Risk factors for central venous catheter thrombotic complications in children and adolescents with cancer


  • S. Revel-Vilk, J. Yacobovich, G. Goldstein, S. Nemet, and G. Kenet were responsible for data collection. S. Revel-Vilk and O. Paltiel were responsible for statistical analysis and manuscript preparation. J. Yacobovich, H. Tamary, O. Paltiel, M. Weintraub, and G. Kenet contributed to the revision of the manuscript.

  • We thank the nurses and physicians who cared for these patients and their families.



The use of central venous catheters (CVCs) has greatly improved the quality of care in children with cancer, yet these catheters may cause serious infectious and thrombotic complications. The aim of this prospective registry study was to assess the host and CVC-related risk factors for CVC-created thrombotic complications.


Patients undergoing CVC insertion for chemotherapy were followed prospectively for CVC complications. At the time of enrollment, demographic, clinical, and CVC-related data, and family history of thrombosis were collected. Survival and Cox regression analyses were performed.


A total of 423 CVCs were inserted into 262 patients for a total of 76,540 catheter days. The incidence of CVC-related deep-vein thrombosis (DVT) was 0.13 per 1000 catheter-days (95% confidence interval [CI], 0.06-0.24). Insertion of peripherally inserted central catheters (PICCs) and insertion in an angiography suite significantly increased the risk of symptomatic CVC-related DVT. The incidence of CVC occlusion was 1.35 per 1000 catheter-days (95% CI, 1.1-1.63). Positive family history of thrombosis significantly increased the risk of CVC occlusion (hazard ratio [HR], 2.16; 95% CI, 1.2-3.8). The CVC-related risk factors were insertion of Hickman catheters, insertion in angiography suite, and proximal-tip location. Patients developing at least 1 episode of both CVC occlusion and infection had an increased risk for developing symptomatic CVC-related DVT (HR, 4.15; 95% CI, 1.2-14.4).


Both patient-related and CVC-related factors are associated with higher risk of symptomatic thrombotic complications. These risk factors could be used in the clinical setting and in developing future studies for CVC thromboprophylaxis. Cancer 2010. © 2010 American Cancer Society.

Children with cancer require intravenous administration of chemotherapy and other therapies for a considerable period of time. Repeated venipunctures are poorly tolerated by children and by their veins. The introduction of central venous catheters (CVCs) in the 1980s significantly improved the quality of life of pediatric oncology patients.1, 2 However, the use of these CVCs has been associated with mechanical, infectious, and thrombotic complications.

The rate of thrombotic complications related to CVCs in children treated with chemotherapy was previously reported in 2 cohort studies.3, 4 The cumulative risk (CR) of CVC-related deep-vein thrombosis (DVT) was 0.06 and 0.18 per 1000 catheter-days, and the CR of CVC occlusion was 0.78 and 2.0 per 1000 catheter-days. In children with acute lymphoblastic leukemia (ALL), high risk of symptomatic and asymptomatic CVC-related DVT was found in children treated with asparaginase and prednisone5, 6 and with insertion to the left side and to the subclavian vein.7 Although CVC-related thrombotic complications in pediatric cancer patients are not rare events, the clinical impact and risk factors for these complications are not clear. The aim of this study was to prospectively assess the host, underlying disease and CVC-related risk factors for symptomatic thrombotic complications in a large cohort of pediatric patients undergoing CVC insertion for chemotherapy.


Study design

A registry of patients undergoing CVC insertion for chemotherapy treatment was initiated in June 2006 in the 3 largest pediatric hematology/oncology centers in Israel. Each patient undergoing CVC insertion for treatment of a newly diagnosed cancer, relapsed cancer, or for bone marrow transplantation (BMT) was eligible for registration in this study, and the only exclusion criterion for this study was the lack of consent to participate. The study was approved by the local institution review board of each participating center.

Data collection


An initial registration form was completed for each participating patient. This form included questions on demographic, clinical, and CVC-related data, and family history of thrombosis. Family history of thrombosis was considered positive when there was a report of any first and/or second degree family member with DVT, pulmonary embolism, myocardial infarction and/or stroke at a young age (<40 years),8 recurrent abortions (≥3) and/or laboratory diagnosis of inherited or acquired thrombophilia. Patients were followed from the day of enrollment until removal of the last CVC, death, last follow-up, or at study closure, when none of the above occurred.


