, survivors and nonsurvivors. Eosinophil counts at admission and after 72 h were studied and compared. Blood was collected in K3 EDTA 5.4 mg tube, approximately 3 ml of blood was required. AEC was done by automated method using Beckman coulter machine. If values were high, manual method was used to calculate AEC.

Sample size calculation

A study entitled “Eosinopenia as an indicator for organ dysfunction” conducted by Syafri Arif et al.[9] has revealed that of 34 patients admitted to the study, 22 patients (64.7%) are likely to die during the course of 7 days of follow-up. Further, it has been indicated that the mean and standard deviation of eosinophil count among the dead and survived patients to be 0.046 ± 0.59 and 0.3 ± 0.55. Based on the above finding of the study, with an alpha error of 1% and power of 99%, it was estimated that at least 87 patients to be included in the study.

Statistical analysis

Categorical data were represented in the form of frequencies and proportions. Chi-square test or Fischer’s exact test (for 2 × 2 tables only) was used as the test of significance for qualitative data. Continuous data were represented as mean and standard deviation. Independent t-test was used as the test of significance to identify the mean difference between two quantitative variables. Correlations were performed with the Pearson’s correlation coefficient. ROC curves were constructed for AEC and mortality. ROC and optimal cutoff points were chosen for the calculation of sensitivity, specificity, and positive and negative predictive values. A test that predicts an outcome no better than chance has an area under the ROC curve of 0.5. An area under the ROC curve above 0.8 indicated good prediction. Graphical representation of data: MS Excel and MS Word were used to obtain various types of graphs (probability that the result is true). P < 0.05 was considered as statistically significant after assuming all the rules of statistical tests. Statistical software: MS Excel, SPSS version 22 (IBM SPSS Statistics, Somers NY, USA) was used to analyze data.

Ethical clearance

The study was conducted after the approval of the institutional ethics committee with reference number MSRMC/EC/PG-27/01-2021. Informed consent was obtained from the patients.

RESULTS

Participants

A total of 100 patients admitted with sepsis were included in the study according to Sepsis 3 guidelines.

Descriptive data

The study included 100 patients, 64% (64) of the subjects were males and 36% (36) were females. The majority of the subjects in our study were in the age group of 51–60 years (30%) with the mean age of 58.72 years [Table 1]. The majority of the patients were diabetic 55 (55%) and hypertensive 51 (51%) in our study [Table 2]. The most common source of sepsis in our study was urinary tract infection in 38%, followed by respiratory infections in 29%, meningitis in 15%, intra‐abdominal infections in 10%, and other sources (8%) such as ulcers, wounds, pelvic inflammatory disease, skin infection, and genital infection [Table 3]. Of 100 patients in our study, 66% of the patients required ionotropic support, 25% of patients required dialysis, and 39% patients required ventilator support. The mortality rate in our study was 37%. AEC counts were calculated and compared among survivors and nonsurvivors. At admission, the mean AEC was 175.97 cells/mm³ and SD of 72.81 in the patients who survived and mean AEC of 130.54 cells/mm³ and SD of 55.83 in nonsurvivors (P < 0.001). After 72 h, the mean AEC was 167.17 cells/mm³ and SD 103.49 in the patients survived and the mean AEC of 40.62 cells/mm³ and SD of 12.84 in patients who died with (P < 0.001). There was also statistically significant difference found between mortality and AEC trend, i.e., 53.7% of patients with decreasing AEC trend died compared to 3% of patients with increasing AEC trend [Table 4].

Table 1.

Distribution of subjects based on age group

Age group (years)	Frequency (%)
<30	6 (6.0)
31–40	5 (5.0)
41–50	15 (15.0)
51–60	30 (30.0)
61–70	20 (20.0)
>70	24 (24.0)
Total	100 (100.0)

Table 2.

Frequency distribution of comorbidities among the subjects

Comorbidities	Frequency (%)
Diabetes	55 (55.6)
Hypertension	51 (51.0)
IHD	22 (22.0)
CKD	15 (7.0)
CVA	13 (15.0)
Thyroid	10 (13.0)
Others	7 (25.0)

IHD=Ischemic heart disease, CKD=Chronic kidney disease, CVA=Cerebrovascular accident

Table 3.

