Abstract
Background/aims
High triglyceride (TG) levels, low high-density lipoprotein cholesterol (HDL-C), and an elevated TG/HDL-C ratio are linked to increased systemic inflammation and impaired immune function. However, the clinical significance of the TG/HDL-C ratio in pyogenic liver abscess (PLA) remains poorly defined.
Methods
In this multicenter, cross-sectional study of 2,201 cases of PLA, logistic regression, smoothed curve fitting, and comprehensive sensitivity analyses were employed to assess the relationship between the baseline TG/HDL-C ratio and positive blood or pus cultures. Subgroup analyses and interaction tests were performed to evaluate the robustness of the findings.
Results
Logistic regression showed that the TG/HDL-C ratio was positively associated with blood or pus culture positivity in PLA [odds ratio (OR) = 1.04, 95% confidence interval (CI): 1.02–1.06, P < 0.001. Smoothed curve fitting revealed a nonlinear dose-response relationship, with an inflection point at 9.2. Below this threshold, the TG/HDL-C ratio remained significantly associated with culture positivity (OR = 1.1, 95% CI: 1.1–1.2, P < 0.001). Sensitivity analyses and interaction tests confirmed these results.
Conclusions
The TG/HDL-C ratio demonstrates a nonlinear relationship with the risk of positive blood or pus cultures in patients with PLA, with a pivotal threshold value identified at 9.2. This readily obtainable ratio may help clinicians individualize management strategies and facilitate timely interventions for PLA patients in this at-risk group, leading to improved health outcomes for this vulnerable population.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12944-025-02682-8.
Keywords: Liver abscess, Pyogenic, Triglycerides, Cholesterol, HDL, Blood culture, Suppuration
Introduction
Pyogenic liver abscess (PLA) is a purulent infection within the hepatic parenchyma. Although uncommon, PLA can be fatal if left untreated. Timely diagnosis and appropriate interventions, including anti-infective therapy and drainage, are crucial for improving clinical outcomes [1, 2]. Pus culture remains the gold standard for diagnosing PLA, whereas blood culture is also used to detect pathogenic bacteria associated with PLA [3]. However, the positivity rate of blood culture is considerably lower than that of pus culture [4]. In a previous study, pus culture demonstrated a sensitivity of 52% to Gram-negative rods and 90% to Gram-positive cocci, whereas blood culture achieved sensitivities of only 30% and 39% for these pathogens. In addition, pus and blood cultures have several other limitations. For instance, pus culture is invasive during sample collection, and both cultures are time-consuming. As such, non-invasive and effective biomarkers are needed to provide rapid results, enabling earlier initiation of interventions, especially in cases that could progress to sepsis [5].
Elevated triglycerides (TG), reduced high-density lipoprotein cholesterol (HDL-C), and their integrated TG/HDL-C ratio are established biomarkers of cardiovascular disease (CVD) and metabolic dysfunction. These lipid alterations are closely linked to systemic inflammation and impaired immune function; lipids participate in both the innate and adaptive immune responses [6]. After adjustment for established CVD risk factors, a higher baseline TG/HDL-C ratio is associated with an increased risk of CVD [7]. Similarly, a higher TG/HDL-C ratio is significantly associated with aggravated severity in patients experiencing recurrent episodes of acute pancreatitis [8]. Critical care populations with sepsis or systemic inflammatory response syndrome (SIRS) often display metabolic changes in their lipid profiles, chiefly characterized by elevated TG levels [9]. Another study demonstrated an inverse dose–response relationship between HDL-C and infection-driven hospitalization, with no comparable findings for TG or low-density lipoprotein cholesterol (LDL-C) [10]. HDL-C concentrations in septic patients declined within 72 h of intensive care unit (ICU) entry and then rose above the day-3 trough by discharge [11]. Furthermore, in septic shock cohorts, low HDL-C at hospital entry was strongly linked to 28-day fatality, whereas no similar trend was observed for TG levels [11]. However, among patients with schizophrenia who were seropositive for Toxoplasma gondii, the only notable alteration in metabolic profiles was a lower TG/HDL-C ratio [12].
Extensive research has characterized PLA, including studies of pus microbiology, comparisons between Klebsiella pneumoniae and Escherichia coli subgroups, risk factors for ESBL-producing Enterobacter, and the use of statins [13–15]; however, empirical data linking the TG/HDL-C ratio to PLA risk remain sparse and inconclusive. He et al. [16] reported that blood-culture evidence of bacteremia was present in 32.9% of non-septic individuals versus 58.1% of septic patients, a difference that achieved statistical significance (P < 0.05). In addition, a predictive model identified blood-culture positivity as an independent predictor of sepsis in PLA cases [17]. Based upon these previous studies, elevated TG/HDL-C ratio may be associated with an increased risk of infection occurrence and progression, including PLA.
The present analysis aimed to elucidate if an augmented TG/HDL-C ratio predicted the occurrence of culture-positive PLA. Detailed subgroup analyses as well as interaction examinations were conducted to clarify the modifying role of confounders in the observed TG/HDL-C versus culture-positive relationship.
Methods
Study population
Eligible patients diagnosed with PLA were retrospectively identified from five Fujian hospitals between January 2015 and June 2024: Fujian Medical University Union Hospital, the First Affiliated Hospital of Fujian Medical University, Fujian Medical University Affiliated Quanzhou Hospital, Fujian Medical University Affiliated Zhangzhou Hospital, and Ningde Municipal Hospital of Ningde Normal University. The inclusion criteria were as follows [18, 19]: (1) symptoms such as fever, fatigue, and upper abdominal discomfort; (2) abdominal ultrasonography, CT, or MRI demonstrating PLA; and (3) either a positive culture from percutaneous drainage or blood, or a significant reduction or disappearance of the abscess after anti-infection treatment. Any subject who satisfied any of the predefined exclusion conditions was removed: (1) age < 14 years; (2) incomplete blood lipid data on admission; (3) fungal, tuberculous, or amoebic abscesses; (4) malignant tumors complicated by PLA after radiofrequency ablation; or (5) receipt of other treatment before admission, such as percutaneous fine needle aspiration, percutaneous catheter drainage, or surgical drainage.
This study received approval from the corresponding institutional ethics committee. Because all data were collected retrospectively and no additional interventions were performed, the need for informed consent was formally exempted.
