Abstract
Background
Patients with hematologic malignancies are at high risk of developing catheter-related thrombosis (CRT) following peripherally inserted central catheter (PICC) placement; however, reliable prediction tools based on inflammation and coagulation biomarkers remain limited. This study aimed to investigate inflammation and coagulation biomarkers associated with CRT and to construct an individualized risk prediction model.
Methods
This retrospective study enrolled 172 patients with hematologic malignancies who underwent PICC placement (April 2022–December 2023). Inflammatory [white blood cell count, C-reactive protein (CRP), interleukin (IL)-6, IL-1β] and coagulation [fibrinogen (FIB), D-dimer (D-D), international normalized ratio (INR), activated partial thromboplastin time (APTT), prothrombin time (PT), thrombin time (TT)] indicators were collected. Patients were classified into the CRT and non-CRT groups according to the occurrence of CRT, defined as both symptomatic and asymptomatic thrombi detected by scheduled upper-extremity duplex ultrasound following PICC placement. Least absolute shrinkage and selection operator (LASSO) regression and multivariate logistic regression were used to identify independent risk factors. A nomogram was built and validated internally using receiver operating characteristic (ROC) curves, decision curve analysis (DCA), calibration curves, and bootstrap.
Results
Overall CRT, including both symptomatic and asymptomatic events detected by scheduled upper-extremity ultrasound screening following PICC placement, occurred in 58 patients (33.7%; 15 symptomatic and 43 asymptomatic). The median time to thrombus was 17.5 days [interquartile range (IQR), 12.5–28.0 days]. Platelet count (PLT), IL-1β, CRP, IL-6, FIB, D-D, PT, and TT were higher in the CRT group (all P<0.05). Multivariate analysis identified PLT [odds ratio (OR) =0.936], IL-1β (OR =1.375), CRP (OR =1.367), IL-6 (OR =1.134), FIB (OR =1.870), and D-D (OR =1.005) as independent risk factors (all P<0.05). The nomogram achieved an area under the receiver operating characteristic curve (AUC) of 0.947 [95% confidence interval (CI): 0.909–0.986] with good calibration.
Conclusions
PLT, IL-1β, CRP, IL-6, FIB, and D-D are independently associated with CRT following PICC placement in patients with hematologic malignancies. The prediction model performed well internally but requires external multi-center validation before clinical application.
Keywords: Hematologic malignancies, catheter-related thrombosis (CRT), inflammation, coagulation, risk prediction model
Highlight box.
Key findings
• In 172 patients with hematologic malignancies undergoing peripherally inserted central catheter (PICC) placement, catheter-related thrombosis (CRT) occurred in 58 patients (33.7%), with most events being asymptomatic and detected through scheduled ultrasound surveillance.
• Platelet count (PLT), interleukin-1β (IL-1β), C-reactive protein (CRP), interleukin-6 (IL-6), fibrinogen (FIB), and D-dimer (D-D) were identified as independent predictors of CRT.
• A nomogram incorporating inflammation and coagulation biomarkers showed excellent discrimination and calibration, with potential value for individualized CRT risk assessment.
What is known and what is new?
• CRT is a common complication after PICC placement in patients with hematologic malignancies. Existing risk assessment approaches mainly rely on clinical characteristics and may not adequately reflect the biological mechanisms underlying thrombosis development.
• This study established a biomarker-based prediction model integrating inflammatory and coagulation pathways. By combining six independent biomarkers, the model provides a more biologically informed and clinically applicable approach for early identification of patients at high risk for CRT.
What is the implication, and what should change now?
• The findings highlight the importance of inflammation-coagulation interactions in CRT development. Risk stratification based on these biomarkers may help clinicians and nursing teams implement targeted surveillance strategies, including closer clinical monitoring and earlier ultrasound assessment in high-risk patients. External multicenter validation is needed before routine clinical application.
Introduction
Patients with hematologic malignancies often exhibit a state of heightened inflammatory response and hypercoagulability due to multiple factors, including tumor burden, chemotherapeutic agents, impaired bone marrow function, and abnormal immune status. Consequently, they are at high risk for venous thromboembolism (VTE) (1,2). The peripherally inserted central catheter (PICC) is widely used in hematology departments because of its safety and suitability for long-term infusion. However, PICC placement can also induce local hemodynamic changes and venous endothelial injury. When combined with the patient’s inherent hypercoagulable tendency, catheter-related thrombosis (CRT), including both symptomatic and asymptomatic thrombi detected by ultrasound screening, becomes a major complication following PICC placement (3-5). CRT may not only cause catheter dysfunction and delay chemotherapy but can also progress to pulmonary embolism, posing a serious threat to patient safety (6). Therefore, identifying key factors underlying CRT following PICC placement in patients with hematologic malignancies and establishing an effective prediction model hold significant clinical value.
Recent studies have increasingly focused on the role of the inflammation-coagulation axis in CRT. Inflammatory cytokines such as interleukin (IL)-6, IL-1β, and C-reactive protein (CRP) can directly accelerate thrombus formation by inducing tissue factor expression, promoting fibrin generation, and causing endothelial damage (7-9). Coagulation parameters such as fibrinogen (FIB) and D-dimer (D-D) reflect coagulation activation and fibrinolytic burden, and can more sensitively indicate thrombotic risk (10,11). Unlike solid tumors, hematologic malignancies are characterized by more fluctuating inflammatory levels and more pronounced coagulation abnormalities. Traditional prediction methods relying solely on clinical variables are limited, and there is a greater need for integrated models incorporating inflammation-coagulation biomarkers. However, current research on CRT, particularly overall CRT detected by scheduled ultrasound screening rather than symptomatic events alone, remains limited in patients with hematologic malignancies. Furthermore, accurate, visual, and clinically applicable predictive tools for individualized risk assessment are still lacking.
To address this gap, we enrolled 172 patients with hematologic malignancies who underwent PICC placement in our hospital between April 2022 and December 2023. The primary outcome was overall CRT, including both asymptomatic and symptomatic cases, confirmed by routine ultrasound screening after catheterization. We systematically collected inflammatory indicators (IL-6, IL-1β, CRP, etc.) and coagulation parameters [FIB, D-D, prothrombin time (PT), thrombin time (TT), etc.], compared differences between the CRT and non-CRT groups, and used multivariate logistic regression to identify independent risk factors. Based on the significant variables, we further constructed a nomogram risk prediction model and evaluated its discrimination, calibration, and clinical utility using receiver operating characteristic (ROC) curves, decision curve analysis (DCA), and calibration curves. Our aim was to develop an operable predictive tool suitable for patients with hematologic malignancies. Leveraging real-world data and multi-indicator comprehensive assessment, this study may provide preliminary evidence to address the limitations of existing prediction models that lack biological basis and practical applicability, thereby improving the safety and continuity of PICC management in hematologic cancer patients and ultimately reducing thrombotic risk and enhancing clinical outcomes. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1203/rc).