Distribution of subjects based on source of sepsis

Source of sepsis	Frequency (%)
Urinary	38 (38.0)
Respiratory	29 (29.0)
Intra-abdominal	15 (15.0)
Meningitis	10 (10.0)
Others	8 (8.0)

Table 4.

Comparison of absolute eosinophil count based on mortality

Outcome	AEC (at admission), mean±SD	AEC (after 72 h), mean±SD
Survivors	175.97±72.81	167.17±103.49
Nonsurvivors	130.54±55.83	40.62±12.84
P	<0.001	<0.001

SD=Standard deviation, AEC=Absolute eosinophil count

There was a negative weak correlation between SOFA score and AEC at admission, which was not statistically significant. However, there was a negative strong correlation between SOFA score and AEC after 72 h in our study, which is statistically significant with P < 0.01 and r value of − 0.736 [Figure 1 and Table 5]. ROC curve was plotted in our study to find out the cutoffs of AEC. The cutoff AEC at admission for predicting mortality was < 174 cells/mm³ with a sensitivity of 81.08%, specificity of 53.87%, AUC 0.705, 95% CI, P < 0.001, and cutoff AEC after 72 h for predicting mortality was <50 cells/mm³ with a sensitivity of 94.59%, specificity of 79.37%, AUC 0.923, 95% CI, P < 0.001 [Figures 2, 3 and Table 6]. There was a statistically significant difference found between the need of inotropes and AEC after 72 h with a mean AEC of 66.5 cells/mm³ in the patients requiring inotropic support and mean AEC of 224.79 cells/mm³ in patients not requiring inotropic support with P < 0.001. There was no statistically significant difference found between the need for dialysis and AEC at admission. However, there was a statistically significant difference found between the need of dialysis and AEC after 72 h with a mean AEC of 54.8 cells/mm³ in the patients requiring dialysis and mean AEC of 142.2 cells/mm³ in patients not requiring dialysis with P < 0.001. There was a statistically significant difference found between the need of ventilator support and AEC at both admission and after 72 h. At admission, the mean AEC was 139.31 cells/mm³ in the patients requiring ventilator support and the mean AEC of 171.85 cells/mm³ in patients not requiring ventilator support. After 72 h, the mean AEC was 50.46 cells/mm³ in the patients requiring ventilator support and the mean AEC of 165.03 cells/mm³ in patients not requiring ventilator support with P < 0.001.

Figure 1.

Scatter plot showing the correlation between absolute eosinophil count and Sequential Organ Failure Assessment score at admission and after 72 h. SOFA: Sequential Organ Failure Assessment

Table 5.

Correlation of sequential organ failure assessment score and absolute eosinophil count at admission and after 72 h

Correlation of SOFA score and AEC at admission and after 72 h
Parameters	Correlation coefficient
SOFA and AEC (at admission)
Pearson correlation (r)	-0.108
P	0.284
SOFA and AEC (after 72 h)
Pearson correlation (r)	-0.736**
P	<0.001

**P < 0.05 is considered statistically significant. SOFA=Sequential organ failure assessment, AEC=Absolute eosinophil count

Figure 2.

ROC curve for AEC at admission for predicting Mortality

Figure 3.

ROC curve for AEC after 72 hours of admission for predicting Mortality

Table 6.

Area under receiver operating characteristic curve, optimal cutoff points, specificity, sensitivity, positive predictive value, and negative predictive value for absolute eosinophil count at admission and after 72 h for predicting mortality among patients with sepsis

	AUC	95% CI	P	Cutoff	Sensitivity	Specificity	PPV	NPV
AEC at admission	0.705	0.606–0.792	<0.001	≤174	81.08	53.97	50.8	82.9
AEC after 72 h	0.923	0.853–0.967	<0.001	≤50	94.59	79.37	72.9	96.2

AUC=Area under curve, PPV=Positive predictive value, NPV=Negative predictive value, CI=Confidence interval, AEC=Absolute eosinophil count

DISCUSSION

Eosinophils are considered as an integral part of the immune and inflammatory network and homeostatic regulation, all of which play an indispensable role in the pathophysiology of sepsis. In our study, we assessed AEC in predicting the prognosis of patients with sepsis and septic shock. We studied the correlation between AEC and SOFA scores in predicting the outcome.