For each CVC insertion, the following data were collected: date of insertion, type of CVC, place of procedure, side of insertion, vein cannulated, and location of the tip of the catheter. Decisions related to CVC insertion were made by the treating physician and were based mainly on personal preferences and underlying patient characteristics. The insertion and maintenance of CVCs was in accordance with institutional protocols. Routine thromboprophylaxis was not used in the participating centers. If the CVC was removed during the study period, the date and cause of removal were recorded. When a new CVC was inserted during the study period, the new CVC-related data were recorded.

For documentation of CVC-related complications, the following events were reported to the registry: (1) occlusion of the CVC, defined as an inability to infuse and/or withdraw blood, requiring medical, ie, local installation of thrombolysis therapy, or surgical intervention, ie, CVC removal; (2) venous thrombosis, suspected by the primary physician and confirmed by diagnostic imaging; (3) infection, including all cases of positive blood culture from the CVC drawn in case of fever or clinical sign suspected for infection. Concurrent peripheral blood culture was not routinely drawn. For coagulase-negative staphylococci, a common skin contaminant, only cases where 2 or more blood cultures were positive and drawn on separate occasions were considered true infection.9

Tests for inherited thrombophilia

Protein C and antithrombin activities were measured by chromogenic assays (Baxter Dade, Bonnstrasse, Switzerland). Free protein-S antigen was measured by enzyme-linked immunosorbent assay (ELISA) (Gradipore Elisa test kit; Riverside Corporate Park, Australia). Only repeatedly low-for-age values established the diagnosis of a coagulation inhibitor deficiency. Factor V Leiden, C677T gene polymorphism in the MTHFR and G20210A substitution in the factor II gene were detected as previously described.10-13

Data analysis

Demographic and clinical characteristics of the study cohort are reported with 95% confidence intervals. The association between the type of CVC and other variables in the analysis was analyzed by using the ANOVA test for continuous data, ie, age, and chi-square test for categorical data, ie, underlying disease, place of procedure, side of insertion, vein cannulated, and place of the tip. As the age variable was found to be skewed to the right, a z = square root of y transformation gave an approximately normal distribution. Groups of underlying diagnosis were predefined including acute lymphoblastic leukemia (ALL), myeloid leukemia (acute and juvenile myelomonocytic leukemia), brain tumors, sarcoma (Ewing sarcoma, osteosarcoma, rhabdomyosarcoma), neuroblastoma, lymphoma (Hodgkin and non-Hodgkin), bone-marrow transplantation (BMT), and other (occurring in ≤5 children). The grouping was based on similar treatment modalities for similar types of cancers.

Analysis of the risks of CVC-related thrombotic complications was performed separately for host-related and CVC-related factors. First, the risk (with 95% CI) of a patient to develop at least 1 episode of the outcome was calculated, the numerator being all patients who had at least 1 episode of outcome and the denominator being patient-days at risk. Patient-days were the sum of follow-up from time of enrollment to time of removal of the last CVC, death, last follow-up, or study closure when none of the above occurred. Second, the risk of each CVC (with 95% CI) to develop the outcome was calculated. The numerator was all CVCs that had at least 1 episode of the outcome, and the dominator was catheter-days at risk. Catheter-days were the sum of follow-up from time of CVC insertion to the CVC removal, death, last follow-up, or study closure when none of the above occurred. For both analyses, a Kaplan-Meier curve was computed, and the 1-year CR to develop the outcome was established. The log-rank test was used to compare survival by categories of study variables. Finally, a Cox regression model was computed to quantify the independent contribution of 1 or more factors of interest on survival, expressed as the hazard ratio (HR). For host-related factors, the following variables were considered: age, underlying diagnosis, family origin, family history of thrombosis and thrombophilia. For CVC-related factors, the following variables were considered: type of CVC, place of procedure, side of insertion, vein cannulated, and location of the tip. To avoid multicolinearity, a univariate analysis for all other CVC-related variables was computed separately for each type of CVC. A separate model was computed for the association between CVC-related DVT, occlusion, and infection.

