Is longer indwelling duration an independent risk factor for UVC-related thrombosis? Insights from a matched case-control analysis Authors: Xuemei Huang, Ba Wei, Shuangjing Pan, Shaohong Zhang, Shumin Wei, Ximing Mo, Haihong Tang, Zhenqin Mo, Jianping Chen Published: 2026 ABSTRACT AbstractBackgroundUmbilical venous catheter (UVC)-associated thrombosis is a major complication in the neonatal intensive care unit (NICU). Whether longer catheter dwell time independently increases this risk remains highly debated. This study aimed to isolate and evaluate specific risk factors for UVC-related thrombosis using a rigorous matched case-control design to control for baseline developmental confounders.MethodsWe conducted a retrospective, 1:2 matched case-control study. Neonates with ultrasound-confirmed UVC-related thrombosis were matched with controls based on gestational age and birth weight. Clinical characteristics, UVC indwelling duration, and laboratory biomarkers (including D-Dimer) were compared between the two cohorts. Receiver Operating Characteristic (ROC) curve analysis was utilized to evaluate the predictive utility of these variables.ResultsThe study included 34 cases and 68 matched controls. Baseline demographics, including gestational age and birth weight, were successfully balanced between groups (all P > 0.05). The median UVC indwelling duration was identical in both the thrombosis and control cohorts (8.0 vs. 8.0 days, P = 0.206). Although D-Dimer levels exhibited greater variance in the thrombosis group, the median difference was not statistically significant (P = 0.808). Furthermore, ROC analysis demonstrated poor predictive value for both indwelling duration (AUC = 0.42) and D-Dimer levels (AUC = 0.55).ConclusionsWhen crucial baseline confounders are strictly controlled, longer UVC indwelling duration does not emerge as an independent risk factor for thrombosis. Clinical management should prioritize correct catheter tip positioning and standardized maintenance protocols rather than adhering to rigid, arbitrary removal timelines. Additionally, D-Dimer levels lack reliability as a standalone predictive biomarker in this specific neonatal population.Keywords: Umbilical venous catheter, Neonatal thrombosis, Indwelling duration, D-Dimer, Neonatal intensive care unit Introduction Umbilical venous catheterization (UVC) is essential for providing parenteral nutrition, medications, and fluid resuscitation to critically ill and premature neonates in the neonatal intensive care unit (NICU). Despite its indispensable clinical utility, UVC placement carries a significant risk of complications, most notably catheter-related thrombosis (CRT). The reported incidence of UVC-associated thrombosis varies widely, ranging from less than 4% in symptom-based retrospective cohorts [1, 2] to as high as 19%–75% in prospective studies utilizing routine serial ultrasound screening [3, 4]. Clinically, CRT is a major concern as it can precipitate severe long-term sequelae, including portal hypertension, hepatic lobar atrophy, and systemic embolization. Current clinical guidelines and consensus statements present conflicting recommendations regarding the optimal dwell time for UVCs. Traditional protocols often recommend early, routine removal (e.g., within 7–10 days) to mitigate the risks of central line-associated bloodstream infections and thrombosis [5–7]. However, recent high-quality evidence challenges these rigid duration limits. For instance, a randomized controlled trial by Hess et al. showed that extending UVC dwell time up to 14 days in very low birth weight infants significantly reduced the need for painful procedures without increasing the overall incidence of catheter-related complications [8]. This aligns with a Cochrane review indicating insufficient evidence to support routine early removal [9], as well as observational data confirming the relative safety of prolonged UVC use [10]. Meanwhile, the clinical utility of biomarkers such as D-Dimer in predicting neonatal thrombosis remains highly debated, primarily due to the physiological variability inherent in the developing neonatal coagulation system [11, 12]. To address these ongoing clinical uncertainties, we conducted a rigorous matched case-control study. By strictly matching for gestational age and birth weight—two major confounders intrinsically linked to the immaturity of neonatal developmental hemostasis [7, 11]—we aimed to isolate and objectively evaluate the