The potential role of serum lipoprotein in children with sepsis

Abstract

To determine whether serum lipoprotein is correlated with sepsis on the day of admission and help with early warning, identification, and intervention for sepsis. This retrospective study involved all children admitted to our pediatric intensive care unit from January 2021 to June 2023. Clinical data of involved patients were collected via inquiring databases of our hospital. The Pediatric Risk of Mortality and Pediatric Index of Mortality II scores were calculated, multivariate analysis was conducted to evaluate the independence of the association between serum lipoprotein and the risk of sepsis on the day of admission. This study involved 88 children (48 with sepsis, 40 non-sepsis). The sepsis group had a lower level of high-density lipoprotein (HDL), total cholesterol, and low-density lipoprotein (LDL). The CRP, white blood cell, and procalcitonin (PCT) were considerably higher than the non-sepsis group. After adjusting for covariates, logistic regression analysis suggested that the CRP, PCT, HDL, Pediatric Risk of Mortality score, Pediatric Index of Mortality II score and LDL were independent risk factors for sepsis. Moreover, the AUC of CRP, PCT, HDL, and LDL were 0.58, 0.76, 0.82, and 0.86, respectively. Our results may indicate that serum lipoprotein is correlated with sepsis on the day of admission and may help with early warning, identification, and intervention for sepsis.

1. Introduction

Sepsis accounts for a significant proportion of pediatric intensive care patients and is a critical factor for mortality and morbidity worldwide.^[1] Although significant advances have been made in dealing with sepsis, it remains an intractable condition with high morbidity and mortality rates in clinical practice worldwide.^[2] Timely identification of bacteremia and prompt antimicrobial therapy are essential to reduce sepsis-related deaths. Blood cultures are the golden standard for the diagnosis of sepsis, but time-consuming and false negatives undermine its clinical use, especially in the pediatric population.^[3] Thus, biomarkers are urgently needed to aid in the diagnosis and prognosis prediction of children with sepsis. Several recent studies have shown that serum lipoprotein concentrations were changed during sepsis in adults, and lipoprotein concentration can be used as a prognostic factor for severe sepsis.^[4,5] Still, their use has not been studied among the pediatric population. In the present study, we collected and analyzed the data of serum lipoproteins metabolism are altered in sepsis and non-sepsis among the pediatric population, and demonstrated that serum lipoprotein may be used as an indicator to evaluate the severity of sepsis among pediatrics.

2. Methods

2.1. Study population

This retrospective study involved 48 children with sepsis who were admitted to the Linping Branch of the Second Affiliated Hospital of Zhejiang University hospital’s pediatric intensive care unit (PICU) from January 2021 to June 2023. The inclusion criteria were: children were older than 1 mouth and younger than 14 year, acute diseases needing PICU admission, and parental consent. The exclusion criteria were: children with severe hepatic and renal dysfunction, children with blood disorders, Children on long-term glucocorticosteroids, taking drugs that affect blood lipids, and incomplete clinical data. In brief, sepsis patients were defined according to the sepsis definition.^[6,7] A cohort of matched 40 healthy individuals with serum lipoprotein results was also involved to serve as a control group. This study was approved by the medical ethics committee of the Linping Branch of the Second Affiliated Hospital of Zhejiang University (Approval number: Linping 2023044). Because of the retrospective nature of this study, the requirement for informed consent was waived.

For each patient, Pediatric Risk of Mortality (PRISM) and Pediatric Index of Mortality II (PIM II) were calculated, namely.^[8,9] PRISM score is automatically calculated within 24 hours of admission. PIM2 is simpler and needs to be calculated within 1 hour of the face-to-face contact with the patient.

The primary outcome measure was the occurrence of death during hospital admission. The secondary outcome measures included the length of PICU stay and the need and duration of mechanical ventilation.

2.2. Data collection

Demographic data and clinical and laboratory parameters of enrolled patients were collected from the electronic medical record system in the hospital. Patients’ demographic characteristics are documented in detail in the records, and vital signs, including body temperature, systolic blood pressure, and heart rate are collected. serum lipoprotein results, blood routine examination results, C-response protein (CRP) and amyloid A results, procalcitonin (PCT) results, as well as ferritin results were collected. Whole blood was collected for testing within the first 3 hours after admission to our PICU. Furthermore, two mortality predictive scores were calculated to determine the severity of disease including PRISM and PIM II.

