Central Venous Lines in Children With Lesser Risk Acute Lymphoblastic Leukemia: Optimal Type and Timing of Placement
http://www.100md.com
▲還散笫雖悝◎
the Departments of Pediatrics and Public Health Sciences, Wake Forest University School of Medicine, Winston-Salem, NC
ABSTRACT
PURPOSE: In pediatric patients with acute lymphoblastic leukemia (ALL), the optimal time for central venous line (CVL) insertion and the optimal type of CVL (internal v external) is unclear. This study was undertaken to compare complication rates between early versus late line insertion, and between internal versus external lines in children with lesser risk ALL.
PATIENTS AND METHODS: We performed a retrospective analysis of patients enrolled onto Pediatric Oncology Group (POG) protocol 9201. Data regarding demographics, CVL types and insertion dates, blood counts, and complications were reviewed through week 25 of therapy.
RESULTS: Of 697 patients enrolled onto POG protocol 9201, 362 patients had sufficient data for analysis. When compared to late line placement (> day 15 of induction), early CVL placement ( day 15 of induction) was associated with an increased risk of having a positive blood culture (odds ratio, 2.2; 95% CI, 1.0 to 5.0; P = .05). When compared with internal CVLs ("ports"), external CVLs were associated with a positive blood culture (odds ratio, 3.1; 95% CI, 1.3 to 7.5; P = .01), thrombosis (odds ratio, 3.9; 95% CI, 1.5 to 10.3; P = .006), and CVL removal (odds ratio, 5.6; 95% CI, 2.7 to 11.6; P < .001).
CONCLUSION: In pediatric patients with lesser risk ALL, internal lines (ports) should be the preferred CVL type due to a lower risk of infectious and thrombotic complications. In addition, CVLs placed early in induction are associated with a higher risk of positive blood culture than those placed later in induction.
INTRODUCTION
Central venous lines (CVLs) are widely used in pediatric oncology, though the optimal timing for insertion and type of CVL in pediatric patients with acute lymphoblastic leukemia (ALL) are unknown. Within the pediatric ALL population, we have observed interinstitutional variation in the timing of CVL insertion. Although individual circumstances vary, some institutions insert the CVL at the beginning of remission induction therapy, while others delay CVL placement until the patient has entered remission (typically 4 weeks from initiation of therapy). Although children with ALL are often neutropenic at diagnosis, early placement is often performed because of the added convenience and comfort of having a CVL in place during remission induction. A number of studies have addressed CVL complications in oncology patients (pediatric and adult), but data are sparse regarding optimal timing of CVL placement in pediatric ALL patients.1每15
The present study was undertaken to address the issues of optimal timing and type of CVL placement, relative to risks such as infection and thrombosis, in pediatric patients with lesser risk ALL. We hypothesized that early CVL placement and external CVLs are associated with higher complication rates compared to late CVL placement and internal CVLs.
PATIENTS AND METHODS
Patients
The treatment protocol was opened June 8, 1992 as a pilot study to assess feasibility. A second pilot study and the phase III study (Pediatric Oncology Group [POG] protocol 9201) enrolled a total of 697 patients through November 15, 1999. Eligibility criteria included patients aged 1 to 10 years with B-precursor ALL with an initial peripheral WBC lower than 50,000/mm3 and no overt CNS disease. In addition, patients' leukemia cells must have had simultaneous trisomy of chromosomes 4 and 10 (or, if genetic testing was uninformative, a DNA index of > 1.16) along with the absence of poor-risk lesions [t(1;19), t(9;22), and 11q23 rearrangements]. POG Protocol 9201 was approved by the Wake Forest University School of Medicine institutional review board, and informed consent was obtained for all patients. Similarly, institutional review board approval and informed consent was obtained by the treating institution for each patient registered onto the study. In addition, the present study was approved by the ALL Committee Chair of the POG (B. Camitta, personal communication, 2001).
Therapy
Patients on POG protocol 9201 were treated with a 4-week induction regimen that consisted of prednisone, vincristine (VCR), L-asparaginase, and intrathecal (IT) chemotherapy (initially methotrexate [MTX], hydrocortisone, and cytarabine; later changed to MTX only). This was followed by a 20-week consolidation phase (weeks 5 to 24) consisting of oral 6-mercaptopurine (6-MP), VCR, prednisone, and intravenous and IT MTX. Continuation therapy (weeks 25 to 130) also consisted of prednisone, VCR, MTX (intramuscular and IT), and 6-MP. The therapy for both pilot studies and the phase III study was essentially the same for all patients, with minor changes in IT scheduling and an adjustment of the 6-MP dose after the 33 patients in the first pilot were enrolled. To date, the therapy has been well tolerated with a 7-year event-free survival of more than 85%. (A.R.C., unpublished results).