Methods
Study population and sample size calculation
This study was a single-center, retrospective, exploratory prediction model analysis. The reported incidence of CRT varies widely in the literature, largely depending on the use of routine ultrasound screening and outcome definitions. A systematic review and meta-analysis by Chopra et al. (12) demonstrated that PICCs are associated with a substantial risk of VTE, particularly among patients with cancer. Prospective ultrasound screening studies have shown (13-15) that the incidence of asymptomatic central venous catheter-related upper extremity deep vein thrombosis (UEDVT) in patients with cancer ranges from approximately 11.7% to 44%, depending on patient characteristics, catheter type, and surveillance strategy. Based on our protocol of routine ultrasound screening on days 7, 14, and 28 after catheterization, and integrating the above literature, we expected an overall CRT (symptomatic + asymptomatic) detection rate of 25–35%, with a conservative estimate of 25%.
We planned to include 6 candidate predictor variables in the prediction model. According to the conservative rule of thumb requiring at least 5 events per variable (EPV =5) in logistic regression analysis, the minimum number of positive events needed was 6×5=30. With an expected overall CRT incidence of 25%, the minimum total sample size was 30/0.25=120. Finally, we enrolled 172 patients with hematologic malignancies who underwent PICC placement in our hospital between April 2022 and December 2023. Among them, scheduled ultrasound screening detected overall CRT in 58 patients (33.7%) yielding an EPV of 58/6≈9.7, which meets the stability requirement for an exploratory prediction model (EPV ≥5).
We selected 172 patients with hematologic malignancies who underwent PICC placement in the hematology department of our hospital between April 2022 and December 2023. Inclusion criteria were: leukemia, lymphoma, or other hematologic malignancies confirmed clinically and pathologically; first PICC placement; ultrasound-guided placement of a single-lumen 4 Fr Bard PICC catheter in the upper extremity; standardized PICC placement procedure and completed follow-up; baseline inflammatory and coagulation parameters measured within 1 week before catheterization with complete data; PICC placement performed under ultrasound guidance. Exclusion criteria included: venous thrombosis unrelated to the PICC catheter;; prior history of deep vein thrombosis (DVT) or significant coagulation abnormalities before catheterization; concurrent active bleeding or coagulation disorders that could affect assessment; development of septic shock or multi-organ failure during catheterization that could influence evaluation; inability to complete follow-up or missing data. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Human Ethics Committee of Guangxi Medical University Cancer Hospital (Ethics approval No. KY2023059). Given the retrospective nature of the study and the use of de-identified clinical data, the requirement for informed consent was waived by the Ethics Committee.
PICC placement
Operators and equipment
Before insertion, ultrasound guidance was performed using a Site-Rite 80 ultrasound system (CHISON Medical Technologies Co., Ltd., Wuxi, China) equipped with a high-frequency linear probe (5–12 MHz) to assess the diameter, course, and branches of the bilateral brachial veins. Based on the patient’s venous anatomy, the vein with the largest diameter and straightest course was selected preferentially. PICC placement was performed by certified intravenous therapy specialist nurses at our hospital.
Insertion procedure
All enrolled patients underwent ultrasound-guided puncture using the modified Seldinger technique and microintroducer sheath technique. The steps were as follows: (I) the patient was placed in a supine position with the puncture-side upper extremity abducted to 90°. The skin was sterilized, and sterile drapes were placed; (II) under ultrasound guidance, a 21-G needle was used to puncture the target vein. After blood return was observed, a 0.018-inch guidewire was inserted; (III) after skin dilation, a tear-away sheath was advanced over the guidewire, and the guidewire and dilator were removed; (IV) the catheter was inserted through the sheath. The catheter used was a 4 Fr single-lumen polyurethane catheter (model: Bard PowerPICC Solo, outer diameter 1.33 mm, length 50–60 cm); (V) blood return was aspirated to confirm the catheter was within the vein. The catheter was flushed with 10 mL of normal saline in a pulsatile manner, and a heparin cap was attached. During the procedure, we recorded the name of the punctured vein, number of puncture attempts (defined as the number of times the needle tip entered the skin), and occurrence of puncture complications (e.g., arterial injury, hematoma, nerve irritation). After insertion, a transparent dressing was used to secure the puncture site. Standardized catheter maintenance was performed during the catheter dwell period, including at least weekly dressing changes, assessment of catheter function, and flushing with normal saline to ensure catheter patency and cleanliness of the puncture site, thereby reducing the risk of infection and thrombosis.
Catheter tip positioning
After catheter placement, all patients underwent bedside chest X-ray to confirm the catheter tip position. The tip was required to be located in the mid-to-lower segment of the superior vena cava (specific anatomical landmarks: between the carina level and the level of the 6th thoracic vertebra, or at the junction of the right paratracheal line and the second anterior rib). The tip position was reviewed jointly by the operating nurse and a radiologist. Abnormal tip position was defined as the tip being located outside the superior vena cava, or in the upper segment of the superior vena cava/right atrium. For patients with abnormal tip position, the catheter was adjusted to the ideal position under X-ray guidance, and the number of adjustments was recorded.
Definition of CRT and patient grouping
Ultrasound screening protocol
All patients underwent routine, scheduled upper extremity color Doppler ultrasound examinations on days 7 (±2 days), 14 (±2 days), and 28 (±3 days) after catheterization, regardless of the presence of clinical symptoms. In addition, if a patient developed any of the following clinical signs suggestive of CRT, an ultrasound examination was performed immediately: (I) swelling of the catheterized upper extremity (circumference increase ≥2 cm compared with the contralateral side); (II) pain or tenderness along the venous pathway; (III) local warmth; (IV) superficial venous dilatation or erythema; (V) limited upper extremity movement or patient discomfort due to swelling. Both symptomatic and asymptomatic thrombi detected through this scheduled ultrasound screening protocol were included in the primary outcome (overall CRT).
Definition of CRT
Ultrasound examinations were independently performed by two imaging physicians with more than 5 years of experience in vascular ultrasound, using a GE color Doppler ultrasound system equipped with a high-frequency linear probe (5–12 MHz). According to the International Society on Thrombosis and Haemostasis (ISTH) consensus criteria for CRT, the definition of CRT required all of the following: (I) the thrombus was located within the segment of the vein containing the catheter (i.e., the vein where the catheter was placed, including the basilic, brachial, cephalic, axillary, and subclavian veins); (II) the thrombus was anatomically adjacent to the catheter (attached to the catheter wall or within the same venous lumen); (III) the ultrasound imaging met the features of DVT (16). All detected thrombi were further classified: According to anatomical relationship with the catheter: CRT (thrombus located in the catheterized vein segment and adjacent to the catheter) versus non-CRT (thrombus in a non-catheterized vein, e.g., contralateral upper extremity or lower extremity veins; such cases were not included as outcomes in this study). According to thrombus morphology: mural thrombus (non-occlusive, attached to one side of the vessel wall, with some blood flow still visible in the lumen) versus occlusive thrombus (completely obstructing the lumen, with no distal blood flow signal). According to thrombus extent: proximal thrombosis (involving the axillary vein or more proximal subclavian vein) versus distal thrombosis (involving only distal veins such as the brachial, basilic, or cephalic veins). According to clinical presentation: symptomatic CRT (presence of at least one of limb swelling, pain, warmth, or superficial venous dilatation) versus asymptomatic CRT (absence of the above clinical manifestations).
