Development and validation of a nomogram for predicting PICC catheter-related bloodstream infection among patients with hematologic malignancies.

Introduction

Patients with hematologic malignancies, encompassing conditions such as leukemia, lymphoma, and multiple myeloma, face complex and often protracted treatment regimens involving intensive chemotherapy, stem cell transplantation, and prolonged supportive care [ 1 , 2 ]. A reliable and durable venous access option is indispensable for the successful delivery of these therapies, including chemotherapy, blood products, antibiotics, and nutritional support [ 3 ]. Peripherally inserted central catheters (PICCs) have emerged as a cornerstone for long-term venous access in this population. PICCs offer significant advantages over conventional central venous catheters, including reduced insertion-related risks (such as pneumothorax and hemorrhage), extended indwelling duration, cost-effectiveness, and the safe administration of vesicant chemotherapeutics, hyperosmolar solutions, and other cytotoxic agents, thereby minimizing the risk of extravasation-related vascular injury and tissue necrosis [ 1 , 4 – 7 ]. However, the benefits of PICCs use are counterbalanced by the significant risk of complications, among which catheter-related bloodstream infection (CRBSI) stands out as one of the most frequent and devastating [ 8 ].

CRBSI is microbiologically defined as concordant bacteremia or fungemia from peripheral and catheter-tip cultures, occurring during catheterization or ≤ 48 h post-removal, with no other identifiable infection source [ 9 – 11 ]. The burden of CRBSI is particularly severe in patients with hematologic malignancies, reaching an incidence of 20.3–22.0 episodes/1000 neutropenic days [ 12 – 14 ], significantly higher than rates observed in solid tumors. CRBSI poses a grave threat in this population, frequently necessitating unplanned catheter removal, disrupting critical anti-neoplastic treatment schedules, precipitating sepsis and septic shock, prolonging hospital stays, substantially escalating healthcare costs, and significantly contributing to morbidity and mortality [ 15 – 20 ]. This unique susceptibility stems fundamentally from the underlying pathophysiology of hematologic malignancies and their treatments, characterized by severe and sustained immunosuppression and neutropenia, frequent breaches in mucosal barriers, and the necessity for repeated transfusions and chemotherapy infusions-collectively creating an environment highly conducive to bloodstream infections [ 17 – 20 ]. This substantial risk underscores the imperative for rigorous infection control protocols to optimize the safe utility of PICCs in this exceptionally vulnerable patient group.

Although several risk prediction models for CRBSI have been developed, the majority focus on general inpatients, critically ill, hemodialysis, and solid tumors populations [ 21 – 24 ]. However, these models are poorly suited for patients with hematologic malignancies due to several critical limitations. For instance, Erica Herc et al. [ 25 ] constructed the MPC model to identify risk factors for PICC- CLABSI(Central-line-associated bloodstream infection); however, this model showed only moderate discriminatory power (AUC: 0.67–0.77) and ignored demographic statistical characteristics (e.g., Age) and other idiosyncratic factors that may significantly influence clinical outcomes. Similarly, the nomograms construct of Wang et al. [ 26 ] and Tang et al. [ 27 ], although incorporating certain biomarkers (e.g., albumin < 30 g/L, platelets<30 × 10⁹/L, total bilirubin>34.2 µmol/L), did not incorporate crucial catheter-specific parameters (e.g., Time of PICC placement, Catheterized vein). Critically, existing models neglect the unique pathophysiology of hematologic malignancies, where synergistic effects of cytotoxic therapy, hematopoietic stem cell transplantation, and prolonged cytopenias create a distinct, high-risk milieu for CRBSI [ 28 ]. Consequently, there is a compelling need for a validated, clinically practical tool designed specifically to stratify the individual risk of PICC-CRBSI in patients with hematologic malignancies.

To address this gap, this study aimed to develop and validate a comprehensive nomogram for predicting the risk of PICC-CRBSI specifically among patients with hematologic malignancies.

Materials and methods

Participants

We retrospectively included patients with hematologic malignancies undergoing PICC catheterization at a tertiary medical center in Hangzhou, China, from January 2021 to December 2024. Data related to patient demographics, diseases characteristics and treatments were extracted from electronic health records (Hospital Information System, HIS). The inclusion criteria of the study were: (1) Age ≥ 18 years; (2) Hospitalized patients diagnosed with hematological malignancies and treated with chemotherapy; (3) Complete electronic medical record including demographic profile, treatment regimens, and laboratory parameters; (4) PICC placement time was ≥ 48 h. The exclusion criteria were: (1) PICC-CRBSI secondary to other sites, such as pulmonary infection, abdominal infection, etc.; (2) Hematopoietic stem cell donors.

In this study, patients with central venous catheters (CVCs) were not included. In clinical practice, PICCs are generally preferred in patients with hematologic malignancies due to their lower platelet count requirements, longer dwell times, and greater feasibility for chemotherapy administration. By contrast, CVCs often require a platelet threshold >50 × 10⁹/L, are associated with higher bleeding and thrombosis risks, and are less suitable for long-term use [ 29 – 31 ]. Furthermore, cervical lymphadenopathy in lymphoma patients frequently complicates CVC insertion. Based on these considerations, our study focused specifically on PICCs rather than CVCs.

PICC insertion and maintenance

The PICCs used in this study were manufactured by Bard (Covington, GA) and were primarily a model 4-French single-lumen catheter (PowerPICC ® Solo² Catheter; No.7617405) and a model 5-French dual-lumen catheter (PowerPICC ® Dual-Lumen Catheter; No.3275118). All PICC placement and maintenance was performed in strict accordance with evidence-based best practice guidelines [ 32 ].

Study endpoints and definitions

The data collection period began 48 h post-PICC placement. The primary endpoint was the occurrence of PICC-CRBSI, defined as the number of days from PICC placement to catheter removal or study cutoff in uninfected individuals. Censored events were defined as PICC removal for non-CRBSI reasons, loss to follow-up, or failure to complete the data collection period.

Vascular catheter-related infections included localized vascular catheter infections and CRBSI, with the former presenting only inflammatory manifestations such as redness, swelling, heat, pain, and exudation, and CRBSI presenting systemic manifestations of infection in addition to local manifestations. To diagnose ‘PICC-CRBSI,’ the same microorganism must be isolated from both the peripheral blood culture and the catheter tip, or there must be a differential time to positivity [ 33 ].The time to positivity was defined by a central line peripheral blood culture turning positive at least 2 h before the peripheral blood culture drawn at the same time [ 34 , 35 ].

Data collection

All data were extracted from the HIS. A comprehensive literature review and group meetings with physicians and nurses were conducted before we decided on potential risk variables [ 28 , 33 , 36 ]. Three independent researchers (GQ, ZXL, and ZX) extracted, reviewed, and validated data from electronic medical records. The following data were collected (1) general patient information (Gender, Age, Body mass index [BMI], Type of disease, Whether hematopoietic stem cell transplantation[HSCT], History of diabetes, History of hypertension, and Carrying Hepatitis B Virus); (2) information on PICC placement (History of PICC placement, Catheter diameter, PICC insertion attempts, Catheterized Vein, and Time of PICC placement); and (3) laboratory examination (48 hours prior to placement of White Blood Cells [WBC], Platelets [PLT], Absolute neutrophil count [ANC], Absolute lymphocyte count [ALC] and D-dimer).

Ethical considerations

Ethical approval for this study was obtained from the Research Ethics Committee of the First Affiliated Hospital of Zhejiang Chinese Medical University (Zhejiang Provincial Hospital of Chinese Medicine) (ref.2025-KLS-109-01). Informed consent was waived due to the retrospective and observational nature of the study. The study was conducted according to the standards set by the Declaration of Helsinki, and patient information was anonymized and de-identified before analysis.

Data analysis

All statistical analyses were performed using SPSS (version 26.0) and R (version 4.5.1) software. Categorical variables were expressed as n (%) and analyzed using the chi-square test when the cell expected frequency was ≥ 5 and the Fisher exact test if the cell expected frequency was<5. Normality of continuous variables was assessed using the Shapiro-Wilk test. Variables that followed a normal distribution were reported as mean ± standard deviation (SD) and compared using the independent samples t-test; continuous variables that did not follow a normal distribution were reported as median [interquartile range (IQR)] and compared using the Mann-Whitney U test.

