Published online February 26, 2007
PEDIATRICS Vol. 119 No. 3 March 2007, pp. e742-e753 (doi:10.1542/peds.2006-2221)
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow P3Rs: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when P3Rs are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow E-mail this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My File Cabinet
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via CrossRef
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Danschutter, D.
Right arrow Articles by Huyghens, L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Danschutter, D.
Right arrow Articles by Huyghens, L.
Related Collections
Right arrow Blood

ARTICLE

Di-(2-ethylhexyl)phthalate and Deep Venous Thrombosis in Children: A Clinical and Experimental Analysis

Dirk Danschutter, CCRN, CP, MSca, Filip Braet, PhDb, Elke Van Gyseghem, PhDc, Said Hachimi-Idrissi, MD, PhD, FCCMa, Benny Van Bruwaene, MDd, Pat Moloney-Harmon, RN, MS, CCNS, CCRN, FAANe and Luc Huyghens, MD, PhD, FCCMa

a Department of Intensive Care
d Department of Medical Registration and Statistics, Free University of Brussels, Brussels, Belgium
c Laboratory of Pharmaceutical Technology and Biopharmacy, Catholic University of Leuven, Leuven, Belgium
b Australian Key Centre for Microscopy and Microanalysis, University of Sydney, Sydney, Australia
e Advanced Nursing Practice, Children's Services, Sinai Hospital, Baltimore, Maryland


   ABSTRACT
TOP
 ABSTRACT
PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 LIMITATIONS
 REFERENCES
 
BACKGROUND. Five children with catheter-related deep venous thrombosis were encountered in our PICU. Three types of polyvinyl chloride tubing for the administration of intravenous solutions were in use (Terumo, Codan, and Perfusend). All were di-(2-ethylhexyl)phthalate plasticized. We suspected problems with the Codan tubing.

METHODS. Different types of tubing at different time intervals in vitro were investigated. Tubing segments were assessed on structural alterations by surface electron microscopy. High-performance liquid chromatography-diode array detection and liquid chromatography-mass spectrometry-diode array detection were performed to identify and to quantify di-(2-ethylhexyl)phthalate. The hospital's minimal clinical data set (coded with the International Classification of Diseases, Ninth Revision, Clinical Modification) was investigated on catheter-related deep venous thrombosis between 2000 and 2004.

RESULTS. Surface electron microscopy demonstrated that the Codan tubing's inner surface was severely altered, showing large particles (34.5 ± 6.1 µm). High-performance liquid chromatography documented that all Codan samples showed a peak at the di-(2-ethylhexyl)phthalate retention time. The analysis of the minimal clinical data set for total catheter-related deep venous thrombosis showed an unusual high incidence in 2001 (52) compared with the expected 36 per year.

CONCLUSIONS. Such occurrence of catheter-related deep venous thrombosis led to the assumption that disintegration of intravenous tubing resulted in intravenous administration of debris. Our data suggested that the particles derived from the tubing are of such size that they might induce catheter-related deep venous thrombosis. The absence of catheter-related deep venous thrombosis caused by the introduction of submicron inline filters outlines the important pathophysiological role of di-(2-ethylhexyl)phthalate-plasticized particles in the onset of catheter-related deep venous thrombosis. Our data indicate that a considerable number of patients might have been exposed to di-(2-ethylhexyl)phthalate, and a major concern is whether this jeopardized the health of the patients at that time.

Key Words: di-(2-ethylhexyl)phthalate

Abbreviations: DVT—deep venous thrombosis • CR-DVT—catheter-related deep venous thrombosis • PVC—polyvinyl chloride • DEHP—di-(2-ethylhexyl)phthalate • PF—peristaltic finger • TOTM— trioctyl trimellitate • HPLC-DAD— high-performance liquid chromatography-diode array detection • LC-MS-DAD—liquid chromatography-mass spectrometry-diode array detection • m/z—mass to charge ratio • SEM—surface electron microscopy • HPLC—high-performance liquid chromatography • MS—mass spectrometry • mAbs—milliabsorbance

Deep venous thrombosis (DVT) is a seldom encountered but potentially dangerous medical condition in children. Factors that may predispose to DVT include a hypercoagulable state, phlebitis, systemic or catheter-related infections, irritation, and drug incompatibility.

Over a 33-day period in 2001, we encountered in our PICU 5 children with catheter-related DVT (CR-DVT) of unknown origin. Having excluded primary medical conditions as a possible cause, we extensively investigated the possible etiologic role of fluid and drug incompatibilities and the intravenous administration tubing for intravenous lines. The type of femoral catheters used by the PICU historically consisted of polyethylene and polyurethane. Before the occurrence of the 5 CR-DVT cases, the hospital and subsequently the PICU had gradually begun introducing new intravenous administration tubing sets, as well as new peristaltic (volumetric) intravenous pumps. In all, 2 types of peristaltic pumps (Terumo [Leuven, Belgium], type STC-503 and Argus [Heimberg, Switzerland], type 414) and 3 types of intravenous-sets were in use (Terumo, Codan [Lensahn, Germany], and Perfusend [Sendal, Almaraz, Spain]) during the occurrence period. All intravenous administration tubing sets were commercially manufactured from polyvinyl chloride (PVC) plasticized with di-(2-ethylhexyl)phthalate (DEHP). The intravenous tubing system was pressure driven by peristaltic finger (PF) pumps. With 1 of the tubings recently introduced (Codan tubing with Terumo PF pumps), we noticed that the dialed-in volume on the infusion pump ran in over a shorter period of time than programmed. These PF pumps were thoroughly checked and found to be functioning correctly. Hence, we hypothesized that there might be a tubing-related problem.

This made the PICU unit eager to investigate whether tubing wear with potential release of plastic particles and DEHP may have played a role in the etiology of CR-DVT in these patients. We decided to investigate 3 types of tubing used in our PICU: Terumo, Codan V86-P-Y ref. 16/03/14054, and Perfusend. From a prospective interest, we included in our study the intravenous tubing of Cardinalhealth (Hampshire, United Kingdom) for Cardinalhealth type 571 PF pumps.