Patient grouping and study outcomes
Based on the above criteria, all enrolled patients were divided into two groups: non-CRT group (114 patients): no CRT detected on ultrasound at any time point. CRT group (58 patients), including asymptomatic CRT subgroup (43 patients) and symptomatic CRT subgroup (15 patients). The primary outcome was overall CRT, defined as thrombus formation within the catheterized upper-extremity vein confirmed by scheduled duplex ultrasound examinations, including both asymptomatic and symptomatic events. The prediction model was developed using overall CRT as the study endpoint. Symptomatic CRT was prespecified as a secondary outcome; however, owing to the limited number of symptomatic events, only exploratory univariate analyses were performed for this outcome, and it was not included in multivariable model development.
Data collection
Baseline data were collected at the time of PICC placement, including: gender, age, body mass index (BMI), tumor type, and comorbidities. PICC insertion-related variables included catheterization method (only ultrasound-guided insertions were included in this study), catheterization side, punctured vein, number of puncture attempts, abnormal tip position, concurrent chemotherapy at catheterization, chemotherapy regimen, and puncture complications. We also collected pre-insertion platelet count (PLT, ×109/L), infection status (presence of active infection within 48 hours before insertion, defined as body temperature ≥38.5 ℃ with elevated procalcitonin or CRP, or microbiological evidence), and disease activity (classified as active/inactive based on serum lactate dehydrogenase, β2-microglobulin, and bone marrow blast percentage).
Laboratory biomarkers were measured using peripheral venous blood collected before PICC placement. Samples were sent to our hospital’s clinical laboratory center, and inflammatory and coagulation indicators were tested under uniform instrument and reagent conditions. Inflammatory biomarkers included white blood cell count (WBC), CRP, IL-6, and IL-1β. IL-1β and IL-6 were measured by electrochemiluminescence using a Roche Cobas e801 automated immunoassay analyzer with corresponding kits. CRP was measured by immunoturbidimetry using a Beckman AU5800 automated biochemical analyzer with corresponding kits. Coagulation indicators included FIB, D-D (D-D concentrations were expressed as µg/L FEU), international normalized ratio (INR), activated partial thromboplastin time (APTT), PT, and TT. FIB was measured by the Clauss method, and D-D by immunoturbidimetry, both using a Sysmex CS5100 automated coagulation analyzer with corresponding reagents. PT, INR, APTT, and TT were measured by coagulation method using the same Sysmex CS5100 analyzer and corresponding reagents. All assays were performed according to strict quality control procedures to ensure reliable results.
Missing data handling
The missing rate for each variable included in the analysis was calculated. In this study, IL-1β had 1 missing value (0.6%) and IL-6 had 2 missing values (1.2%); all other continuous and categorical variables had no missing data. For continuous variables with a missing rate <5%, median imputation was used. For categorical variables, mode imputation was used. No categorical variables contained missing data; therefore, no imputation was required for categorical variables. After imputation, complete data were available for all 172 patients, and all cases were included in the subsequent univariate analysis.
Statistical analysis
Statistical analyses were performed using SPSS 26.0 and R 4.5.3. Normally distributed continuous variables were expressed as mean ± standard deviation, and group comparisons were performed using independent samples t-test or analysis of variance. Non-normally distributed variables were expressed as median [interquartile range (IQR)], with comparisons conducted using Mann-Whitney U test or Kruskal-Wallis H test. Categorical variables were expressed as percentages, and group comparisons were performed using χ2 test or Fisher’s exact test. All candidate variables were first evaluated by univariate analysis for their association with CRT. Variables with P<0.05 were subsequently entered into a least absolute shrinkage and selection operator (LASSO) regression model for variable selection. Variables retained by the LASSO regression were then entered into a multivariable logistic regression model using a forward stepwise selection procedure, with overall CRT as the dependent variable (1= CRT, 0= non-CRT), to identify independent predictors of CRT. Regression coefficients obtained from the final multivariable logistic regression model were used to construct the nomogram. Model discrimination was assessed using receiver operating characteristic (ROC) curves, and the area under the curve (AUC) was reported. Model calibration was assessed using the Hosmer-Lemeshow test and the bootstrap method (200 resamples), and the optimism-corrected AUC was calculated. The cumulative incidence of CRT according to levels of different biomarkers was plotted using the Kaplan-Meier method, with the catheterization date as time zero, and group comparisons were performed using the log-rank test. All tests were two-sided, and P<0.05 was considered statistically significant.
Results
Baseline characteristics
A total of 172 patients with hematologic malignancies were included in this study. Overall CRT was detected in 58 patients (CRT group), whereas 114 patients had no evidence of CRT (non-CRT group), corresponding to an overall incidence of 33.7%. The two groups showed no statistically significant differences in age (55.92±7.03 vs. 55.22±7.28 years, P=0.07), BMI (22.75±3.95 vs. 22.52±4.01 kg/m2, P=0.29), catheter length (42.51±3.06 vs. 42.21±3.13 cm, P=0.07), gender, tumor type, hypertension, diabetes, coronary heart disease, catheterization side, punctured vessel, puncture complications, catheter infection status, chemotherapy regimen, number of puncture attempts (≥2 times), abnormal tip position, or prophylactic anticoagulation use (all P>0.05). However, PLT was significantly higher in the CRT group [(84.97±15.97) ×109/L] compared with the non-CRT group [(80.04±14.02) ×109/L], with a statistically significant difference (P=0.04). Among the 58 patients with CRT, 15 (25.86%) presented with symptoms suggestive of CRT, whereas 43 (74.14%) were asymptomatic and were identified through scheduled ultrasound screening. The median time from PICC placement to CRT detection was 17.50 days (IQR, 12.50–28.00 days; Table 1). According to the primary thrombus location identified on the initial ultrasound examination, thrombi were most frequently located in the basilic vein (30/58, 51.72%), followed by the brachial vein (12/58, 20.69%), cephalic vein (6/58, 10.34%), axillary vein (6/58, 10.34%), and subclavian vein (4/58, 6.90%).
Table 1. Baseline characteristics and comparative analysis.