To address class imbalance (PICC-CRBSI prevalence: 6.02%), we applied the Synthetic Minority Over-sampling Technique for Nominal and Continuous features (SMOTE-NC) exclusively to the training cohort ( n = 534), while preserving the natural outcome distribution in the validation set ( n = 230). This strict segregation prevents data leakage and aligns with TRIPOD guidelines [ 37 ]. SMOTE-NC—an advanced resampling method—generates synthetic minority-class samples by interpolating between nearest neighbors using modified Euclidean distance metrics, effectively handling datasets with mixed categorical and continuous variables [ 38 – 41 ]. We selected this approach over cost-sensitive learning or ensemble methods due to its validated efficacy in medical prediction studies [ 40 , 41 ]and suitability for our feature typology.

LASSO regression was selected for identifying potential characteristics of PICC-CRBSI occurrence in hematological oncology patients and the optimal λ value was determined by 10-fold cross-validation. Subsequently, the potential predictors identified through lasso regression techniques were used to build a predictive model using multivariate stepwise logistic regression, which eliminated highly correlated factors. We further used the statistically significant predictors obtained from stepwise logistic regression to construct a nomogram, which was created using the R package ‘regplot’, which can directly generate logistic regression-based results. The discrimination, calibration, and clinical utility of the model were assessed using subject work characteristics (ROC) curves, calibration curves, and decision curve analysis (DCA).

Results

Patient characteristics

We initially identified 822 patients with hematologic malignancies. Of these, 30 were<18 years of age; 18 were hematopoietic stem cell transplant donors; 6 had a PICC implantation time of less than 48 h; and 4 patients had PICC-CRBSI secondary to another site. As a result, a total of 764 patients with hematologic neoplasms met the inclusion criteria and were included in the analysis (Fig. 1 ). Baseline demographic and clinical characteristics are summarized in Table 1 . Fig. 1 Study flow diagram

Table 1 Demographic information and feature distribution for the study sample Variables Total n = 764 No PICC-CRBSI n = 718 PICC-CRBSI n = 46 Gender, n (%) Female 529 (69.24) 502 (69.92) 27 (58.70) Male 235 (30.76) 216 (30.08) 19 (41.30) Age, M (IQR) 60.00 (44.00, 70.00) 60.00 (44.25, 70.00) 58.00 (44.25, 71.75) BMI, M (IQR) 22.26 (19.63, 24.61) 22.39 (19.78, 24.74) 20.17 (16.88, 23.37) Type of disease, n (%) Leukemia 347 (45.42) 327 (45.54) 20 (43.48) Lymphoma 260 (34.03) 244 (33.98) 16 (34.78) Multiple myeloma 157 (20.55) 147 (20.47) 10 (21.74) HSCT, n (%) No 406 (53.14) 387 (53.90) 19 (41.30) Yes 358 (46.86) 331 (46.10) 27 (58.70) History of diabetes, n (%) No 675 (88.35) 647 (90.11) 28 (60.87) Yes 89 (11.65) 71 (9.89) 18 (39.13) History of hypertension, n (%) No 633 (82.85) 602 (83.84) 31 (67.39) Yes 131 (17.15) 116 (16.16) 15 (32.61) Carrying hepatitis B virus, n (%) No 696 (91.10) 660 (91.92) 36 (78.26) Yes 68 (8.90) 58 (8.08) 10 (21.74) History of PICC placement, n (%) No 497 (65.05) 469 (65.32) 28 (60.87) Yes 267 (34.95) 249 (34.68) 18 (39.13) Catheter Diameter, n (%) Single-lumen 320 (41.88) 304 (42.34) 16 (34.78) Dual-lumen 444 (58.12) 414 (57.66) 30 (65.22) PICC insertion attempts, n (%) 1 attempt 678 (88.74) 646 (89.97) 32 (69.57) ≥ 2 attempts 86 (11.26) 72 (10.03) 14 (30.43) Catheterized Vein, n (%) basilic vein 487 (63.74) 467 (65.04) 20 (43.48) brachial vein 258 (33.77) 236 (32.87) 22 (47.83) cephalic vein 19 (2.49) 15 (2.09) 4 (8.70) Time of PICC placement, n (%) <60 days 213 (27.88) 206 (28.69) 7 (15.22) ≥ 60 days 551 (72.12) 512 (71.31) 39 (84.78) WBC, M (IQR) 4.50 (2.40, 7.10) 4.40 (2.40, 7.10) 3.95 (1.95, 9.48) PLT, M (IQR) 32.00 (17.00, 79.00) 31.00 (16.25, 75.00) 66.50 (23.75, 126.25) ANC, M (IQR) 2.09 (0.40, 4.90) 3.00 (0.40, 5.00) 0.55 (0.12, 3.32) ALC, M (IQR) 0.80 (0.40, 1.90) 0.80 (0.40, 1.90) 0.65 (0.23, 1.00) D-dimer, M (IQR) 0.73 (0.31, 2.19) 0.73 (0.31, 2.26) 0.74 (0.34, 1.41) The quantitative data were skewed and expressed as medians (25th-75th percentile); Qualitative data were expressed in n% IQR Interquartile range, BMI Body mass index, HSCT Hematopoietic Stem Cell Transplantation, PICC Peripherally Inserted Central Catheter, WBC White Blood Cell, PLT Platelet, ANC Absolute Neutrophil Count, ALC Absolute Lymphocyte Count

Incidence of PICC-CRBSI among hematologic malignancies

Among 46 patients diagnosed with PICC-CRBSI, a total of 98 microbial isolates were identified (Table 2 ). Gram-negative bacteria predominated (57.14%, 56/98), with Pseudomonas aeruginosa (12.24%) and Stenotrophomonas maltophilia (12.24%) being the most frequently isolated species. Gram-positive bacteria accounted for 24.49% (24/98) of the isolates, primarily represented by Staphylococcus hominis (4.08%) and Enterococcus faecium (4.08%). In contrast, fungal infections (6.12%,6/98) and unclassified pathogens (12.24%,12/98) were less prevalent (Fig. 2 A). Monomicrobial infections involving a single pathogen were observed in 19 patients. In contrast, polymicrobial infections (i.e., co-infections with two or more pathogens) were detected in 27 cases, including one case with eight distinct bacterial species (Fig. 2 B).

Table 2 Distribution of pathogenic bacteria classification bacteria Number of strains Percentage(%) Gram-negative bacteria 56 57.14 Pseudomonas aeruginosa 12 12.24 Stenotrophomonas maltophilia 12 12.24 Klebsiella pneumoniae 10 10.20 Enterobacter cloacae 5 5.10 Escherichia coli 4 4.08 Klebsiella oxytoca 3 3.06 Klebsiella variicola 3 3.06 Acinetobacter baumannii 3 3.06 Burkholderia multivorans 1 1.02 Serratia marcescens 1 1.02 Morganella morganii 1 1.02 Ralstonia pickettii 1 1.02 Gram-positive bacteria 24 24.49 Staphylococcus aureus 3 3.06 Staphylococcus haemolyticus 1 1.02 Staphylococcus capitis 2 2.04 Staphylococcus hominis 4 4.08 Staphylococcus epidermidis 3 3.06 Enterococcus faecalis 2 2.04 Enterococcus faecium 4 4.08 Enterococcus avium 1 1.02 Micrococcus luteus 1 1.02 Streptococcus oralis 3 3.06 Fungi 6 6.12 Candida albicans 2 2.04 Candida tropicalis 1 1.02 Fungal spores and hyphae 3 3.06 Others/Unclassified pathogens 12 12.24 Clostridioides difficile (toxin-producing) 1 1.02 Gram-negative cocci 2 2.04 Gram-negative bacilli 3 3.06 Gram-positive cocci 2 2.04 Capnocytophaga sputigena 1 1.02 Enterobacter xiangfangensis 1 1.02 Elizabethkingia miricola 1 1.02 Corynebacterium pseudodiphtheriticum 1 1.02 Total 98 100

Fig. 2 A :Distribution of pathogens in PICC-CRBSI patients by classification（n=98)； B : Number of bacterial strains per PICC-CRBSI patient (n=46)

Results of LASSO regression and multivariate analyses

A cross-validation error plot was generated for the LASSO regression (Fig. 3 A). As shown by the vertical dashed line in Fig. 3 B, nine candidate variables were retained when the cross-validation error reached one standard error; these candidates were entered into a multivariate logistic regression model. The final multivariate analysis identified nine independent predictors of PICC-CRBSI (Table 3 ): History of diabetes (OR = 8.204, 95% CI: 5.115–13.520, P < 0.001), Age (per year increase: OR = 0.977, 95% CI: 0.965–0.988, P < 0.001), Time of PICC placement > 60 days (OR = 6.213, 95% CI: 3.668–10.933, P < 0.001), ≥ 2 PICC insertion attempts (OR = 2.917, 95% CI: 1.803–4.778, P < 0.001), Dual-lumen catheter (OR = 2.833, 95% CI: 1.901–4.270, P < 0.001), Catheterized vein — Cephalic vein (vs. Basilic vein) (OR = 7.288, 95% CI: 3.168–18.533, P < 0.001; Brachial vein vs. Basilic vein was not significant: OR = 0.874, 95% CI: 0.601–1.266, P = 0.478), ANC < 1.5 × 10⁹/L (OR = 1.864, 95% CI: 1.294–2.701, P < 0.001), ALC < 1.0 × 10⁹/L (OR = 5.476, 95% CI: 3.619–8.458, P < 0.001), and D-dimer ≥ 0.5 mg/L (OR = 7.029, 95% CI: 4.515–11.203, P < 0.001). These results indicate that both patient-level factors (e.g., diabetes, immune cell counts, coagulation markers) and catheter-related factors (e.g., dwell time, insertion attempts, lumen number, insertion site) independently contribute to PICC-CRBSI risk in patients with hematologic malignancies (Table 3 ).