   PATIENTS AND METHODS
TOP
 ABSTRACT
 PATIENTS AND METHODS
RESULTS
 DISCUSSION
 LIMITATIONS
 REFERENCES
 
Patients
All 5 patients were admitted to a tertiary pediatric care center and required central line infusions. Using a Seldinger technique, the catheters were introduced in the right or left femoral vein by using either a Vygon (Ecoven, France) single-lumen catheter or a Cook (Bloomington, IN) double-lumen catheter. All patients had additional peripherical inserted catheters and an arterial line. The central access served for the administration of inotropes and total parenteral nutrition. As a routine, the compatibility of solutions infused was checked for each patient against the recommendations of the Handbook on Injectable Drugs.1

Tubing
We investigated the 3 types of tubing used in the PICU (and through the entire hospital): Terumo, Codan, and Perfusend. From a prospective interest the Cardinalhealth tubing used in the NICU was investigated by surface electron microscopy (SEM) alone. The investigators were blind to the type of tubing being investigated.

To assess the accuracy of the different types of PF pumps, a 1-liter bag was connected to Codan, Terumo, and Perfusend tubing and driven by Terumo for Codan and Terumo and Argus PF pumps for Perfusend. The reference solution infused in the tubing during the experiment consisted of 0.9% sodium chloride in water for injection (Baxter, Lessines, Belgium). The volume of solution, administered over 24 hours for all tubing, was calculated to match an infusion for a child weighing approximately 10 kg and amounted to 42 mL per hour. The experiment was allowed to run for 96 hours.

During another experiment, paired samples of 10 mL of fluid (sample "Cod1" and "Cod2") were collected at 0, 24, 48, 72, and 96 hours from 2 different Codan tubings. The 2 Codan samples collected at 96 hours from 2 different Codan tubings first passed an inline screen filter with a pore size of 0.2 µm. A sample of 10 mL of fluid was collected at 0, 24, 48, and 72 hours from the Perfusend tubing. The distal end of the different tubing was allowed to drain freely into a collector without contact with the latter.

The collected samples were used for identification and quantification of DEHP or trioctyl trimellitate (TOTM) by using high-performance liquid chromatography-diode array detection (HPLC-DAD) and liquid chromatography-mass spectrometry-diode array detection (LC-MS-DAD).

Segments from all types of tubing (Codan, Terumo, Perfusend, and Cardinalhealth) in contact with the PF pumps were removed at 0 and 72 hours for SEM investigation. After the initial SEM results of these segments, additional pump segments of the Codan tubing were removed at 6, 12, and 18 hours for SEM investigation.

Surface Electron Microscopy
All types of intravenous tubing pump segments (Codan, Terumo, Perfusend, and Cardinalhealth) were carefully sliced with a razor blade to reveal the inner surface of the tubing. Samples were transferred to a desiccator at room temperature for 24 hours to avoid water contamination. Subsequently, the tubes were mounted on 1-in aluminum pin stubs (G399, Agar Scientific, Essex, United Kingdom) with the aid of double-sided tape, and then sputter coated with 10-nm gold. The specimens were examined with a Philips (Eindhoven, Netherlands) SEM 505 at an accelerating voltage of 30 kV and a spot size of 20 nm to obtain images with low noise content, as previously described.2 The magnification of the SEM was regularly calibrated with a cross-grating replica (Polaron, Watford, United Kingdom; 30000 lines/in) with the specimen in eucentric position. The obtained images were transferred to the UTHSCSA Image Tool 2.0 software (Department of Dental Diagnostic Science, University of Texas Health Science Center, San Antonio, TX) and was used to determine the particle size and for figure assembly.

HPLC-DAD and LC-MS-DAD Investigation of the Presence of DEHP in Collection Fluids
Reagents
Navigated by SEM investigation, solution samples from Codan and Perfusend were analyzed for the presence of either DEHP or TOTM with HPLC-DAD. To avoid environmental contamination, the samples were collected in borosilicate glass tubes with phthalate-free stopcocks. Two more Codan samples from the 2 different Codan tubings were collected after 96 hours runtime while having crossed an inline screen filter with a pore size of 0.2 µm.

A saturated solution of bis-(2-ethylhexyl)phthalate for synthesis (Merck-Schuchardt, Schuchardt, Hohenbrun bei München, Germany) in Milli-Q water was used to obtain the ultraviolet spectrum of DEHP. Milli-Q water was prepared with the Millipore purification system (Millipore, Molsheim, France). In addition, a solution of bis-(2-ethylhexyl)phthalate for synthesis (Merck-Schuchardt) was prepared in acetonitrile, Hypersolv for high-performance liquid chromatography (HPLC; BDH, Poole, United Kingdom), because the solubility of DEHP is higher in the organic solvent. For the same reason, the tubing was immersed for 2 hours in acetonitrile.

For the HPLC-DAD experiments, the solutions and mobile phases were prepared by using acetonitrile, Hypersolv for high-performance liquid chromatography (HPLC; BDH) and Milli-Q water.

To be sure of the nature of the component behind the retention time, additional structural characterization was necessary. Therefore, the experiments were repeated by using LC-MS-DAD, so that through mass spectrometry (MS) the mass to charge ratio (m/z) of the compound could also be determined. The mobile phase of LC-MS-DAD was prepared by using acetonitrile, for HPLC far UV (Acros Organics, Geel, Belgium).

Instruments
The ultraviolet spectrum was recorded ranging from 200 to 400 nm by using a Perkin Elmer-Lambda 20 UV/VIS spectrophotometer (Perkin Elmer, Norwalk, CT). A Bransonic 5210E-MT (Branson Ultrasone Cooperation, Danbury, CT) ultrasonic bath was used for degassing.

The HPLC-DAD instrument consisted of a model 5000 liquid chromatograph pump (Varian, Palo Alto, CA), a 100-µL loop, a CTO-10A column oven, and an SPD-M10A diode array detector (both Shimadzu, Kyoto, Japan). The chromatographic methods and data were created and treated with the Class-M10A LC workstation software (Shimadzu). The diode array detector scanned in the range of 200 to 400 nm.

The stationary phase used was Discovery RP-AmideC16 (100 x 4.6 mm inner diameter; mesh size: 5 µm; Supelco, Bellefonte, PA). The column can withstand 100% aqueous conditions without showing phase collapse because it is polar-embedded. The flow rate was 1.0 mL/min. The temperature of the oven was kept at 40°C.