Data Collection
We retrospectively examined data from patients on POG protocol 9201. Data were abstracted from patient-specific protocol flow sheets regarding demographics and blood counts (WBC, absolute neutrophil count [ANC], hemoglobin, and platelets) at the time of diagnosis and CVL placement, presence of CVL, CVL type (internal v external), and the dates of diagnosis, and CVL placement. CVL placement was defined as "early" if placed on or before treatment day 15, and "late" if placed on treatment day 16 or later. We chose this cutoff point between the early and late groups based on the distribution of known placement times, which was bimodal as reported on the flow sheets. In addition, occurrences of infectious and thrombotic complications were recorded. These data included a positive blood culture, infection of any type, thrombosis or occlusion associated with the CVL, and CVL removal. Data regarding these complications were recorded for weeks 1 through 25 of therapy, by which time the CVL complication rate for all groups had dropped dramatically. From all enrolled patients, a subset of patients was identified as having sufficient data for analysis.
Data Analysis
Statistical analysis was used to assess associations between the three response variables (positive blood culture, thrombosis, and CVL removal) and explanatory variables (age, sex, race, CVL type, early v late CVL placement, blood culture status, and CVL-associated thrombosis status) after accounting for dependence of observations within institutions. ANC at the time of placement was included as an additional explanatory variable to assess the effect of ANC and any change in other effects after adjusting for ANC. All analyses were performed using either simple or multiple logistic regression models as implemented in Stata Statistical Software 6.0 (Stata Corporation, College Station, TX). A backward stepwise approach was used to remove statistically nonsignificant effects of CVL type, CVL placement timing, blood culture status, and thrombosis status from the final multiple logistic regression models. Results are reported as unadjusted or adjusted odds ratios, 95% CIs, and P values. P .05 was considered statistically significant.
RESULTS
Patients
Flow sheets for all 697 patients were reviewed in detail through week 25 of therapy. Toxicity reporting, including infections and thromboses, was required by each treating institution. Reporting data on the presence or absence of CVL, type of CVL, and date of CVL placement was not required, but many patients' flow sheets did contain these data. Only the patients whose flow sheets contained data from these items were included in this report. Of 697 protocol-registered patients, 362 patients (52%) had sufficient data available for analysis (Fig 1). Patient characteristics are listed in Table 1.
Of the 362 eligible patients, all achieved remission after 4 weeks of induction chemotherapy. One hundred eighty-two patients (50%) had a documented fever during the first 25 weeks of therapy. During this time period, 22 patients (6%) had a documented thrombosis associated with their CVL. Fifty patients (14%) had their CVL removed. Sixty patients (17%) had a positive blood culture, and 85 patients (23%) had a documented infection of any type (including positive blood cultures). All positive blood cultures were bacterial in etiology except in two patients who had fungal isolates: one of those two patients had an external CVL placed early, and the other had a port placed early. Two hundred eighty-five patients had data on both CVL type and timing of CVL placement. The majority of patients (53%) had an internal line (port) placed early. The primary outcomes assessed for this study were positive blood culture, CVL-associated thrombus, and CVL removal for any reason. There were no deaths associated with CVL placements or CVL complications.
Internal Versus External CVLs
Of the 362 patients known from flow sheets to have had a CVL, 308 patients had data available regarding internal versus external CVLs. Of these 308 patients, 245 patients (80%) had an internal line (port) and 63 patients (20%) had an external line. Of those with an external CVL, no information was reported as to whether the CVL was tunneled or nontunneled. Patients who had no known CVL or a CVL of unknown type (internal or external) were excluded from this analysis. External CVLs were more likely than internal CVLs to be associated with a positive blood culture (odds ratio, 3.1; 95% CI, 1.3 to 7.5; P = .01), thrombosis (odds ratio, 3.9; 95% CI, 1.5 to 10.3; P = .006), and CVL removal (odds ratio, 5.6; 95% CI, 2.7 to 11.6; P < .001), even when adjusted for age, sex, race, institution, and timing of placement (Table 2). These risks were also independent of ANC at the time of CVL placement (data not shown).