Fig. 3 LASSO regression analysis result chart. A : LASSO regression cross validation plot; B : LASSO regression coefficient path diagram. LASSO, least absolute shrinkage and selection operator

Table 3 Multivariate analysis of risk factors for PICC-CRBSI among hematologic malignancies Variable β SE Z value OR (95% CI) P value Intercept −4.747 0.521 −9.109 – < 0.001 *** History of diabetes (Yes) 2.105 0.247 8.505 8.204(5.115, 13.520) < 0.001 *** Age (Years) −0.024 0.006 −4.017 0.977 (0.965–0.988) < 0.001 *** Time of PICC placement ( ≥ 60 days) 1.827 0.278 6.579 6.213 (3.668–10.933) < 0.001 *** PICC insertion attempts (≥ 2 attempt) 1.070 0.248 4.314 2.917 (1.803–4.778) < 0.001 *** Catheter diameter (Dual-lumen) 1.041 0.206 5.052 2.833 (1.901–4.270) < 0.001 *** Catheterized vein (Basilic vein) Brachial vein −0.135 0.190 −0.710 0.874 (0.601–1.266) 0.478 Cephalic vein 1.986 0.448 4.436 7.288 (3.168–18.533) < 0.001 *** ANC (<1.5 × 10⁹/L) 0.623 0.187 3.323 1.864 (1.294–2.701) < 0.001 *** ALC (<1.0 × 10⁹/L) 1.700 0.216 7.867 5.476 (3.619–8.458) < 0.001 *** D-dimer (≥ 0.5 mg/L) 1.950 0.231 8.426 7.029 (4.515–11.203) < 0.001 *** Abbreviations : SE S standard error, OR Odds ratio, CI Confidence interval, ANC Absolute Neutrophil Count, ALC Absolute Lymphocyte Count *** P <0.001

Development of the nomogram

Based on the nine independent predictors identified in the multivariate logistic regression analysis, a nomogram was constructed to estimate the individualized risk of PICC-CRBSI in patients with hematologic malignancies. The predictors incorporated into the model were: History of diabetes (yes/no), Age (Continuous), Time of PICC placement (< 60 days vs. ≥60 days), PICC insertion attempts (1attempt vs. ≥2 attempts), Catheter diameter (single-lumen vs. dual-lumen), Catheterized vein (Basilic vein, Brachial vein, or Cephalic vein), ANC (< 1.5 × 10⁹/L vs. ≥1.5 × 10⁹/L), ALC (< 1.0 × 10⁹/L vs. ≥1.0 × 10⁹/L), and D-dimer (< 0.5 mg/L vs. ≥0.5 mg/L).As shown in Fig. 4 A, each variable was assigned a corresponding point value according to its regression coefficient, and the sum of these points generated a total risk score, which was then mapped to the predicted probability of developing PICC-CRBSI. To facilitate clinical application, the assigned point values for each predictor are detailed in Table 4 . The higher the total score, the greater the likelihood of PICC-CRBSI. An illustrative example of nomogram application is provided in Fig. 4 B, demonstrating how individual patient profiles can be translated into risk probabilities for clinical decision-making.

Fig. 4 A Nomogram for predicting the probability of PICC- CRBSI in patients with hematologic malignancies. B Example application: A 44-year-old patient with a hematologic malignancy, a history of diabetes, time of PICC placement ≥60 days, PICC insertion 1 attempt, catheter placement in the basilic vein, ANC <1.5 × 10⁹/L, ALC <1.0 × 10⁹/L, and D-dimer ≥0.5 mg/L achieves a total predictor score of 364.4 points on the nomogram, corresponding to a predicted probability of PICC-CRBSI of 62%

Table 4 Nomogram model scores for predicting PICC-CRBSI risk Variables Score History of diabetes No 0 Yes 100 Age (Years) 1.1 (100 = 0) a Time of PICC placement <60days 0 ≥ 60days 86.8 PICC insertion attempts 1 attempt 0 ≥ 2 attempts 50.9 Catheter Diameter Single-lumen 0 Dual-lumen 49.5 Catheterized Vein Brachial vein 0 Basilic vein 6.4 Cephalic vein 94.4 ANC ≥ 1.5 × 10 9 /L 0 <1.5 × 10 9 /L 29.6 ALC ≥ 1.0 × 10 9 /L 0 <1.0 × 10 9 /L 80.8 D-dimer <0.5 mg/L 0 ≥ 0.5 mg/L 92.7 ANC Absolute Neutrophil Count, ALC Absolute Lymphocyte Count a 100 years = 0 points, and nomogram score increases by 1.1 points for each 10-year decrease in age

Performance of the CRBSI prediction nomogram

The area under the curve (AUC) was 0.883 (95% CI: 0.863–0.903) in the training cohort (Fig. 5 A) and 0.822 (95% CI: 0.719–0.924) in the validation cohort (Fig. 5 B). Calibration curves demonstrated good agreement between nomogram-predicted probabilities and observed outcomes across both cohorts (Figs. 5 C, D). The DCA revealed potential for clinical utility, with the net benefit curve exceeding the extreme lines over a threshold probability range of 0–1 in the derivation and validation datasets (Figs. 5 E, F).

Fig. 5 Performance evaluation of nomograms to predict PICC-CRBSI with the hematological malignancies; A. Receiver operating characteristic in the training cohort; B . Receiver operating characteristic in the validation cohort; C . Calibration curves in the training cohort; D . Calibration curves in the validation cohort; E . Decision curve analysis in the training cohort; F . Decision curve analysis in the validation cohort

Discussion

This study developed and validated a novel nomogram integrating nine independent predictors: History of diabetes, Age, Time of PICC placement, PICC insertion attempts, Catheter diameter, Catheterized vein, ANC, ALC, and D-dimer—to stratify the risk of PICC-CRBSI among patients with hematological malignancies. The model demonstrated strong discrimination and calibration in both training and validation cohorts, with an AUC of 0.883 and 0.822, respectively. Importantly, our nomogram addresses key limitations of prior CRBSI prediction tools by incorporating catheter-specific parameters, hematology-relevant biomarkers, and host-specific immune vulnerabilities, thereby providing a more tailored risk assessment for this unique patient population.

Microbial epidemiology

The microbial spectrum observed in our cohort underscores the distinctive vulnerability of hematologic malignancy patients. Although Gram-positive bacteria have historically dominated CRBSI etiology [ 42 ], we found a predominance of Gram-negative bacteria (57.1%), with Pseudomonas aeruginosa and Stenotrophomonas maltophilia as the most frequent isolates. This finding is consistent with recent evidence indicating a rising incidence of Gram-negative bacteria CRBSI (10~36%) in immunocompromised populations [ 43 – 45 ]. Alarmingly, Multidrug-resistant (MDR) pathogens may constitute 50% of these cases [ 46 ], substantially compromising therapeutic efficacy and clinical outcomes [ 47 , 48 ]. As demonstrated by Calò et al. [ 49 ], both timely catheter removal and adequate empiric antimicrobial therapy are protective against mortality, reinforcing the clinical imperative for early recognition, environmental infection control, and antimicrobial stewardship in hematology units.

Mechanistic insights into predictors

In multivariable analysis, nine factors were independently associated with PICC-CRBSI: history of diabetes (OR = 8.204), younger age (per-year OR = 0.977), time of PICC placement ≥ 60 days (OR = 6.213), ≥ 2 PICC insertion attempts (OR = 2.917), dual-lumen catheter (OR = 2.833), cephalic vein insertion (vs. basilic; OR = 7.288), ANC < 1.5 × 10⁹/L (OR = 1.864), ALC < 1.0 × 10⁹/L (OR = 5.476), and D-dimer ≥ 0.5 mg/L (OR = 7.029). These predictors can be broadly categorized into device- and procedure-related factors, host immune competence, and thrombo-inflammatory status.