The study was funded for 2 years by a grant from the Israeli Cancer Association. In these 2 years, we planned to recruit 300 patients undergoing CVC insertion for chemotherapy. Statistical analysis were performed by using the SPSS 14.0 for Windows (release 14.0.1, November 2005; SPSS, Chicago, Ill) and the WINPEPI (PEPI-for-Windows, version 2.8, March 2007; PEPI, Programs for EPIdemiologists; freeware for Windows, developed for DOS computers by Joseph H Abramson, School of Public Health and Community Medicine, Hebrew University, Jerusalem, Israel).14 Missing data were not included in the analysis. Taking into account the multiple comparisons in this study, only a 2-tailed P value of <.01 was considered statistically significant. Bonferroni post hoc correction was performed for the subgroup analysis of study subjects who had undergone thrombophilia testing.



A total of 262 patients were enrolled to the registry with a median age of 7.42 years (range, 28 days to 28 years). Eighteen young adult patients (older than 18 years of age) were included because they were treated in a pediatric oncology center for pediatric types of tumors, mainly sarcoma. Patients were followed for a total of 77,798 patient-days. The main patient characteristics are presented in Table 1. Questions about the family history of thrombosis were completed in 234 (89%) patients and were positive for thrombosis in 24 (10.3%) patients (95% CI, 6.7%-14.9%).

Table 1. Clinical Characteristics of the Entire Study Cohort
CharacteristicsNo. (%)
 Total no.262
 Age, y, median [range]7.4 [0.08-28.3]
 Jewish origin164 (62.6)
 Positive Family history of thrombosis24 (10.3)
Underlying disease 
 Acute lymphoblastic leukemia73 (27.9)
 Sarcoma48 (18.3)
 Lymphoma36 (13.7)
 Myeloid leukemia32 (12.2)
 Brain tumor25 (9.5)
 Neuroblastoma15 (5.7)
 Bone marrow transplantation6 (2.3)
 Other diagnosis27 (10.3)
Central venous catheters (CVCs) 
 Total no.423
 1 CVC per patient162 (61.6)
 2 CVCs per patient64 (24.7)
 ≥3 CVCs per patient36 (13.7)
Duration of CVC/pt, median {95% CI}, mo4.8 {3.9-5.6}
 Peripherally inserted central catheters3.1 {2.2-4.0}
 Hickman catheters4.1 {2.7-5.6}
 Port-a-Caths9.8 {8.1-11.5}

CVC type and insertion

The types of CVC inserted were Hickman catheters in 104 (24%), Port-a-Caths in 126 (30%), and peripherally inserted central catheters (PICCs) in 188 (45%). In 5 cases, the type of CVC was not recorded. The CVC duration ranged from 1 day to 744 days for a total of 76,540 catheter-days. The number and median duration of CVC per patient is presented in Table 1. Age, underlying disease, and insertion data were significantly associated with the type of CVC (Table 2).

Table 2. Association Between the Type of Central Venous Catheter and Variables in the Analysis
 HickmansPort-a-CathsPICCsPallPHickmans vs Ports
No. (%)No. (%)No. (%)
  1. PICCs indicate periphery inserted central catheters; NA, not applicable.