2.3. Data analyses and statistics

Statistical analyses were performed using SPSS for Windows Version 26 (IBM Corp., Armonk, NY). Statistical data are expressed as n (%), and the χ² test or Fisher’s exact test was used for comparison between groups. Normally distributed measures are expressed as mean ± standard deviation, and the t test for 2 independent samples was used for comparison between groups. Non-normally distributed measures are expressed as median (interquartile range), and the Kruskal–Wallis rank sum test was used for comparison between groups. In the multifactor logistic regression model, the risk factors with statistical significance in the univariate analysis were selected as covariates, and sepsis and non-sepsis subgroups were used as dependent variables to analyze the independent risk factors for BPD. A receiver operating characteristic curve (ROC) was used to analyze the predictive value of CRP, PCT, high-density lipoprotein (HDL) and low-density lipoprotein (LDL) on the first day of admission for sepsis, Differences were considered statistically significant at P < .05.

3. Results

In total, 88 children were enrolled during the study period. Of all those patients, 48 were diagnosed with sepsis whereas 40 did not have sepsis (Table 1). The basic clinical characteristics of the population in the sepsis group and non-sepsis group are summarized in Table 1. With regard to the laboratory data, the levels of CRP (64.45 ± 18.26 vs 4.26 ± 3.28), PCT (6.48 ± 2.33 vs 0.54 ± 0.14) neutrophils (10.28 ± 1.68 vs 6.12 ± 1.26), white blood cell (15.64 ± 3.28 vs 9.25 ± 2.44) and monocytes (3.28 ± 1.52 vs 1.57 ± 0.65) in the sepsis group were significantly higher. In contrast, the levels of LDL (1.32 ± 0.49 vs 1.85 ± 0.64), total cholesterol (2.26 ± 0.53 vs 3.24 ± 0.42), and HDL (0.64 ± 0.26 vs 1.26 ± 0.31) were markedly lower than those of the control group. Otherwise, there was no significant difference in sex, age, body mass index, systolic blood pressure, and triglycerides level between the sepsis and non-sepsis groups.

Table 1.

Baseline characteristics of studied population.

	Sepsis (n = 48)	Non-sepsis (n = 40)	P
Sex (male, %)	20 (41.7)	18 (45)	.481
Age, yr	8.42 ± 2.65	7.43 ± 3.67	.287
BMI, kg/m2	22.82 ± 2.84	22.68 ± 2.71	.646
SBP, mm Hg	98.54 ± 15.63	94.68 ± 10.27	.351
CRP, mg/L	64.45 ± 18.26	4.26 ± 3.28	.001
PCT, ng/mL	6.48 ± 2.33	0.54 ± 0.14	.001
Neutrophil, ×109/L	10.28 ± 1.68	6.12 ± 1.26	.001
Lymphocyte, ×109/L	3.28 ± 1.52	1.57 ± 0.65	.001
WBC, ×109/L	15.64 ± 3.28	9.25 ± 2.44	.001
HDL, mmol/L	0.64 ± 0.26	1.26 ± 0.31	.001
LDL, mmol/L	1.32 ± 0.49	1.85 ± 0.64	.001
TG, mmol/L	1.12 ± 0.26	1.08 ± 0.24	.154
TC, mmol/L	2.26 ± 0.53	3.24 ± 0.42	.001
Mechanical ventilation, n (%)	18 (37.5)	–	–
PICU stay, d	10.62 ± 3.63	–	–
PRISM score	9.56 ± 8.23	–	–
PIM II	7.62 ± 10.25	–	–
Mortality in the PICU	11 (22.9%)	–	–

Data are presented as mean ± standard deviation, n (%), or median (interquartile range).

BMI = body mass index, CRP = C-reactive protein, HDL = high-density lipoprotein, LDL = low-density lipoprotein, PCT = procalcitonin, PICU = pediatric intensive care unit, PIM II = Pediatric Index of Mortality II, PRISM = Pediatric Risk of Mortality, SBP = systolic blood pressure, TC = total cholesterol, TG = triglycerides, WBC = white blood cell.