First, device and procedure-related parameters emerged as major contributors to infection risk. Prolonged dwell time was one of the strongest predictors, consistent with reports that more than 70% of exit-site infections occur beyond 30 days post-insertion, highlighting that risk accrues primarily from long-term maintenance rather than the insertion procedure itself [ 50 ].Extended indwelling duration also acts synergistically with repeated insertion attempts, both of which cause endothelial disruption, fibrin sheath formation, and vascular injury, thereby creating a favorable substrate for microbial adherence and colonization [ 51 ]. Device configuration further influenced outcomes: dual-lumen catheters were associated with nearly a threefold higher risk of CRBSI, plausibly due to increased internal surface area and more frequent hub manipulation, both of which are established entry points for microbial contamination. Vein selection also proved critical; Cephalic vein cannulation carried a markedly higher risk compared with the Basilic vein, consistent with its smaller caliber, tortuous course, and more challenging catheter advancement, which predispose to suboptimal tip positioning and local thrombosis. By contrast, the Basilic vein offers anatomical advantages—including larger diameter, straighter trajectory, and greater distance from neurovascular structures—that are associated with higher procedural success and lower complication rates [ 50 , 52 – 56 ].

Second, Host-level vulnerabilities significantly potentiated catheter-related risks. Diabetes mellitus conferred the greatest single effect (OR = 8.204), reflecting its multifactorial impact on infection susceptibility, including impaired neutrophil chemotaxis and phagocytosis, diminished microvascular perfusion, and hyperglycemia-induced microbial overgrowth and skin colonization [ 57 – 60 ]. Hematologic immunosuppression was also central: both severe lymphocytopenia (ALC < 1.0 × 10⁹/L) and neutropenia (ANC < 1.5 × 10⁹/L) independently increased PICC-CRBSI risk, underscoring the essential roles of both adaptive and innate immune pathways in preventing microbial persistence.

Third, abnormalities in coagulation and systemic inflammation were closely linked to infection risk, with elevated D-dimer (≥ 0.5 mg/L) emerging as a strong independent predictor; as an established marker of hypercoagulability and subclinical inflammation, heightened D-dimer signifies the potential for nascent catheter-associated thrombus formation, likely reflecting underlying catheter-related thrombus formation and fibrin sheath development, which promote microbial adherence and shield biofilms from host defenses and antimicrobial agents [ 51 , 61 , 62 ].

Finally, the inverse association between age and CRBSI (per-year OR < 1) requires cautious interpretation. Rather than reflecting a true protective effect of older age, this relationship may be confounded by treatment intensity: younger patients are more likely to undergo aggressive induction chemotherapy, allogeneic transplantation, or salvage regimens that induce profound cytopenias and necessitate prolonged or repeated central venous access. Larger multicenter studies with stratified analyses are warranted to clarify this relationship.

Overall, these mechanistic insights highlight the interplay of vascular injury, immune suppression, and thrombosis in driving PICC-CRBSI risk and suggest actionable targets for prevention.

Disease type and transplantation-related considerations

Clinically. Our findings elucidate hematologic malignancy-specific vulnerabilities absent in general ICU or solid tumor CRBSI models [ 21 – 24 ]. In theory, acute leukemias, which are typically managed with more intensive chemotherapy regimens, confer deeper and more sustained marrow suppression compared with lymphomas, thereby increasing susceptibility to infection. Similarly, among transplant recipients, the degree of immunosuppression varies by modality: allogeneic HSCT generally entails the most profound immune dysfunction (due to conditioning intensity and graft-versus-host disease prophylaxis), followed by autologous HSCT, with CAR-T therapy generally resulting in comparatively shorter-term immune suppression (primarily lymphodepletion with fludarabine/cyclophosphamide). Physiologically, these gradations would be expected to translate into differential CRBSI risk.

In our cohort, however, no statistically significant association was observed between disease type (e.g., leukemia vs. lymphoma) or HSCT and the incidence of PICC-CRBSI. This is likely attributable to the limited sample size within these subgroups, as well as overlapping risk pathways already captured by our retained predictors (e.g., ANC, ALC, D-dimer). Nevertheless, prior studies have demonstrated higher infection risks in allo-HSCT recipients compared to auto-HSCT, largely due to more intensive conditioning regimens and graft-versus-host disease–related immunosuppression [ 63 , 64 ]. Similarly, acute leukemias undergoing induction chemotherapy have been consistently associated with profound and prolonged neutropenia, predisposing to invasive infections [ 28 ]. While our model did not retain these variables, their well-established biological relevance suggests that future large-scale, multi-center studies should explicitly examine the incremental prognostic value of disease subtype and transplantation modality.

Our findings hold direct implications for clinical practice. First, preferential use of the basilic vein for PICC placement should be emphasized to minimize infection risk [ 50 , 56 ]. Second, patients with diabetes and severe cytopenias warrant heightened surveillance, with routine monitoring of biomarkers such as D-dimer to guide prophylactic interventions (e.g., early thrombus detection, anticoagulation strategies). Third, training programs for nurses on long-term catheter care have been shown to significantly reduce infection rates [ 65 ] and should be integrated into standard care. Ultimately, the developed nomogram will provide clinicians with an evidence-based, individualized risk assessment at the bedside, facilitating targeted preventive interventions such as enhanced catheter care protocols, intensified monitoring, or even alternative access strategies for those at highest risk.

Limitations

Despite its robust validation, this study has several limitations. First, the single-center retrospective design may restrict generalizability, and external validation across diverse healthcare settings is warranted to confirm broader applicability. In addition, our prediction model was specifically developed and validated for patients with Hematologic Malignancies receiving PICCs, using Bard (Covington, GA) 4-French single-lumen PowerPICC ® Solo² Catheters (No. 7617405) and 5-French dual-lumen PowerPICC ® Dual-Lumen Catheters (No. 3275118), with placement and maintenance performed according to evidence-based best practice guidelines. Therefore, the applicability of our model to other types of central venous access devices, such as tunneled catheters or Powerlines, cannot be assumed. Second, biomarker measurements (e.g., ALC, D-dimer) were obtained pre-insertion and may not reflect dynamic changes during catheter dwell time. Third, unmeasured confounders—such as detailed chemotherapy regimens, antibiotic prophylaxis, or neutrophil recovery kinetics-were not incorporated, potentially omitting influential variables. Moreover, our study was unable to comprehensively assess the depth and duration of neutropenia as independent risk factors for PICC-CRBSI due to retrospective data constraints. Clarifying these effects requires more granular longitudinal monitoring in future investigations. Fourth, due to the heterogeneity of chemotherapy regimens and frequent adjustments based on individual patient conditions, the impact of specific chemotherapy agents on PICC-CRBSI risk could not be assessed. Certain agents may exert greater mucosal or tissue toxicity and thus influence infection risk, but this level of detail was beyond the scope of the current retrospective study and warrants further investigation in future research. Fifth, the study did not separately analyze patients receiving immunotherapies. As immunotherapies are generally less cytotoxic than conventional chemotherapy, their impact on PICC-CRBSI risk may differ and warrants further investigation in future studies. Sixth, the study did not assess the antimicrobial resistance patterns of Gram-negative pathogens, including multidrug-resistant (MDR) strains, which have been increasingly recognized as emerging causes of PICC-CRBSI. Finally, prospective multicenter studies are needed to validate the nomogram’s real-world utility and confirm its impact on clinical decision-making.

Conclusion

This study established and validated a practical nomogram integrating patient comorbidities (History of diabetes), demographic factors (Age), catheter-related variables (Time of PICC placement, PICC insertion attempts, Catheter Diameter, Catheterized vein), and disease-specific biomarkers (ANC, ALC and D-dimer) to predict PICC-CRBSI in hematologic malignancy patients. Its methodological rigor, leveraging SMOTE-NC to address class imbalance and LASSO regression for feature selection, ensures robustness for this high-risk population. This tool may provide clinicians with a practical means to stratify patients with hematologic malignancies into differential risk categories, guiding personalized surveillance and preemptive strategies (e.g., intensified catheter care protocols, early diagnostic evaluation) to mitigate PICC-CRBSI-associated morbidity and mortality in hematologic oncology practice.