Additional experiments were performed by using an LC-MS-DAD instrument consisting of a Waters 2695 separations module (= alliance) HPLC compartment (Waters, Milford, MA), a Mistral column oven (Spark Holland Instrumenten, Emmen, the Netherlands), a column switcher (VICI AG, Schenkon, Switzerland), a Waters 996 Photodiode Array Detector (Waters, Milford, MA), and a quadrupole time-of-flight mass spectrometer (Waters, Manchester, England, United Kingdom). The mass spectrometer data were obtained by using a scan range of 80 to 1000 Dalton (m/z), a scan time of 1 second, a resolution of ± 9000, the multiple channel plate at 2.2 kV, applying electron spray ionization in the positive ion mode as mass spectrometer parameters, and the source temperature of the ion source at 100°C. The chromatographic methods were created and the data treated by using both Millennium 4.0 software (Waters) for the spectral data and MassLynx 3.5 software (Micromass, Cary, NC) for the MS data.

The solution samples were researched by using a linear gradient at a flow rate of 1.0 mL/min. All experiments were conducted at 40°C. The injection volume was 5 µL. The experiments were run on a Hypersil C18-BDS (100 x 4.0 mm inner diameter; mesh size: 3 µm) and a Zorbax SB-C8 (150 x 4.6 mm inner diameter; mesh size: 5 µm) stationary phase, both from Agilent (Palo Alto, CA).

Minimal Clinical Data Set
In Belgium, as in most of the European countries, all diagnoses that patients present during their hospital stay are coded with the International Classification of Diseases. In Belgium, the International Classification of Diseases, Ninth Revision, Clinical Modification, is used.

In the database of the University Hospital of Free University of Brussels, all admissions between 2000 and 2004 with DVT of the femoral vein (code 451.11) as secondary diagnosis were selected and counted in the nominator. The number of patients of the diagnosis-related groups in which at least 1 DVT was noticed was used as denominator. The incidence of CR-DVT was calculated for the PICU, for the ICUs, and for hospitalization wards for adult patients. Because the Department of Neonatology did not use the Codan tubes, this department was excluded.


   RESULTS
TOP
 ABSTRACT
 PATIENTS AND METHODS
 RESULTS
DISCUSSION
 LIMITATIONS
 REFERENCES
 
Patients
In the PICU, 5 of 117 patients presented with a CR-DVT in 2001, semester 1. Patient 1 (1.25 years old) was admitted after cardiac arrest after accidental strangulation. Patient 2 (4 years old) was hospitalized for encephalitis and deep comatose condition caused by Epstein-Barr viral infection. Patient 3 (1.75 years old) with multiple congenital abnormalities and cerebral cysts was admitted for respiratory syncytial virus pneumonia and respiratory failure attributable to status epilepticus. Patient 4 (15 years old) was admitted for respiratory failure attributable to Duchenne myodystrophy and was transferred to the PICU for postoperative and septic shock after major surgery on the respiratory system. Patient 5 (0.75 years old) was admitted for respiratory failure complicating respiratory syncytial virus pneumonia. Demographic and clinical data of the 4 patients are represented in Table 1. The coagulation profile of the 5 patients is shown in Table 2. The type of solutions infused before onset of the CR-DVT is shown in Table 3. The site of venous thrombosis was established by using echo-Doppler ultrasound evaluation and found to be at the tip of the catheter and/or short distance upstream. In patient 4 (Table 1), a massive CR-DVT was asymptomatic and subsequently diagnosed by a computed tomography scan of the pelvic region for another condition (acute abdominal pain). Immediately after the diagnosis of thrombosis, all 5 patients were treated with low molecular weight heparin. Two were additionally treated with coumarin (Table 3).


View this table:
[in this window]
[in a new window]

  		TABLE 1 Patient Demographic and Clinical Data

 

View this table:
[in this window]
[in a new window]

  		TABLE 2 Coagulation Profile of the Patients

 

View this table:
[in this window]
[in a new window]

  		TABLE 3 Solutions Administered Before Onset of CR-DVT, Anticoagulant Treatment, and Patient Outcome

 
Because of the fifth incidence of CR-DVT at the PICU (patient 5), inline screen filters with a pore size of 0.2 µm were introduced with not a single occurrence of CR-DVT observed in the PICU until June 2006.

Bedside nurses throughout the hospital occasionally observed and reported that Terumo PF pumps did not administer intravenous fluids as accurately as in the past, whereas service engineers were unable to detect any malfunctioning of the equipment.

Surface Electron Microscopy
SEM examination of the Perfusend tubes at 0 hours showed multiple parallel-organized, worm-like structures, which seemed to be packed into multiple layers on top of each other. In contrast, the same Perfusend tubes exposed at 72 hours to the PF pump mechanism showed a more flattened outlook, ie, the worm-like impressions showed a strong parallel organization and multiple layers seemed to be absent. In addition, these tubes showed large surface indentations that were positioned with a periodicity of ~0.14 ± 0.1 mm (n = 20), indicating that these topology changes are derived from the segments of the PF pump-driving mechanism. These indentations could already be observed by macroscopic visual inspection.

Codan tubes at 0 hours revealed smaller worm-like impressions with a random organization (Fig 1) compared with Perfusend tubes at 0 hours (Fig 2). Cardinalhealth and Terumo control tubes were revealed to have a rather smooth surface with no worm-like structure.


Figure 1
View larger version (144K):
[in this window]
[in a new window]

  		FIGURE 1 SEM of the inner surface of the Codan control tube. Note the worm-like structures. Scale bars, 0.1 mm.

 

Figure 2
View larger version (173K):
[in this window]
[in a new window]

  		FIGURE 2 SEM of the inner surface of the Perfusend control tube. Note the worm-like structures. Scale bars, 0.1 mm.

 
Codan tubes that were placed in the PF pump mechanism showed a dramatic change in surface topology, ie, large and deep grooves, as early as 6 hours after pumping. These large artifactual grooves or cracks gave the impression that the polymer's surface had lost its integrity and was severely damaged (Fig 3). In addition to these artifactual grooves, the surface lacked the small worm-like structures seen at 0 hours and instead large ribbon-like surface corrugations appeared. Moreover, large particles (34.5 ± 6.1 µm [n = 20]) became apparent (Fig 4); these seemed to be fragments of the Codan tube.


Figure 3
View larger version (145K):
[in this window]
[in a new window]

  		FIGURE 3 PF pump–exposed Codan tube. The white arrow denotes grooves or cracks derived from the tube segments exposed to the PF pump, and the gray arrow shows large ribbon-like surface corrugations. Scale bars, 0.1 mm.