Timing of CVL Placement
Of the 362 patients, 339 patients had sufficient data regarding timing of CVL placement. Of these 339 patients, 234 patients (69%) had a CVL placed early, (defined as on or before day 15 of induction therapy; range, day 每3 to day 15); 105 patients (31%) had CVLs placed late (defined as day 16 or later of therapy; range, day 16 to day 133). By multivariate analysis, when compared to late line placement, early CVL placement was associated with an increased risk of having a positive blood culture (odds ratio, 2.2; 95% CI, 1.0 to 5.0; P = .05), even when adjusted for age, sex, race, institution, and type of CVL (internal v external; Table 2). Early CVL placement was not independently associated with thrombosis or CVL removal. The association of a positive culture with early CVL placement was also independent of ANC at the time of CVL placement (data not shown).
ANC
The mean ANC at diagnosis for all patients was 679/mm3 (standard deviation [SD], 1,053/mm3; range, 0/mm3 to 9,396/mm3); the mean ANC at time of CVL placement was 1,291/mm3 (SD, 1,999/mm3; range, 0/mm3 to 16,498/mm3); and the mean ANC at the time of remission was 3,162/mm3 (SD, 2,950/mm3; range, 35/mm3 to 16,544/mm3). The mean ANC at the time of CVL placement was significantly lower in the CVLs that were removed (mean, 601/mm3) than in those that were not removed (mean = 1,391/mm3; P = .004). Neutropenia was not associated with CVL thrombosis. Hemoglobin and platelet counts were not associated with CVL complications.
DISCUSSION
This study defines the independent contribution of CVL type to the risk of infectious and thrombotic complications among a large group of pediatric patients with lesser risk ALL, all of whom received essentially identical treatment. We found that, when compared with external CVLs, internal CVLs had a lower rate of infection, thrombosis, and need for removal. This finding was independent of the timing of CVL placement (early v late) and ANC at the time of CVL placement. We also found that CVLs placed later in induction therapy (on or after day 16 of induction) were associated with a lower rate of positive blood culture than those placed early (day 每3 to day 15).
The literature comparing CVL type (internal v external) and the risk of infection is controversial,1,15每19 but most institutions currently favor ports over external CVLs. In our study, 17% of assessable patients had external CVLs placed. Our results support a lower rate of infection for internal CVLs. Although the exact reasons are unknown, infection risk is likely related to several characteristics of external CVLs: a permanent breach in the skin integrity, increased manipulation (both intentional and unintentional), and the practice of more frequent flushing. The issues of perioperative antibiotic prophylaxis14 and antibiotic-impregnated catheters20每22 have been raised but not well studied in this population. Other strategies to prevent CVL complications have recently been reviewed,23 as have conditions associated with CVL infection.24
Our data are in agreement with the findings of Ingram et al18 that external CVLs have more occlusion or thrombosis than internal CVLs. In that study, catheter diameter was suggested as one possibility for the observed difference between external and internal CVLs. Like infection risk, the amount of manipulation required of external CVLs may be a contributing factor to thrombosis risk. A recent study showed that CVL thrombosis risk is influenced by location of the CVL and insertion technique,6 neither of which were assessed in our study. The overall risk of thrombosis in children with ALL is exacerbated by L-asparaginase therapy.25 L-asparaginase therapy would explain any increased risk of thrombosis for CVLs placed early, but it would not explain the increased risk for external CVLs compared to internal CVLs. While CVL-related thrombosis has been well documented in children,12 primary thrombotic prophylaxis is not current standard practice in children with ALL, because no controlled trial has been published documenting safety and efficacy.
Evidence indicates that neutropenia is an independent risk factor for infection related to CVLs, and for sepsis of unknown origin.14,26 Shaul et al14 indicate that neutropenia and the failure to administer prophylactic antibiotics are risk factors for the development of early CVL infection in pediatric patients. To avoid early infection and possible premature CVL removal, the authors recommend that placement of permanent CVLs be postponed until the ANC is higher than 1,000/mm3. Perioperative antibiotic administration is also recommended.14 Our data support neutropenia as a contributing factor to the risk of CVL removal, though it is difficult to determine an independent effect of ANC because the patients' ANC generally improved during induction.
To our knowledge only one previous study19 has assessed the timing of CVL placement as a risk factor in a similar population. In that study, "early" was defined as during the first 3 weeks of induction therapy, and was not found to be a risk factor for infection. However, the study reported an unusually high rate of infections, was relatively small (n = 148), and was from a single institution.19 For some patients, for example, those with difficult intravenous access, the benefits of inserting a port early in induction therapy may outweigh the risks. The benefits of early CVL placement include reduced pain of phlebotomy and reliable access for medication administration (including at least one vesicant, vincristine) during the first 4 weeks of therapy. Certainly, the risks and benefits for individual patients must be considered. In general, however, the results of our study show that external CVLs and early placement of CVLs are associated with a higher risk of complications compared with internal CVLs and late placement of CVLs.