Data collection
Data, encompassing demographics, medical histories, comorbidities, laboratory test results, and pus and blood cultures, were collected retrospectively. Pus and blood cultures were defined as follows. (a) Pus culture: Ultrasound was used to locate the liver abscess, and the puncture site on the body surface was identified. A sterile syringe was used to aspirate 10–20 ml of turbid pus. Dual aerobic–anaerobic culturing was applied to all collected samples. (b) Blood Culture: Under sterile conditions, blood samples (5–10 ml) were collected from both hands for aerobic–anaerobic culturing. A 7-day regimen of daily subcultures was applied to all blood culture bottles to detect delayed or platenegative growth [20]. A blood sample was drawn within 24 h of patient admission. Complete blood counts were performed using flow cytometry (XN1000), biochemical tests were conducted with the chemiluminescence method (Roche Cobas e702), immunetransmission turbidimetry (XN1000) was used for C-reactive protein (CRP) determination, and procalcitonin (PCT) levels were determined by electrochemiluminescence (Roche Cobas e801). Pus and blood cultures were processed with an automated microbial identification system (BACTEC FX; Becton, Dickinson and Company). Lipid test results from the preceding two years (obtained under non-infectious conditions) were retrieved; when available, these served as preadmission lipid data.
Statistical testing
Statistical computations were performed in R (v4.2.0 ) complemented by Empower(R) (http://www.empowerstats.com). Continuous variables are summarized as mean (SD) or median [IQR]; intergroup comparisons used the t-test or Mann–Whitney U test. Categorical data are presented as frequency (percentage), and the χ² test was applied for between group comparisons. The association between baseline lipid levels and positive blood or pus cultures in PLA cases was evaluated with a dose–response model based on smoothing curve fitting. A segmented (two-phase) linear model was then built to quantify the effect at the identified breakpoint. Multivariable logistic regression analyses yielded odds ratio (OR) and 95% CI describing the independent relationship of TG/HDL-C with blood or abscess fluid culture positivity in PLA. Sensitivity analyses were conducted through stratified and interaction analyses. A two-tailed P < 0.05 was defined as statistical significance.
Results
Demographics and clinical features
The study included 2,201 patients diagnosed with PLA; the mean age was 59.2 years, and 65.1% were male. The population was divided into four quartiles according to baseline TG/HDLC levels, with their characteristics shown in Table 1. No significant age differences were observed among the TG/HDL-C quartiles. An elevated TG/HDL-C ratio (Q4) was more common in male patients and in those with diabetes mellitus, fatty liver, and non-tumor disease. In addition, WBC, CRP, PCT, total bilirubin, gamma-glutamyl transpeptidase, aspartate aminotransferase, alkaline phosphatase, alanine aminotransferase, TG, Apo-B, creatinine, uric acid, blood glucose, HbA1c, and BMI were markedly greater in the highest quartile of TG/HDL-C, whereas albumin, TC, Apo-A1, HDL-C, and LDL-C were lower (all P < 0.05).
Table 1.
Baseline profiles of participants in different TG/HDL-C quartiles
| Variable | TG/HDL-C (Q1) (0.19–1.23)(n = 534) |
TG/HDL-C (Q2) (1.24–2.06)(n = 538) |
TG/HDL-C (Q3) (2.07–3.82)(n = 561) |
TG/HDL-C (Q4) (3.83–69.25)(n = 568) |
P-value |
|---|---|---|---|---|---|
| Age (years, mean ± sd) | 59.13 ± 15.27 | 59.41 ± 13.44 | 59.46 ± 12.87 | 59.05 ± 13.25 | 0.949 |
| WBC (109/L, mean ± sd) | 12.28 ± 5.67 | 11.93 ± 5.47 | 12.80 ± 6.03 | 14.72 ± 7.58* | < 0.001 |
| CRP [mg/L, Median (Q1-Q3)] | 96.92 (61.92-157.93) | 103.78 (70.70–166.00) | 112.00 (77.39–187.00)* | 148.00 (90.00-204.50)* | < 0.001 |
| PCT [ng/mL, Median (Q1-Q3)] | 0.88 (0.22–6.30) | 1.01 (0.30–6.54) | 4.60 (0.76–22.08)* | 15.64 (3.97–52.56)* | < 0.001 |
| TBIL [µmol/L, Median (Q1-Q3)] | 14.65 (9.50-22.33) | 13.95 (9.00-22.20) | 15.90 (10.10–27.10)* | 20.25 (12.47–42.62)* | < 0.001 |
| ALB (g/L, mean ± sd) | 34.73 ± 5.08 | 32.73 ± 5.43* | 30.81 ± 5.46* | 27.85 ± 5.64* | < 0.001 |
| ALT [IU/L, Median (Q1-Q3)] | 46.01 (24.00-80.01) | 48.01 (27.00-80.75) | 54.00 (32.00–93.00)* | 77.50 (46.00-128.25)* | < 0.001 |
| AST[IU/L, Median (Q1-Q3)] | 34.02 (22.01–65.25) | 35.01 (21.92-64.00) | 40.00 (25.00–81.00) | 63.00 (35.08–131.00)* | < 0.001 |