Abbreviations

Absolute lymphocyte count

Absolute neutrophil count

Area under the receiver operating characteristic curve

Body mass index

Central venous catheter

Catheter-related bloodstream infection

Decision curve analysis

Hospital Information System

Hematopoietic stem cell transplantation

Interquartile range

Institutional Review Board

Least absolute shrinkage and selection operator

Multidrug-resistant

Odds ratio

Peripherally inserted central catheter

Platelets

Receiver operating characteristic

Standard deviation

White blood cells

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Qian Guo and Yuting Liu contributed equally to this work and are the co-first authors.

Acknowledgements

The authors sincerely thank all patients with hematologic malignancies who participated in this study.

Authors’ contributions

GQ and LYT conceived and designed the study. GQ: Writing–original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. LYT: Writing–review & editing, Methodology, Software, Investigation, Conceptualization. ZXL: Investigation, Formal analysis. GYT: Investigation, Visualization, Software. ZY: Writing–review & editing, Visualization, Validation. ZX: Writing–review & editing, Investigation, Project administration.All authors read and approved the final manuscript.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Medical Ethics Committee of the First Affiliated Hospital of Zhejiang Chinese Medical University (Zhejiang Provincial Hospital of Chinese Medicine) (ref: 2025-KLS-109-01). The study was conducted in accordance with the ethical principles of the Declaration of Helsinki. Given the retrospective and observational design, the requirement for informed consent was waived by the Medical Ethics Committee of the First Affiliated Hospital of Zhejiang Chinese Medical University. Patient information was anonymized and de-identified prior to analysis.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

References

1. Yue J Zhang Y Xu F A clinical study of peripherally inserted central catheter-related venous thromboembolism in patients with hematological malignancies Sci Rep 2022 12 1 9871 35701467 10.1038/s41598-022-13916-5 PMC9197841 Yue J, Zhang Y, Xu F, et al. A clinical study of peripherally inserted central catheter-related venous thromboembolism in patients with hematological malignancies. Sci Rep. 2022;12(1):9871. 35701467 10.1038/s41598-022-13916-5 PMC9197841

2. Dos Santos BN Beato BVG Ferreira EB Prevalence of PICC-related thrombosis in patients with hematological malignancies: a systematic review[J] Support Care Cancer 2024 32 7 462 38922487 10.1007/s00520-024-08636-7 Dos Santos BN, Beato BVG, Ferreira EB, et al. Prevalence of PICC-related thrombosis in patients with hematological malignancies: a systematic review[J]. Support Care Cancer. 2024;32(7):462. 38922487 10.1007/s00520-024-08636-7

3. Cui L, Cao J. Nursing strategies for PICC in patients with hematologic malignancies [J]. Asian J Surg. 2024;S1015-9584(24):2776-3. 10.1016/j.asjsur.2024.11.149 39632185

4. Cotogni P Mussa B Degiorgis C Comparative complication rates of 854 central venous access devices for home parenteral nutrition in cancer patients: a prospective study of over 169,000 catheter-days JPEN J Parenter Enteral Nutr 2021 45 4 768 76 32511768 10.1002/jpen.1939 Cotogni P, Mussa B, Degiorgis C, et al. Comparative complication rates of 854 central venous access devices for home parenteral nutrition in cancer patients: a prospective study of over 169,000 catheter-days. JPEN J Parenter Enteral Nutr. 2021;45(4):768–76. 32511768 10.1002/jpen.1939

5. Morano SG Latagliata R Girmenia C Catheter-associated bloodstream infections and thrombotic risk in hematologic patients with peripherally inserted central catheters (PICC)[J] Supportive Care Cancer: Official J Multinational Association Supportive Care Cancer 2015 23 11 3289 95 10.1007/s00520-015-2740-7 25910751 Morano SG, Latagliata R, Girmenia C, et al. Catheter-associated bloodstream infections and thrombotic risk in hematologic patients with peripherally inserted central catheters (PICC)[J]. Supportive Care Cancer: Official J Multinational Association Supportive Care Cancer. 2015;23(11):3289–95. 10.1007/s00520-015-2740-7 25910751

6. Jaffray J Witmer C O’Brien SH Peripherally inserted central catheters lead to a high risk of venous thromboembolism in children[J] Blood 2020 135 3 220 6 31909784 10.1182/blood.2019002260 Jaffray J, Witmer C, O’Brien SH, et al. Peripherally inserted central catheters lead to a high risk of venous thromboembolism in children[J]. Blood. 2020;135(3):220–6. 31909784 10.1182/blood.2019002260

7. Balsorano P Virgili G Villa G Peripherally inserted central catheter-related thrombosis rate in modern vascular access era-when insertion technique matters: a systematic review and meta-analysis J Vasc Access 2020 21 1 45 54 31177939 10.1177/1129729819852203 Balsorano P, Virgili G, Villa G, et al. Peripherally inserted central catheter-related thrombosis rate in modern vascular access era-when insertion technique matters: a systematic review and meta-analysis. J Vasc Access. 2020;21(1):45–54. 31177939 10.1177/1129729819852203

8. Thomsen SL Boa R Vinter-Jensen L Safety and efficacy of midline vs peripherally inserted central catheters among adults receiving IV therapy: a randomized clinical trial[J] JAMA Netw Open 2024 7 2 e2355716 38349655 10.1001/jamanetworkopen.2023.55716 PMC10865154 Thomsen SL, Boa R, Vinter-Jensen L, et al. Safety and efficacy of midline vs peripherally inserted central catheters among adults receiving IV therapy: a randomized clinical trial[J]. JAMA Netw Open. 2024;7(2):e2355716. 38349655 10.1001/jamanetworkopen.2023.55716 PMC10865154

9. Beckers MMJ Ruven HJT Seldenrijk CA Risk of thrombosis and infections of central venous catheters and totally implanted access ports in patients treated for cancer[J] Thromb Res 2010 125 4 318 21 19640573 10.1016/j.thromres.2009.06.008 Beckers MMJ, Ruven HJT, Seldenrijk CA, et al. Risk of thrombosis and infections of central venous catheters and totally implanted access ports in patients treated for cancer[J]. Thromb Res. 2010;125(4):318–21. 19640573 10.1016/j.thromres.2009.06.008

10. Gahlot R Nigam C Kumar V Catheter-related bloodstream infections Int J Crit Illn Inj Sci 2014 4 2 162 7 25024944 10.4103/2229-5151.134184 PMC4093967 Gahlot R, Nigam C, Kumar V, et al. Catheter-related bloodstream infections. Int J Crit Illn Inj Sci. 2014;4(2):162–7. 25024944 10.4103/2229-5151.134184 PMC4093967

11. Central line–associated blood stream infections - StatPearls. - NCBI bookshelf[EB/OL]. [2025-09-13]. https://www.ncbi.nlm.nih.gov/books/NBK430891/ .

12. Ram R Gafter-Gvili A Raanani P Surveillance of infectious complications in hemato-oncological patients[J] Isr Med Association Journal: IMAJ 2009 11 3 133 7 19544700 Ram R, Gafter-Gvili A, Raanani P, et al. Surveillance of infectious complications in hemato-oncological patients[J]. Isr Med Association Journal: IMAJ. 2009;11(3):133–7. 19544700

13. Chaberny IF Ruseva E Sohr D Surveillance with successful reduction of central line-associated bloodstream infections among neutropenic patients with hematologic or oncologic malignancies Ann Hematol 2009 88 9 907 12 19139891 10.1007/s00277-008-0687-1 Chaberny IF, Ruseva E, Sohr D, et al. Surveillance with successful reduction of central line-associated bloodstream infections among neutropenic patients with hematologic or oncologic malignancies. Ann Hematol. 2009;88(9):907–12. 19139891 10.1007/s00277-008-0687-1

14. Apostolopoulou E Raftopoulos V Terzis K Infection probability score, APACHE II and KARNOFSKY scoring systems as predictors of bloodstream infection onset in hematology-oncology patients BMC Infect Dis 2010 10 135 20504343 10.1186/1471-2334-10-135 PMC2890000 Apostolopoulou E, Raftopoulos V, Terzis K, et al. Infection probability score, APACHE II and KARNOFSKY scoring systems as predictors of bloodstream infection onset in hematology-oncology patients. BMC Infect Dis. 2010;10:135. 20504343 10.1186/1471-2334-10-135 PMC2890000

15. Lacostena-Pérez ME Buesa-Escar AM Gil-Alós AM Complications related to the insertion and maintenance of peripheral venous access central venous catheter Enferm Intensiva 2019 30 3 116 26 10.1016/j.enfi.2018.05.002 30190250 Lacostena-Pérez ME, Buesa-Escar AM, Gil-Alós AM. Complications related to the insertion and maintenance of peripheral venous access central venous catheter. Enferm Intensiva. 2019;30(3):116–26. 10.1016/j.enfi.2018.05.002 30190250