| Variables | Total (n=172) | Non-CRT group (n=114) | CRT group (n=58) | t/χ2 | P |
|---|---|---|---|---|---|
| Age (years) | 55.92±7.03 | 55.22±7.28 | 57.31±6.34 | −1.859 | 0.07 |
| BMI (kg/m2) | 22.75±3.95 | 22.52±4.01 | 23.20±3.83 | −1.067 | 0.29 |
| Catheter length (cm) | 42.51±3.06 | 42.21±3.13 | 43.10±2.87 | −1.802 | 0.07 |
| PLT (×109/L) | 81.70±14.84 | 80.04±14.02 | 84.97±15.97 | −2.075 | 0.04 |
| Gender | 0.039 | 0.84 | |||
| Female | 70 (40.70) | 47 (41.23) | 23 (39.66) | ||
| Male | 102 (59.30) | 67 (58.77) | 35 (60.34) | ||
| Tumor type | 0.767 | 0.68 | |||
| Leukemia | 73 (42.44) | 47 (41.23) | 26 (44.83) | ||
| Malignant lymphoma | 61 (35.47) | 43 (37.72) | 18 (31.03) | ||
| Multiple myeloma | 38 (22.09) | 24 (21.05) | 14 (24.14) | ||
| Hypertension | 0.253 | 0.62 | |||
| No | 117 (68.02) | 79 (69.30) | 38 (65.52) | ||
| Yes | 55 (31.98) | 35 (30.70) | 20 (34.48) | ||
| Diabetes mellitus | 0.059 | 0.81 | |||
| No | 144 (83.72) | 96 (84.21) | 48 (82.76) | ||
| Yes | 28 (16.28) | 18 (15.79) | 10 (17.24) | ||
| Coronary heart disease | 0.181 | 0.67 | |||
| No | 158 (91.86) | 104 (91.23) | 54 (93.10) | ||
| Yes | 14 (8.14) | 10 (8.77) | 4 (6.90) | ||
| Catheterization side | 0.420 | 0.52 | |||
| Right | 56 (32.56) | 39 (34.21) | 17 (29.31) | ||
| Left | 116 (67.44) | 75 (65.79) | 41 (70.69) | ||
| Punctured vessel | 0.996 | 0.80 | |||
| Brachial vein | 22 (12.79) | 15 (13.16) | 7 (12.07) | ||
| Basilic vein | 133 (77.33) | 86 (75.44) | 47 (81.03) | ||
| Median cubital vein | 8 (4.65) | 6 (5.26) | 2 (3.45) | ||
| Cephalic vein | 9 (5.23) | 7 (6.14) | 2 (3.45) | ||
| Puncture complications | 0.027 | 0.87 | |||
| No | 158 (91.86) | 105 (92.11) | 53 (91.38) | ||
| Yes | 14 (8.14) | 9 (7.89) | 5 (8.62) | ||
| Catheter infection status | 0.783 | 0.38 | |||
| No | 123 (71.51) | 84 (73.68) | 39 (67.24) | ||
| Yes | 49 (28.49) | 30 (26.32) | 19 (32.76) | ||
| Chemotherapy regimen | 1.389 | 0.50 | |||
| Standard chemotherapy | 114 (66.28) | 74 (64.91) | 40 (68.97) | ||
| Targeted therapy-containing | 41 (23.84) | 30 (26.32) | 11 (18.97) | ||
| Other regimens | 17 (9.88) | 10 (8.77) | 7 (12.07) | ||
| Number of puncture attempts | 2.343 | 0.13 | |||
| 1 attempt | 136 (79.07) | 94 (82.46) | 42 (72.41) | ||
| ≥2 attempts | 36 (20.93) | 20 (17.54) | 16 (27.59) | ||
| Abnormal tip position | 3.809 | 0.051 | |||
| No | 112 (65.12) | 80 (70.18) | 32 (55.17) | ||
| Yes | 60 (34.88) | 34 (29.82) | 26 (44.83) | ||
| Prophylactic anticoagulation | 1.657 | 0.20 | |||
| No | 89 (51.74) | 55 (48.25) | 34 (58.62) | ||
| Yes | 83 (48.26) | 59 (51.75) | 24 (41.38) | ||
| CRT symptoms | |||||
| No | – | – | 43 (74.14) | – | – |
| Yes | – | – | 15 (25.86) | – | – |
| Time to thrombosis (d), | – | – | 17.50 (12.50, 28.00) | – | – |
Data are presented as mean ± standard deviation, n (%) or median (IQR). BMI, body mass index; CRT, catheter-related thrombosis; IQR, interquartile range; PLT, platelet count.
Inflammatory indicators
WBC showed no significant difference between the two groups (7.82×109/L vs. 8.05×109/L, P=0.62). However, IL-1β (9.52 vs. 10.53 ng/L), CRP (11.32 vs. 17.60 mg/L), and IL-6 (45.26 vs. 55.25 ng/L) were significantly higher in the CRT group (all P<0.001; Table 2).
Table 2. Inflammatory indicators.
| Variables | Total (n=172) | Non-CRT group (n=114) | CRT group (n=58) | t | P |
|---|---|---|---|---|---|
| WBC (×109/L) | 7.90±2.88 | 7.82±2.94 | 8.05±2.76 | −0.504 | 0.62 |
| IL-1β (ng/L) | 9.86±2.15 | 9.52±2.43 | 10.53±1.22 | −3.623 | <0.001 |
| CRP (mg/L) | 13.44±5.35 | 11.32±3.33 | 17.60±6.11 | −7.294 | <0.001 |
| IL-6 (ng/L) | 48.63±10.13 | 45.26±8.21 | 55.25±10.34 | −6.899 | <0.001 |
Data are presented as mean ± standard deviation. CRP, C-reactive protein; CRT, catheter-related thrombosis; IL-1β, interleukin-1β; IL-6, interleukin-6; WBC, white blood cell count.
Coagulation indicators
FIB and D-D were significantly higher in the CRT group (FIB: 4.65 vs. 6.23 g/L, t=−6.378, P<0.001; D-D: 672.37 vs. 920.00 µg/L, t=−5.945, P<0.001). PT and TT were also significantly prolonged in the CRT group (PT: 11.67 vs. 12.73 s, P=0.003; TT: 16.95 vs. 18.22 s, P=0.02). No significant differences were observed in INR or APTT between the two groups (P>0.05; Table 3).
Table 3. Coagulation indicators.
| Variables | Total (n=172) | Non-CRT group (n=114) | CRT group (n=58) | t | P |
|---|---|---|---|---|---|
| FIB (g/L) | 5.18±1.70 | 4.65±1.48 | 6.23±1.62 | −6.378 | <0.001 |
| D-D (μg/L) | 755.87±262.74 | 672.37±210.21 | 920.00±279.53 | −5.945 | <0.001 |
| INR | 1.08±0.25 | 1.10±0.26 | 1.04±0.23 | 1.471 | 0.14 |
| APTT (s) | 30.42±5.46 | 29.91±5.33 | 31.42±5.61 | −1.733 | 0.09 |
| TT (s) | 17.37±3.27 | 16.95±3.28 | 18.22±3.12 | −2.440 | 0.02 |
| PT (s) | 12.03±2.25 | 11.67±2.11 | 12.73±2.36 | −2.980 | 0.003 |
Data are presented as mean ± standard deviation. APTT, activated partial thromboplastin time; CRT, catheter-related thrombosis; D-D, D-dimer; FIB, fibrinogen; INR, international normalized ratio; PT, prothrombin time; TT, thrombin time.
LASSO regression analysis
Variables with P<0.05 in the univariate analysis (36 variables in total) were entered into the LASSO regression model for variable selection to reduce multicollinearity and avoid model overfitting (Figure 1A). The LASSO algorithm shrank the coefficients of relatively unimportant variables toward zero while retaining the most informative predictors. Ten-fold cross-validation and the minimum criterion were used to determine the optimal value of the regularization parameter λ (ln(λ)) (Figure 1B). Variables retained by the LASSO regression were subsequently entered into the multivariable logistic regression model for identification of independent predictors of CRT. The LASSO procedure retained PLT, IL-1β, CRP, IL-6, FIB, D-D, PT, and TT as candidate predictors for subsequent multivariable analysis (Figure 1).
