Exploration of the Clinical Parameters as Predictive Biomarkers and Their Potential Roles in the Regulation of Inflammation in Elderly Septic Patients with Pneumonia

Introduction

The global prevalence of infection in intensive care units (ICU) is estimated to be 30%. As an acute infection of the lung parenchyma, pneumonia is characterized by alveolar inflammation resulted from a predominantly bacterial or viral pathogen and listed as one of the vital causes of morbidity and mortality around the world.¹ It may cause sepsis and septic shock in the hospitalized patients, with a dramatically increased short-term mortality rate of up to 50% in ICU patients.^2,3 Sepsis is a clinical condition defined as life-threatening organ dysfunction caused by a dysregulated host response to infection.⁴ As a major public health problem worldwide with a high incidence, sepsis remains one of the most prominent causes of death.^4,5 Age is an independent predictor for mortality in septic patients, especially for those over 65 years.⁶ Elderly patients own fewer specific symptoms and lower immunity during sepsis, which result in higher mortality.⁷ Early diagnosis and treatment of sepsis can reduce the patient mortality.⁴ As pneumonia is also an important risk factor for sepsis, understanding the characteristics of sepsis complicated by pneumonia, is of great value for elderly patients in the clinical setting.⁸

Several parameters have been shown to be correlated with community-acquired pneumonia (CAP) or sepsis. Hemoglobin (HGB) plays a vital role in oxygen delivery, tissue perfusion and the management of critically ill patients with sepsis.^9,10 As an indicator for impaired immunological function, low levels of HGB have been observed in about 30% of patients with CAP and found to be related to increased mortality in hospitalized patients with CAP.^11,12 Albumin (ALB) acts as an antioxidant scavenger in the tissues and contributes to nearly three-quarters of antioxidant capacity in the serum.^13,14 It breaks down at an accelerated rate in inflammatory conditions and has been shown to be closely correlated with inflammation with respect to that hypoalbuminemia was frequently seen in patients with inflammatory diseases.^15–17 Serum lipids are critical in cell-building and changes in lipid metabolism correlate with inflammation.¹⁸ Total cholesterol (CHOL) metabolism has been shown to be directly regulated by the interferon cytokine response, with upstream and downstream metabolites of the pathway directing immune effector functions and anti-infective activity.^19,20 Neutrophil (NEU) plays vital roles in the innate immune response to infection and its function is impaired with age. The aberrant NEU function seen in the elderly is further impaired by pulmonary infection, where increased severity of infection is associated with enhanced impairment of NEU function.²¹ A number of ratio indicators, including neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), glucose-to-lymphocyte ratio (GLR), and lymphocyte-to-monocyte ratio (LMR), have been investigated for the prognosis of critically ill patients with CAP.^10,22 Also, procalcitonin-to-albumin ratio (PAR) can be used as an indicator for infection, inflammation, and nutritional status and a predictive biomarker of mortality in adult patients with sepsis-induced acute kidney injury.²³

Despite the above findings, significance of the clinical parameters and their ratios as diagnostic and prognostic markers for elderly septic patients with and without pneumonia remains unknown. This study aimed to investigate the values of the clinical parameters and their ratios, along with the correlations between them, which may provide insights into their potential roles in the regulation of inflammation in sepsis complicated by pneumonia.

Patients and Methods

Ethics

This study received ethical approval (No. 2021-ke-636) from the Ethical Committee of Beijing Chaoyang Hospital Affiliated to Capital Medical University. In compliance with the Declaration of Helsinki, this study gained informed consent from all patients or their families.

Patients and Inclusion Criteria

Patients with an acute change in total SOFA score ≥2 points and suspicion of infection were enrolled according to the Sepsis-3 criteria.⁴ Those under the age of 60, or with malignant tumors, connective tissue diseases, hematological diseases, and those receiving immunosuppressive therapy were excluded. This single-center retrospective study was conducted at the emergency department of Beijing Chaoyang Hospital and a total of 220 septic patients (132 male and 88 female) entered the eligibility assessment of which 82 patients were excluded (Figure 1).

Figure 1 Flow chart of the patients in this study.

The diagnosis of pneumonia was based on a new pulmonary infiltrate on chest radiography, accompanied by symptoms and signs of a lower respiratory tract infection. A total of 138 patients (97 in the non-pneumonia group and 41 in the pneumonia group) were included in this study and further divided into four subgroups according to the occurrence of pneumonia or their survival status at the 28 day interval upon admission (Figure 1).