| GGT [IU/L, Median (Q1-Q3)] | 99.00 (55.00-178.00) | 130.00 (75.00-224.50)* | 122.00 (74.00-214.00)* | 135.00 (74.00-247.00)* | 0.002 |
| ALP [IU/L, Median (Q1-Q3)] | 121.00 (88.50–175.00) | 140.85 (97.00-209.00)* | 148.00 (109.00-223.50)* | 166.00 (116.00-246.00)* | < 0.001 |
| TG (mmol/L, mean ± sd) | 0.85 ± 0.26 | 1.19 ± 0.35* | 1.60 ± 0.53* | 2.67 ± 1.70* | < 0.001 |
| TC (mmol/L, mean ± sd) | 3.59 ± 1.00 | 3.54 ± 1.04 | 3.35 ± 1.02* | 3.19 ± 1.16* | < 0.001 |
| HDL-C (mmol/L, mean ± sd) | 1.06 ± 0.44 | 0.74 ± 0.22* | 0.58 ± 0.19* | 0.34 ± 0.18* | < 0.001 |
| LDL-C (mmol/L, mean ± sd) | 2.28 ± 0.81 | 2.33 ± 0.90 | 2.08 ± 0.91* | 1.51 ± 0.89* | < 0.001 |
| Apo-A1 (g/L, mean ± sd) | 0.90 ± 0.26 | 0.73 ± 0.23* | 0.62 ± 0.33* | 0.44 ± 0.19* | < 0.001 |
| Apo-B (g/L, mean ± sd) | 0.79 ± 0.32 | 0.86 ± 0.26* | 0.86 ± 0.31* | 0.86 ± 0.30* | < 0.001 |
| TG/HDLC(mean ± sd) | 0.86 ± 0.25 | 1.62 ± 0.23* | 2.82 ± 0.51* | 9.71 ± 8.85* | < 0.001 |
| Cr [µmol/L, Median (Q1-Q3)] | 67.80 (53.30–81.00) | 66.00 (54.30–79.00) | 66.00 (51.10–84.00) | 73.85 (57.00-101.00)* | < 0.001 |
| UA [µmol/L, Median (Q1-Q3)] | 232.50 (169.75-293.25) | 228.00 (162.27-293.93) | 219.00 (161.50-296.30) | 250.00 (187.00-360.12)* | < 0.001 |
| Glu[mmol/L, Median (Q1-Q3)] | 5.91 (4.89–7.90) | 6.81 (5.23–9.59)* | 7.22 (5.48–11.09)* | 8.81 (6.09–13.66)* | < 0.001 |
| HBA1c (%,mean ± sd) | 7.17 ± 2.47 | 7.78 ± 2.58* | 8.23 ± 2.69* | 9.09 ± 2.90* | < 0.001 |
| BMI (kg/m2,mean ± sd) | 23.03 ± 3.46 | 23.50 ± 3.26 | 23.29 ± 3.42 | 23.85 ± 3.44* | 0.011 |
|
Length of hospitalization (days, mean ± sd) |
16.18 ± 9.43 | 16.01 ± 9.11 | 17.12 ± 10.00 | 18.37 ± 10.44* | 0.002 |
| Gender, n (%) | 0.016 | ||||
| Male | 341 (63.86%) | 332 (61.71%) | 360 (64.29%) | 400 (70.42%) | |
| Female | 193 (36.14%) | 206 (38.29%)* | 200 (35.71%) | 168 (29.58%)* | |
| Diabetes mellitus, n (%) | < 0.001 | ||||
| No | 386 (72.28%) | 309 (57.43%) | 290 (51.69%) | 260 (45.77%) | |
| Yes | 148 (27.72%) | 229 (42.57%)* | 271 (48.31%)* | 308 (54.23%)* | |
| Hypertension, n (%) | 0.531 | ||||
| No | 381 (71.35%) | 373 (69.33%) | 404 (72.01%) | 389 (68.49%) | |
| Yes | 153 (28.65%) | 165 (30.67%) | 157 (27.99%) | 179 (31.51%) | |
| Fatty liver, n (%) | 0.014 | ||||
| No | 454 (85.02%) | 429 (79.74%) | 444 (79.14%) | 441 (77.64%) | |
| Yes | 80 (14.98%) | 109 (20.26%)* | 117 (20.86%)* | 127 (22.36%)* | |
| Biliary disease, n (%) | 0.883 | ||||
| No | 335 (62.73%) | 349 (64.87%) | 362 (64.53%) | 361 (63.56%) | |
| Yes | 199 (37.27%) | 189 (35.13%) | 199 (35.47%) | 207 (36.44%) | |
| Cardiovascular diseases, n (%) | 0.435 | ||||
| No | 442 (82.77%) | 450 (83.64%) | 467 (83.24%) | 489 (86.09%) | |
| Yes | 92 (17.23%) | 88 (16.36%) | 94 (16.76%) | 79 (13.91%) | |
| Statin use, n (%) | 0.147 | ||||
| No | 514 (96.25%) | 524 (97.40%) | 550 (98.04%) | 558 (98.24%) | |
| Yes | 20 (3.75%) | 14 (2.60%) | 11 (1.96%) | 10 (1.76%) | |
| Malignancy, n (%) | 0.003 | ||||
| No | 459 (85.96%) | 486 (90.33%) | 508 (90.55%) | 526 (92.61%) | |
| Yes | 75 (14.04%) | 52 (9.67%)* | 53 (9.45%)* | 42 (7.39%)* | |
WBC white blood cell count, CRPC-reactive protein, PCT procalcitonin, TBIL total bilirubin, ALB albumin, ALT alanine aminotransferase, AST aspartate aminotransferase, GGT gamma-glutamyl transpeptidase, ALP alkaline phosphatase, TG triglycerides, TC total cholesterol, HDL-C high-density lipoprotein cholesterol, LDL-C low-density lipoprotein cholesterol, Apo-A1 Apolipoprotein-A1, Apo-B Apolipoprotein-B, Cr creatinine, UA uric acid, Glu glucose, HBA1c hemoglobin A1c, BMI Body mass index
Values are n (%) or mean ± SD or median (quartile)
Q1 serves as the reference. * P < 0.05
Logistic regression analysis of relationship linking TG/HDL-C ratio to positive blood or pus culture
Table 2 presents the ORs and 95% CIs for culture-positive outcomes across TG/HDL-C quartiles. Unadjusted analyses revealed a significant correlation between greater TG/HDL-C ratios and positive blood or pus cultures in PLA patients (P < 0.001). This relationship persisted in Model I, which was minimally adjusted for age and sex (P < 0.001) and persisted in Model II after full adjustment for age, sex, diabetes mellitus, fatty liver, and statin use (P = 0.002). When TG/HDL-C was treated as quartiles in sensitivity analyses, the risk of a positive culture increased by 25% with each higher quartile (P for trend < 0.001).
Table 2.