Figure 1.
LASSO regression analysis. LASSO, least absolute shrinkage and selection operator.
Multivariate logistic regression for independent risk factors for CRT
The variables retained by the LASSO regression model were subsequently entered into a multivariable logistic regression model using a forward stepwise selection procedure. Using the occurrence of CRT (assigned 1= CRT, 0= non-CRT) as the dependent variable, six variables were identified as independent predictors of CRT in patients with hematologic malignancies: PLT [P=0.01, odds ratio (OR) =0.936, 95% confidence interval (CI): 0.889−0.986], IL-1β (P=0.04, OR =1.375, 95% CI: 1.021−1.851), CRP (P<0.001, OR =1.367, 95% CI: 1.161−1.608), IL-6 (P<0.001, OR =1.134, 95% CI: 1.062−1.211), FIB (P=0.001, OR =1.870, 95% CI: 1.280−2.732), and D-D (P<0.001, OR =1.005, 95% CI: 1.002−1.007). PT and TT, although retained by the LASSO regression model, were not retained in the final multivariable logistic regression model (P=0.09 and P=0.10, respectively) and therefore were not identified as independent predictors of CRT (Table 4).
Table 4. Independent risk factors for CRT.
| Variables | β | S.E | Wald χ2 | P | OR (95% CI) |
|---|---|---|---|---|---|
| PLT | −0.066 | 0.026 | 6.310 | 0.01 | 0.936 (0.889–0.986) |
| IL-1β | 0.318 | 0.152 | 4.394 | 0.04 | 1.375 (1.021–1.851) |
| CRP | 0.312 | 0.083 | 14.156 | <0.001 | 1.367 (1.161–1.608) |
| IL-6 | 0.126 | 0.033 | 14.079 | <0.001 | 1.134 (1.062–1.211) |
| FIB | 0.626 | 0.193 | 10.464 | 0.001 | 1.870 (1.280–2.732) |
| D-D | 0.005 | 0.001 | 14.298 | <0.001 | 1.005 (1.002–1.007) |
CI, confidence interval; CRP, C-reactive protein; CRT, catheter-related thrombosis; D-D, D-dimer; FIB, fibrinogen; IL-1β, interleukin-1β; IL-6, interleukin-6; OR, odds ratio; PLT, platelet count; S.E, standard error.
Preliminary construction of a risk prediction model for CRT
Based on the regression coefficients obtained from the final multivariable logistic regression model (PLT, IL-1β, CRP, IL-6, FIB, and D-D), we constructed a nomogram model for individualized risk assessment (Figure 2A). By correlating the scores of each variable with the total score, the model allows intuitive quantitative prediction of the probability of the target event. The model’s discrimination was excellent, with an AUC of 0.947 (95% CI: 0.909−0.986), indicating high discriminative ability for the target event (Figure 2B). Bootstrap internal validation using 200 resamples yielded an optimism-corrected AUC of 0.934, indicating minimal optimism and good model stability. The calibration curve showed good agreement between predicted probabilities and observed probabilities, with the curve closely fitting the ideal calibration line. The Hosmer-Lemeshow test yielded χ2=8.3 (df =8, P=0.41), further indicating no significant systematic bias and acceptable calibration (Figure 2C). DCA showed that within a wide range of threshold probabilities, the net benefit of the model was consistently higher than the two extreme strategies of “intervening for all” or “intervening for none”, suggesting that the model has potential value for clinical decision-making (Figure 2D). To improve the reproducibility of the prediction model, the final multivariable logistic regression equation, including the intercept and regression coefficients for all predictors, is provided in Appendix 1. However, further external validation is needed to assess its practical utility.
Figure 2.
Construction of the risk prediction model for CRT. (A) Nomogram for predicting post-PICC thrombosis risk in patients with hematologic malignancies. (B) ROC curve for internal validation of the prediction model. (C) Calibration curve of the prediction model. (D) Decision curve of the prediction model. AUC, area under the curve; CI, confidence interval; CRP, C-reactive protein; CRT, catheter-related thrombosis; D-D, D-dimer; FIB, fibrinogen; HL, Hosmer-Lemeshow; IL-1β, interleukin-1β; IL-6, interleukin-6; PICC, peripherally inserted central catheter; PLT, platelet count; ROC, receiver operating characteristic.
Stratified analysis of symptomatic CRT
To assess whether symptomatic versus asymptomatic status influenced the analysis of factors associated with CRT, we performed a stratified analysis using CRT type (symptomatic vs. asymptomatic) as the dependent variable. Since only 15 patients had symptomatic CRT (8.7% of the total sample, 25.9% of the CRT group), the event count was insufficient to support multivariate regression (EPV <2). Therefore, we performed only an exploratory univariate analysis, using the asymptomatic CRT group (43 patients) as the reference. The results (Table 5) showed that inflammatory and coagulation markers (IL-1β, CRP, IL-6, FIB, D-D, PT, TT, etc.) were not significantly different between symptomatic and asymptomatic CRT (all P>0.05). Regarding tumor type, compared with leukemia, the OR for malignant lymphoma was 0.170 (95% CI: 0.032−0.899, P=0.04), and for multiple myeloma was 0.227 (95% CI: 0.042−1.228, P=0.09). No other clinical variables showed significant differences between the two groups (all P>0.05).
Table 5. Sensitivity analysis of symptomatic vs. asymptomatic CRT.