16. Tilton D Central venous access device infections in the critical care unit Crit Care Nurs Q 2006 29 2 117 22 16641647 10.1097/00002727-200604000-00004 Tilton D. Central venous access device infections in the critical care unit. Crit Care Nurs Q. 2006;29(2):117–22. 16641647 10.1097/00002727-200604000-00004

17. Baier C Linke L Eder M Incidence, risk factors and healthcare costs of central line-associated nosocomial bloodstream infections in hematologic and oncologic patients PLoS ONE 2020 15 1 e0227772 31978169 10.1371/journal.pone.0227772 PMC6980604 Baier C, Linke L, Eder M, et al. Incidence, risk factors and healthcare costs of central line-associated nosocomial bloodstream infections in hematologic and oncologic patients. PLoS ONE. 2020;15(1):e0227772. 31978169 10.1371/journal.pone.0227772 PMC6980604

18. Peris A Zagli G Bonizzoli M Implantation of 3951 long-term central venous catheters: performances, risk analysis, and patient comfort after ultrasound-guidance introduction Anesth Analg 2010 111 5 1194 201 20829559 10.1213/ANE.0b013e3181f333c1 Peris A, Zagli G, Bonizzoli M, et al. Implantation of 3951 long-term central venous catheters: performances, risk analysis, and patient comfort after ultrasound-guidance introduction. Anesth Analg. 2010;111(5):1194–201. 20829559 10.1213/ANE.0b013e3181f333c1

19. Mollee P Jones M Stackelroth J Catheter-associated bloodstream infection incidence and risk factors in adults with cancer: a prospective cohort study[J] J Hosp Infect 2011 78 1 26 30 21459476 10.1016/j.jhin.2011.01.018 Mollee P, Jones M, Stackelroth J, et al. Catheter-associated bloodstream infection incidence and risk factors in adults with cancer: a prospective cohort study[J]. J Hosp Infect. 2011;78(1):26–30. 21459476 10.1016/j.jhin.2011.01.018

20. Touré A Vanhems P Lombard-Bohas C Totally implantable central venous access port infections in patients with digestive cancer: incidence and risk factors Am J Infect Control 2012 40 10 935 9 22633131 10.1016/j.ajic.2012.01.024 Touré A, Vanhems P, Lombard-Bohas C, et al. Totally implantable central venous access port infections in patients with digestive cancer: incidence and risk factors. Am J Infect Control. 2012;40(10):935–9. 22633131 10.1016/j.ajic.2012.01.024

21. Zhang Y Wang Y Sheng Z Incidence rate, pathogens and economic burden of catheter-related bloodstream infection: a single-center, retrospective case-control study[J] Infect Drug Resist 2023 16 3551 60 37305736 10.2147/IDR.S406681 PMC10256568 Zhang Y, Wang Y, Sheng Z, et al. Incidence rate, pathogens and economic burden of catheter-related bloodstream infection: a single-center, retrospective case-control study[J]. Infect Drug Resist. 2023;16:3551–60. 37305736 10.2147/IDR.S406681 PMC10256568

22. Hoggard JG Blair RD Montero M Clinical outcomes associated with the use of the nexsite Hemodialysis catheter with new exit barrier technology: results from a prospective, observational multi-center registry study[J] PLoS ONE 2019 14 10 e0223285 31589644 10.1371/journal.pone.0223285 PMC6779244 Hoggard JG, Blair RD, Montero M, et al. Clinical outcomes associated with the use of the nexsite Hemodialysis catheter with new exit barrier technology: results from a prospective, observational multi-center registry study[J]. PLoS ONE. 2019;14(10):e0223285. 31589644 10.1371/journal.pone.0223285 PMC6779244

23. Kumar A Sharma RM Jaideep CN Diagnosis of central venous catheter-related bloodstream infection without catheter removal: a prospective observational study Med J Armed Forces India 2014 70 1 17 21 24623941 10.1016/j.mjafi.2013.08.001 PMC3946418 Kumar A, Sharma RM, Jaideep CN, et al. Diagnosis of central venous catheter-related bloodstream infection without catheter removal: a prospective observational study. Med J Armed Forces India. 2014;70(1):17–21. 24623941 10.1016/j.mjafi.2013.08.001 PMC3946418

24. Lee JH Kim MU Kim ET Prevalence and predictors of peripherally inserted central venous catheter associated bloodstream infections in cancer patients: a multicentre cohort study Medicine 2020 99 6 e19056 32028425 10.1097/MD.0000000000019056 PMC7015558 Lee JH, Kim MU, Kim ET, et al. Prevalence and predictors of peripherally inserted central venous catheter associated bloodstream infections in cancer patients: a multicentre cohort study. Medicine. 2020;99(6):e19056. 32028425 10.1097/MD.0000000000019056 PMC7015558

25. Herc E Patel P Washer LL A model to predict central-line-associated bloodstream infection among patients with peripherally inserted central catheters: the MPC score Infect Control Hosp Epidemiol 2017 38 10 1155 66 28807074 10.1017/ice.2017.167 Herc E, Patel P, Washer LL, et al. A model to predict central-line-associated bloodstream infection among patients with peripherally inserted central catheters: the MPC score. Infect Control Hosp Epidemiol. 2017;38(10):1155–66. 28807074 10.1017/ice.2017.167

26. Wang J Wang M Zhao A Microbiology and prognostic prediction model of bloodstream infection in patients with hematological malignancies Front Cell Infect Microbiol 2023 13 1167638 37457950 10.3389/fcimb.2023.1167638 PMC10347389 Wang J, Wang M, Zhao A, et al. Microbiology and prognostic prediction model of bloodstream infection in patients with hematological malignancies. Front Cell Infect Microbiol. 2023;13:1167638. 37457950 10.3389/fcimb.2023.1167638 PMC10347389

27. Tang Y Cheng Q Yang Q Prognostic factors and scoring model of hematological malignancies patients with bloodstream infections Infection 2018 46 4 513 21 29767394 10.1007/s15010-018-1151-3 Tang Y, Cheng Q, Yang Q, et al. Prognostic factors and scoring model of hematological malignancies patients with bloodstream infections. Infection. 2018;46(4):513–21. 29767394 10.1007/s15010-018-1151-3

28. Heidenreich D Hansen E Kreil S The insertion site is the main risk factor for central venous catheter-related complications in patients with hematologic malignancies Am J Hematol 2022 97 3 303 10 34978721 10.1002/ajh.26445 Heidenreich D, Hansen E, Kreil S, et al. The insertion site is the main risk factor for central venous catheter-related complications in patients with hematologic malignancies. Am J Hematol. 2022;97(3):303–10. 34978721 10.1002/ajh.26445

29. Schalk E Panse J Which trial do we need? Scheduled central venous catheter removal vs. routine clinical care for prevention of central venous catheter-related bloodstream infections in patients with haematologic malignancies Clin Microbiol Infect 2023 29 4 417 8 36639031 10.1016/j.cmi.2023.01.003 Schalk E, Panse J. Which trial do we need? Scheduled central venous catheter removal vs. routine clinical care for prevention of central venous catheter-related bloodstream infections in patients with haematologic malignancies. Clin Microbiol Infect. 2023;29(4):417–8. 36639031 10.1016/j.cmi.2023.01.003

30. Panse J Tölle D Fiegle E Scheduled removal of central venous catheters (CVC) to prevent CVC-related bloodstream infections in patients with hematological disease or autologous stem cell transplantation: a registry-based randomized simulation-study Ann Hematol 2022 101 10 2317 24 35978182 10.1007/s00277-022-04958-w PMC9463193 Panse J, Tölle D, Fiegle E, et al. Scheduled removal of central venous catheters (CVC) to prevent CVC-related bloodstream infections in patients with hematological disease or autologous stem cell transplantation: a registry-based randomized simulation-study. Ann Hematol. 2022;101(10):2317–24. 35978182 10.1007/s00277-022-04958-w PMC9463193

31. Heybati K Seeger R Thyagu S Catheter management across patients with hematologic malignancies and catheter-related blood stream infections: a systematic review Ann Hematol 2022 101 11 2515 24 36107204 10.1007/s00277-022-04969-7 Heybati K, Seeger R, Thyagu S, et al. Catheter management across patients with hematologic malignancies and catheter-related blood stream infections: a systematic review. Ann Hematol. 2022;101(11):2515–24. 36107204 10.1007/s00277-022-04969-7