Data Collection and Laboratory Tests

The baseline data of the patients, such as gender and age, were collected and documented at the time of admission. Vital signs, such as systolic pressure (SBP), diastolic pressure (DBP), mean arterial pressure (MAP), were also collected. Scores, such as Sequential Organ Failure Assessment (SOFA) and Acute Physiology and Chronic Health Evaluation (APACHE) II, were calculated on admission for all enrolled patients.

Blood samples were collected for all patients within 24 hours of admission to the emergency department for the measurement of the following parameters: lactate (Lac), total cholesterol (CHOL), high-density lipoprotein cholesterol (HDL), low-density lipoprotein cholesterol (LDL), hemoglobin (HGB), hematocrit (HCT), platelet counts (PLT), total bilirubin (TBIL), aspartate transaminase (AST), alanine transaminase (ALT), albumin (ALB), procalcitonin (PCT), C-reactive protein (CRP), white blood cell (WBC), lymphocyte count (LYM), monocyte count (MON), and neutrophil count (NEU).

Statistical Analysis

The SPSS 19.0 software (SPSS Inc., Chicago, IL, USA) was used for all statistical calculations and analyses. The results were presented as medians with interquartile ranges or means with standard deviation where appropriate. Categorical variables were presented with absolute values and relative proportions. The Kolmogorov–Smirnov goodness of fit test was used for the examination of normality of the continuous variables. The Student’s t test or Mann–Whitney U-test was performed accordingly for the data analysis. Categorical data were analyzed by the Pearson’s chi square test. Statistical differences between the four subgroups for the clinical parameters were assessed with ANOVA, with post-hoc tests for two-group comparison. Correlations between different parameters were calculated by using Pearson correlation coefficients. Binary logistic regression analyses were performed to assess the independent relationship and screen the risk factors for occurrence of pneumonia in elderly septic patients. Receiver operating characteristic (ROC) curve analysis was conducted to evaluate the diagnostic performance of the biomarkers, and the area under the receiver operating characteristic curve (AUROC) was used to evaluate their predictive values. P<0.05 was considered statistically significant.

Results

Comparison of the Routine Clinical Parameters and the Parameter Ratios Between the Pneumonia Group and Non-Pneumonia Group

In total, 138 patients (79 male and 59 female) were included in this study, of whom 55 survived and 83 died within a 28-day period. As shown in Table 1, the mortality rate in the pneumonia group was higher than the non-pneumonia group (68.3% vs 56.7%, P>0.05). No significant differences existed in gender, SOFA and APACHE II between these two groups (P>0.05).

Table 1 Clinical Data of the Septic Patients from the Non-Pneumonia Group and Pneumonia Group

As shown in Table 1, there were no significant differences in the parameters of lactate, CHOL, HDL, LDL, SBP, DBP, WBC, HCT, PLT, TBIL, AST, CRP, LYM, MON, NEU between the pneumonia group (N=41) and the non-pneumonia group (N=97) (P>0.05). There was a significant difference in age between these two groups (median value: 82.0 vs 71.0, P<0.05). MAP, HGB, ALT in the pneumonia group were significantly lower than those of the non-pneumonia group (mean or median value: 94.2 vs 108.5, 115.0 vs 135.5, 18.5 vs 24.5, P<0.05). The median values of PCT in the pneumonia group were higher than those of the non-pneumonia group (P<0.05). Compared with the non-pneumonia group (N=97), there were significantly higher levels in the ratios of PCT/ALB and PCT/CHOL in the pneumonia group (N=41) (both P<0.05) (Table 1).

Results from the Logistic Regression and ROC Curve Analysis

The logistic regression analysis found that age was independent risk factor associated with the occurrence of pneumonia in elderly patients with sepsis (OR=1.038, 95% CI [1.008, 1.070], P<0.05) (Table 2). HGB and ALB were protective factors associated with the occurrence of pneumonia in elderly patients with sepsis (OR=0.982, 0.914, 95% CI [0.969, 0.996], [0.868, 0.962], P<0.05) (Table 2).