independent impact of UVC indwelling duration and laboratory biomarkers on thrombus formation in a high-risk NICU population. Methods Study design and setting This retrospective, matched case-control study was conducted in the Neonatal Intensive Care Unit (NICU) of Liuzhou Maternity and Child Healthcare Hospital, a tertiary referral center in Guangxi, China, between January 2022 and December 2024. The study protocol was approved by the Institutional Ethics Committee of Liuzhou Maternity and Child Healthcare Hospital (Approval No. Quick-Review-Scientific-Research-2022-032). The requirement for informed maternal/parental consent was waived due to the retrospective nature of the data collection. To minimize potential confounding factors related to developmental hemostasis, we employed a rigorous 1:2 matched case-control design, strictly matching controls to cases based on gestational age (GA) and birth weight (BW). The detailed patient recruitment and selection process is illustrated in Fig. 1. The NICU is a 30-bed, Level III regional referral unit admitting approximately 535 neonates annually, providing comprehensive medical care for neonates from 22 + 2 weeks of gestation onwards, with surgical care provided in collaboration with the pediatric surgical team. Flowchart of study population selection. From an initial cohort of 526 neonates with UVC placement (January 2022–December 2024), 82 were excluded based on predefined criteria: major congenital anomalies (n = 18), very short hospitalization < 72 hours (n = 18), and incomplete laboratory or serial tip position records (n = 46), leaving 444 eligible neonates. Of 36 neonates with UVC-related thrombosis, 2 were excluded (thrombosis detected early with withdrawal of care within 72 hours), yielding 34 cases. From 408 eligible non-thrombotic neonates, 68 were selected as matched controls by gestational age and birth weight at a 1:2 ratio, while 340 were not selected. The final analytical cohort comprised 34 cases and 68 controls (n = 102) Participants and eligibility criteria Neonates admitted to the NICU who required UVC placement during the study period were screened for inclusion. To ensure the internal validity of the thrombosis assessment, the following predefined exclusion criteria were applied: (1) Major congenital malformations: Infants presenting with abdominal wall defects (e.g., gastroschisis, omphalocele) or complex congenital heart diseases were excluded. These conditions may inherently alter venous hemodynamics or catheter placement trajectories, thus acting as independent confounders for thrombogenesis [8, 13]. (1) Major congenital malformations: Infants presenting with abdominal wall defects (e.g., gastroschisis, omphalocele) or complex congenital heart diseases were excluded. These conditions may inherently alter venous hemodynamics or catheter placement trajectories, thus acting as independent confounders for thrombogenesis [8, 13]. (2) Very short hospitalization (< 72 h): Neonates who were discharged or whose families withdrew care within 72 h of admission were excluded, as such short durations were insufficient for adequate catheter-related outcome assessment and typically reflected early withdrawal of care or acute transfer, introducing competing risk bias. (2) Very short hospitalization (< 72 h): Neonates who were discharged or whose families withdrew care within 72 h of admission were excluded, as such short durations were insufficient for adequate catheter-related outcome assessment and typically reflected early withdrawal of care or acute transfer, introducing competing risk bias. (3) Incomplete records: Neonates with substantial missing laboratory data or incomplete serial tip position follow-up documentation were excluded (n = 46). (3) Incomplete records: Neonates with substantial missing laboratory data or incomplete serial tip position follow-up documentation were excluded (n = 46). Catheter management and tip positioning UVC insertion was performed by trained neonatologists or specialized nurse practitioners under strict sterile conditions. The target tip position was the junction of the inferior vena cava and the right atrium (IVC-RA). In accordance with established international guidelines and previous protocols, the optimal tip location was radiographically or sonographically confirmed at the T8–T9 vertebral level above the diaphragm [6, 14]. To maintain catheter patency and prevent occlusion, a continuous infusion of heparinized saline (0.5–1.0 