Total104 (25%)126 (30%)188 (45%)
Age at time of insertion, median [range], years3.9 [0.08-18.1]5.45 [0.48-28.3]8.95 [0.31-28.3]<.0001<.0001
Underlying disease     
 Acute lymphoblastic leukemia20 (19%)47 (37%)54 (29%).01.002
 Sarcoma10 (10%)26 (21%)28 (15%).07.02
 Lymphoma8 (8%)11 (9%)33 (17%).02.78
 Myeloid leukemia33 (32%)6 (5%)24 (13%)<.0001<.0001
 Brain tumor5 (5%)12 (10%)25 (13%).05.17
 Neuroblastoma9 (9%)8 (6%)10 (5.5%).54.51
 Bone marrow transplantation10 (10%)1 (1%)5 (3%).001.002
 Other10 (10%)16 (13%)9 (5%).04.46
Place of procedure     
 Operation room103 (99%)116 (93.5%)28 (15%)<.0001.034
 Angiography suite1 (1%)8 (6.5%)159 (85%)  
Side of insertion, right50 (60%)54 (50%)96 (57%).5.19
Vein cannulated     
 Internal jugular vein26 (31%)24 (22%)NA.19
 External jugular vein35 (41%)40 (37%)   
 Subclavian vein24 (28%)43 (40%)   
Place of tip     
 Right atrium46 (69%)59 (74%)43 (25%)<.0001.168
 Superior vena cava18 (27%)13 (16%)2 (1%)  
 Right atrium/superior vena cava3 (4.5%)8 (10%)128 (74%)  

CVC occlusion

Per patient

Ninety-one patients had at least 1 episode of occlusion with an incidence of 1.17 (95% CI, 0.94-1.44) per 1000 patient-days. The 1-year CR for the first occlusion per patient from enrollment was 38%. Family history of thrombosis was significantly associated with at least 1 episode of occlusion (Fig. 1). The HR (95% CI) for positive family history of thrombosis was 2.16 (1.21-3.84;) P = .007). Other host-related variables, such as age, underlying diagnosis, and family origin were not significant.

Figure 1.

Cumulative risk of first central venous catheter occlusion for patients with positive family history of thrombosis compared with patients with negative family history of thrombosis.


Occlusion of the CVC occurred in 102 CVCs with an incidence of 1.35 (95% CI, 1.1-1.63) per 1000 catheter-days. The 1-year CR for first occlusion per catheter from insertion was 37%. Obstruction of the CVC was the cause for its removal in 21 cases. The risk of at least 1 occlusion was associated with the type of CVC with a 1-year CR of 54%, 40%, and 18% for Hickman catheters, Port-a-Caths, and PICCs, respectively (P = .002). The HR (95% CI) for occlusion of Port-a-Caths and Hickman catheters compared with PICCs was 1.8 (1.11-2.96) and 2.3 (1.38-3.82), respectively. A univariate analysis of the CVC-related risk factors for occlusion was performed for each type of CVC separately. Insertion in the angiography suite compared with the operating room increased the risk of occlusion in Port-a-Caths and Hickman catheters with an HR (95% CI) of 3.87 (1.62-9.24) and 51.17 (4.64-564.3), respectively. Tip of the CVC in the superior vena cava (SVC) compared with the right atrium (RA)/RA-SVC junction increased the risk of occlusion only in Port-a-Caths with an HR (95% CI) of 2.7 (1.23-5.93). Other risk factors were not associated with CVC occlusion.

Venous thrombosis

Fourteen events of venous thrombosis occurred in 13 children during the study period (Table 3). CVC-related DVT occurred in 10 patients with an incidence of 0.13 (95% CI, 0.06-0.24) per 1000 patient-days. The 1-year CR for CVC-related DVT per patient was 4.6%. Age, underlying diagnosis, family origin, and family history of thrombosis were not associated with CVC-related DVT. A higher rate of CVC-related DVT was found in PICCs compared with Hickman catheters and Port-a-Caths with a 1-year CR of 5.3% for PICCs and 1.9% for other CVC types (Fig. 2). The risk of CVC-related DVT was also associated with place of procedure with 1-year CR of 11.3% and 0.5% for the angiography suite and operating room, respectively (Table 4). Other CVC-related variables were not significant (Table 4). As the type of the CVC and place of procedure were found to be highly correlated (Table 2); multivariate analysis was not performed.

Figure 2.

Cumulative risk of central venous catheter -related deep vein thrombosis for peripherally inserted central catheters (PICCs) compared with other central venous catheters, ie, Hickman catheters and Port-a-Caths.