TG/HDL-C and its link to positive pus or blood culture in different logistic regression models
| Unadjusted model | Adjusted model I | Adjusted model II | |
|---|---|---|---|
|
TG/HDLC [OR (95%CI) P] |
1.04 (1.02, 1.06) < 0.001 | 1.04 (1.02, 1.06) < 0.001 | 1.04 (1.01, 1.06) 0.002 |
| TG/HDLC quartiles [OR (95%CI) P] | |||
| Q1 | 1.0 | 1.0 | 1.0 |
| Q2 | 0.94 (0.71, 1.23) 0.636 | 0.94 (0.71, 1.23) 0.633 | 0.90 (0.68, 1.18) 0.440 |
| Q3 | 1.22 (0.93, 1.60) 0.159 | 1.21 (0.92, 1.59) 0.168 | 1.15 (0.87, 1.51) 0.329 |
| Q4 | 1.94 (1.46, 2.58) < 0.001 | 1.94 (1.46, 2.57) < 0.001 | 1.80 (1.35, 2.41) < 0.001 |
| P for trend | < 0.001 | < 0.001 | < 0.001 |
Adjusted model I: adjusted for gender and age, Adjusted model II: adjusted for gender, age, diabetes mellitus, fatty liver, and statin use
TG/HDL-C dose–response curve for culture-positive status
Smoothed splines were used to visualize and quantify the dose–response association between TG/HDL-C and positive blood or pus cultures in the PLA population (Fig. 1). When a nonlinear relationship was detected, threshold-effect analysis identified an inflection point at 9.2. This finding indicates that the strength and/or direction of the TG/HDLC–culture association in PLA is not uniform across this threshold. Notably, values below this cut-off exhibited a pronounced effect (OR = 1.1, 95% CI: 1.1–1.2; P < 0.001). In contrast, when the TG/HDL-C ratio exceeded the inflection value, its link to positive cultures in blood or pus disappeared (P = 0.461). These findings were further validated when adjusting for age, gender, or when additionally adjusting for age, gender, diabetes mellitus, fatty liver, the use of statins, WBC, CRP, and PCT (Table 3).
Fig. 1.
Smoothed curve fitting to delineate how incremental changes in TG/HDL-C influence culture-positive outcomes. Smooth curve fitting was employed to generate smooth curves, including (A) unadjusted; (B) gender and age-adjusted; (C) accounted for gender, age, diabetes mellitus, fatty liver, and statin use; and (D) accounted for gender, age, diabetes mellitus, fatty liver, statin use, white blood cells, C-reactive protein, and procalcitonin
Table 3.
Threshold effects of TG/HDL-C ratio on the risk of positive pus or blood culture in segmented linear regression models
| Threshold of TG/HDL-C ratio | OR (95%CI) | P value |
|---|---|---|
| Crude (unadjusted) model | ||
| < 9.2 | 1.1(1.1–1.2) | < 0.001 |
| > 9.2 | 1.0(1.0–1.0) | 0.461 |
| Adjusted model I | ||
| < 9.4 | 1.1(1.1–1.2) | < 0.001 |
| > 9.4 | 1.0(1.0–1.0) | 0.437 |
| Adjusted model II | ||
| < 9.4 | 1.1(1.1–1.2) | < 0.001 |
| > 9.4 | 1.0(1.0–1.0) | 0.419 |
| Adjusted model III | ||
| < 8.9 | 1.1(1.0-1.2) | < 0.001 |
| > 8.9 | 1.0(1.0–1.0) | 0.760 |
Adjusted model I: gender and age-adjusted, Adjusted model II: controlled for gender, age, diabetes mellitus, fatty liver, and statin use, Adjusted model III: controlled for gender, age, diabetes mellitus, fatty liver, and statin use, white blood cells, C-reactive protein, procalcitonin
OR odds ratio, CI confidence interval
The dose-response association between TG/HDL-C and the likelihood of a positive pus culture was also evaluated. In both unadjusted and fully adjusted two-segment models (Fig. 2; Table 4), TG/HDL-C displayed a positive but nonsignificant association with culture positivity across both segments (all P > 0.05).
Fig. 2.
Smoothed curve fitting of the dose-response impact of TG/HDL-C on the likelihood of positive pus culture. Smooth curve fitting was employed to generate smooth curves, including (A) unadjusted (B) gender and age-adjusted (C) accounted for gender, age, diabetes mellitus, fatty liver, statin use and (D) accounted for gender, age, diabetes mellitus, fatty liver, statin use, white blood cells, C-reactive protein, and procalcitonin
Table 4.
Dose-response analysis of TG/HDL-C ratio with positive pus culture risk via two-piecewise linear regression
| Models | OR (95%CI) | P value |
|---|---|---|
| Crude (unadjusted) model | 1.0 (1.0-1.1) | 0.175 |
| Adjusted model I | 1.0 (1.0-1.1) | 0.184 |
| Adjusted model II | 1.0 (1.0-1.1) | 0.297 |
| Adjusted model III | 1.0 (1.0-1.1) | 0.448 |
Adjusted model I: gender and age-adjusted, Adjusted model II: controlled for gender, age, diabetes mellitus, fatty liver, and statin use, Adjusted model III: controlled for gender, age, diabetes mellitus, fatty liver, and statin use, white blood cells, C-reactive protein, procalcitonin
OR odds ratios, CI confidence interval
Similarly, smoothed curve fitting for the relationship between TG/HDL-C and the risk of a positive blood culture suggested a nonlinear pattern (Fig. 3). Without covariate adjustment, a TG/HDL-C value of 7.6 was linked to a 1.2-fold increase in the odds of blood-culture positivity (95%CI 1.1–1.2, P < 0.001). Above 7.6, higher TG/HDL-C ratios did not confer additional risk (OR = 1.0, 95%CI 0.9-1.0, P = 0.248). The dose–response curve remained virtually unchanged after minimal adjustment (age and gender) and after full adjustment for eight additional covariates (Fig. 3B-D; Table 5), highlighting the robustness of these findings.
Fig. 3.
Smoothed curve fitting of the dose-response impact of TG/HDL-C on the probability of positive blood culture. Apply smoothed curve fitting to generate smooth curves, including (A) unadjusted; (B) gender and age-adjusted; (C) accounted for gender, age, diabetes mellitus, fatty liver, statin use; and (D) accounted for gender, age, diabetes mellitus, fatty liver, statin use, white blood cells, C-reactive protein, and procalcitonin
Table 5.