Table 2 Univariate Logistic Regression Analysis of Risk Factors for the Occurrence of Pneumonia in Elderly Patients with Sepsis

As shown in Table 3, the AUROCs of HGB, ALB and age were 0.638, 0.677, 0.640, respectively. ALB possessed the highest sensitivity of 84.0%, while HGB possessed a more balanced sensitivity and specificity of 62.8% and 67.5%, respectively.

Table 3 Receiver Operating Characteristic Curve (ROC) Analysis and Predictive Values of HGB, ALB and Age as Risk Factors for the Occurrence of Pneumonia in Elderly Patients with Sepsis

Correlation Analyses of the Routine Clinical Parameters Between Different Subgroups

The parameter pairs with significance in correlation analyses were demonstrated in Tables 4 and 5. In the pneumonia group, lactate was correlated with HDL, HGB, HCT, PCT (r=0.717, −0.358, 0.672, 0.592, all P<0.05) and HDL was correlated with HGB, HCT, PCT (r=−0.466, 0.645, 0.595, all P<0.01). PLT was correlated with CRP and NEU (r=0.666, 0.552, both P<0.001) in the pneumonia group while PLT was correlated with LYM (r=0.252, P<0.05) in the non-pneumonia group (Table 4).

Table 4 Correlation Analysis Between Lactate, HGB, Lipids and Other Parameters for Septic Patients from the Non-Pneumonia Group and Pneumonia Group

Table 5 Correlation Analysis Between Lactate, HGB, Lipids and Other Parameters for the Septic Survivals or Non-Survivals from the Non-Pneumonia Group and Pneumonia Group

In the pneumonia group of non-survivors (N=28), lactate was correlated with HDL, HGB, PCT (r=0.800, −0.398, 0.638, all P<0.05) and HDL was correlated with HGB, HCT, PCT (r=−0.486, 0.771, 0.697, all P<0.01). In the pneumonia group of survivors (N=13), lactate was correlated with NEU (r=0.869, P=0.001) and CRP was correlated with CHOL, HDL, PLT (r=−0.620, −0.792, 0.907, all P<0.05). In the non-pneumonia group of survivors (N=42), lactate was correlated with HDL, LDL, HCT (r=0.621, 0.517, −0.391, all P<0.05) while PLT was correlated with ALB, HCT, NEU (r=−0.437, −0.360, 0.630, all P<0.05). PLT was correlated with NEU both in the non-pneumonia group of survivors (N=42) and the non-survival group with pneumonia (N=28) (P<0.01) (Table 5). The PCT/ALB ratio was found to be correlated with the APACHE II scores in the pneumonia group of survivors (N=13) (P<0.01) (Table 5). A diagram which reflects the correlations between the parameters of different subgroups has been shown in Figure 2.

Figure 2 Correlation network between the clinical parameters in the different subgroups.

Comparison of the Routine Clinical Parameters and the Parameter Ratios Between Different Subgroups

Compared with the group of non-pneumonia survivors (N=42), there was significantly lower levels of ALB in the group of pneumonia survivors (N=13) (median value: 28.7 vs 36.3, P<0.05) (Table 6). The SOFA and APACHE II scores were both shown to be significantly increased in the non-pneumonia group of non-survivors (P<0.05) (Table 6).

Table 6 Clinical Data with Significant Differences of Between Septic Patients from Different Subgroups

As shown in Table 6, compared with the non-pneumonia group of survivors (N=42), there were significantly higher levels of the ratios of PCT/CHOL and NEU/CHOL in the pneumonia group of survivors (N=13) (both P<0.05). The pneumonia group of non-survivors (N=28) possessed lower levels of the NEU/CHOL ratio compared with the pneumonia group of survivors (N=13) (P<0.05).

ROC Curve Analysis of the Parameter Ratios Between Different Subgroups

As shown in Table 7, the PCT/ALB ratio and PCT/CHOL ratio can be used to distinguish the non-pneumonia group (N=97) and pneumonia group (N=41) with the AUROCs of 0.694 and 0.643, respectively (both P<0.05). And the sensitivity and specificity of the PCT/ALB ratio were 75.6% and 69.1% for distinguishing these two groups. The ratios of PCT/ALB, PCT/CHOL and NEU/CHOL can be used to distinguish the non-pneumonia group of survivors (N=42) and pneumonia group of survivors (N=13) (all P<0.05, AUROC value: 0.761, 0.744, 0.733). The specificity of PCT/CHOL ratio and the sensitivity of PCT/ALB in distinguishing these two groups can be 95.0% and 77.8%, respectively. The AUROC of the NEU/CHOL ratio in distinguishing the pneumonia group of survivors (N=13) and non-survivors (N=28) can reach 0.761 with the sensitivity and specificity of 70.0% and 82.1%, respectively (Table 7).