| Variables | β | S.E | Z | P | OR (95% CI) |
|---|---|---|---|---|---|
| Gender | |||||
| Female | 1.000 (reference) | ||||
| Male | 0.778 | 0.660 | 1.179 | 0.24 | 2.177 (0.597–7.933) |
| Tumor type | |||||
| Leukemia | 1.000 (reference) | ||||
| Malignant lymphoma | −1.769 | 0.849 | −2.085 | 0.04 | 0.170 (0.032–0.899) |
| Multiple myeloma | −1.482 | 0.861 | −1.721 | 0.09 | 0.227 (0.042–1.228) |
| Hypertension | |||||
| No | 1.000 (reference) | ||||
| Yes | −0.488 | 0.664 | −0.736 | 0.46 | 0.614 (0.167–2.253) |
| Diabetes mellitus | |||||
| No | 1.000 (reference) | ||||
| Yes | −0.396 | 0.855 | −0.463 | 0.64 | 0.673 (0.126–3.594) |
| Coronary heart disease | |||||
| No | 1.000 (reference) | ||||
| Yes | −0.049 | 1.196 | −0.041 | 0.97 | 0.952 (0.091–9.922) |
| Catheterization side | |||||
| Right | 1.000 (reference) | ||||
| Left | −0.662 | 0.632 | −1.047 | 0.30 | 0.516 (0.149–1.781) |
| Punctured vessel | |||||
| Brachial vein | 1.000 (reference) | ||||
| Basilic vein | 0.830 | 1.128 | 0.736 | 0.46 | 2.294 (0.251–20.941) |
| Median cubital vein | −14.774 | 1,696.735 | −0.009 | 0.99 | 0.000 (0.000–Inf) |
| Cephalic vein | 1.792 | 1.780 | 1.007 | 0.31 | 6.000 (0.183–196.280) |
| Puncture complications | |||||
| No | 1.000 (reference) | ||||
| Yes | 1.634 | 0.970 | 1.685 | 0.09 | 5.125 (0.766–34.311) |
| Catheter infection status | |||||
| No | 1.000 (reference) | ||||
| Yes | 0.431 | 0.623 | 0.692 | 0.49 | 1.538 (0.454–5.216) |
| Chemotherapy regimen | |||||
| Standard chemotherapy | 1.000 (reference) | ||||
| Targeted therapy-containing | −0.134 | 0.760 | −0.176 | 0.86 | 0.875 (0.197–3.880) |
| Other regimens | −16.719 | 1,495.296 | −0.011 | 0.99 | 0.000 (0.000–Inf) |
| Number of puncture attempts | |||||
| 1 attempt | 1.000 (reference) | ||||
| ≥2 attempts | −0.550 | 0.726 | −0.758 | 0.45 | 0.577 (0.139–2.393) |
| Abnormal tip position | |||||
| No | 1.000 (reference) | ||||
| Yes | 0.100 | 0.602 | 0.166 | 0.87 | 1.105 (0.340–3.595) |
| Prophylactic anticoagulation | |||||
| No | 1.000 (reference) | ||||
| Yes | −0.077 | 0.611 | −0.126 | 0.90 | 0.926 (0.280–3.067) |
| Age | −0.041 | 0.049 | −0.836 | 0.40 | 0.960 (0.872–1.057) |
| BMI | −0.013 | 0.079 | −0.164 | 0.87 | 0.987 (0.845–1.153) |
| Catheter length | −0.140 | 0.112 | −1.250 | 0.21 | 0.869 (0.698–1.083) |
| Time to thrombosis | −0.018 | 0.031 | −0.587 | 0.56 | 0.982 (0.924–1.044) |
| PLT | −0.025 | 0.020 | −1.254 | 0.21 | 0.975 (0.937–1.014) |
| WBC | 0.136 | 0.111 | 1.224 | 0.22 | 1.146 (0.921–1.425) |
| IL-1β | 0.398 | 0.264 | 1.510 | 0.13 | 1.489 (0.888–2.496) |
| CRP | −0.075 | 0.054 | −1.378 | 0.17 | 0.928 (0.834–1.032) |
| IL-6 | −0.022 | 0.030 | −0.748 | 0.45 | 0.978 (0.922–1.037) |
| FIB | 0.198 | 0.192 | 1.034 | 0.30 | 1.219 (0.837–1.775) |
| D-D | −0.001 | 0.001 | −1.222 | 0.22 | 0.999 (0.996–1.001) |
| INR | −0.705 | 1.329 | −0.531 | 0.60 | 0.494 (0.036–6.687) |
| APTT | 0.058 | 0.055 | 1.047 | 0.30 | 1.060 (0.951–1.181) |
| TT | 0.117 | 0.101 | 1.165 | 0.24 | 1.124 (0.923–1.370) |
| PT | −0.135 | 0.133 | −1.015 | 0.31 | 0.874 (0.673–1.134) |
APTT, activated partial thromboplastin time; BMI, body mass index; CI, confidence interval; CRP, C-reactive protein; CRT, catheter-related thrombosis; D-D, D-dimer; FIB, fibrinogen; IL-1β, interleukin-1β; IL-6, interleukin-6; Inf, infinity; INR, international normalized ratio; OR, odds ratio; PLT, platelet count; PT, prothrombin time; S.E, standard error; TT, thrombin time; WBC, white blood cell count.
Discussion
In this study based on follow-up data from 172 patients with hematologic malignancies after PICC placement, overall CRT, including both symptomatic and asymptomatic events detected by scheduled ultrasound screening, occurred in 58 patients (33.7%). Of these, 43 patients (74.1% of the CRT group) were asymptomatic and identified only through routine ultrasound surveillance, whereas 15 patients (25.9%) presented with clinical symptoms. This incidence is comparable to the 20−40% reported in previous studies and highlights the substantial thrombotic burden in this population, particularly the high proportion of clinically occult thrombosis that would likely have been missed without routine ultrasound screening. Our findings further demonstrated that PLT, D-D, FIB, IL-6, CRP, and IL-1β were independent predictors of CRT. These biomarkers collectively represent key components of the inflammation-coagulation axis, suggesting that CRT is not a random complication but rather the consequence of a complex interaction among malignancy-associated hypercoagulability, systemic inflammatory activation, therapeutic interventions, and catheter-related endothelial injury. Furthermore, the nomogram developed in this study demonstrated excellent discrimination (AUC =0.947) and good calibration, indicating that integrating inflammation and coagulation biomarkers may provide an effective approach for individualized prediction of CRT risk in patients with hematologic malignancies undergoing PICC placement. The present study adopted a sequential modeling strategy consisting of univariate analysis, LASSO regression, and multivariable logistic regression, which enabled preliminary variable screening, reduced multicollinearity and model overfitting, and improved the stability and interpretability of the final prediction model. Although the model demonstrated excellent internal performance, it should be regarded as an exploratory model until externally validated in independent multicenter cohorts.
From an inflammatory perspective, elevated IL-6, IL-1β, and CRP indicate significant inflammatory responses in patients. IL-6 is a central link between inflammation and the coagulation system; it stimulates the liver to synthesize FIB, increases blood viscosity, and induces tissue factor expression, thereby activating the extrinsic coagulation pathway (17,18). IL-1β, as a key pro-inflammatory cytokine in innate immunity, acts directly on vascular endothelium, causing endothelial injury and a prothrombotic tendency (19,20). Additionally, CRP, a classic acute-phase protein, is markedly elevated in inflammatory states. CRP is not only a sensitive indicator of inflammation but also participates in thrombosis by promoting monocyte expression of coagulation factors (21,22). In our study, levels of IL-6, IL-1β, and CRP were significantly higher in the CRT group than in the non-CRT group, and multivariate analysis confirmed them as independent risk factors, indicating that inflammatory storms have a decisive impact on thrombus formation in patients with hematologic malignancies.
Regarding the coagulation system, we found that FIB and D-D were significantly elevated in the CRT group and were both independent risk factors. FIB is a core substance in the coagulation cascade. Under the stimulation of IL-6, its marked increase substantially raises plasma viscosity and promotes red blood cell aggregation, resulting in a hypercoagulable state (23,24). Elevated D-D reflects an ongoing dynamic process of fibrin formation and lysis, representing a comprehensive manifestation of thrombus formation and fibrinolytic activation (25,26). In patients with hematologic malignancies, microparticles released by tumor cells can further promote coagulation activation, leading to persistently elevated D-D (27). Although the OR for a single unit increase in D-D was modest in our regression analysis, the absolute changes in D-D levels in patients are sensitive and clinically meaningful, making it a key indicator for early recognition of thrombotic activity.