Threshold effects of TG/HDL-C ratio on the risk of positive blood culture in segmented linear regression models
| Threshold of TG/HDL-C ratio | OR (95%CI) | P value |
|---|---|---|
| Unadjusted model | ||
| < 7.6 | 1.2(1.1–1.2) | < 0.001 |
| > 7.6 | 1.0(0.9-1.0) | 0.248 |
| Adjusted model I | ||
| < 7.9 | 1.2(1.1–1.2) | < 0.001 |
| > 7.9 | 1.0(0.9-1.0) | 0.195 |
| Adjusted model II | ||
| < 7.9 | 1.2(1.1–1.2) | < 0.001 |
| > 7.9 | 1.0(0.9-1.0) | 0.204 |
| Adjusted model III | ||
| < 6.5 | 1.1(1.0-1.2) | 0.007 |
| > 6.5 | 1.0(0.9-1.0) | 0.156 |
Adjusted model I: gender and age-adjusted, Adjusted model II: accounted for gender, age, diabetes mellitus, fatty liver, and statin use. Adjusted model III: controlled for gender, age, diabetes mellitus, fatty liver, and statin use, white blood cells, C-reactive protein, procalcitonin
OR odds ratios, CI confidence interval
Subgroup analyses
Credibility checks included stratified analyses and formal interaction tests, both presented in Table 6. Except for WBC (P < 0.05), there were no significant interactions for any other variable. The association remained robust and significant (all P < 0.05) across all examined strata—male and female, younger age, nondiabetic status, fatty liver and non–fatty liver, lower BMI, lower WBC, elevated CRP, and elevated PCT—highlighting the reliability of the TG/HDL-C signal in PLA. The TG/HDL-C ratio was also converted to a categorical variable by dividing it into quartiles. Within the Q4 group, more pronounced associations were observed among subgroups of patients with PLA, including male and female, younger and older age, nondiabetic status, non–fatty liver, lower BMI, both lower and higher WBC, elevated CRP, and elevated PCT (all OR > 1, P < 0.05).
Table 6.
Subgroup analysis of TG/HDL-C in relation to positive pus or blood cultures
| Confounding factors | TG/HDL-C OR(95%CI)P |
Q1 | Q2 OR(95%CI)P |
Q3 OR(95%CI)P |
Q4 OR(95%CI)P |
P for interaction |
|---|---|---|---|---|---|---|
| Gender | 0.058 | |||||
| Male | 1.03 (1.00, 1.05) 0.026 | 1.0 | 0.87 (0.62, 1.23) 0.424 | 1.29 (0.91, 1.82) 0.147 | 1.88 (1.33, 2.66) < 0.001 | |
| Female | 1.08 (1.03, 1.14) 0.002 | 1.0 | 1.06 (0.68, 1.65) 0.799 | 1.09 (0.69, 1.70) 0.718 | 2.08 (1.26, 3.42) 0.004 | |
| Age (years) | 0.127 | |||||
| <60 | 1.06 (1.02, 1.10) 0.001 | 1.0 | 0.73 (0.49, 1.08) 0.116 | 1.04 (0.70, 1.54) 0.843 | 2.37 (1.55, 3.63) < 0.001 | |
| ≥60 | 1.02 (1.00, 1.05) 0.089 | 1.0 | 1.17 (0.81, 1.71) 0.401 | 1.41 (0.96, 2.07) 0.076 | 1.60 (1.09, 2.35) 0.016 | |
| Diabetes mellitus | 0.299 | |||||
| No | 1.05 (1.01, 1.09) 0.007 | 1.0 | 0.96 (0.69, 1.34) 0.805 | 1.50 (1.06, 2.12) 0.022 | 2.17 (1.49, 3.16) < 0.001 | |
| Yes | 1.03 (1.00, 1.05) 0.065 | 1.0 | 0.70 (0.42, 1.15) 0.156 | 0.70 (0.43, 1.15) 0.158 | 1.23 (0.75, 2.01) 0.419 | |
| Fatty liver | 0.160 | |||||
| No | 1.03 (1.01, 1.06) 0.006 | 1.0 | 0.92 (0.68, 1.25) 0.604 | 1.25 (0.93, 1.70) 0.142 | 1.94 (1.41, 2.65) < 0.001 | |
| Yes | 1.08 (1.01, 1.16) 0.018 | 1.0 | 0.98 (0.51, 1.90) 0.957 | 1.08 (0.56, 2.07) 0.823 | 1.94 (0.99, 3.80) 0.051 | |
| BMI (kg/m2) | 0.215 | |||||
| <25 | 1.07 (1.02, 1.12) 0.002 | 1.0 | 0.90 (0.62, 1.30) 0.566 | 1.07 (0.73, 1.58) 0.716 | 2.42 (1.56, 3.75) < 0.001 | |
| ≥25 | 1.03 (0.98, 1.08) 0.221 | 1.0 | 0.79 (0.44, 1.41) 0.423 | 1.04 (0.55, 1.95) 0.909 | 1.19 (0.66, 2.15) 0.560 | |
| WBC (109/L) | 0.006 | |||||
| <10 | 1.11 (1.05, 1.18) 0.001 | 1.0 | 1.01 (0.64, 1.58) 0.974 | 1.62 (1.01, 2.58) 0.044 | 4.23 (2.41, 7.40) < 0.001 | |
| ≥10 | 1.02 (1.00, 1.05) 0.073 | 1.0 | 0.96 (0.67, 1.35) 0.799 | 1.07 (0.76, 1.50) 0.716 | 1.41 (1.01, 1.98) 0.047 | |
| CRP (mg/L) | 0.168 | |||||
| <8 | 0.64 (0.30, 1.39) 0.259 | 1.0 | 0.37 (0.03, 4.71) 0.447 | 0.27 (0.04, 1.95) 0.196 | 0.00 (0.00, Inf) 0.995 | |
| ≥8 | 1.06 (1.03, 1.09) < 0.001 | 1.0 | 0.76 (0.56, 1.03) 0.079 | 1.19 (0.87, 1.62) 0.281 | 1.86 (1.34, 2.59) < 0.001 | |
| PCT (ng/mL) | ||||||
| <0.05 | 0.52 (0.16, 1.71) 0.281 | 1.0 | 0.33 (0.02, 6.65) 0.472 | 0.33 (0.03, 4.19) 0.395 | 0.00 (0.00, Inf) 0.997 | 0.200 |
| ≥0.05 | 1.05 (1.02, 1.08) < 0.001 | 1.0 | 0.85 (0.64, 1.13) 0.269 | 1.20 (0.90, 1.61) 0.215 | 1.87 (1.38, 2.53) < 0.001 | |
OR odds ratio, CI confidence interval, Q quartile, TG triglycerides, HDL-C high-density lipoprotein cholesterol, BMI body mass index, WBC white blood cells , CRPC-reactive protein, PCT procalcitonin
Premorbid blood-lipid data were available for 246 patients and were analyzed retrospectively. Table 7 illustrates the change
Table 7.
