JPPT Clinical Investigation Characterizing the Risk Factors Associated With Venous Thromboembolism in Pediatric Patients After Central Venous Line Placement Sarah Wisecup, PharmD,1 Shannan Eades, PharmD,2 and Yuliya Turiy, MD2,3 1 Children’s Hospital of San Antonio—Christus Santa Rosa, San Antonio, Texas, 2Children’s Memorial Hermann Hospital, Houston, Texas, 3University of Texas Health Science Center, Houston, Texas
OBJECTIVES: With the apparent increase in venous thromboembolism noted in the pediatric population, it is important to define which children are at risk for clots and to determine optimal preventative therapy. The purpose of this study was to determine the risk factors for venous thromboembolism in pediatric patients with central venous line placement. METHODS: This was an observational, retrospective, case-control study. Control subjects were patients aged 0 to 18 years who had a central venous line placed. Case subjects had a central line and a radiographically confirmed diagnosis of venous thromboembolism. RESULTS: A total of 150 patients were included in the study. Presence of multiple comorbidities, particularly the presence of a congenital heart defect (34.7% case vs. 14.7% control; p < 0.005), was found to put pediatric patients at increased risk for thrombosis. Additionally, the administration of parenteral nutrition through the central line (34.7% case vs. 18.7% control; p = 0.03) and location of the line increased the risk for clot formation. CONCLUSIONS: With increased awareness of central venous line–related thromboembolism, measures should be taken to reduce the number and duration of central line placements, and further studies addressing the need for thromboprophylaxis should be conducted. INDEX TERMS: central venous lines, clot, pediatrics, venous thromboembolism J Pediatr Pharmacol Ther 2015;20(5):358–366
INTRODUCTION As pediatric survival rates continue to increase with the introduction of new procedures and greater use of invasive vascular devices, so does the incidence of venous thromboembolism (VTE) in children.1 A recent retrospective cohort study of hospitalized children showed that the incidence of VTE rose from 34 per 10,000 hospital admissions to 58 per 10,000 hospital admissions during the period 2001-2007, yielding a 70% increase in the reported rate of thrombus formation.2 While the development of VTE seems to be multifactorial, central venous lines (CVLs) appear to be a causal risk factor.3 Potential mechanisms for clot formation include flow disruption, vessel wall trauma, and endothelial damage caused by the CVL.4 In another study, more than 50% of children and 80% of infants with a CVL developed clots.1 358
Despite the availability of a published clinical guideline by the American College of Chest Physicians for the prevention and treatment of VTE in the pediatric population,5 there is still a lack of randomized controlled trials identifying risk factors and evaluating the utility and safety of thromboprophylaxis in a general pediatric population with CVL placement.6,7 With the apparent increase in VTE noted in the pediatric population, it is important to define which children are at risk for clots and to determine optimal preventative therapy. Prevention of clot formation and subsequent exposure to prophylactic doses of anticoagulants may decrease morbidity associated with postthrombotic syndrome and mortality associated with pulmonary embolism. With that objective in mind, the aim of this retrospective case-control study was to identify risk factors for developing VTE in pediatric patients with J Pediatr Pharmacol Ther 2015 Vol. 20 No. 5 • www.jppt.org
JPPT
Thromboembolism in Pediatric Patients With Central Lines
a CVL. The information gathered in this study will assist us in developing future protocols to provide prophylaxis against VTE in high-risk pediatric patients with CVL.