 

Figure 4
View larger version (124K):
[in this window]
[in a new window]

  		FIGURE 4 PF pump–exposed Codan tube. Note the large particles (white arrowhead) and small particles (gray arrowhead). Scale bars, 0.1 mm.

 
Particles with a smaller size (4.5 ± 1.2 µm [n = 20]) could also be noticed on the Terumo tubes (Fig 5) exposed to the PF pump mechanism, although they were present in a lower quantity. Like the Terumo tubes, the Perfusend (Fig 6) and Cardinalhealth PF pump segments kept their wall integrity, even after 96 hours of pumping time.


Figure 5
View larger version (154K):
[in this window]
[in a new window]

  		FIGURE 5 PF pump–exposed Terumo tubes that have lost their smooth surface and show a corrugated surface in the form of bumps (black arrow). Scale bars, 0.1 mm.

 

Figure 6
View larger version (146K):
[in this window]
[in a new window]

  		FIGURE 6 PF pump–exposed Terumo tubes. Note the particles with a smaller size on the inner surface of a PF pump–exposed Perfusend tube (white arrow). Scale bars, 0.1 mm.

 
HPLC-DAD and LC-MS-DAD
The ultraviolet spectrum of DEHP was recorded for a saturated solution of the substance in water to investigate whether the compound was exhibiting a large extinction coefficient.3–5 It seemed to have a very high absorbance, which implies that the presence of DEHP in the samples would be easily detected (figure not shown). The spectrum shows a wavelength of maximal absorbance at 274 nm.

Injection of the solution of DEHP in acetonitrile applying HPLC-DAD using 95%/5% (vol/vol) acetonitrile/water isocratically as the mobile phase gave rise to a peak at 1.712 minutes/1.691 minutes (220 nm/254 and 274 nm) retention time. Knowing now the retention time of the DEHP, as well as its ultraviolet spectrum in both water and acetonitrile, 2 different Perfusend samples collected at 72 hours from 2 different PF pumps were injected using the same chromatographic conditions, but no peak was detected at that retention time. The reference solution, however, gave rise to 2 peaks of ~180 milliabsorbance (mAbs) (220 nm) at elution times of 1.545 minutes and 1.737 minutes (35 mAbs [254 nm] at 1.55 minutes and 1.737 minutes, and 28 mAbs [274 nm] at 1.544 minutes and 1.737 minutes; figure not shown). Comparable peak heights and peak areas were observed for the Codan samples, whereas for 1 Codan sample, significantly higher values for peak heights (244 mAbs at 220 nm, 48 mAbs at 254 nm, and 38 mAbs at 274 nm) and peak areas were encountered at those retention times (Fig 7).


Figure 7
View larger version (32K):
[in this window]
[in a new window]

  		FIGURE 7 Chromatogram at 3 different wavelengths (220, 254, and 274 nm) of the Codan sample with the highest peak areas and heights.

 
The absorbance leads to the conclusion that there is a higher concentration in Codan samples. The diode array detector output showed the same spectrum each time as that recorded for the saturated solution of DEHP in water and the reference solution.

The unambiguous identification of TOTM from other organic impurities is difficult, because 2 substances differing in structure can exhibit the same ultraviolet spectrum. Six Codan samples were investigated with HPLC-DAD and LC-MS-DAD. The tubings used to collect the samples were immersed in 200 mL of acetonitrile for 2 hours and placed on the ultrasonic bath. Both Codan and Perfusend tubings consisting of PVC-DEHP were used as softening agents, and both components dissolved in acetonitrile. To obtain better peak shapes, the mobile phase conditions were changed to gradient elution (Table 4).


View this table:
[in this window]
[in a new window]

  		TABLE 4 Conditions for the Gradient Run

 
Injection of the solution of the tubing using HPLC-DAD gave rise to a very large peak at 18.298 minutes elution time. The spectrum (figure not shown) is identical to the saturated solution. Afterward, the Codan samples were injected using the same chromatographic conditions. All samples showed a peak at that retention time, exhibiting the same ultraviolet spectrum as shown in Fig 8.


Figure 8
View larger version (11K):
[in this window]
[in a new window]

  		FIGURE 8 Ultraviolet spectrum recorded at 19.49 minutes elution time on the Hypersil C18-BDS using gradient elution conditions for the injection of the solution of the tubing in acetonitrile.

 
The Codan fluid samples collected after 24 hours showed the highest absorbances among them, and their spectra were similar to that recorded for the solution of the Codan tubing itself. The 2 filtered (via 0.2-µm inline screen filter) Codan samples collected after 96 hours only showed intermediate peak heights. These observations suggest that the Codan tubing displays an impressive leakage of an organic substance into the infusion solution compared with the Perfusend tubing or reference solution, moderately mitigated by the presence of a submicron inline filter.

To be sure of the nature of the component behind the observed retention time, additional structural characterization was necessary. Therefore, the experiments were repeated by using LC-MS-DAD, so that through MS the m/z value of DEHP could be also determined. DEHP has a molecular weight of 390,6 whereas TOTM exhibits a molecular mass of 546.7 Two stationary phases with similar selectivity were used, applying the same gradient conditions as for the HPLC-DAD experiment.

For the solution of the tubing immersed in acetonitrile injected on the Hypersil C18-BDS, a large peak is shown at retention time 19.49 minutes, of which the ultraviolet spectrum is visualized in Fig 8. The MS scan showed a peak of 100% abundance for the m/z value of 391, whereas no abundance was measured at m/z equal to 547 (Fig 9), which strongly suggests the presence of DEHP.


Figure 9
View larger version (8K):
[in this window]
[in a new window]

  		FIGURE 9 MS spectrum with 100% abundance for m/z value of 391 (identifying DEHP) on the Hypersil C18-BDS using gradient elution conditions for the injection of the solution of the tubing in acetonitrile.

 
Similarly, the solution of the tubing immersed in acetonitrile injected on the Zorbax SB-C8 stationary phase gave a large peak at retention time 15.68 minutes (Fig 10). Scanning the MS spectrum revealed a peak of 100% abundance at the m/z value or 391 at that elution time, whereas for the m/z value of 547, no abundance could be detected (figure not shown).