Comparison of TG/HDL-C levels at the onset of PLA and before the onset in PLA patients with positive pus or blood cultures
| TG/HDL-C [median(Q1,Q3)] |
Onset of PLA | Before PLA | t-test | P-value |
|---|---|---|---|---|
| Pus or blood cultrue | 2.26(1.24,5.07) | 1.39(0.67,2.31) | 4.541 | < 0.001 |
| Pus cultrue | 2.31(1.38,5.09) | 1.44(0.76,2.33) | 4.162 | < 0.001 |
| blood cultrue | 2.96(1.12–5.77) | 1.46(0.77,2.41) | 2.921 | 0.005 |
PLA pyogenic liver abscess, TG triglyceride, HDL-C high-density lipoprotein cholesterol
in TG/HDL-C from the premorbid period to the acute episode in culture-positive PLA. Regardless of culture source (blood, pus, or both), TG/HDL-C levels at disease onset were consistently higher than premorbid values (all comparisons P < 0.05).
Discussion
For the first time, a multicenter, 10-year longitudinal investigation has delineated how the TG/HDL-C ratio relates to culture-confirmed PLA in a Chinese population. The TG/HDL-C ratio was closely correlated with positive blood or pus cultures in patients with PLA (P < 0.001), independent of age, sex, BMI, diabetes mellitus, fatty liver, CRP, and PCT. Smoothed curve fitting suggested that the relationship between TG/HDL-C and culture positivity followed a curvilinear, dose-dependent pattern, and a similar curvilinear relationship was observed for positive blood cultures alone. Sensitivity and subgroup analyses confirmed the consistency and robustness of the findings.
TG is the body’s primary form of energy storage, accumulating in adipocytes, whereas HDL-C transports excess cholesterol to the liver for disposal or recycling [21]. The TG/HDL-C ratio is a valuable composite index in predicting liver and infectious diseases. In obese adults, a biopsyproven case–control analysis identified TG/HDL-C as a robust indicator for metabolic dysfunction–associated steatotic liver disease (MASLD) [22]. Emerging evidence also suggests that, among infected patients, elevated TG/HDL-C correlates with more severe episodes of recurrent acute pancreatitis [8]. Compared with nonsevere cases, PLA patients who developed septic shock or multiple organ dysfunction syndromes exhibited significantly reduced HDL-C (P < 0.001) and elevated TG (P = 0.003) [23]. Additionally, a low TC concentration at admission was an autonomous determinant of severe PLA. Compared with survivors, COVID-19 fatalities presented with higher WBC, neutrophils, CRP, and TG/HDL-C (all P < 0.05), and TG/HDL-C positively correlated with WBC, neutrophils, and CRP (all P < 0.05) [24]. Moreover, elevated TG/HDL-C values paralleled rising SYNTAX scores, reflecting greater coronary lesion severity and complexity (P < 0.001) [25]. In logistic modeling, BMI and TG/HDL-C emerged as independent factors influencing the SYNTAX score [25].
To date, no reports have explored whether TG/HDL-C affects the etiologic outcomes of PLA, although dyslipidemia has been documented in various infectious diseases, including PLA and sepsis. A prospective cohort study found no link of TG levels to the risk of infectious diseases—gastrointestinal infections, bacterial pneumonia, urinary tract infections, and sepsis—whereas HDL-C showed a U-shaped relationship with disease risk across these conditions [26]. During episodes of sepsis, HDL-C dropped markedly, whereas TG levels were not associated with prognostic indicators such as ICU mortality, ICU duration of stay, or Sequential Organ Failure Assessment (SOFA) scores exceeding 6 [27]. Notably, HDL-C was negatively correlated with the risk of infection related hospitalization (HR = 0.84 per mmol/L, P = 0.004), whereas TG levels showed a positive correlation. A Mendelian randomization analysis further indicated that HDL-C, rather than TG, was causally linked to the risk of infectious hospitalization [10].
Feng et al. reported that patients with severe PLA (septic shock or multiple organ dysfunction syndrome) had significantly greater TG levels and markedly lower HDL-C than nonsevere patients (all P < 0.05) [19]. In the present study, WBC, CRP, and PCT all peaked in Q4, coinciding with a TG/HDL-C ratio significantly greater than that observed in Q1–Q3; this elevation was accompanied by more pronounced liver injury and inflammation. These findings support the view that PLA itself destabilizes lipid homeostasis, leading to hypertriglyceridemia and hypo-α-lipoproteinemia and, consequently, a higher TG/HDL-C ratio. Because the predominant bacteria causing PLA are gramnegative and lipopolysaccharide is a major component of their outer membrane, previous research demonstrated that lipopolysaccharide activity was positively associated with serum cholesterol and TG levels but negatively associated with HDL-C levels [28]. Based on the interaction analysis, it is evident that WBC count significantly influences the relationship between TG/HDL-C and culture results in PLA. Peripheral WBC count is closely correlated with bloodlipid levels [29], and TG levels are strongly, positively correlated with total WBC count [30].
Several mechanisms could explain the observed nonlinear dose–response linking TG/HDL-C to culture-confirmed PLA: (1) Lipotoxicity: Acting as an index of lipid imbalance, the TG/HDL-C ratio might promote free-fatty-acid buildup when it surpasses a particular threshold. These fatty acids interact with reactive oxygen species (ROS) to produce toxic lipid metabolites, typically through lipid peroxidation. This process ultimately causes hepatocyte apoptosis or damage; (2) Inflammatory response: The overactivation of Kupffer cells (hepatic macrophages) during sepsis may contribute to liver dysfunction and damage to hepatic sinusoidal endothelial cells. Upon activation, Kupffer cells secrete substantial amounts of ROS, and tumor necrosis factor-alpha (TNF-α), causing oxidative stress and damage within cells. During the occurrence of PLA, the clearance rate of plasma TG was diminished due to the stimulation of catecholamines, endotoxins, and cytokines, alongside an increase in TG produced by the liver. This cascade of events ultimately led to elevated plasma TG levels. Binding of bacterial lipopeptides and diacylated peptides to HDL-C underlies their role in immunomodulation. By interrupting Toll-like receptors(TLR)−2/TLR-4 pathways [36], quenching macrophage activation, and curbing cytokine production, these processes precipitate a reduction in HDL-C [36].