32. Emoli A Cappuccio S Marche B [the ISP (safe insertion of PICCs) protocol: a bundle of 8 recommendations to minimize the complications related to the peripherally inserted central venous catheters (PICC)][J] Assist Inferm Ric 2014 33 2 82 9 25002059 10.1702/1539.16813 Emoli A, Cappuccio S, Marche B, et al. [the ISP (safe insertion of PICCs) protocol: a bundle of 8 recommendations to minimize the complications related to the peripherally inserted central venous catheters (PICC)][J]. Assist Inferm Ric. 2014;33(2):82–9. 25002059 10.1702/1539.16813

33. Böll B Schalk E Buchheidt D Central venous catheter-related infections in hematology and oncology: 2020 updated guidelines on diagnosis, management, and prevention by the infectious diseases working party (AGIHO) of the German society of hematology and medical oncology (DGHO) Ann Hematol 2021 100 1 239 59 32997191 10.1007/s00277-020-04286-x PMC7782365 Böll B, Schalk E, Buchheidt D, et al. Central venous catheter-related infections in hematology and oncology: 2020 updated guidelines on diagnosis, management, and prevention by the infectious diseases working party (AGIHO) of the German society of hematology and medical oncology (DGHO). Ann Hematol. 2021;100(1):239–59. 32997191 10.1007/s00277-020-04286-x PMC7782365

34. Mermel LA Allon M Bouza E Clinical practice guidelines for the diagnosis and management of intravascular catheter-related infection: 2009 update by the Infectious Diseases Society of America Clin Infect Dis 2009 49 1 1 45 19489710 10.1086/599376 PMC4039170 Mermel LA, Allon M, Bouza E, et al. Clinical practice guidelines for the diagnosis and management of intravascular catheter-related infection: 2009 update by the Infectious Diseases Society of America. Clin Infect Dis. 2009;49(1):1–45. 19489710 10.1086/599376 PMC4039170

35. O’Grady NP Alexander M Burns LA Guidelines for the prevention of intravascular catheter-related infections[J] Clin Infect Diseases: Official Publication Infect Dis Soc Am 2011 52 9 e162 193 10.1093/cid/cir257 PMC3106269 21460264 O’Grady NP, Alexander M, Burns LA, et al. Guidelines for the prevention of intravascular catheter-related infections[J]. Clin Infect Diseases: Official Publication Infect Dis Soc Am. 2011;52(9):e162–193. 10.1093/cid/cir257 PMC3106269 21460264

36. Elangovan S Lo JJ Xie Y Impact of central-line-associated bloodstream infections and catheter-related bloodstream infections: a systematic review and meta-analysis J Hosp Infect 2024 152 126 37 39151801 10.1016/j.jhin.2024.08.002 Elangovan S, Lo JJ, Xie Y, et al. Impact of central-line-associated bloodstream infections and catheter-related bloodstream infections: a systematic review and meta-analysis. J Hosp Infect. 2024;152:126–37. 39151801 10.1016/j.jhin.2024.08.002

37. Collins GS Reitsma JB Altman DG Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement BMJ 2015 350 g7594 25569120 10.1136/bmj.g7594 Collins GS, Reitsma JB, Altman DG, et al. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement. BMJ. 2015;350:g7594. 25569120 10.1136/bmj.g7594

38. Sui Y Wei Y Zhao D Computer-aided lung nodule recognition by SVM classifier based on combination of random undersampling and SMOTE Comput Math Methods Med 2015 2015 368674 25977704 10.1155/2015/368674 PMC4419492 Sui Y, Wei Y, Zhao D. Computer-aided lung nodule recognition by SVM classifier based on combination of random undersampling and SMOTE. Comput Math Methods Med. 2015;2015:368674. 25977704 10.1155/2015/368674 PMC4419492

39. SMOTE. synthetic minority over-sampling technique | journal of artificial intelligence research[EB/OL]. [2025-09-13]. https://www.jair.org/index.php/jair/article/view/10302 .

40. Sung SF Hung LC Hu YH Developing a stroke alert trigger for clinical decision support at emergency triage using machine learning Int J Med Inform 2021 152 104505 34030088 10.1016/j.ijmedinf.2021.104505 Sung SF, Hung LC, Hu YH. Developing a stroke alert trigger for clinical decision support at emergency triage using machine learning. Int J Med Inform. 2021;152:104505. 34030088 10.1016/j.ijmedinf.2021.104505

41. An insight into Classification with imbalanced data: empirical results and current trends on using data intrinsic characteristics[J] Inf Sci 2013 250 113 41 An insight into. Classification with imbalanced data: empirical results and current trends on using data intrinsic characteristics[J]. Inf Sci. 2013;250:113–41.

42. Novosad SA Fike L Dudeck MA Pathogens causing central-line-associated bloodstream infections in acute-care hospitals-United states, 2011–2017 Infect Control Hosp Epidemiol 2020 41 3 313 9 31915083 10.1017/ice.2019.303 PMC13108540 Novosad SA, Fike L, Dudeck MA, et al. Pathogens causing central-line-associated bloodstream infections in acute-care hospitals-United states, 2011–2017. Infect Control Hosp Epidemiol. 2020;41(3):313–9. 31915083 10.1017/ice.2019.303 PMC13108540

43. Muñoz P Cruz AF Rodríguez-Créixems M Gram-negative bloodstream infections[J] Int J Antimicrob Agents 2008 32 Suppl 1 S10 14 18775648 10.1016/j.ijantimicag.2008.06.015 Muñoz P, Cruz AF, Rodríguez-Créixems M, et al. Gram-negative bloodstream infections[J]. Int J Antimicrob Agents. 2008;32(Suppl 1):S10–14. 18775648 10.1016/j.ijantimicag.2008.06.015

44. Parameswaran R Sherchan JB Varma DM Intravascular catheter-related infections in an Indian tertiary care hospital[J] J Infect Developing Ctries 2011 5 6 452 8 10.3855/jidc.1261 21727644 Parameswaran R, Sherchan JB, Varma DM, et al. Intravascular catheter-related infections in an Indian tertiary care hospital[J]. J Infect Developing Ctries. 2011;5(6):452–8. 10.3855/jidc.1261 21727644

45. Nørgaard M Larsson H Pedersen G Short-term mortality of bacteraemia in elderly patients with haematological malignancies[J] Br J Haematol 2006 132 1 25 31 16371016 10.1111/j.1365-2141.2005.05845.x Nørgaard M, Larsson H, Pedersen G, et al. Short-term mortality of bacteraemia in elderly patients with haematological malignancies[J]. Br J Haematol. 2006;132(1):25–31. 16371016 10.1111/j.1365-2141.2005.05845.x

46. Ruiz-Ruigómez M Aguado JM Duration of antibiotic therapy in central venous catheter-related bloodstream infection due to gram-negative bacilli Curr Opin Infect Dis 2021 34 6 681 5 34261908 10.1097/QCO.0000000000000763 Ruiz-Ruigómez M, Aguado JM. Duration of antibiotic therapy in central venous catheter-related bloodstream infection due to gram-negative bacilli. Curr Opin Infect Dis. 2021;34(6):681–5. 34261908 10.1097/QCO.0000000000000763

47. Lee YM Moon C Kim YJ Clinical impact of delayed catheter removal for patients with central-venous-catheter-related gram-negative bacteraemia[J] J Hosp Infect 2018 99 1 106 13 29330016 10.1016/j.jhin.2018.01.004 Lee YM, Moon C, Kim YJ, et al. Clinical impact of delayed catheter removal for patients with central-venous-catheter-related gram-negative bacteraemia[J]. J Hosp Infect. 2018;99(1):106–13. 29330016 10.1016/j.jhin.2018.01.004

48. Bouza E Eworo A Fernández Cruz A Catheter-related bloodstream infections caused by gram-negative bacteria[J] J Hosp Infect 2013 85 4 316 20 24183319 10.1016/j.jhin.2013.08.008 Bouza E, Eworo A, Fernández Cruz A, et al. Catheter-related bloodstream infections caused by gram-negative bacteria[J]. J Hosp Infect. 2013;85(4):316–20. 24183319 10.1016/j.jhin.2013.08.008

49. Calò F Retamar P Martínez Pérez-Crespo PM Catheter-related bloodstream infections: predictive factors for gram-negative bacteria aetiology and 30 day mortality in a multicentre prospective cohort J Antimicrob Chemother 2020 75 10 3056 61 32688386 10.1093/jac/dkaa262 Calò F, Retamar P, Martínez Pérez-Crespo PM, et al. Catheter-related bloodstream infections: predictive factors for gram-negative bacteria aetiology and 30 day mortality in a multicentre prospective cohort. J Antimicrob Chemother. 2020;75(10):3056–61. 32688386 10.1093/jac/dkaa262