Table 7 Receiver Operating Characteristic Curve (ROC) Analysis and Predictive Values of Parameter Ratios for the Occurrence of Pneumonia or Survival in Septic Patients from Different Subgroups

Discussion

Sepsis is estimated to account for worldwide incident cases of about 50 million, resulting in approximately 5 million deaths each year.^24–26 Age is an independent predictor of mortality for patients with sepsis, of whom the mortality rate of patients between the age of 60–64 years is 26% and that of patients aged over 85 years reaches 38%.^27,28 And age was an independent risk factor associated with the occurrence of pneumonia in elderly patients with sepsis, which may reflect a deterioration of immune function with the increase of age. As it is difficult to diagnose elderly patients with sepsis due to their relatively few specific symptoms, it is vital to evaluate potential biological markers with diagnostic or prognostic values for the elderly. As CAP is the leading cause of sepsis and also a risk factors for sepsis, understanding the clinical characteristics of sepsis complicated by pneumonia is of great value for elderly patients. And there is an urgent demand to unravel the molecular processes that underlie the disease severity and determine adverse outcomes.²⁹ Our study for the first time explored the clinical parameters and their ratios as potential diagnostic and prognostic markers for elderly septic patients with and without pneumonia. The correlations between them may provide insights into their potential roles in the regulation of inflammation in sepsis complicated by pneumonia.

HGB is a main carrier of oxygen and its level decrease correlates with oxygen deficiency as well as the hypoxic and ischemic injury in the body. Several studies have investigated the prognostic values of HGB in sepsis. The level of HGB has been identified as an indicator for predicting long-term outcomes in sepsis. And septic patients with early HGB levels below 8 g/dL had significantly lower survival rates.³⁰ For the ICU patients with sepsis, the HGB level was strongly correlated with the in-hospital mortality, and significant associations were observed between HGB levels and the 14-day mortality risk or 1-year mortality risk in CAP patients.^31,32 In our study, the level of HGB was found to be an indicator of the occurrence of pneumonia in elderly patients with sepsis. We conclude that several possible mechanisms may explain the relationship between HGB and pneumonia in sepsis. First of all, HGB exerts a vital role in anti-inflammation by defending against bacteria and enhancing the function of leucocytes.³³ Secondly, HGB impacts on the absorption and metabolism of antibiotics with respect to that decreased level of HGB significantly attenuate the anti-infection of antibiotics.³⁴ Thirdly, HGB is zinc-containing enzyme related to low-grade inflammation.³⁵ All these factors may influence the clinical progression to pneumonia and the outcome of the elderly septic patients.

ALB plays important roles in acid-base physiology and fluid distribution.³⁶ It also acts as an antioxidant and free radical scavenger in the blood and tissues.¹⁴ In inflammatory conditions, however, ALB breaks down in an accelerated rate which might result in hypoalbuminemia.^15–17 Septic patients possess lower plasma ALB levels than non-septic patients.³⁷ Serum lipids are critical in cell-building and changes in lipid metabolism correlate with inflammation.¹⁸ HDL aids with the removal of bacterial toxins and participates in the pathophysiology of sepsis. CHOL is the key component of enveloped virus membranes and host cells which has been shown to play a significant role in SARS-CoV-2 virus entry into cells.³⁸ A recent research based on systematic review, meta-analysis, and perspective of observational studies has shown that HDL, LDL, and CHOL levels either prior to or early during sepsis and critical illness are adversely prognostic.²² A retrospective analysis on bacterial pneumonia and COVID-19 infection showed that patients with CAP had lower ALB and CHOL levels than healthy controls and those with non-severe COVID-19 infection. The serum ALB, triglyceride (TG), CHOL, and LDL levels were considerably lower in the severe community acquired pneumonia patients (SCAP) group than in the non-severe community-acquired pneumonia patients (NSCAP) group.³⁹ In our research, serum ALB levels were significantly lower in the pneumonia group than those in the non-pneumonia group of elderly septic patients, indicating the diagnostic value of ALB for the occurrence of pneumonia.