Table 3. Clinical Characteristics and Management of Patients Diagnosed With Venous Thrombotic Events
No.Age, yType of CancerCVC InvolvedPresentationVein InvolvedCVC RemovedOther ManagementInherited Thrombophilia
  • CVC indicates central venous catheter; ALL, acute lymphoblastic leukemia; PICC, peripherally inserted central catheters; Lt, left; RA, right atrium; FVL, heterozygote Factor-V Leiden; FII, heterozygote FII G20210A; NBL, neuroblastoma; Rt, right; MTHFR, homozygote MTHFR; NHL, non-Hodgkin lymphoma; PE, pulmonary embolism; RT, retinoblastoma; PNET, peripheral neuro-ectodermal tumor; ES, Ewing sarcoma; CT, computer tomography.

  • a

    Recurrent event in the same child on anticoagulation prophylaxis.

12.73ALLPICCEvaluation for pain/swellingLt basilicNoNoNegative
23.08ALLPICCEvaluation for feverRANoAnticoagulationFVL, FII
34.11NBLPICCEvaluation for pain/swellingRt basilicYesNoMTHFR
44.70ALLPICCEvaluation for pain/swellingRt brachial, Rt axillaryNoAnticoagulationNegative
55.36NHLEvaluation for pain/swellingPE, bilateral femoralAnticoagulationFII
66.13RTHickmanEvaluation for obstructionLt inominateNoAnticoagulationNegative
76.44PNETPICCEvaluation for pain/swellingRt axillaryNoAnticoagulationNegative
86.99ESPICCRoutine echocardiographyRANoAnticoagulationNegative
913.28ALLPICCEvaluation for pain/swellingRt brachial, RAYesAnticoagulationNegative
1013.28aALLEvaluation for painRt sagittal sinusAnticoagulationNegative
1115.61ALLPortEvaluation for pain/swellingPENoAnticoagulationNegative
1216.22Brain tumorPICCEvaluation for pain/swellingRt brachialYes (infection)AnticoagulationNegative
1317.32GerminomaEvaluation for pain/swellingPE, Rt femoralUmbrella for 2 wkNot available
1417.63NHLCT at diagnosisLt internal jugularAnticoagulationFVL
Table 4. Univariate Analysis of Central Venous Catheter-Related Risk Factors for Venous Thrombosis
Related Risk FactorsHR (95.0% CI)P
  1. PICCs indicate peripherally inserted central catheters; RA, right atrium; SCV, superior vena cava.

Type of catheter 0.006
 Not PICCs1 (ref) 
 PICCs7.03 (1.46-34.12) 
Place of procedure <.0001
 Operation room1 (ref) 
 Angiography suite16.25 (2.04-129.4) 
Side of insertion 0.36
 Left side1 (ref) 
 Right side0.56 (0.16-1.98) 
Vein cannulated 0.13
 Basilic vein1 (ref) 
 Internal jugular vein0.37 (0.04-3.01) 
 External jugular vein0.00 (0.00-1.0E+27) 
 Subclavian vein0.21 (0.25-2.14) 
Place of tip 0.53
 RA/RA and SVC junction1 (ref) 
 SVC1.64 (0.35-7.82) 

Of 10 patients with CVC-related DVT, 6 had at least 1 episode of occlusion, and 8 had at least 1 episode of infection. Patients developing at least 1 episode of both occlusion and infection had a trend for an increased rate for developing CVC-related DVT (Fig. 3). The HR (95% CI) was 4.15 (1.2-14.37) for patients with both infection and occlusion compared with patients with 1 type or no previous event (P = .039).

Figure 3.

Cumulative risk of central venous catheter -related deep vein thrombosis for patients with both occlusion and infection compared with those with a single type of complication and those with no complications.

Thrombophilia studies

Because of financial constraints, blood test for thrombophilia were performed in a subgroup of 85 (32%) patients that included the majority of patients seen in 1 of the centers and patients with ALL and DVT from all centers. As such, thrombophilia screening was performed more often in patients with ALL (47% vs 28% with other malignancies), with DVT (85% vs 30.4% without), with at least 1 CVC occlusion (42% vs 28% without), and less often in patients with sarcoma (17% vs 37% with other malignancies). The median age, family origin, and positive family history of thrombosis were similar between those with and without thrombophilia screening.