IU/mL) was administered throughout the entire indwelling period [15, 16]. In our unit, a standardized protocol for serial UVC tip position monitoring was implemented. Tip position was confirmed at initial placement and was subsequently reassessed by ultrasound on days 1, 3, 7, 10, and 14 of catheterization, together with systematic evaluation of the heart, liver, upper and lower extremity veins, and documentation of catheter tip position and any catheter-related thrombus, including its location and size. For neonates who developed thrombosis, catheter tip position was documented at the time of thrombus detection. All UVCs used during the study period were single-lumen, 3.5 Fr catheters. Outcome definition: UVC-related thrombosis The primary outcome of this study was UVC-related thrombosis, objectively confirmed via Doppler ultrasound. Based on established diagnostic criteria, a confirmed thrombus was defined as an echogenic mass or lesion within the vessel lumen (including the umbilical vein, ductus venosus, portal vein, or IVC) or the right atrium, frequently accompanied by partial or complete occlusion of regional blood flow [4, 17, 18]. Ultrasound examinations were conducted either as part of routine screening protocols for high-risk neonates or prompted by clinical suspicion of catheter-related complications, ensuring the capture of both asymptomatic and symptomatic thrombi. High-risk neonates prioritized for screening were defined by one or more of the following: extreme prematurity or very low birth weight, suspected catheter malposition, catheter dysfunction, unexplained abdominal distension or hepatomegaly, abnormal liver function, infection or sepsis, or coagulopathy. Doppler ultrasound served as the primary modality for thrombus screening and diagnosis, performed on days 1, 3, 7, 10, and 14 of catheterization and whenever complications were clinically suspected, whereas chest X-ray was used only for initial tip position confirmation. Both cases and controls were drawn from this same high-risk population managed under an identical institutional protocol, which was therefore applied uniformly across both groups. Data collection and statistical analysis Maternal and neonatal baseline characteristics, catheter-related variables (specifically indwelling duration), and laboratory biomarkers (including platelet counts and D-Dimer levels) were systematically extracted from the hospital’s electronic medical records. Continuous variables were initially assessed for normality of distribution. Depending on the distribution pattern, these variables were compared between the case and control groups using either the independent Student’s t-test or the Mann-Whitney U test. To evaluate the predictive utility of the identified risk factors, Receiver Operating Characteristic (ROC) curve analysis was performed. The Area Under the Curve (AUC) was calculated to quantify the diagnostic performance of D-Dimer levels and catheter dwell time in discriminating between the thrombotic and non-thrombotic cohorts, a methodological approach strongly supported by recent literature on venous thrombosis research [19, 20]. A two-sided P-value of < 0.05 was considered to indicate statistical significance. All statistical analyses were conducted to rigorously test the hypothesis regarding independent risk factors. Results Baseline characteristics of the study population A total of 526 neonates who underwent UVC placement between January 2022 and December 2024 were initially screened. After applying the predefined exclusion criteria, 102 neonates were enrolled in this matched case-control study, comprising 34 infants in the thrombosis group (overall incidence: 6.46%) and 68 matched controls (1:2 ratio). As detailed in Table 1, the strict matching process successfully achieved a well-balanced distribution of key demographic variables. No statistically significant differences were observed between the two cohorts regarding gestational age (31.8 ± 4.2 vs. 31.8 ± 4.1 weeks, P = 0.988) or birth weight (1621.5 ± 879.5 vs. 1625.7 ± 756.4 g, P = 0.980). Furthermore, gender distribution, delivery mode, and the prevalence of major comorbidities—including neonatal asphyxia and infection—remained comparable across both groups (all P > 0.05). Among the 34 thrombosis cases, 8 (23.5%) were symptomatic (presenting with abdominal distension, or fever with tachycardia on the day of detection), while 26 (76.5%) were detected through routine high-risk ultrasound