Strengths
This study has some strengths. First, the TG/HDL-C ratio was modeled both categorically and continuously to evaluate its relationship with the risk of positive pus or blood cultures in PLA. This approach minimizes information loss and enhances the quantitative assessment of the association. Second, the study is the first to document nonlinear correlations between these variables, providing new insights into the dose–response relationship between the TG/HDL-C ratio and culture positivity in PLA. The stability and reliability of the association were further supported by sensitivity, subgroup, and interaction analyses.
Limitations
Nevertheless, the research comes with several limitations. To begin with, being a retrospective cross-sectional study, it provides relatively weak evidence regarding the relationship between exposure and outcomes, making it challenging to establish causality. To address this limitation, rigorous statistical methods were employed for subgroup and sensitivity analyses, which yielded consistent results that enhanced the reliability of the research outcomes. Secondly, some studies suggest that cholesterol-lowering drugs, such as statins, may reduce the risk of PLA [15]. However, statin use was observed in only 60 patients, accounting for 2.22% of the study population. Thus, this study cannot eliminate the possibility of minor bias in the research results due to statin use.
Taken together, the TG/HDL-C ratio exhibited a threshold-style, nonlinear association with the odds of positive cultures among PLA cases. In resource constrained settings where culture confirmation is challenging, TG/HDL-C may serve as an inexpensive early warning tool for identifying patients at elevated risk. Specifically, a TG/HDL-C threshold of < 9.2 could prompt clinicians to initiate broader spectrum antibiotics or implement closer monitoring while awaiting culture results, even if the cultures ultimately prove negative. This approach may aid clinicians in tailoring management strategies and reducing treatment delays, leading to enhanced clinical outcomes and improved healthcare for individuals with PLA.
Supplementary Information
Acknowledgements
Not applicable.
Authors’ contributions
Qian Huang, Zi-Xuan Huang, Hou-Xiong Lin, Yong Xu contributed equally to this work and share the first authorship. Qian Huang, Xiao-Xu Xie, Meng-Xin Lin, Da-Wu Zeng participated in the conception and design of this study. Qian Huang, Xiao-Xu Xie analyzed the data. Meng-Xin Lin, Da-Wu Zeng performed the study. Qian Huang, Zi-Xuan Huang, Hou-Xiong Lin, Yong Xu, Meng-Xin Lin partipated in the collection or generation of the study data. Qian Huang, Zi-Xuan Huang wrote the manuscript. Qian Huang, Hou-Xiong Lin, Yong Xu were involved in the analyses or interpretation of the data.
Funding
This study was supported by the Fujian Provincial Health Technology Project (No. 2023CXA023).
Data availability
No datasets were generated or analysed during the current study.
Declarations
Ethics approval and consent to participate
The protocol for this study received approval from the ethics committees of the participating hospitals. Written informed consent for participation was not required for this retrospective analysis without any treatment intervention.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Xiao-Xu Xie, Email: xxx@fjmu.edu.cn.
Meng-Xin Lin, Email: 305222711@qq.com.
Da-Wu Zeng, Email: Zengdw1980@fjmu.edu.cn.
References
- 1.Lübbert C, Wiegand J, Karlas T. Therapy Liver Abscesses Viszeralmedizin. 2014;30:334–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Longworth S, Han J. Pyogenic liver abscess. Clin Liver Dis. 2015;6:51–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Long B, Koyfman A. Best clinical practice: blood culture utility in the emergency department. J Emerg Med. 2016;51:529–39. [DOI] [PubMed] [Google Scholar]
- 4.Khim G, Em S, Mo S, Townell N. Liver abscess: diagnostic and management issues found in the low resource setting. Br Med Bull. 2019;132:45–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Chemaly RF, Procop GW, Procop GW. Microbiology of liver abscesses and the predictive value of abscess gram stain and associated blood cultures. Diagnostic microbiology and infectious disease. 2003;46:245–8. [DOI] [PubMed] [Google Scholar]
- 6.Wendel M, Paul R, Fau - Heller AR, Heller AR. Lipoproteins in inflammation and sepsis. II. Clinical aspects. Intensive Care Med. 2007;33:25–35. [DOI] [PubMed] [Google Scholar]
- 7.Che B, Zhong C, Zhang R, Pu L, Zhao T, Zhang Y, et al. Triglyceride-glucose index and triglyceride to high-density lipoprotein cholesterol ratio as potential cardiovascular disease risk factors: an analysis of UK biobank data. Cardiovasc Diabetol. 2023;22:34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Jia X, Zhang X, Sun D, Yang N, Li R, Luo Z. Triglyceride to HDL-C ratio is associated with plasma D-dimer levels in different types of pancreatitis. Sci Rep. 2022;12:12952. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Golucci APBSMF, Ribeiro AF, Nogueira RJN. Lipid profile associated with the systemic inflammatory response syndrome and sepsis in critically ill patients. Nutrition. 2018;55:7–14. [DOI] [PubMed] [Google Scholar]
- 10.Trinder M, Walley KR, Boyd JH, Brunham LR. Causal inference for genetically determined levels of high-density lipoprotein cholesterol and risk of infectious disease. Arterioscler Thromb Vasc Biol. 2020;40:267–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Tanaka SSJ, Bouzid D, Robert T, Dehoux M, Snauwaert A, et al. Relationship between lipoprotein concentrations and short-term and 1-year mortality in intensive care unit septic patients: results from the HIGHSEPS study. Ann Intensive Care. 2021;11:11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Sagud M, Vlatkovic S, Svob Strac D, Sviben M, Zivkovic M, Vilibic M, et al. Latent Toxoplasma gondii infection is associated with decreased serum triglyceride to high-density lipoprotein cholesterol ratio in male patients with schizophrenia. Compr Psychiatry. 2018;82:115–20. [DOI] [PubMed] [Google Scholar]