A positive thrombophilia screen was found in 21 (25%) patients (95% CI, 16%-35%), more frequently in patients of Arab origin (43%) compared with patients of Jewish origin (13%) (P = .006). Positive thrombophilia was not associated with CVC-related DVT or CVC occlusion. As in the total study population, in the subgroup of children who had thrombophilia testing, patients with a family history of thrombosis were more likely to have at least 1 episode of occlusion, 100%, compared with patients with no family history of thrombosis, 37% (P = .01, after Bonferroni correction).

Scoring algorithm

Based on the results of our cohort, we attempted to develop a scoring algorithm for prediction of CVC occlusion (Table 5, part A). Of the 418 CVCs (5 other types of CVCs were excluded), 170 (41%) CVCs scored 0, 216 (51.5%) CVCs scored 1, and 32 (7.5%) scored 2. Final scores of 2 and 1 were associated with a higher risk for CVC occlusion (Table 5, part B).

Table 5. Scoring Algorithm for Predication of Central Venous Catheter Occlusion
A. VariableScore
  • PICC, Peripherally inserted central catheter; RA, right atrium; SVC, superior vena cava; CVC, central venous catheter; HR, hazard ratio; CI, confidence interval.

  • a

    Missing data for place of the tip and family history of thrombosis were scored as 0.

Type of central venous catheter 
 Hickman catheter2
Place of the tipa 
 Superior vena cava1
 RA and RA/SVC0
Family history of thrombosisa 
Calculated initial score 
B. Final Score Calculated Based on the Initial ScoreFinal Score
 For initial score of 3-42
 For initial score of 1-21
 For initial score of 00
Prediction (HR) of CVC Occlusion Based on the Final Score, P = .00195% CI
 1 (ref)


This study was designed to assess the host and CVC-related risk factors for the development of CVC-related thrombotic complications in patients treated with chemotherapy for cancer and BMT. We found that the risk of occlusion was strongly associated with a family history of thrombosis, type of CVC, place (eg, angio suite) of procedure, and placement of the tip. The risk of CVC-related DVT was also associated with type of CVC and place of procedure. We believe that the recognition of these risk factors is important for future prevention and risk reduction of CVC-related thrombotic complications.

The use of PICCs was associated with higher rates of CVC-related DVT and lower rates of occlusions compared with Port-a-Caths and Hickman catheters. A lower rate of occlusion with PICCs could be related to their shorter duration (Table 1) compared with Port-a-Caths and Hickman catheters. An association between PICCs and DVT was found in cancer patients15, 16 and acute-care surgical patients.17 Yet, in children without cancer, symptomatic and asymptomatic CVC-related DVT were not associated with the type of CVC.18 The difference could be related to the different study populations and the inclusion of asymptomatic cases. The increased rate of PICC-related DVT can be viewed according to 2 perspectives. The first assumes that the increased rate is real because the PICCs cause more endothelial damage due to a relatively high catheter:vein diameter ratio.19 Another possibility is that the PICCs were inserted in sicker patients, who had a higher risk for DVT. The second assumes that the increased rate in PICCs is due to diagnostic bias; upper-arm DVT is usually symptomatic, whereas DVTs in internal veins, ie, jugular or subclavian veins, are less likely to be symptomatic.6 Because we did not perform systematic screening for asymptomatic DVT, this question could not be assessed in our study. For clinical purposes, currently, only symptomatic CVC-related DVTs are treated with anticoagulant therapy,20 a major therapeutic issue in patients with chemotherapy-induced thrombocytopenia. Emerging evidence for higher risk for post-thrombotic syndrome after asymptomatic CVC-related DVT could change this recommendation.21 Meanwhile, we believe that because of the higher risk for symptomatic CVC-related DVT the use of PICCs should be carefully considered.

Traditionally, CVCs were inserted by a surgical service that used either a percutaneous puncture or a cut-down technique guided by anatomic landmarks. Increasingly, interventional radiology services are inserting CVCs.22, 23 It was recommended that an ultrasound, image-guided, percutaneous technique for radiologic insertion of CVC should be used to decrease thrombotic complications.24, 25 As an ultrasound, image-guided, percutaneous technique was not routinely used, and its use was not recorded, we cannot conclude that the lack of its use is the cause for increased rates of thrombotic complications in angiographically placed CVC.