Moreover, PLT, as an important cellular component linking inflammation and thrombosis, was also an independent risk factor in our study, possibly related to tumor-associated inflammation-induced megakaryocyte proliferation or platelet activation. In our multivariate regression, although PT and TT showed significant between-group differences in univariate analysis, they were not retained as independent predictors after LASSO screening, suggesting their independent contribution to thrombosis is limited, potentially due to collinearity within the inflammation-coagulation network. In summary, elevated FIB and D-D constitute a linked chain of coagulation activation-fibrinolytic response, which together with inflammatory factors promote thrombus formation.
Further exploring the mechanism of PICC placement in the inflammation-coagulation network, PICC itself acts as a local irritant. The placement process causes local mechanical injury, stretching and compressing epithelial cells, leading to exposure of tissue factor and activation of the extrinsic coagulation pathway (8). Additionally, the long-term presence of the catheter in the venous lumen can reduce blood flow velocity and shear stress, promoting vortex formation and thrombus cores around the catheter (7,28). In patients with hematologic malignancies, the tumor microenvironment already exhibits inflammatory activation; when combined with catheter irritation, the inflammation-coagulation system is easily further disrupted, leading to accelerated thrombus formation (29). From our results, although traditional risk factors such as age, gender, BMI, and puncture method did not emerge as significant factors, inflammatory and coagulation biomarkers showed strong associations, indicating that in patients with hematologic malignancies, biological markers better reflect true thrombotic risk than traditional clinical variables.
Our prediction model, based on six core risk factors (PLT, IL-1β, CRP, IL-6, FIB, D-D) and visualized as a nomogram, is highly operable. The ROC curve showed an AUC of 0.947, indicating excellent discriminative ability. The calibration curve showed good agreement between predicted and observed probabilities. DCA showed that within a wide range of threshold probabilities, the model provided clinical net benefit, suggesting that it can be applied in clinical nursing practice. After PICC placement, nursing staff can risk-stratify patients based on model scores. If a patient’s predicted score exceeds the optimal cutoff, enhanced monitoring measures should be initiated, including early re-assessment of blood indicators, bedside ultrasound screening, and frequent assessment of limb status. Furthermore, high-risk patients should minimize venous injury on the affected side, promote limb activity, avoid dehydration, and, if necessary, consult physicians for early intervention to achieve early prevention and management of thrombosis.
Moreover, from a nursing perspective, this study proposes intervention strategies targeting the inflammation-coagulation network. In patients with hematologic malignancies, inflammation remains at persistently high levels, and thrombotic risk fluctuates dynamically. Therefore, a single assessment is insufficient for nursing needs. Based on high-risk model results, nursing staff should recheck inflammatory and coagulation indicators within 24−48 hours after catheterization to identify rapid rises. Our results suggest that if PLT, IL-6, CRP, IL-1β, FIB, and D-D show a continuous upward trend, immediate ultrasound screening should be performed to detect early thrombus formation. In long-term management, the nursing team should strengthen health education to encourage patients to proactively report early warning signs such as limb numbness, pain, varicose veins, and swelling, so that thrombosis can be identified before clinical symptoms develop (30). It is particularly noteworthy that asymptomatic CRT accounted for 74.1% of the CRT group in this study, and stratified analysis showed no statistically significant differences in inflammation-coagulation indicators (IL-1β, CRP, IL-6, FIB, D-D, PT, TT, etc.) between symptomatic and asymptomatic CRT, suggesting that the risk factor profiles of the two types of CRT may be similar. However, due to the small number of symptomatic CRT events (only 15), the statistical power of this stratified analysis was insufficient, and the conclusion needs verification in larger studies. Therefore, our model is primarily suitable for predicting overall CRT (asymptomatic + symptomatic) under ultrasound screening; its predictive ability for symptomatic CRT alone requires further evaluation.
Strengths of this study include: (I) a relatively adequate sample size with strict inclusion/exclusion criteria, enhancing the reliability of the prediction model; (II) using the inflammation-coagulation network as the research entry point, which is more mechanistic than previous models based solely on clinical factors; (III) multi-angle validation of the model using nomograms, ROC curves, DCA, and bootstrap calibration, significantly enhancing the credibility of the results; (IV) the findings can be directly used to construct clinical nursing pathways, providing practical value.
However, this study has several limitations. First, it is a single-center retrospective study lacking external validation, and selection bias cannot be excluded. Therefore, the proposed nomogram should be considered an exploratory prediction model, and external validation in independent multicenter cohorts is required before routine clinical application. Second, only pre-insertion inflammation and coagulation biomarkers were evaluated, and dynamic changes during catheterization were not assessed, which may have underestimated their predictive value. Third, additional coagulation-related biomarkers, such as genetic factors, circulating microparticles, and other emerging molecular markers, were not included. Future studies integrating multi-omics data may further improve model performance. Fourth, the primary outcome of this study was overall CRT, including both symptomatic and asymptomatic events detected by scheduled ultrasound screening. Because asymptomatic thrombosis accounted for the majority of events, the model's ability to predict clinically significant symptomatic CRT requiring therapeutic intervention remains uncertain. Moreover, the limited number of symptomatic events precluded robust subgroup analyses. Despite these limitations, this study provides an initial framework for individualized CRT risk assessment based on inflammation and coagulation biomarkers and may inform the design of future prospective multicenter validation studies and targeted thrombosis prevention strategies.
Conclusions
In summary, this study developed a CRT prediction model based on inflammation and coagulation biomarkers in patients with hematologic malignancies undergoing PICC placement. The results demonstrated that PLT, IL-6, IL-1β, CRP, FIB, and D-D were independent predictors of CRT, highlighting the important role of the inflammation-coagulation axis in thrombosis development. The proposed nomogram demonstrated excellent discrimination and good calibration, providing a practical and user-friendly tool for early identification of patients at high risk of CRT and supporting individualized clinical management and nursing interventions. Nevertheless, because this was a single-center retrospective study with internal validation only, external validation in larger multicenter cohorts is warranted to further evaluate the generalizability and clinical applicability of the model. Such efforts may contribute to more precise risk stratification and optimized thrombosis prevention strategies for patients with hematologic malignancies undergoing PICC placement.
Supplementary
The article’s supplementary files as
Acknowledgments
None.
Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Human Ethics Committee of Guangxi Medical University Cancer Hospital (Ethics approval No. KY2023059). Given the retrospective nature of the study and the use of de-identified clinical data, the requirement for informed consent was waived by the Ethics Committee.
Footnotes
Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1203/rc
Funding: This research was supported by Guangxi Zhuang Autonomous Region Health Commission self-funded scientific research project (No. Z-A20230758) and Zhongguancun Precision Medicine Foundation Medical & Health Public Welfare Initiative-Hospital Management Special Research Project (No. ZGC-YXKY-54).
Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1203/coif). The authors have no conflicts of interest to declare.