- 13.Yin D, Ji C, Zhang S, Wang J, Lu Z, Song X, et al. Clinical characteristics and management of 1572 patients with pyogenic liver abscess: a 12-year retrospective study. Liver Int. 2021;41:810–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Shi SH, Feng XN, Lai MC, Kong HS, Zheng SS. Biliary diseases as main causes of pyogenic liver abscess caused by extended-spectrum beta-lactamase-producing Enterobacteriaceae. Liver Int. 2017;37:727–34. [DOI] [PubMed] [Google Scholar]
- 15.Liao KF, Cheng KC, Lin CL, Lai SW. Statin use correlates with reduced risk of pyogenic liver abscess: a population-based case-control study. Basic Clin Pharmacol Toxicol. 2017;121:144–9. [DOI] [PubMed] [Google Scholar]
- 16.He SY, J.Wang H, et al. Association of blood culture with carbapenem use in pyogenic liver abscess: a two-center retrospective study. BMC Emerg Med. 2021;21:58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Li J, Wang Y, Luo J, Yin Z, Huang W, Zhang J. Development and validation of a nomogram for predicting sepsis in patients with pyogenic liver abscess. Sci Rep. 2023;13:10849. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Yoon JH, Kim YJ, Kim SI. Prognosis of liver abscess with no identified organism. BMC Infect Dis. 2019;19:488. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Feng T, Zhang W, Hou X, Yuan H, Cai J, Jiang Z, et al. Clinical significance of dynamic variation of low cholesterol and its prognostic value in patients with pyogenic liver abscesses: a retrospective study. BMC Infect Dis. 2023;23:70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Shyam K, Kumar NP, Balram J, Omar Shyam K, Kumar, et al. Pyogenic liver abscess: clinical features and microbiological profiles in tertiary care center. J Family Med Prim Care. 2020;9:4337–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Iwani NA, Jalaludin MY, Zin RM, Fuziah MZ, Hong JY, Abqariyah Y, et al. Triglyceride to HDL-C ratio is associated with insulin resistance in overweight and obese children. Sci Rep. 2017;7:40055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Martínez-Montoro JA-O, Martínez-Sánchez MA, Balaguer-Román AA-O, Fernández-Ruiz VE, Hernández-Barceló JE, Ferrer-Gómez M, et al. Triglyceride to HDL cholesterol ratio for the identification of MASLD in obesity: A liver Biopsy-Based Case-Control study. Nutrients. 2024;16:1310. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Feng TTZW, Hou XX, Yuan H, Cai JY, et al. Clinical signifcance of dynamic variation of low cholesterol and its prognostic value in patients with pyogenic liver abscesses:a retrospective study. BMC Infect Dis. 2023;23:70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Peng FL, Zhang S, Zhong Q, Wu Y. Triglyceride/high-density lipoprotein cholesterol ratio is associated with the mortality of COVID-19: a retrospective study in China. Int J Gen Med. 2022;15:985–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Demir ÖFK. The relationship between triglyceride/high-density lipoprotein cholesterol ratio and the severity of coronary artery disease in patients presenting with acute coronary syndrome. Coron Artery Dis. 2024;35:564–71. [DOI] [PubMed] [Google Scholar]
- 26.Madsen CM, Varbo A, Tybjærg-Hansen A, Frikke-Schmidt R, Nordestgaard BG. U-shaped relationship of HDL and risk of infectious disease: two prospective population-based cohort studies. Eur Heart J. 2018;39:1181–90. [DOI] [PubMed] [Google Scholar]
- 27.Tanaka SA-O, Labreuche J, Drumez E, Harrois A, Hamada S, Vigué B, et al. Low HDL levels in sepsis versus trauma patients in intensive care unit. Ann Intensive Care. 2017;7: 60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Pussinen PJ, Havulinna As Fau -, Lehto M, Lehto M, Fau - Sundvall J, Sundvall J, Fau - Salomaa V, Salomaa V. Endotoxemia is associated with an increased risk of incident diabetes. Diabetes Care. 2011;34:392–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Hu WZ, Su P, Li Q, Hang D. Y. Peripheral leukocyte counts vary with lipid levels, age and sex in subjects from the healthy population. Atherosclerosis 2020;308:15–21. [DOI] [PubMed]
- 30.Tucker BS, McDonald S, Rye H, Patel KA, Ong S, Cochran KL. B. J. The association of serum lipid and lipoprotein levels with total and differential leukocyte counts: results of a cross-sectional and longitudinal analysis of the UK biobank. Atherosclerosis 2021;319:1–9. [DOI] [PubMed]
- 31.Engin AB. What Is Lipotoxicity? Nutrients. 2017;960:197–220. [DOI] [PubMed] [Google Scholar]
- 32.Gai ZA-O, Wang T, Visentin M, Kullak-Ublick GA, Fu X, Wang Z. Lipid Accumulation Chronic Kidney Disease Nutrients. 2019;11:722. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Wen YA-O, Lambrecht J, Ju C, Tacke F. Hepatic macrophages in liver homeostasis and diseases-diversity, plasticity and therapeutic opportunities. Cell Mol Immunol. 2021;18:45–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.van Leeuwen HJ, van Strijp JAG, van Strijp Ja Fau -, Verhoef J, Verhoef J, van Fau KPM, van Kessel KP. Lipoprotein metabolism in patients with severe sepsis. Critical care medicine. 2003;31:1359–66. [DOI] [PubMed] [Google Scholar]
- 35.Sorokin AV, Karathanasis SK, Yang ZH, Freeman L, Kotani K, Remaley AT. COVID-19-associated dyslipidemia: implications for mechanism of impaired resolution and novel therapeutic approaches. FASEB J. 2022;34:9843–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.van Bergenhenegouwen J, Kraneveld AD, Rutten L, Garssen J, Vos AP, Hartog A. Lipoproteins attenuate TLR2 and TLR4 activation by bacteria and bacterial ligands with differences in affinity and kinetics. BMC Immunol. 2016;17:42. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
No datasets were generated or analysed during the current study.