Both the PCT/ALB ratio (PAR) and CRP/ALB ratio (CAR) have been shown to possess the ability to identify sepsis in neonates with pneumonia.^40,41 Despite these findings, no reports have been made on elderly septic patients with pneumonia. In this study, several parameter ratios have been evaluated based on the insights from the correlation analysis of the different subgroups and proven to have predictive values either for the disease subtypes, severity or adverse outcomes. The PCT/ALB ratio plays a role in distinguishing the non-pneumonia group and pneumonia group. Also, correlated with the APACHE II scores in the pneumonia group of survivors (N=13), the PCT/ALB ratio possesses the predictive value for the disease severity for this subgroup of patients.

As NEU and CHOL are both correlated with control of infection and regulation of inflammation, we wondered whether the ratios of NEU to CHOL and PCT to CHOL can be used as diagnostic and prognostic markers for elderly septic patients with and without pneumonia. Our results demonstrated that the NEU/CHOL ratio was adversely prognostic in the pneumonia group of elderly septic patients. We conclude that the NEU/CHOL ratio can be a novel indicator of the inflammation and infection status of the elderly septic patients with pneumonia, which greatly influences the disease progression and mortality. Also, our data confirmed the diagnostic values of the NEU/CHOL ratio and the PCT/CHOL ratio in distinguishing the pneumonia group and non-pneumonia group of survivors.

Serum levels of lactate are elevated as a result of increased catecholamine secretion, impaired pyruvate dehydrogenase activity, tissue hypoperfusion, and increased immune cell activation in sepsis.⁴² Increased serum levels and impaired clearance of lactate are independently associated with higher mortality in patients with sepsis.^43,44 The role of lactate in inflammation is not fully clear for the fact that the concurrent pro-inflammatory and anti-inflammatory effects of lactate acid were both reported in sepsis.^45,46

As the clinical parameters are cross-linked at different levels in vivo in the regulation of inflammation in sepsis, we performed the correlation analysis aiming at investigating the relationships between these clinical parameters. Correlations between several parameters in each subgroup have been shown in this study. The characteristic correlation between HGB-CHOL was confirmed both in the non-pneumonia group of survivors and non-survivors. The correlations between HCT-CHOL-LYM have been shared by the pneumonia group of survivors and non-survivors. The correlation between WBC-PLT has been shared by the pneumonia and non-pneumonia group of non-survivors. In the pneumonia group of non-survivors, there were characteristic correlations between lactate, HDL, HGB and PCT, compared with the other subgroups. In this study, interestingly, we observed strong correlations between lactate and NEU in the pneumonia group of survivors (r=0.869) and between lactate and HDL in the pneumonia group of non-survivors (r=0.800). One previous study has found that the neutrophil count in sputum is associated with increased sputum glucose and sputum L-lactate in cystic fibrosis.⁴⁷ L-Lactate might be a marker of inflammation, as it is produced from glucose by polymorphonuclear neutrophils (PMNs) in cystic fibrosis (CF) lungs.⁴⁷ This reflects that lactate and NEU may have a cause–effect relationship, which may influence the clinical outcome when the correlation between lactate and NEU was unbalanced or out of control. The detailed mechanisms still need further investigation by animal model studies. However, the association between lactate and HDL has not yet been mentioned by other reports. Altogether, these characteristic correlation network between the parameters in different subgroups, especially those between lactate, HGB, HDL, NEU and CHOL, may provide novel insights into their potential roles in the regulation of inflammation in sepsis with pneumonia.

This study also has limitations, such as the relatively small sample size, lack of healthy controls and the fact that it was a single-center study. Also, sample size calculation was not conducted for the lack of an expected estimate of Spearman correlation coefficient either from previous similar research or expert opinion. Further multi-center studies with larger sample sizes are still needed to the validation of the values of the parameter ratios. Future research should focus on the cause–effect relationships between these clinical parameters and more studies should be performed to verify the regulatory mechanisms in in vitro cell models or in vivo animal models.

Conclusion

Our study for the first time revealed that levels of HGB and ALB, along with the ratios of PCT/ALB, PCT/CHOL, and NEU/CHOL, can serve as potential diagnostic and prognostic markers for elderly septic patients with and without pneumonia. The link between these clinical parameters may provide insights into their potential roles in the regulation of inflammation in sepsis.