Figure 10
View larger version (8K):
[in this window]
[in a new window]

  		FIGURE 10 Chromatogram with peak at 15.68 minutes' retention time on the Zorbax SB-C8 using gradient elution conditions for the injection of the solution of the tubing in acetonitrile.

 
Minimal Clinical Data Set
For the PICU, the data revealed 3 patients of 117 admissions in 2001 semester 1 and 1 patient of 142 admissions in 2000 semester 1. In the other semesters, no CR-DVT was registered (Fig 8).

The incidence of CR-DVT in the 2001 semester 1 was tested with an exact test for binomial distributions and was significantly high (P < .0001).

For the adult ICUs and hospitalization wards for adults, a peak in CR-DVT could be observed in 2001 semester 2 (Fig 9). A total of 52 CR-DVT incidents were counted on 5336 admissions (1.0%) against an average of 0.6% in the other semesters between 2000 and 2004. This is 16 events more than expected (P = .002).


   DISCUSSION
TOP
 ABSTRACT
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
LIMITATIONS
 REFERENCES
 
During a 33-day period in 2001, we had 5 cases of CR-DVT in children admitted to the PICU. We were unable to identify a pathophysiological cause. We could not implicate the femoral catheters because these have been in use and unchanged in composition for years. None of the patients had evidence for hereditary or acquired hypercoagulability.

These thromboses occurred during a period in which the PICU (following hospital budgetary politics and recommendation) changed to using new types of intravenous administration sets and intravenous volumetric PF pumps. This change occurred gradually with a blending of both the different types of PF pumps (Terumo and Argus) and the intravenous tubing sets (Codan, Terumo, and Perfusend). The Terumo PF pumps were progressively replaced with the newer type of Argus PF pumps, while the Codan tubings had invaded local unit stocks hospitalwide, gradually becoming prominently available as the hybrid alternative to the more expensive Terumo tubing. As Argus PF pumps (Perfusend tubing) progressively became more available, there was no additional need to use the Terumo hybrid tubing (Codan) that by now had completely replaced the Terumo tubing. From a retrospective investigator's point of view, an insuperable shortcoming that inhibits all attempts to statistical analysis is that the type of PF pump and the type of intravenous tubing used for the administration of intravenous fluids were never recorded in the patient or nursing charts. In fact, an amalgam of 2 different tubings combined with 2 different types of PF pumps and even more types of syringe pumps were simultaneously being used bedside at the time of the occurrence of the 5 CR-DVT incidents. In a best-case scenario, these 5 patients at the PICU may well have never been infused over Codan tubings and suffered CR-DVT purely because of coincidence.

The results of this in vitro experiment, however indisputable, establish that the Codan tubing disintegrated during normal use with a PF pump. This now explains why bedside nurses throughout the hospital occasionally observed and reported that Terumo PF pumps did not administer intravenous fluids as accurately as in the past, whereas service engineers were unable to detect malfunctioning of the equipment. SEM documented that the disintegration of the Codan tubing's pump segment occurred as soon as after 6 hours of running time. The in vitro HPLC experiment determined that the disintegration product in the Codan PVC tubing is DEHP. The loss of PVC wall integrity and shedding of plastic into passing fluid undoubtedly resulted in altered diameters of the calibrated Codan tubing's pump segment. As the Codan tubing flattened and the inner diameter increased, more fluid was pumped at the initial volume set, causing the PF pumps to run inadequately.

The incentive to perform this investigation followed the highly unusual occurrence of CR-DVT in 4 children in such short period. This led to the assumption that disintegration of intravenous tubing into intravenous fluid resulted in actual intravenous administration of plasticized debris to patients. A fifth patient was reported soon after this hypothesis was formed but before inline filters were introduced systematically. The in vitro experiment confirmed that the particles shed from the intravenous tubing into the intravenous fluid are of such large size (34.5 ± 6.1 µm) that hypothetically they can easily induce mechanical thrombosis. Moreover, the absence of additional occurrences of CR-DVT at the PICU because of the introduction of 0.2-µm inline filters suggests that the intravenous administration of large plastic particles and DEHP might actually have played a pathophysiological role in the formation of CR-DVT. It is also probable that phthalate plasticized tubing with a higher shore and better elasticity characteristics is likely to release fewer particles. The presence of very large particles on SEM, migrating freely on the surface of the intravenous Codan tubing with disintegration of the inner wall, while other types of investigated tubings (Terumo, Perfusend, and Cardinalhealth) kept their wall integrity for a longer period than 96 hours (Figs 3–6), enforces this hypothesis. HPLC documented that the in vitro insertion of a commercial 0.2-µm inline filter mitigates the phthalate dose to intermediate values, which also may explain the absence of CR-DVT after the filters were introduced in vivo.

Whether or not these 5 patients were infused with considerable amounts of phthalate and PVC particles cannot be confirmed. The administration of medication and fluids via different types of PF and syringe pumps or different types of tubing has never been systematically registered in the PICU nursing plans or in the PICU medical files. More surprising, however, is the observation that during the period 1997 to 2002, Codan tubings were used extensively throughout the hospital. So-called r41 tubings of Baxter without pump segment were used for off-pump gravitational infusion therapy, which implies that almost all Codan tubings were used in the devastating combination with a PF pump. Through the hospital and, therefore, including the PICU, an unrecorded percentage of patients have been infused with medication or intravenous solutions via 44568 Codan tubings and PF pumps and, therefore, might have been infused with DEHP plasticized PVC particles. According to our data, the Codan tubing is a very morbid one but the most important shifts in the intravenous tubing field occurred in 1998; however, the 5 CR-DVT cases at the PICU were observed in January 2001, which is ~2 years out of date (Fig 11). However, analyzing the hospital Minimal Clinical Data Set (coded with International Classification of Diseases, Ninth Revision, Clinical Modification) for total CR-DVT occurrence shows a surprisingly high incidence in 2001 (semester 2 [n = 52] > semester 1 [n = 40]) compared with the expected n = 36 on a yearly basis. The short delay in peak of CR-DVT is probably because of the fact that the Codan tubings remained at use at the adult wards until exemption of stock, whereas the PICU systematically introduced inline filters combined with an early and complete ban of the Codan tubings (Fig 12).


Figure 11
View larger version (19K):
[in this window]
[in a new window]

  		FIGURE 11 Percentage of CR-DVT on the PICU for the period 2000–2004. s indicates semester.