This study is limited by its retrospective nature and the fact that it did not include all patients on POG protocol 9201. Some patients had missing data regarding CVL type and timing of placement. Thus, it is possible that a selection bias exists in these data, with patients who had complications more likely to have data than those who did not have complications. It is also possible that thrombolytic agents such as urokinase or alteplase (t-PA) were used in some patients but not recorded on the data flow sheets. Thus, the rate of thrombosis may be underestimated in this study.
Data regarding the complications were recorded only through week 25 of therapy. However, it is our opinion that CVL complications beyond week 25 are unlikely to be related to the timing of insertion. This theory is supported by the rare occurrence of CVL complications at week 25 in our patients. Patients who had their CVLs placed early obviously had a longer time period to develop complications, and our results suggest that this extra time (during induction therapy) is the high-risk period. Therefore, we believe that CVLs placed early are at higher risk of complication due to the additional CVL exposure time during the first weeks of therapy. Although it could be argued that the distinction between early and late placement (day 15 of induction) is arbitrary, this cutoff was chosen based on the bimodal distribution of the data. Because it is generally hoped that a CVL will last throughout treatment (130 weeks in POG 9201), the results of this study should be considered when scheduling CVL placement in this patient population.
Another limitation of this study is that only lesser risk ALL patients were included. As remission rates are similar for standard and even high-risk ALL patients, it is possible that these results may be applicable to other pediatric patients with ALL. It should be emphasized, however, that patients receiving more intensive therapy were not included in this study.
In summary, these data show that for pediatric patients with lesser risk ALL, CVLs placed early and external CVLs are associated with higher complication rates. We recommend that internal CVLs (ports) should be the line of choice. To minimize infectious and thrombotic complications, consideration should be given to delaying port placement until the end of remission induction therapy. Prospective studies are needed to further clarify optimal timing for CVL placement in these and other ALL patients.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
NOTES
Presented in part at the 15th Annual Meeting of the American Society of Pediatric Hematology/Oncology and Pediatric Academic Society, Baltimore, MD, May 2-5, 2002.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Wiener ES, McGuire P, Stolar CJ, et al: The CCSG prospective study of venous access devices: An analysis of insertions and causes for removal. J Pediatr Surg 27:155每163, 1992
Johnson EM, Saltzman DA, Suh G, et al: Complications and risks of central venous catheter placement in children. Surgery 124:911每916, 1998
van Hoff J, Berg AT, Seashore JH: The effect of right atrial catheters on infectious complications of chemotherapy in children. J Clin Oncol 8:1255每1262, 1990
Cairo MS, Spooner S, Sowden L, et al: Long-term use of indwelling multipurpose silastic catheters in pediatric cancer patients treated with aggressive chemotherapy. J Clin Oncol 4:784每788, 1986
Rackoff WR, Ge J, Sather HN, et al: Central venous catheter use and the risk of infection in children with acute lymphoblastic leukemia: A report from the Children's Cancer Group. J Pediatr Hematol Oncol 21:260每267, 1999
Male C, Chait P, Andrew M, et al: Central venous line-related thrombosis in children: Association with central venous line location and insertion technique. Blood 101:4273每4278, 2003
Dawson S, Pai MK, Smith S, et al: Right atrial catheters in children with cancer: A decade of experience in the use of tunnelled, exteriorized devices at a single institution. Am J Pediatr Hematol Oncol 13:126每129, 1991
Kappers-Klunne MC, Degener JE, Stijnen T, et al: Complications from long-term indwelling central venous catheters in hematologic patients with special reference to infection. Cancer 64:1747每1752, 1989
Abrahm JL, Mullen JL: A prospective study of prolonged central venous access in leukemia. JAMA 248:2868每2873, 1982
Reilly JJ Jr, Steed DL, Ritter PS: Indwelling venous access catheters in patients with acute leukemia. Cancer 53:219每223, 1984
Gorelick MH, Owen WC, Seibel NL, et al: Lack of association between neutropenia and the incidence of bacteremia associated with indwelling central venous catheters in febrile pediatric cancer patients. Pediatr Infect Dis J 10:506每510, 1991
Massicotte MP, Dix D, Monagle P, et al: Central venous catheter related thrombosis in children: Analysis of the Canadian Registry of Venous Thromboembolic Complications. J Pediatr 133:770每776, 1998