screening in the absence of overt symptoms. The median time from UVC insertion to thrombus detection was 7.7 ± 3.2 days. Table 1Demographic and clinical characteristics of the matched study populationCharacteristicThrombosis (n = 34)Control (n = 68)P-valueGestational Age (weeks), mean ± SD31.8 ± 4.231.8 ± 4.10.988Birth Weight (g), mean ± SD1621.5 ± 879.51625.7 ± 756.40.980Male Gender, n (%)21 (61.8%)43 (63.2%)0.887UVC Duration (days), median (IQR)8.0 (5.2–10.0)8.0 (7.0–11.0)0.206Infection, n (%)31 (91.2%)58 (85.3%)0.395Confirmed early-onset sepsis, n (%)7(20.6%)7(10.3%)0.221D-Dimer (mg/L), median (IQR)3.5 (2.6–10.0)3.4 (2.1–7.4)0.808Data are presented as mean ± SD or median (IQR)EOS early-onset sepsis, IQR interquartile rangeP-values for categorical variables were calculated using the chi-square or Fisher’s exact test Demographic and clinical characteristics of the matched study population Data are presented as mean ± SD or median (IQR) EOS early-onset sepsis, IQR interquartile range P-values for categorical variables were calculated using the chi-square or Fisher’s exact test Catheter-related and laboratory variables The association between UVC indwelling duration and thrombus formation was rigorously evaluated. The median indwelling duration was 8.0 days (interquartile range [IQR]: 5.2–10.0) in the thrombosis group, compared to 8.0 days (IQR: 7.0–11.0) in the control group. As illustrated by the violin plot (Fig. 2A), the distribution of indwelling time was highly consistent between the cohorts, demonstrating no statistically significant difference (P = 0.206). Distribution of risk factors. A Violin plot of UVC indwelling duration. The width represents data density; the median duration was 8.0 days in both groups (P > 0.05). B Box plot of D-Dimer levels. The thrombosis group showed higher variability and extreme outliers, but the median difference was not significant Regarding laboratory biomarkers, platelet counts at birth were comparable between the groups (P = 0.524). Although D-Dimer levels exhibited higher variability and more pronounced extreme outliers in the thrombotic cases (Fig. 2B), the difference in median levels did not reach statistical significance (3.5 mg/L vs. 3.4 mg/L, P = 0.808). Predictive utility of indwelling duration and D-Dimer levels Receiver Operating Characteristic (ROC) curve analysis was performed to evaluate the clinical predictive utility of these parameters (Fig. 3). The Area Under the Curve (AUC) for UVC indwelling duration was 0.42 (95% CI: 0.30–0.54), and the AUC for D-Dimer levels was 0.55 (95% CI: 0.40–0.69). These findings indicate that neither longer indwelling duration nor elevated D-Dimer levels possess sufficient sensitivity or specificity to serve as reliable, independent predictors for UVC-associated thrombosis in this strictly matched neonatal cohort. Receiver Operating Characteristic (ROC) curves for prediction of thrombosis. ROC analysis was performed to evaluate the predictive value of UVC indwelling duration (AUC = 0.42) and D-Dimer levels (AUC = 0.55). Both parameters demonstrated low sensitivity and specificity as standalone predictors for catheter-related thrombosis in this cohort Discussion Principal findings In this strictly matched case-control study, we sought to elucidate the specific risk factors for UVC-associated thrombosis in a high-risk neonatal cohort. Our primary finding demonstrates that, when key baseline confounders such as gestational age and birth weight are rigorously controlled, longer UVC indwelling duration (median 8.0 days) is not significantly associated with an elevated risk of thrombosis. Furthermore, although D-Dimer levels exhibited broader variance among thrombotic cases, they demonstrated inadequate predictive value as a standalone screening modality (AUC = 0.55). Interpretation of indwelling duration The association between UVC indwelling duration and the risk of thrombosis remains a subject of ongoing debate. Several historical observational studies have proposed a time-dependent paradigm, postulating that the risk of thrombosis increases linearly with prolonged catheterization, particularly beyond 5–7 days [10, 18]. However, critical appraisal of these earlier reports reveals notable methodological limitations, primarily the absence of rigorous adjustment for pivotal baseline confounders, specifically gestational age (GA) and birth weight (BW) [7]. As highlighted by recent cohort data, infants requiring prolonged central venous access are