Clinico-demographic parameters between survived and non-survived groups are shown in Table 2. Non-survived group had significantly higher PRISM and PIM II score; longer PICU stay; and increased mechanical ventilation need compared with survived group. Besides, the levels of CRP (76.34 ± 18.86 vs 41.26 ± 19.37), PCT (7.26 ± 1.28 vs 5.53 ± 3.24) in the sepsis group were significantly higher. In contrast, the levels of LDL (1.18 ± 0.26 vs 1.46 ± 0.36), total cholesterol (1.58 ± 0.38 vs 2.08 ± 0.46), and HDL (0.58 ± 0.16 vs 0.78 ± 0.21) were markedly lower than those of the control group.

Table 2.

Comparison between survivors and non-survivor groups regarding demographic data and clinical characteristics.

	Study groups (n = 48)		P value
	Survivous (n = 37)	Non-survivous (n = 11)
Sex (male, %)	16 (43.2)	4 (36.4)	.481
Age, yr	9.26 ± 3.24	7.53 ± 1.57	.126
BMI, kg/m2	23.14 ± 3.26	22.18 ± 2.71	.268
SBP, mm Hg	99.26 ± 13.83	93.16 ± 12.25	.348
CRP, mg/L	41.26 ± 19.37	76.34 ± 18.86	.001
PCT, ng/mL	5.53 ± 3.24	7.26 ± 1.28	.001
Neutrophil, ×109/L	9.66 ± 2.14	10.63 ± 1.38	.168
Lymphocyte, ×109/L	3.16 ± 1.28	3.32 ± 1.36	.468
WBC, ×109/L	15.28 ± 3.42	15.78 ± 2.42	.827
HDL, mmol/L	0.78 ± 0.21	0.58 ± 0.16	.001
LDL, mmol/L	1.46 ± 0.36	1.18 ± 0.26	.001
TG, mmol/L	1.16 ± 0.14	1.18 ± 1.20	.364
TC, mmol/L	2.08 ± 0.46	1.58 ± 0.38	.001
Mechanical ventilation, n (%)	7 (18.9%)	11 (100%)	.001
PICU stay, d	8.64 ± 2.32	11.64 ± 1.62	.001
PRISM score	6.89 ± 6.18	15.28 ± 4.56	.001
PIM II score	4.26 ± 5.65	15.69 ± 4.56	.001

Data are presented as mean ± standard deviation, n (%), or median (interquartile range).

BMI = body mass index, CRP = C-reactive protein, HDL = high-density lipoprotein, LDL = low-density lipoprotein, PCT = procalcitonin, PICU = pediatric intensive care unit, PIM II = Pediatric Index of Mortality II, PRISM = Pediatric Risk of Mortality, SBP = systolic blood pressure, TC = total cholesterol, TG = triglycerides, WBC = white blood cell.

The logistic regression analysis showed that the CRP, PCT, PRISM and PIM II score, HDL, and LDL were independent risk factors for sepsis (Table 3). The optimal cutoff value of the CRP was 58.42, with 66.5% sensitivity, 68.2% specificity, and an area under the ROC curve of 0.58. The optimal cutoff value of the PCT was 6.27, with 70.2% sensitivity, 66.5% specificity, and an area under the ROC curve of 0.76. The optimal cutoff value of the HDL was 0.66, with 73.2% sensitivity, 75.3% specificity, and an area under the ROC curve of 0.82. The optimal cutoff value of the LDL was 1.30, with 76.2% sensitivity, 79.3% specificity, and an area under the ROC curve of 0.86 (Fig. 1).

Table 3.

Logistic regression analysis showing independent predictors of sepsis.

Factors	β	S b	Wald	OR	95% CI	P
CRP	2.45	0.68	12.89	8.63	3.05–24.38	.001
PCT	1.99	0.84	5.62	7.30	1.41–36.81	.018
Neutrophil	0.37	0.21	3.09	1.84	1.30–2.74	.079
Lymphocyte	2.29	4.15	0.31	2.97	1.64–5.40	.84
WBC	0.94	0.72	1.70	2.56	0.62–10.51	.193
HDL	1.95	0.85	5.29	0.85	0.81–0.92	.022
LDL	3.13	1.24	6.3	0.88	0.86–0.93	.012
TC	0.92	0.48	3.76	0.92	0.88–0.96	.053
PRISM score	1.49	0.71	9.1	4.42	2.46–8.96	.001
PIM II score	1.16	1.36	0.73	3.82	1.96–7.28	.036

CI = confidence interval, CRP = C-reactive protein, HDL = high-density lipoprotein, LDL = low-density lipoprotein, OR = odds ratio, PCT = procalcitonin, PIM II = Pediatric Index of Mortality II, PRISM = Pediatric Risk of Mortality, TC = total cholesterol, TG = triglycerides, WBC = white blood cell.