Data Sharing Statement
Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1203/dss
References
- 1.Wang TF, Leader A, Sanfilippo KM. Thrombosis and bleeding in hematological malignancy. Best Pract Res Clin Haematol 2022;35:101353. 10.1016/j.beha.2022.101353 [DOI] [PubMed] [Google Scholar]
- 2.Kekre N, Connors JM. Venous thromboembolism incidence in hematologic malignancies. Blood Rev 2019;33:24-32. 10.1016/j.blre.2018.06.002 [DOI] [PubMed] [Google Scholar]
- 3.Dos Santos BN, Beato BVG, Ferreira EB, et al. Prevalence of PICC-related thrombosis in patients with hematological malignancies: a systematic review. Support Care Cancer 2024;32:462. 10.1007/s00520-024-08636-7 [DOI] [PubMed] [Google Scholar]
- 4.Puri A, Dai H, Giri M, et al. The incidence and risk of venous thromboembolism associated with peripherally inserted central venous catheters in hospitalized patients: A systematic review and meta-analysis. Front Cardiovasc Med 2022;9:917572. 10.3389/fcvm.2022.917572 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Elias A, Debourdeau P, Espitia O, et al. Central venous catheter associated upper extremity deep vein thrombosis in cancer patients: Diagnosis and therapeutic management. Arch Cardiovasc Dis 2024;117:72-83. 10.1016/j.acvd.2023.11.011 [DOI] [PubMed] [Google Scholar]
- 6.Cavallaro C, Santini P, Leoni L, et al. Long term anticoagulation for Catheter-Related deep vein thrombosis of the upper extremities in women with cancer: retrospective analysis of effectiveness and safety outcomes. J Thromb Thrombolysis 2025;58:1127-34. 10.1007/s11239-025-03182-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Wang P, He L, Yuan Q, et al. Risk factors for peripherally inserted central catheter-related venous thrombosis in adult patients with cancer. Thromb J 2024;22:6. 10.1186/s12959-023-00574-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.López-Rubio M, Lago-Rodríguez MO, Ordieres-Ortega L, et al. A Comprehensive Review of Catheter-Related Thrombosis. J Clin Med 2024;13:7818. 10.3390/jcm13247818 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Yin Q, Zheng X, Han L, et al. Outcome of antithrombotic therapy in cancer patients with catheter-related thrombosis: a systematic review. Front Cardiovasc Med 2023;10:1290822. 10.3389/fcvm.2023.1290822 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Vladić N, Englisch C, Berger JM, et al. Validation of risk assessment models for venous thromboembolism in patients with cancer receiving systemic therapies. Blood Adv 2025;9:3340-9. 10.1182/bloodadvances.2025016044 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Liu Q, Chen L, Wang Z, et al. The role of D-dimer and fibrinogen testing in catheter-directed thrombolysis with urokinase for deep venous thrombosis. Phlebology 2023;38:389-97. 10.1177/02683555231176911 [DOI] [PubMed] [Google Scholar]
- 12.Chopra V, Anand S, Hickner A, et al. Risk of venous thromboembolism associated with peripherally inserted central catheters: a systematic review and meta-analysis. Lancet 2013;382:311-25. 10.1016/S0140-6736(13)60592-9 [DOI] [PubMed] [Google Scholar]
- 13.Verso M, Agnelli G, Kamphuisen PW, et al. Risk factors for upper limb deep vein thrombosis associated with the use of central vein catheter in cancer patients. Intern Emerg Med 2008;3:117-22. 10.1007/s11739-008-0125-3 [DOI] [PubMed] [Google Scholar]
- 14.Evans RS, Sharp JH, Linford LH, et al. Risk of symptomatic DVT associated with peripherally inserted central catheters. Chest 2010;138:803-10. 10.1378/chest.10-0154 [DOI] [PubMed] [Google Scholar]
- 15.Fallouh N, McGuirk HM, Flanders SA, et al. Peripherally Inserted Central Catheter-associated Deep Vein Thrombosis: A Narrative Review. Am J Med 2015;128:722-38. 10.1016/j.amjmed.2015.01.027 [DOI] [PubMed] [Google Scholar]
- 16.Linnemann B, Beyer-Westendorf J, Espinola-Klein C, et al. Management of Deep Vein Thrombosis: An Update Based on the Revised AWMF S2k Guideline. Hamostaseologie 2024;44:97-110. 10.1055/a-2178-6574 [DOI] [PubMed] [Google Scholar]
- 17.Dobson DA, Holle LA, Lin FC, et al. Novel genetic regulators of fibrinogen synthesis identified by an in vitro experimental platform. J Thromb Haemost 2023;21:522-33. 10.1016/j.jtha.2022.10.027 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Liu D, Xu C, Gong Z, et al. GRSF1 antagonizes age-associated hypercoagulability via modulation of fibrinogen mRNA stability. Cell Death Dis 2023;14:717. 10.1038/s41419-023-06242-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Immanuel J, Yun S. Vascular Inflammatory Diseases and Endothelial Phenotypes. Cells 2023;12:1640. 10.3390/cells12121640 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Chi J, Wang Q, Wang Z, et al. S100 calcium-binding protein A8 exacerbates deep vein thrombosis in vascular endothelial cells. Sci Rep 2025;15:831. 10.1038/s41598-025-85322-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Dix C, Zeller J, Stevens H, et al. C-reactive protein, immunothrombosis and venous thromboembolism. Front Immunol 2022;13:1002652. 10.3389/fimmu.2022.1002652 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Han Z, Liu Q, Li H, et al. The role of monocytes in thrombotic diseases: a review. Front Cardiovasc Med 2023;10:1113827. 10.3389/fcvm.2023.1113827 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhang Y, Li Z, Zhang J, et al. Fibrinogen: A new player and target on the formation of pre-metastatic niche in tumor metastasis. Crit Rev Oncol Hematol 2025;207:104625. 10.1016/j.critrevonc.2025.104625 [DOI] [PubMed] [Google Scholar]
- 24.Ma C, Yang B, Mao Q. Prognostic value of fibrinogen change value in adrenocortical carcinoma patients. Discov Oncol 2024;15:320. 10.1007/s12672-024-01197-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Tayal D, Jain P, Goswami B. D-dimer - a multifaceted molecule. Horm Mol Biol Clin Investig 2024;45:75-84. 10.1515/hmbci-2022-0093 [DOI] [PubMed] [Google Scholar]
- 26.Chang Y, Lu J, Chen C. Biomarkers in Venous Thrombosis: Diagnostic Potential and Limitations. Biology (Basel) 2025;14:800. 10.3390/biology14070800 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Kim AS, Khorana AA, McCrae KR. Mechanisms and biomarkers of cancer-associated thrombosis. Transl Res 2020;225:33-53. 10.1016/j.trsl.2020.06.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Hu Z, He R, Zhao Y, et al. Risk assessment models for PICC-related venous thrombosis in adult patients with cancer: A network meta-analysis. Thromb Res 2024;239:109030. 10.1016/j.thromres.2024.05.003 [DOI] [PubMed] [Google Scholar]
- 29.Gasparyan AY, Ayvazyan L, Mukanova U, et al. The Platelet-to-Lymphocyte Ratio as an Inflammatory Marker in Rheumatic Diseases. Ann Lab Med 2019;39:345-57. 10.3343/alm.2019.39.4.345 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Sun CJ, Ma L, Wang SY, et al. A Quasi-Experimental Study on the Preventive Effect of Risk-Stratified Nursing Interventions for PICC-Related Thrombosis in Cancer Patients. Risk Manag Healthc Policy 2025;18:3209-20. 10.2147/RMHP.S521571 [DOI] [PMC free article] [PubMed] [Google Scholar]