 

Figure 12
View larger version (24K):
[in this window]
[in a new window]

  		FIGURE 12 Percentage of CR-DVT in the AZ-VUBadult population for the period 2000–2004. s indicates semester.

 
The percentage of mechanically pumped intravenous therapies has gradually increased over the years. The hospital material databases (hardware, orders, and medical disposables) reveal how the number of PF and syringe pumps increased and subsequently how the use of intravenous disposables proliferated, while the turnover of patients is still increasing. The overall intravenous tubing order increased 41% over the past 9 years (114180 in 1996 vs 193404 in 2005). The hospital as a whole has become more intravenous technical and likely more prone to intravenous technical-related adverse effects. Although this might partially explain the sudden increase of hospital CR-DVT in 2001, it does not explain why until today, the incidence fell back to the expected baseline coincidentally with the Codan tubings and Terumo PF pumps disappearing from the hospital scene. All these findings suggest that the problems started either as the Terumo tubing completely was replaced or when the Codan tubing for unidentified reasons started to disintegrate right before 2001. When the SEM results became available, the Terumo PF pumps and Codan tubings were immediately removed from the PICU, and the hospital management was informed of the preliminary results.

Our in vitro experiments strongly suggest that a considerable number of patients, both at the PICU and throughout the hospital, have been intravenously injected with DEHP plasticized particles. This occurred involuntarily because of wear of the Codan PVC tubing and happened insidiously. A major concern is whether and how this jeopardized the health of the unidentified patient population of that time.

DEHP is the only plasticizer approved by the US Food and Drug administration for medical use.8 It is a colorless oil used in high concentrations (20%–50%) in medical disposables to soften a commercial resin in white powder form that is called PVC.9 Despite this approval, serious concerns have been postulated over the potential of adverse effects to health when individuals are exposed to phthalates,10–24 hence the ban on some specific phthalates in children's toys instituted by the European Commission in June 2005 and fully supported by the European Parliament.25 To be more specific, 6 of the currently used phthalates to soften PVC toys and childcare articles are now permanently banned for reasons of adverse health effects and reprotoxicity. Under the directive, 3 of these phthalates are classified as undoubtedly toxic: DEHP, dibutyl phthalate, and butyl benzyl phthalate. As for the 3 other phthalates, di-isononyl phthalate, di-iso-decyl phthalate, and di-n-octyl phthalate, there is uncertainty of the risks they present. Regardless of age, children should no longer be exposed to toys or childcare articles softened with 1 of these 6 plasticizers (in a concentration >0.1%). From now on, there is a stabile legal situation (2005/84/EEC) that will enable industry to plan in conditions of certainty. The earliest possible deadline by which the companies will be required to comply with these new restrictions is likely to be October 2006.26

Because phthalates are not covalently bound, they passively leak from the plastic matrix influenced by factors like temperature, exercised pressure, and storage time.27–34 This leakage is influenced by contact with fluids and by flow rates, and there is documented interaction with body fluids like blood, mucus, and saliva,14–19,22,35 hence, the ban for DEHP plasticized toys that can be chewed on and put in a child's mouth. Remarkably however, and despite this evidence, there is no such ban for the use of DEHP-softened PVC medical devices, which are also put into children's mouths or even beyond.15,18,22 The PF pump mechanism is a chewing one because it continuously squeezes the softened PVC tubing between its peristaltic moving fingers and a solid metal plate. Blood as a body fluid is known to interact strongly with spallated or shed DEHP and particularly its monoester metabolites.8,15 Protein adsorption occurs within seconds of blood–material contact and plays an important role in the subsequent blood response. Increased fibrinogen adsorption can be regarded as the initial index of blood reactions and leads to higher trombogenicity. This also reflects in higher thrombin-antithrombin III complex and complement C3a values.28 This overall relationship among blood and biomaterials may explain the sudden occurrence of catheter-related DVT in the 5 patients of the PICU or the sharp increase of CR-DVT incidence in the hospital in 2001. Furthermore, DEHP and its metabolites are known endocrine disruptors modifying carcinogenesis in rats and mice19,24,36,37 and are genotoxic to human lymphocytes and human mucosal cells.38 DEHP and its metabolites are also known xenobiotics mimicking or antagonizing sex hormones, hence their established reprotoxicity.19,39–44 It is also remarkable that since plasticized products became ubiquitous in the developed world, asthma and allergies evolved to the status of major health care problems. Evidence now reveals that, besides lifestyle and demographic factors, asthma and certain allergies can reflect a biological response to phthalates and especially DEHP.20–23

The clinical International Classification of Diseases coded hospital data recorded the unusual high incidence of CR-DVT quite simultaneously with their real-time occurrence. In the PICU, however, there was a documented underscore compared with the real frequency of CR-DVT. The accurate and a more systematic use of this data set for the monitoring of sentinel events should be considered to reveal problems in patient safety at the earliest moment.45–52 Sentinel function of whatever tool based on entered diagnosis implies that the diagnosis is actually made and put in the registration system. Occult occurrences of conditions or missed diagnosis will not add to any database with new input. In our observations, the oldest (15 years old) of the 5 patients with CR-DVT remained asymptomatic and the thrombosis was coincidentally documented via a computed tomography scan of the pelvic region for another indication (acute abdominal pain). With underdiagnosed CR-DVT, the sentinel function will have its limitations and the true extent of the medical problem remains uncertain.

The in vitro experiments set up retrospectively to investigate the sudden occurrence of 5 CR-DVT incidents at the PICU should have been performed before the Codan tubing was introduced for medical use. The prospective in vitro protocols run by the hospital engineering department focused on PF pump accuracy and may be too scanty. When nurses start to report about PF pump performance inaccuracy, experiments such as those described previously should be initiated to determine the exact cause for this observation. After all, we are dealing with fluids that will be administered intravenously over medical devices softened with hazardous plasticizers.


   LIMITATIONS
TOP
 ABSTRACT
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 LIMITATIONS
REFERENCES
 
Because of limited financial resources, the sample size was small. Additional investigation via Fourier transform infrared microscopic identification or energy dispersive radiograph analysis could not be budgeted. No screening for immunoglobulin E anti-DEHP was performed in patients.

To meet with these restrictions, all tubings were collected randomly from different sites in the hospital. All PF pumps involved in these experiments were collected randomly from different sites in the hospital and the serial numbers were registered.