PATIENTS AND METHODS This study was submitted to and approved by the Institutional Review Board of Memorial Hermann Hospital/Children’s Memorial Hermann Hospital in the Texas Medical Center and The University of Texas Health Science Center in Houston, Texas. Data were collected via a retrospective chart review from an internal, searchable database that contains inpatient data (Cerner Millenium Database, Memorial Hermann, System Services, Houston, TX). Subjects The study population included pediatric patients (0 to 18 years of age) who were admitted to the pediatric intensive care unit, intermediate medical care unit, or general medical unit at Children’s Memorial Hermann Hospital between July 1, 2007 and January 31, 2012, and who had a CVL placed during hospital admission. Subjects with CVL placement during inpatient admission were identified by the following ICD-9 CM codes: 38.93, 38.95, 38.97, and 86.07. To best identify patients of interest for a more homogenous population of study patients, subjects were excluded if they were admitted to the neonatal intensive care unit, were pregnant, or had a CVL placed at an outside facility. Patients with cancer also were excluded, as these patients are not typically treated at our facility. Of note, patients with a history of a clotting disorder or previous VTE were included in the study. Since patients are not routinely screened for clots, only patients with symptoms such as limb swelling/discoloration, poor pulses, fever, or elevation in white blood cell count (not associated with infection) were assessed with upper and/or lower extremity ultrasonography to confirm presence of a VTE. Eligible patients were stratified by the presence or absence of a radiographically confirmed diagnosis of VTE into cases (presence of VTE) and controls (absence of VTE). Owing to resource limitations, 150 patient cases were reviewed. Seventy-five case subjects were selected for analysis by a random number generator and matched to 75 randomly selected control subjects. J Pediatr Pharmacol Ther 2015 Vol. 20 No. 5 • www.jppt.org
Data Collection and Outcomes Patient demographics including age, sex, weight, admitting diagnoses, and comorbid conditions were collected from electronic medical records. Information including central line location, type, size, number of central line lumens, and duration of central line placement was recorded, as well as medications administered through the line, and any history of prior central line placement at our facility or at an outside facility before current admission. For case patients, the type and location of thrombus, as well as the drug used to treat the thrombus, were documented. Data for each subject were collected from the time of admission to the time of discharge from the hospital. The primary endpoint of this study was to determine the risk factors for thrombus formation in pediatric patients after CVL placement. The secondary endpoint was to describe clinical characteristics and annual incidence rates of thrombotic events. Statistical Analyses A comparison of baseline characteristics between case and control subjects was carried out by multiple tests. The Student t test was used for continuous data following a normal distribution, and Wilcoxon rank sum test for continuous data following a non-normal distribution. Chi-squared test or Fisher exact test was used to compare categorical data between the 2 groups. Risk factors for VTE after CVL placement were identified by bivariate analysis, if p value was ≤0.12. Using risk factors identified through the bivariate analysis, a logistic regression model was used to identify risk factors with a p value ≤ 0.05. We estimated that a sample size of 150 patients (75 cases, 75 controls) was needed for multiple regression analysis, assuming α = 0.05, β = 0.8, and a moderate effect size of 0.15. All statistical analyses were conducted by The University of Texas Health Science Center (Houston, TX).
RESULTS Patient Demographics/Characteristics During the study period, a total of 2,388 central lines were placed. Of the potential study subjects, 125 were found to have a radiographically confirmed VTE during their admissions (overall incidence of 5.2% during the study period). Of 359
JPPT
S Wisecup, et al
Table 1. Patient Demographics and Characteristics
Age, mean ± SD, yr Male, % Weight, mean ± SD, kg
CVL With VTE (n = 75)
CVL Without VTE (n = 75)
p Value
4.4 ± 5.6
4.3 ± 4.6
0.94
63
56
0.41
21.9 ± 28.4
22.8 ± 24.5
0.82
29 (38.6)
29 (38.6)
1.000
Admitting diagnosis* Infection Bacteremia
10 (13.3)
6 (8)
1.000
Other
19 (25.3)
23 (30.7)
1.000
Respiratory
15 (20)
14 (18.6)
1.000
Cardiovascular Neurological Gastrointestinal
14 (18.6)
7 (9.4)
0.157
6 (8)
12 (16)
0.208
2 (2.7)
6 (8)
0.276
Renal
4 (5.4)
1 (1.4)
0.367
Hematologic
2 (2.7)
1 (1.4)
1.000
3 (4)
5 (6.6)
0.719
54 (72)
41 (55)
0.003
Other† Comorbidity present*
CVL, central venous line; VTE, venous thromboembolism * Values are No. (%) † Other: overdose, vascular access, trauma, endocrine disorder