invariably younger, smaller, and more critically ill, creating an inherent baseline imbalance [1]. Furthermore, as explicitly acknowledged by Dubbink-Verheij et al. (2020), identifying appropriately matched, uncatheterized controls within extremely preterm cohorts is inherently challenging, rendering unmatched studies highly susceptible to confounding bias [4]. Consequently, the putative “time-risk” correlation observed in previous unmatched cohorts may predominantly reflect the physiological vulnerability and inherent prothrombotic state of severe prematurity (e.g., developmental deficiencies in natural anticoagulants) rather than an independent, direct effect of catheter dwell time. By employing a strict 1:2 matching strategy for gestational age and birth weight, our study effectively controlled for these critical developmental confounders. In this matched cohort, our analysis revealed no significant difference in indwelling duration between the thrombosis and control groups (P = 0.206). This finding aligns well with recent high-quality evidence. Notably, a randomized controlled trial by Hess et al. showed that extending UVC dwell time up to 14 days in very low birth weight infants did not significantly increase the overall incidence of catheter-related complications compared to early removal strategies [8]. Similarly, Hwang et al. reported that, after adjusting for confounding variables, indwelling duration was not an independent risk factor for asymptomatic UVC-related thrombosis [3]. These findings suggest that the quality of catheter management is more critical than the duration of placement. Recent literature emphasizes that suboptimal catheter tip positioning (e.g., non-central or intrahepatic placement) is the primary risk factor for mechanical and thrombotic complications [6, 17]. Implementing standardized care, particularly using real-time ultrasound to verify optimal tip positioning, effectively reduces thrombosis risk regardless of dwell time [21, 22]. Therefore, in NICUs with rigorous maintenance protocols, the decision to remove a functioning UVC should be based on ongoing clinical necessity and correct tip placement, rather than rigid time limits. Diagnostic value of D-Dimer Although D-Dimer serves as a classic biomarker for fibrinolysis, its predictive value in our cohort was remarkably low (AUC = 0.55). This limited utility is intrinsically tied to the unique physiological characteristics of this population, often conceptualized as “developmental hemostasis.” The physiological activation of the coagulation cascade at birth naturally elevates baseline D-Dimer levels, making adult-derived diagnostic thresholds unreliable for neonates [11]. Compounding this issue is the high prevalence of concurrent clinical morbidities. While a broad clinical diagnosis of infection was documented in the majority of our cohort (91.2% of cases vs. 85.3% of controls, P = 0.395), it is important to note that no cases met the criteria for central line-associated bloodstream infection (CLABSI), and most positive blood cultures (e.g., coagulase-negative staphylococci) were consistent with contamination. When restricting the analysis to confirmed early-onset sepsis, the proportion was numerically higher in the thrombosis group (20.6% vs. 10.3%), although this did not reach statistical significance (P = 0.221), likely reflecting the limited sample size. This observation nonetheless suggests that genuine systemic infection, rather than the broad clinical infection label, may contribute to the inflammatory–coagulation dysregulation underlying thrombosis. As Solgun et al. noted, septicemia triggers robust, non-specific D-Dimer elevations through systemic inflammatory and fibrinolytic pathways [12]. Consequently, the substantial overlap in D-Dimer levels between groups confirms its inadequacy as a standalone screening tool, and underscores that D-Dimer interpretation must account for the underlying infectious and inflammatory context in this NICU population. Strengths and limitations The principal strength of this study is the rigorous 1:2 matched case-control design. Prematurity and low birth weight are well-established independent risk factors for neonatal thrombosis [7, 16], primarily driven by the developmental immaturity of the neonatal coagulation system (e.g., physiological deficiencies in natural anticoagulants such as antithrombin) [11]. By matching for these critical variables—an approach validated by previous high-quality