   ACKNOWLEDGMENTS
 
This study was approved by the Bioethical Committee of the Free University of Brussels, Belgium; Ethical Committee ref 2005/157.

We thank K. Gillisjans (Free University Brussels) and W. Janssens (Johnson and Johnson Pharmaceutical Research and Development, a Division of Janssen Pharmaceutica N.V.) for their excellent technical assistance. The technical assistance by the group of E. Maynard and colleagues (Pall Industries, Portsmouth, United Kingdom) is highly appreciated, as are the technical reviews by P. Bauwin (Belgian Federal Government: Public Health, Medical Drugs/Medical Devices Department).


   FOOTNOTES
 
Accepted Sep 26, 2006.

Address correspondence to Dirk Danschutter, RN, CCRN, CP, ANP, MSc, Department of Intensive Care, Free University of Brussels, Laarbeeklaan 101, BE 1090 Brussels, Belgium. E-mail: dirk.danschutter{at}vub.ac.be

The authors have indicated they have no financial relationships relevant to this article to disclose.


   REFERENCES
TOP
 ABSTRACT
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 LIMITATIONS
 REFERENCES
 

  1. Trissel LA. Handbook on Injectable Drugs. 13th ed. Bethesda, MD: American Society of Health-System Pharmacists; 2004
  2. Braet F, De Zanger R, Wisse F. Drying cells for SEM, AFM and TEM by hexamethyldisilazane: a study on hepatic endothelial cells. J Microsc.1997;186 :84 –87[ISI][Medline]
  3. Davis R, Frearson M. Mass Spectrometry: Analytical Chemistry by Open Learning. Chichester, England: John Wiley and Sons Ltd; 1987
  4. Huber L. Validation of computerized analytical systems, part 3: installation and operational qualification. LC-GC Magazine.1996; 14;9 :806 –812
  5. Huber L, George S. Diode-array detection in high-performance liquid chromatography. New York, NY: Marcel Dekker; 1993
  6. Wikipedia. Di(2-ethylhexyl)phthalate. Available at: www.en.wikipedia.org/wiki/DEHP. Accessed August 1, 2006
  7. Tri octyl trimellitate. Available at: www.kadakiagroup.com/NIL_EXP1.htm. Accessed December 28, 2006
  8. Calafat A, Needham L, Silva M, Lambert G. Exposure to di-(2-ethylhexyl) phthalate among premature neonates in a neonatal intensive care unit. Pediatrics. 2004;113 :429 –434[Free Full Text]
  9. Fugit JL, Taverdet JL. Decreasing pollution of plasticized PVC packaging: a comparison of three plastic treatments. J Appl Polymer Sci. 2001;80 :1841 –1847[CrossRef]
  10. Lapinskas P, Brown S, Leesnitzer L, et al. Role of PPAR{alpha} in mediating the effects of phthalates and metabolites in the liver. Toxicology. 2005;207 :149 –163[CrossRef][ISI][Medline]
  11. Roberts R, Chevalier S, Hasmall S, James N, Cosulich S, MacDonald N. PPAR{alpha} and the regulation of cell division and apoptosis. Toxicology. 2002;181–182 :167 –170
  12. McKee R. The role of gap junctional intercellular communication in rodent liver tumor induction by phthalates: review of data on selected phthalates and the potential relevance to man. Regul Toxicol Pharmacol. 2000;32 :51 –55[CrossRef][ISI][Medline]
  13. Dybing E. Development and implementation of the IPCS conceptual framework for evaluating mode of action of chemical carcinogens. Toxicology. 2002;181 :121 –125[CrossRef][ISI][Medline]
  14. Bouma K, Schakel D. Migration of phthlates from PVC toys into saliva simulant by dynamic extraction. Food Addit Contam. 2002;19 :602 –610[CrossRef][ISI][Medline]
  15. Kambia K, Dine T, Gressier B, Dupin-Spriet T, Luyckx M, Brunet C. Evaluation of childhood exposure to di(2-ethylhexyl) phthalate from perfusion kits during long term parenteral nutrition. Int J Pharm. 2003;262 :83 –91[CrossRef][ISI][Medline]
  16. Magliozzi R, Nardacci R, Scarsella G, Di Carlo V, Stefanini S. Effects of the plasticizer DEHP on lung of newborn rats: catalase immunocytochemistry and morphometric analysis. Histochem Cell Biol. 2003;120 :1:41 –49
  17. Yang Q, Xie Y, Depierre J. Effects of peroxisome proliferators on the thymus and spleen of mice. Clin Exp Immunol. 2000;122; 2:219 –226
  18. Roth B, Herkenrath P, Lehmann HJ, et al. Di-(2-ethylhexyl)-phthalate as plasticizer in PVC respiratory tubing systems: indications of hazardous effects on pulmonary function in mechanically ventilated preterm infants. Eur J Pediatr. 1988;147 :41 –46[CrossRef][ISI][Medline]
  19. Tsuda H, Naito A, Kim C, Fukamachi K, Nomoto H, Moore M. Carcinogenesis and its modification by environmental endocrine disruptors: in vivo experimental and epidemiological findings. Jpn J Clin Oncol. 2003;33 :259 –270[Abstract/Free Full Text]
  20. Viegi G, Annesi I, Matteelli A. Epidemiology of asthma. Eur Respir Mon. 2003;23 :1 –25
  21. von Mutius E, Sears M. Risk factors for development of asthma. Eur Respir Mon. 2003;23 :57 –73
  22. Shea K; American Academy of Pediatrics Committee on Environmental Health. Pediatric exposure and potential toxicology of phthalate plasticizers. Pediatrics. 2003;111 :1467 –1474[Abstract/Free Full Text]
  23. Bornehag CG, Sundell J, Weschler J, et al. The association between asthma and allergic symptoms in children and phthalates in house dust: a nested case-control study. Environ Health Prospect. 2004;112 :1393 –1397
  24. Kleinsasser N, Harreus UA, Kastenbauer ER, et al. Mono(2-ethylhexyl)phthalate exhibits genotoxic effects in human lymphocytes and mucosal cells of the upper aerodigestive tract in the comet assay. Toxicol Lett. 2004;148 :83 –90[CrossRef][ISI][Medline]
  25. European Commission, Directorate General for Enterprise and Industry. Chemicals: legislation and market restrictions. Available at: http://europa.eu.int/comm/enterprise/chemicals/legislation/markrestr/preparation_en.htm. Accessed April 25, 2006