The incidence of CVC occlusion was relatively high, especially in Hickman catheters and in Port-a-Caths. In our study and others, the location of the catheter tip in the superior vena cava was found to be a significant risk factor, especially in Port-a-Caths.26-28 Thus, when the CVC is inserted, a major effort should be made to ensure that the tip is not located in the superior vena cava. Family history of thrombosis was recently proven to be an independent risk factor for venous thrombosis.29 Similarly, in our study, a positive family history of thrombosis was associated with occlusion. No association was found with CVC-related DVT, possibly because of the lower rate of events. Although thrombophilia is an important risk factor for venous thrombosis in children,30 thrombophilia status seems to be less contributory for CVC-related thrombotic complications.31 In a subgroup analysis of patients tested for inherited thrombophilia, family history of thrombosis, and not the results of thrombophilia tests, was a predictor for CVC occlusion. Although not all in the cohort were included in the thrombophilia analysis, we opine that the results of this analysis are valid as similar rates of family history of thrombosis were found in those children with and without thrombophilia testing. Thus, in clinical practice, family history of thrombosis may be more useful for risk assessment than a thrombophilia test.29 Family history of thrombosis (definition slightly different from ours) was included in a scoring schema in a study of thromboprophylaxis for children with ALL.32

The clinical impact of CVC occlusion is somewhat controversial. Some suggest that CVC occlusion has little impact on CVC longevity because it is resolved in most cases by installation of local fibrinolytic therapy.3 Others have found that CVC occlusion is associated with infection and with a higher risk for CVC-related DVT.4, 33, 34 Similarly, we found an association between occlusion, infection, and later occurrence of CVC-related DVT. Currently, low dose (10-100 IU/mL) heparin flushes are used to prevent occlusions. A variety of additional regimens were studied in the attempt to prevent occlusions with different rates of success.35-40 Which regimen would be the best for reducing the rate of occlusions and infections and whether reducing the rate of these complications will result in fewer CVC-related DVTs still need to be studied in prospective studies. A scoring algorithm including data on the type of CVC, placement of the catheter tip, and family history of thrombosis was found to be predictive for CVC occlusion. Before physicians use this scoring algorithm to make clinical decisions or in clinical trials for prevention of occlusion, the score needs validation in other prospective cohorts.

Our study had some limitations. First, based on the Israeli Pediatric Cancer registry the enrollment rate was around 70%. The major reason for nonenrollment was lack of communication to the study coordinators that a new patient was to have a CVC insertion. All cases known to the study coordinators were enrolled to this study, and the lack of enrollment was not differential among cases according to study variables. Despite this flaw, because the age distribution and the frequency of the different underlying diseases represent a typical population treated in pediatric oncology centers, we believe that the external validity of our study is acceptable. Second, the ascertainment of study outcomes was based on daily reports given by the nurses. Although, we must assume that not all outcome events were captured, we believe that the number of unreported events is similar across the study population. Finally, the study captured only symptomatic CVC-related DVT, mostly due to clinical symptoms and some due to radiographic evaluation for fever, obstruction, or routine testing; thus, the rate of CVC-related DVT was probably an underestimate. As a consequence, some of the “negative” findings for CVC-related DVT could be due to lack of power given the low number of events (ie, type II error). The short-term and long-term effects of asymptomatic CVC-related DVT is still controversial21, 41, 42 and should be assessed in a future studies when the cancer survivors will be tested for signs and symptoms of post-thrombotic syndrome.

In conclusion, this study reports the incidence and risk factors for CVC-related thrombotic complications in a large, pediatric, oncology cohort. The association of CVC occlusion with positive family history of thrombosis and with CVC-related DVT supports the concept that CVC occlusion is, at least in part, a thrombotic event. The results of this study could help physicians, nurses, and families during the decision process involved in choosing the type of central venous line to be inserted and could be used in developing future studies for reducing the rate of CVC-related complications.


The study was performed with the support of an unrestricted grant from the Israel Cancer Association.