Figure 1.

ROC analysis of parameters for predicting sepsis. CRP = C-reactive protein, PCT = procalcitonin, HDL = high-density lipoprotein, LDL = low-density lipoprotein, ROC = receiver operating characteristic curve.

4. Discussion

Recently, sepsis has been defined as a syndrome of physiological, pathological, and biochemical abnormalities. This induces an uncontrollable host reaction to inflammation which can cause fatal multiple organ dysfunction.^[10] Ultimately, inflammation and oxidative stress are supposed to be the essence of the pathophysiological process. The incidence of sepsis is progressively on the rise and sepsis may be the leading cause of death in patients admitted to pediatric intensive care units.^[11,12] Therefore, identifying markers to predict the risk of sepsis can help in the clinical prevention, diagnosis, and treatment of sepsis at an earlier stage, thus improving the prognosis and quality of survival of children. In our study, we analyzed the relationship between serum lipoprotein and sepsis on the first day of admission among children. We found that the HDL, LDL, CRP and PCT were independent risk factors for sepsis on the first day of admission and that the HDL, LDL, CRP and PCT were valuable in predicting the development of sepsis.

It is demonstrated that sepsis is induced by an uncontrollable host reaction to inflammation. Our study indicated that the decrease of HDL and LDL among children is correlated with sepsis. Several research have demonstrated that serum levels of HDL and LDL were related with the severity of sepsis.^[13–16] Lower HDL levels are related with poor prognosis of septic patients. If infected, the production of pro-inflammatory cytokines promptly induces a reduction of HDL and LDL serum levels, mainly due to the reduction in the cholesterol ester content in these lipoproteins.^[17] These changes have been described to be related to the degree of inflammation measured by pro-inflammatory and anti-inflammatory cytokine levels, showing that the greater the degree of inflammation, the greater the changes in lipoprotein metabolism.^[18,19]

Oxidative stress is considered to be another crucial element involved in the progression of sepsis. Increasing evidence suggests that both the incapacity of cells to consume oxygen and the accumulation of peroxide may severely exacerbate the pathological process of sepsis.^[20] During the process of oxidative stress, HDL inhibits the hyperoxidation of LDL and can exert a protective effect on the endothelium. Moreover, HDL inhibits the activation and transformation of monocytes, thereby resulting in a suppression of inflammatory response.^[21]

In total, we also analyzed CRP and PCT in our research, and their predictive abilities appeared lower compared with HDL and LDL. Previous research indicated that Many biomarkers, such as CRP, and PCT, could be used in the auxiliary diagnosis of sepsis, but their serum levels may be affected by many other stimulations other than infection.^[22–24] therefore, compared with HDL and LDL, their predictive abilities appeared low. That’s consistent with our research.

Regarding clinico-demographic parameters between survived and non-survived groups, no difference was found between 2 groups as regard to age, sex, and body mass index. This was matched to the previous study.^[25,26]

Some study demonstrated that ICU stay, PRISM, PIM mortality rate, and need for mechanical ventilation was significantly more in non-survivors than survivors.^[27]

The retrospective and single-center-based characteristics of the current study have several limitations. Large-sample studies are needed to increase the level of evidence for the relationship of sepsis with serum lipoprotein.

5. Conclusions

Together, our study reveals the serum lipoprotein levels and their potential use in sepsis among children. HDL and LDL may be recommended to help diagnose sepsis. However, further studies are required to confirm this hypothesis.

Acknowledgments

All authors read and approved the final version of the manuscript.

Author contributions

Conceptualization: Chang Su, Yueyan Mao.

Data curation: Junsheng Jiang, Yueyan Mao.

Funding acquisition: Chang Su.

Investigation: Jiabo Wu, Chang Su.

Methodology: Jiabo Wu, Chang Su, Junsheng Jiang.

Project administration: Chang Su.

Resources: Jiabo Wu.

Supervision: Chang Su, Yueyan Mao.

Validation: Yueyan Mao.

Visualization: Yueyan Mao.

Writing – original draft: Jiabo Wu.

Writing – review & editing: Jiabo Wu.