Our study included 100 patients, 64% (64) of the subjects were males and 36% (36) were females. The most common source of sepsis in our study was urinary tract infection (38%), followed by respiratory infections (29%), and meningitis (15%), which is like the study by Joy et al.,[10] where the most common source of infection was urinary tract infection (33%), followed by respiratory infection (23%).

The mortality rate in our study was 37% (37) which is comparable with the mortality rates in the studies done by Joy et al.[10] (44%) and Abidi et al.[11] (33%). AEC at admission and after 72 h were correlated with SOFA score. There was a negative strong correlation between SOFA score and AEC after 72 h in our study. This is comparable to the study by Arif et al.[9] which also has a negative correlation between SOFA score and eosinophil count after 72 h with P = 0.043 and r value of − 0.350. This is also like the study by Tinoco-Sanchez et al.[12] which had a strong negative correlation between SOFA score and eosinophil count after 72 h with P < 0.01. ROC curve was plotted in our study to find out the cutoffs of AEC. The cutoff AEC at admission for predicting mortality was <174 cells/mm³ with AUC 0.705, 95% CI, P < 0.001, and cutoff AEC after 72 h for predicting mortality was <50 cells/mm³ with AUC 0.923, 95% CI, P < 0.001. This is comparable with the studies by Merino et al.[13] and Joy et al.,[10] in which the AEC cut off <50 cells/mm³ was found to be associated with increased mortality rate with the sensitivity of 89% and 20.8% and specificity of 56.4% and 95.65%, respectively. This is also like the study by Abidi et al.,[11] where eosinophil count cut off <50 cells/mm³ was associated with severe sepsis and septic shock with the sensitivity of 80% and specificity of 91%. Whereas, in the study by Wilson et al.,[14] the AEC cutoff <198.5 cells/mm³ was associated with increased mortality with the sensitivity of 75% and specificity of 38.1%.

To the best of our knowledge, our study is the first to assess the correlation between AEC and individual outcomes including the need of inotropic support, dialysis, and ventilator. This is like study by Arif et al.,[9] in which patients with organ dysfunction had decreased eosinophil count (eosinopenia) after 72 h in comparison with patients without organ dysfunction with the sensitivity of 92.8%, specificity of 66.6%, and P < 0.001.

There was a statistically significant difference found between mortality and AEC at both admission and after 72 h in our study with the mean AEC of 130.54 cells/mm³ at admission and 40.62 cells/mm³ after 72 h in patients who died with P < 0.001, which is comparable to study by Tinoco-Sanchez et al.[12] where the mean AEC after 72 h was 61.5 ± 96.1 in patients who died.

The comparison was also done between AEC trend (AEC at admission and AEC after 72 h). There was a statistically significant difference found between mortality and AEC trend, i.e., 53.7% of patients with decreasing AEC trend died compared to 3% of patients with increasing AEC trend which is like the study by Arif et al.[9] in which a decrease in eosinophil count from admission to after 72 h was associated with increased mortality compared to the patients with an increase in eosinophil count.

Eosinopenia can happen in acute infections. Eosinophils are activated by type 2 inflammation, their absence (eosinopenia) may indicate immune imbalance. Furthermore, eosinopenia may be the consequence of eosinophil consumption and eosinophils may contribute to dysregulated host response in infection. Eosinophil recruitment into the inflamed tissue causes tissue damage by generating oxidative stress.[11,13]

Thus, eosinophil count can be used as an inexpensive marker to assess prognosis in sepsis and as a useful tool in guiding physicians in the management of sepsis.

The limitation of the study is relatively smaller sample size and the study done in a single center. Large-scale and multicentric studies are further needed to generalize the results.

CONCLUSION

A decreasing trend in AEC is associated with poor outcomes among patients with sepsis. Falling eosinophil counts after 72 h of admission were found to be associated with increased mortality. Thus, eosinophil count can be used as an inexpensive and easily accessible marker of sepsis severity, which can guide physicians in managing patients with sepsis, especially in resource-limited settings.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.