the subjects with a CVL and a radiographically confirmed VTE, 75 were selected by a random number generator to be case subjects for analysis. Of the patients who had a CVL placed during the study period but who did not have a radiographically confirmed VTE, 75 were randomly selected to be included as control subjects for comparison. Patient demographics including age, sex, weight and admitting diagnoses were similar between the two groups (Table 1). The mean age of the case and control subjects was similar, 4.4 ± 5.6 years and 4.3 ± 4.6 years, respectively. The number of patients with comorbid conditions present at admission was significantly higher in the case group than the control group (72% vs. 55%; p = 0.003) (Table 2). The presence of a congenital heart defect was found to be an independent risk factor for the development of VTE (34.7% of cases vs. 14.7% of controls; p < 0.005), as subanalysis of case and control subjects with congenital heart defects did not detect differences in other factors between the groups that may have been responsible for the higher incidence of clot formation (e.g., mean age, duration of line placement, and mean total comorbidities). Eight patients in the control group had a history of prematurity, defined as a gestational age < 37 360
weeks at birth, compared to 4 patients in the case group; however, the number of patients younger than 37 weeks postconceptional age at the time of central line placement was similar between groups (1 in the control group vs. 0 in the case group). History of prior central line placement did not differ significantly between the groups (38.7% of cases vs. 28% of controls). CVL Characteristics The reason for line placement differed significantly between groups, with more patients in the case group requiring central access to infuse parenteral nutrition (34.7% vs. 18.7%; p = 0.03) (Table 3). Patients receiving parenteral nutrition in the case and control groups were further analyzed to assess for confounding factors that could have contributed to the incidence of clot formation. Mean age, duration of line placement, presence of other infusates, and mean total comorbidities were similar between the case and control groups receiving parenteral nutrition; therefore, the use of parenteral nutrition appeared to be an independent risk factor for VTE. The location of the line also differed significantly between the groups, with case subjects having more lines placed via the internal jugular vein than control J Pediatr Pharmacol Ther 2015 Vol. 20 No. 5 • www.jppt.org
JPPT
Thromboembolism in Pediatric Patients With Central Lines
Table 2. Comorbid Conditions Comorbidities Congenital heart defect
CVL With VTE* (n = 75)
CVL Without VTE* (n = 75)
p Value
26 (34.7)
11 (14.7)
OBJECTIVES: With the apparent increase in venous thromboembolism noted in the pediatric population, it is important to define which children are at risk for clots and to determine optimal preventative therapy. The purpose of this study was to determine the risk factors for venous thromboembolism in pediatric patients with central venous line placement. METHODS: This was an observational, retrospective, case-control study. Control subjects were patients aged 0 to 18 years who had a central venous line placed. Case subjects had a central line and a radiographically confirmed diagnosis of venous thromboembolism. RESULTS: A total of 150 patients were included in the study. Presence of multiple comorbidities, particularly the presence of a congenital heart defect (34.7% case vs. 14.7% control; p < 0.005), was found to put pediatric patients at increased risk for thrombosis. Additionally, the administration of parenteral nutrition through the central line (34.7% case vs. 18.7% control; p = 0.03) and location of the line increased the risk for clot formation. CONCLUSIONS: With increased awareness of central venous line–related thromboembolism, measures should be taken to reduce the number and duration of central line placements, and further studies addressing the need for thromboprophylaxis should be conducted. INDEX TERMS: central venous lines, clot, pediatrics, venous thromboembolism J Pediatr Pharmacol Ther 2015;20(5):358–366
INTRODUCTION As pediatric survival rates continue to increase with the introduction of new procedures and greater use of invasive vascular devices, so does the incidence of venous thromboembolism (VTE) in children.1 A recent retrospective cohort study of hospitalized children showed that the incidence of VTE rose from 34 per 10,000 hospital admissions to 58 per 10,000 hospital admissions during the period 2001-2007, yielding a 70% increase in the reported rate of thrombus formation.2 While the development of VTE seems to be multifactorial, central venous lines (CVLs) appear to be a causal risk factor.3 Potential mechanisms for clot formation include flow disruption, vessel wall trauma, and endothelial damage caused by the CVL.4 In another study, more than 50% of children and 80% of infants with a CVL developed clots.1 358
Despite the availability of a published clinical guideline by the American College of Chest Physicians for the prevention and treatment of VTE in the pediatric population,5 there is still a lack of randomized controlled trials identifying risk factors and evaluating the utility and safety of thromboprophylaxis in a general pediatric population with CVL placement.6,7 With the apparent increase in VTE noted in the pediatric population, it is important to define which children are at risk for clots and to determine optimal preventative therapy. Prevention of clot formation and subsequent exposure to prophylactic doses of anticoagulants may decrease morbidity associated with postthrombotic syndrome and mortality associated with pulmonary embolism. With that objective in mind, the aim of this retrospective case-control study was to identify risk factors for developing VTE in pediatric patients with J Pediatr Pharmacol Ther 2015 Vol. 20 No. 5 • www.jppt.org
JPPT
Thromboembolism in Pediatric Patients With Central Lines
a CVL. The information gathered in this study will assist us in developing future protocols to provide prophylaxis against VTE in high-risk pediatric patients with CVL.