research [23]—we effectively isolated the specific impact of catheter-related factors, thereby reinforcing the reliability of our findings. We acknowledge several limitations, however. First, the retrospective, single-center nature of the study inherently restricts the sample size and limits broader generalizability. Second, because routine serial ultrasound was not universally performed on all asymptomatic controls, there is a potential for missed subclinical thrombi, although our unit maintains a low clinical threshold for targeted screening. Third, although serial tip position monitoring was performed, the retrospective design and the relatively low frequency of confirmed tip migration in our cohort limited the statistical power to formally evaluate tip migration as an independent risk factor for thrombosis. Larger prospective studies are warranted to specifically quantify this association. Fourth, infection was recorded as any clinically diagnosed infection during the entire hospitalization rather than being restricted to the UVC indwelling period, and no cases met the criteria for central line-associated bloodstream infection (CLABSI). The positive blood cultures (e.g., coagulase-negative staphylococci) were most consistent with contamination. Therefore, the high infection prevalence largely reflects coexisting neonatal morbidities (e.g., pneumonia, sepsis, meconium aspiration) rather than catheter-associated bloodstream infection. Future studies should restrict infection ascertainment to the catheter indwelling period and incorporate inflammatory biomarkers to more precisely delineate the role of systemic infection in catheter-related thrombosis. Conclusions Our findings suggest that longer UVC indwelling duration is not an independent risk factor for thrombosis when developmental baseline confounders are strictly controlled. Therefore, the clinical decision to remove a functioning UVC should be driven by ongoing patient needs, standardized maintenance, and correct tip positioning, rather than adherence to arbitrary time limits. Furthermore, D-Dimer levels offer limited diagnostic utility and should be interpreted with extreme caution in this population. Acknowledgements We would like to thank the Liuzhou Key Laboratory of Refine Management for Preterm Infants for providing the research platform and technical support that facilitated this study. We also thank the medical and nursing staff of the NICU at Liuzhou Maternity and Child Healthcare Hospital for their contributions to patient care and data maintenance. Clinical trial number Not applicable. This study was a retrospective observational analysis. Abbreviations Umbilical venous catheter Neonatal intensive care unit Catheter-related thrombosis Gestational age Birth weight Platelets C-reactive protein Receiver operating characteristic Area under the curve Authors’ contributions XH, BW, and SP contributed equally to this work; they conceptualized and designed the study, performed the statistical analysis, and drafted the initial manuscript. SZ, SW, XM, HT, and ZM contributed to patient recruitment, data collection, and clinical assessment. XH and JC jointly supervised the overall research process, acquired funding, and critically revised the manuscript for important intellectual content. All authors read and approved the final manuscript. Funding This study was supported by the Self-funded Scientific Research Project of Health Commission of Guangxi Zhuang Autonomous Region (Grant No. Z-B20221589); the Liuzhou Science and Technology Plan Project (Grant No. 2024YB0103A014); the Wu Jieping Medical Foundation Clinical Research Special Fund (Project No. 320.6750.2025-23-42). 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 conducted in accordance with the Declaration of Helsinki and all relevant institutional, national, and international guidelines and regulations. The study protocol was approved by the Institutional Ethics Committee of Liuzhou Maternity and Child Healthcare Hospital (Approval No. Quick-Review-Scientific-Research-2022-032). The requirement for informed consent from parents or legal guardians was waived by the ethics committee due to the retrospective nature of the study and the use of de-identified clinical data. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Xuemei Huang, Ba Wei and Shuangjing Pan contributed equally to this work and share first authorship. References Associated Data Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.