50. Campagna S Gonella S Berchialla P Can peripherally inserted central catheters be safely placed in patients with cancer receiving chemotherapy? A retrospective study of almost 400,000 catheter-days[J] Oncologist 2019 24 9 e953 9 30755503 10.1634/theoncologist.2018-0281 PMC6738314 Campagna S, Gonella S, Berchialla P, et al. Can peripherally inserted central catheters be safely placed in patients with cancer receiving chemotherapy? A retrospective study of almost 400,000 catheter-days[J]. Oncologist. 2019;24(9):e953–9. 30755503 10.1634/theoncologist.2018-0281 PMC6738314

51. Raad II Luna M Khalil SA The relationship between the thrombotic and infectious complications of central venous catheters[J] JAMA 1994 271 13 1014 6 8139059 Raad II, Luna M, Khalil SA, et al. The relationship between the thrombotic and infectious complications of central venous catheters[J]. JAMA. 1994;271(13):1014–6. 8139059

52. Bertoglio S Faccini B Lalli L Peripherally inserted central catheters (PICCs) in cancer patients under chemotherapy: a prospective study on the incidence of complications and overall failures J Surg Oncol 2016 113 6 708 14 27020965 10.1002/jso.24220 Bertoglio S, Faccini B, Lalli L, et al. Peripherally inserted central catheters (PICCs) in cancer patients under chemotherapy: a prospective study on the incidence of complications and overall failures. J Surg Oncol. 2016;113(6):708–14. 27020965 10.1002/jso.24220

53. Kang J Chen W Sun W Peripherally inserted central catheter-related complications in cancer patients: a prospective study of over 50,000 catheter days J Vasc Access 2017 18 2 153 7 28218366 10.5301/jva.5000670 Kang J, Chen W, Sun W, et al. Peripherally inserted central catheter-related complications in cancer patients: a prospective study of over 50,000 catheter days. J Vasc Access. 2017;18(2):153–7. 28218366 10.5301/jva.5000670

54. Liu Y Gao Y Wei L Peripherally inserted central catheter thrombosis incidence and risk factors in cancer patients: a double-center prospective investigation[J] Ther Clin Risk Manag 2015 11 153 60 25673995 10.2147/TCRM.S73379 PMC4321638 Liu Y, Gao Y, Wei L, et al. Peripherally inserted central catheter thrombosis incidence and risk factors in cancer patients: a double-center prospective investigation[J]. Ther Clin Risk Manag. 2015;11:153–60. 25673995 10.2147/TCRM.S73379 PMC4321638

55. Luo L Jing XM Wang GR Peripherally inserted central catheter-related upper extremity venous thrombosis in oncology patients: a prospective study based on doppler sonography[J] J Ultrasound Medicine: Official J Am Inst Ultrasound Med 2016 35 8 1759 63 10.7863/ultra.15.08019 27371372 Luo L, Jing XM, Wang GR, et al. Peripherally inserted central catheter-related upper extremity venous thrombosis in oncology patients: a prospective study based on doppler sonography[J]. J Ultrasound Medicine: Official J Am Inst Ultrasound Med. 2016;35(8):1759–63. 10.7863/ultra.15.08019 27371372

56. Chinnock B Thornton S Hendey GW Predictors of success in nurse-performed ultrasound-guided cannulation[J] J Emerg Med 2007 33 4 401 5 17976752 10.1016/j.jemermed.2007.02.027 Chinnock B, Thornton S, Hendey GW. Predictors of success in nurse-performed ultrasound-guided cannulation[J]. J Emerg Med. 2007;33(4):401–5. 17976752 10.1016/j.jemermed.2007.02.027

57. Chopra V Ratz D Kuhn L PICC-associated bloodstream infections: prevalence, patterns, and predictors Am J Med 2014 127 4 319 28 24440542 10.1016/j.amjmed.2014.01.001 Chopra V, Ratz D, Kuhn L, et al. PICC-associated bloodstream infections: prevalence, patterns, and predictors. Am J Med. 2014;127(4):319–28. 24440542 10.1016/j.amjmed.2014.01.001

58. Kim K Kim Y Peck KR Previous peripherally inserted central catheter (PICC) placement as a risk factor for PICC-associated bloodstream infections Am J Infect Control 2020 48 10 1166 70 31937457 10.1016/j.ajic.2019.12.014 Kim K, Kim Y, Peck KR. Previous peripherally inserted central catheter (PICC) placement as a risk factor for PICC-associated bloodstream infections. Am J Infect Control. 2020;48(10):1166–70. 31937457 10.1016/j.ajic.2019.12.014

59. He K Wan Y Xian S Risk analysis on infection caused by peripherally inserted central catheter for bone tumor patients J Cancer Res Ther 2018 14 1 90 3 29516966 10.4103/jcrt.JCRT_777_17 He K, Wan Y, Xian S. Risk analysis on infection caused by peripherally inserted central catheter for bone tumor patients. J Cancer Res Ther. 2018;14(1):90–3. 29516966 10.4103/jcrt.JCRT_777_17

60. Liu X Tao S Ji H Risk factors for peripherally inserted central catheter (PICC)-associated infections in patients receiving chemotherapy and the preventive effect of a self-efficacy intervention program: a randomized controlled trial Ann Palliat Med 2021 10 9 9398 405 34628865 10.21037/apm-21-1848 Liu X, Tao S, Ji H, et al. Risk factors for peripherally inserted central catheter (PICC)-associated infections in patients receiving chemotherapy and the preventive effect of a self-efficacy intervention program: a randomized controlled trial. Ann Palliat Med. 2021;10(9):9398–405. 34628865 10.21037/apm-21-1848

61. Richters A van Vliet M Peer PGM Incidence of and risk factors for persistent gram-positive bacteraemia and catheter-related thrombosis in haematopoietic stem cell transplantation Bone Marrow Transplant 2014 49 2 264 9 24185587 10.1038/bmt.2013.172 Richters A, van Vliet M, Peer PGM, et al. Incidence of and risk factors for persistent gram-positive bacteraemia and catheter-related thrombosis in haematopoietic stem cell transplantation. Bone Marrow Transplant. 2014;49(2):264–9. 24185587 10.1038/bmt.2013.172

62. Lordick F Hentrich M Decker T Ultrasound screening for internal jugular vein thrombosis aids the detection of central venous catheter-related infections in patients with haemato-oncological diseases: a prospective observational study Br J Haematol 2003 120 6 1073 8 12648081 10.1046/j.1365-2141.2003.04199.x Lordick F, Hentrich M, Decker T, et al. Ultrasound screening for internal jugular vein thrombosis aids the detection of central venous catheter-related infections in patients with haemato-oncological diseases: a prospective observational study. Br J Haematol. 2003;120(6):1073–8. 12648081 10.1046/j.1365-2141.2003.04199.x

63. Meyer E Beyersmann J Bertz H Risk factor analysis of blood stream infection and pneumonia in neutropenic patients after peripheral blood stem-cell transplantation Bone Marrow Transplant 2007 39 3 173 8 17245425 10.1038/sj.bmt.1705561 Meyer E, Beyersmann J, Bertz H, et al. Risk factor analysis of blood stream infection and pneumonia in neutropenic patients after peripheral blood stem-cell transplantation. Bone Marrow Transplant. 2007;39(3):173–8. 17245425 10.1038/sj.bmt.1705561

64. McDonald MK Culos KA Gatwood KS Defining incidence and risk factors for catheter-associated bloodstream infections in an outpatient adult hematopoietic cell transplantation program Biol Blood Marrow Transplant 2018 24 10 2081 7 29753159 10.1016/j.bbmt.2018.04.031 McDonald MK, Culos KA, Gatwood KS, et al. Defining incidence and risk factors for catheter-associated bloodstream infections in an outpatient adult hematopoietic cell transplantation program. Biol Blood Marrow Transplant. 2018;24(10):2081–7. 29753159 10.1016/j.bbmt.2018.04.031

65. Tian G Zhu Y Qi L Efficacy of multifaceted interventions in reducing complications of peripherally inserted central catheter in adult oncology patients Support Care Cancer 2010 18 10 1293 8 19787379 10.1007/s00520-009-0747-7 Tian G, Zhu Y, Qi L, et al. Efficacy of multifaceted interventions in reducing complications of peripherally inserted central catheter in adult oncology patients. Support Care Cancer. 2010;18(10):1293–8. 19787379 10.1007/s00520-009-0747-7