PATIENTS AND METHODS This study was submitted to and approved by the Institutional Review Board of Memorial Hermann Hospital/Children’s Memorial Hermann Hospital in the Texas Medical Center and The University of Texas Health Science Center in Houston, Texas. Data were collected via a retrospective chart review from an internal, searchable database that contains inpatient data (Cerner Millenium Database, Memorial Hermann, System Services, Houston, TX). Subjects The study population included pediatric patients (0 to 18 years of age) who were admitted to the pediatric intensive care unit, intermediate medical care unit, or general medical unit at Children’s Memorial Hermann Hospital between July 1, 2007 and January 31, 2012, and who had a CVL placed during hospital admission. Subjects with CVL placement during inpatient admission were identified by the following ICD-9 CM codes: 38.93, 38.95, 38.97, and 86.07. To best identify patients of interest for a more homogenous population of study patients, subjects were excluded if they were admitted to the neonatal intensive care unit, were pregnant, or had a CVL placed at an outside facility. Patients with cancer also were excluded, as these patients are not typically treated at our facility. Of note, patients with a history of a clotting disorder or previous VTE were included in the study. Since patients are not routinely screened for clots, only patients with symptoms such as limb swelling/discoloration, poor pulses, fever, or elevation in white blood cell count (not associated with infection) were assessed with upper and/or lower extremity ultrasonography to confirm presence of a VTE. Eligible patients were stratified by the presence or absence of a radiographically confirmed diagnosis of VTE into cases (presence of VTE) and controls (absence of VTE). Owing to resource limitations, 150 patient cases were reviewed. Seventy-five case subjects were selected for analysis by a random number generator and matched to 75 randomly selected control subjects. J Pediatr Pharmacol Ther 2015 Vol. 20 No. 5 • www.jppt.org
Data Collection and Outcomes Patient demographics including age, sex, weight, admitting diagnoses, and comorbid conditions were collected from electronic medical records. Information including central line location, type, size, number of central line lumens, and duration of central line placement was recorded, as well as medications administered through the line, and any history of prior central line placement at our facility or at an outside facility before current admission. For case patients, the type and location of thrombus, as well as the drug used to treat the thrombus, were documented. Data for each subject were collected from the time of admission to the time of discharge from the hospital. The primary endpoint of this study was to determine the risk factors for thrombus formation in pediatric patients after CVL placement. The secondary endpoint was to describe clinical characteristics and annual incidence rates of thrombotic events. Statistical Analyses A comparison of baseline characteristics between case and control subjects was carried out by multiple tests. The Student t test was used for continuous data following a normal distribution, and Wilcoxon rank sum test for continuous data following a non-normal distribution. Chi-squared test or Fisher exact test was used to compare categorical data between the 2 groups. Risk factors for VTE after CVL placement were identified by bivariate analysis, if p value was ≤0.12. Using risk factors identified through the bivariate analysis, a logistic regression model was used to identify risk factors with a p value ≤ 0.05. We estimated that a sample size of 150 patients (75 cases, 75 controls) was needed for multiple regression analysis, assuming α = 0.05, β = 0.8, and a moderate effect size of 0.15. All statistical analyses were conducted by The University of Texas Health Science Center (Houston, TX).