  26. SGS. Inspection, verification, testing and certification. Rouen, France, 1878. Geneva, Switzerland, 1919. Available at: www.sgs.com. Accessed December 28, 2006
  27. Takehisa H, Naoko F, Masahiko S, et al. Release behavior of DEHP from the PVC tubing used for intravenous administration and the plasticized PVC membrane. Int J Pharm. 2005;297 :30 –37[ISI][Medline]
  28. Yin H, Zhao X, Courtney J, Blass C, West R, Lowe G. Blood interactions with plasticized PVC: relevance of plasticizer selection. J Mater Sci Mater Med. 1999;10 :527 –531[CrossRef][ISI][Medline]
  29. Fischer F, Machleidt C, Rettenmeier A, Kuhlmann U, Mettang T. Plasticizers and inhibition of leucocyte function in vitro. Perit Dial Int. 1998;18 :6:620 –625
  30. Han J, Beeton A, Long P, et al. Plasticizer DEHP release in wet-primed extracorporeal membrane oxygenation (ECMO) circuits. Int J Pharm. 2005;294 :157 –159[CrossRef][ISI][Medline]
  31. Kim W, Yoon C. Roller pump induced tubing wear of PVC and silicone rubber tubing: phase contrast and scanning electron microscopic studies. Artif Organs. 1998;22 :10:892 –897
  32. David R, Moore M, Finney D, Guest D. Chronic toxicity of DEHP in rats. Toxicol Sci. 2000;55 :433 –443[Abstract/Free Full Text]
  33. Subramanian P, Ainsworth P, Cassey J, Phelan D. Lost bits: particle shedding with PVC intravenous administration sets. Pediatr Surg Int. 2002;18 :658 –661[ISI][Medline]
  34. Dewan PA, Ehall H, Edwards G, Middleton D, Terlet J. Plastic particle migration during intravenous infusion assisted by a peristaltic finger pump in an animal model. Pediatr Surg Int. 1997;12 :49 –53[ISI][Medline]
  35. Buchta C, Bittner C, Heinzl H, et al. Transfusion-related exposure to the plasticizer DEHP in patients receiving plateletpheresis concentrates. Transfusion. 2005;45 :798 –802[CrossRef][ISI][Medline]
  36. Wong J, Gill S. Gene expression changes induced in mouse liver by DEHP. Toxicol Appl Pharmacol. 2002;185 :3:180 –196
  37. Gourlay T, Samartzis I, Stefanou D, Taylor K. Inflammatory response of rat and human neutrophils exposed to DEHP plasticized PVC. Artif Organs. 2003;27 :256 –260[CrossRef][ISI][Medline]
  38. Seo K, Kim K, Kim Y, Choi J, Lee K, Choi K. Comparison of oxidative stress and changes of xenobiotic metabolizing enzymes induced by phthalates in rats. Food Chem Toxicol. 2004;42 :107 –114[CrossRef][ISI][Medline]
  39. Ljungvall K, Tienpont B, David F, Magnusson U, Törneke K. Kinetics of orally administered DEHP and its metabolite monoEHP in male pigs. Arch Toxicol. 2004;78 :384 –389[ISI][Medline]
  40. Awal M, Kurohmaru M, Andriana B, Kanai Y, Hayashi Y. MonoEHP induces testicular alterations in male guinea pigs at preburtal stage. Tissue Cell. 2005;37 :167 –175[CrossRef][ISI][Medline]
  41. Turner K, Sharpe R. Environmental oestrogens: present understanding. Rev Reprod. 1997;2 :69 –73[Abstract]
  42. Kavlock R, Boekelheide K, Chapin R, et al. NTP center for the evaluation of risks to human reproduction: phthalates expert panel report on the reproductive and developmental toxicity of di-n-octyl phthalate. Reprod Toxicol. 2002;16 :721 –734[CrossRef][ISI][Medline]
  43. Sharpe R. The oestrogen hypothesis: where do we stand now? Int J Androl. 2003;26 :2 –15[CrossRef][ISI][Medline]
  44. Fisher J. Environmental anti-androgens and male reproductive health: focus on phthalates and testicular dysgenesis syndrome. Reproduction. 2004;127 :305 –315[Abstract/Free Full Text]
  45. Barbour G. Usefulness of a discharge diagnosis of sepsis in detecting iatrogenic infection and quality care problems. Am J Med Qual. 1993;81 :1 –25
  46. Berlowitz D, Brand H, Perkins C. Geriatric syndromes as outcome measures of hospital care: can administrative data be used? J Am Geriatr Soc. 1999;47 :692 –696[ISI][Medline]
  47. Best W, Khuri S, Phelan M, et al. Identifying patient preoperative risk factors and postoperative adverse events in administrative databases: results from the Department of Veterans Affairs National Surgical Quality Improvement Program. J Am Coll Surg. 2002;194 :257 –266[CrossRef][ISI][Medline]
  48. Brennan T, Leape L, Laird N, et al. Incidence of adverse events and negligence in hospitalized patients: results of the Harvard practice study. N Engl J Med. 1991;324 :370 –376[Abstract]
  49. De Paepe L, Quaethove P, Sermeus W, Vleugels A. Klinische indicatoren voor het meten van klinische performantie. Toepassing: het quality indicator project. Acta Hospitalia. 2000;40 :55 –67
  50. Kohn L, Corrigan J, Donaldson M. To Err Is Human: Building a Safer Health System. Washington DC: National Academy Press; 2000
  51. Vincent C, Neale G, Woloshynowych M. Adverse events in British hospitals: preliminary retrospective record review [published correction appears in BMJ. 2001;322:1395). BMJ.2001;322 :517 –519[Abstract/Free Full Text]
  52. Vincent C. Understanding and responding to adverse events. N Engl J Med. 2003;348 :1051 –1056[Free Full Text]
PEDIATRICS (ISSN 1098-4275). ©2007 by the American Academy of Pediatrics




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow P3Rs: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when P3Rs are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow E-mail this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My File Cabinet
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via CrossRef
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Danschutter, D.
Right arrow Articles by Huyghens, L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Danschutter, D.
Right arrow Articles by Huyghens, L.
Related Collections
Right arrow Blood