RESULTS Patient Demographics/Characteristics During the study period, a total of 2,388 central lines were placed. Of the potential study subjects, 125 were found to have a radiographically confirmed VTE during their admissions (overall incidence of 5.2% during the study period). Of 359
JPPT
S Wisecup, et al
Table 1. Patient Demographics and Characteristics
Age, mean ± SD, yr Male, % Weight, mean ± SD, kg
CVL With VTE (n = 75)
CVL Without VTE (n = 75)
p Value
4.4 ± 5.6
4.3 ± 4.6
0.94
63
56
0.41
21.9 ± 28.4
22.8 ± 24.5
0.82
29 (38.6)
29 (38.6)
1.000
Admitting diagnosis* Infection Bacteremia
10 (13.3)
6 (8)
1.000
Other
19 (25.3)
23 (30.7)
1.000
Respiratory
15 (20)
14 (18.6)
1.000
Cardiovascular Neurological Gastrointestinal
14 (18.6)
7 (9.4)
0.157
6 (8)
12 (16)
0.208
2 (2.7)
6 (8)
0.276
Renal
4 (5.4)
1 (1.4)
0.367
Hematologic
2 (2.7)
1 (1.4)
1.000
3 (4)
5 (6.6)
0.719
54 (72)
41 (55)
0.003
Other† Comorbidity present*
CVL, central venous line; VTE, venous thromboembolism * Values are No. (%) † Other: overdose, vascular access, trauma, endocrine disorder
the subjects with a CVL and a radiographically confirmed VTE, 75 were selected by a random number generator to be case subjects for analysis. Of the patients who had a CVL placed during the study period but who did not have a radiographically confirmed VTE, 75 were randomly selected to be included as control subjects for comparison. Patient demographics including age, sex, weight and admitting diagnoses were similar between the two groups (Table 1). The mean age of the case and control subjects was similar, 4.4 ± 5.6 years and 4.3 ± 4.6 years, respectively. The number of patients with comorbid conditions present at admission was significantly higher in the case group than the control group (72% vs. 55%; p = 0.003) (Table 2). The presence of a congenital heart defect was found to be an independent risk factor for the development of VTE (34.7% of cases vs. 14.7% of controls; p < 0.005), as subanalysis of case and control subjects with congenital heart defects did not detect differences in other factors between the groups that may have been responsible for the higher incidence of clot formation (e.g., mean age, duration of line placement, and mean total comorbidities). Eight patients in the control group had a history of prematurity, defined as a gestational age < 37 360
weeks at birth, compared to 4 patients in the case group; however, the number of patients younger than 37 weeks postconceptional age at the time of central line placement was similar between groups (1 in the control group vs. 0 in the case group). History of prior central line placement did not differ significantly between the groups (38.7% of cases vs. 28% of controls). CVL Characteristics The reason for line placement differed significantly between groups, with more patients in the case group requiring central access to infuse parenteral nutrition (34.7% vs. 18.7%; p = 0.03) (Table 3). Patients receiving parenteral nutrition in the case and control groups were further analyzed to assess for confounding factors that could have contributed to the incidence of clot formation. Mean age, duration of line placement, presence of other infusates, and mean total comorbidities were similar between the case and control groups receiving parenteral nutrition; therefore, the use of parenteral nutrition appeared to be an independent risk factor for VTE. The location of the line also differed significantly between the groups, with case subjects having more lines placed via the internal jugular vein than control J Pediatr Pharmacol Ther 2015 Vol. 20 No. 5 • www.jppt.org
JPPT
Thromboembolism in Pediatric Patients With Central Lines
Table 2. Comorbid Conditions Comorbidities Congenital heart defect
CVL With VTE* (n = 75)
CVL Without VTE* (n = 75)
p Value
26 (34.7)
11 (14.7)