Landoy Z, Rotstein C, Lucey J, et al: Hickman-Broviac catheter use in cancer patients. J Surg Oncol 26:215每218, 1984
Shaul DB, Scheer B, Rokhsar S, et al: Risk factors for early infection of central venous catheters in pediatric patients. J Am Coll Surg 186:654每658, 1998
Wurzel CL, Halom K, Feldman JG, et al: Infection rates of Broviac-Hickman catheters and implantable venous devices. Am J Dis Child 142:536每540, 1988
Mueller BU, Skelton J, Callender DP, et al: A prospective randomized trial comparing the infectious and noninfectious complications of an externalized catheter versus a subcutaneously implanted device in cancer patients. J Clin Oncol 10:1943每1948, 1992
Decker MD, Edwards KM: Central venous catheter infections. Pediatr Clin North Am 35:579每612, 1988
Ingram J, Weitzman S, Greenberg ML, et al: Complications of indwelling venous access lines in the pediatric hematology patient: A prospective comparison of external venous catheters and subcutaneous ports. Am J Pediatr Hematol Oncol 13:130每136, 1991
Abbas AA, Fryer CJ, Paltiel C, et al: Factors influencing central line infections in children with acute lymphoblastic leukemia: Results of a single institutional study. Pediatr Blood Cancer 42:325每331, 2004
Raad I, Darouiche R, Dupuis J, et al: Central venous catheters coated with minocycline and rifampin for the prevention of catheter-related colonization and bloodstream infections: A randomized, double-blind trial〞The Texas Medical Center Catheter Study Group. Ann Intern Med 127:267每274, 1997
Maki DG, Stolz SM, Wheeler S, et al: Prevention of central venous catheter-related bloodstream infection by use of an antiseptic-impregnated catheter: A randomized, controlled trial. Ann Intern Med 127:257每266, 1997
Hanna H, Benjamin R, Chatzinikolaou I, et al: Long-term silicone central venous catheters impregnated with minocycline and rifampin decrease rates of catheter-related bloodstream infection in cancer patients: A prospective randomized clinical trial. J Clin Oncol 22:3163每3171, 2004
McGee DC, Gould MK: Preventing complications of central venous catheterization. N Engl J Med 348:1123每1133, 2003
Castagnola E, Molinari AC, Fratino G, et al: Conditions associated with infections of indwelling central venous catheters in cancer patients: A summary. Br J Haematol 121:233每239, 2003
Mitchell LG, Andrew M, Hanna K, et al: A prospective cohort study determining the prevalence of thrombotic events in children with acute lymphoblastic leukemia and a central venous line who are treated with L-asparaginase: Results of the Prophylactic Antithrombin Replacement in Kids with Acute Lymphoblastic Leukemia Treated with Asparaginase (PARKAA) Study. Cancer 97:508每516, 2003
Howell PB, Walters PE, Donowitz GR, et al: Risk factors for infection of adult patients with cancer who have tunnelled central venous catheters. Cancer 75:1367每1375, 1995(Thomas W. McLean, Christe)
ABSTRACT
PURPOSE: In pediatric patients with acute lymphoblastic leukemia (ALL), the optimal time for central venous line (CVL) insertion and the optimal type of CVL (internal v external) is unclear. This study was undertaken to compare complication rates between early versus late line insertion, and between internal versus external lines in children with lesser risk ALL.
PATIENTS AND METHODS: We performed a retrospective analysis of patients enrolled onto Pediatric Oncology Group (POG) protocol 9201. Data regarding demographics, CVL types and insertion dates, blood counts, and complications were reviewed through week 25 of therapy.
RESULTS: Of 697 patients enrolled onto POG protocol 9201, 362 patients had sufficient data for analysis. When compared to late line placement (> day 15 of induction), early CVL placement ( day 15 of induction) was associated with an increased risk of having a positive blood culture (odds ratio, 2.2; 95% CI, 1.0 to 5.0; P = .05). When compared with internal CVLs ("ports"), external CVLs were associated with a positive blood culture (odds ratio, 3.1; 95% CI, 1.3 to 7.5; P = .01), thrombosis (odds ratio, 3.9; 95% CI, 1.5 to 10.3; P = .006), and CVL removal (odds ratio, 5.6; 95% CI, 2.7 to 11.6; P < .001).
CONCLUSION: In pediatric patients with lesser risk ALL, internal lines (ports) should be the preferred CVL type due to a lower risk of infectious and thrombotic complications. In addition, CVLs placed early in induction are associated with a higher risk of positive blood culture than those placed later in induction.
INTRODUCTION
Central venous lines (CVLs) are widely used in pediatric oncology, though the optimal timing for insertion and type of CVL in pediatric patients with acute lymphoblastic leukemia (ALL) are unknown. Within the pediatric ALL population, we have observed interinstitutional variation in the timing of CVL insertion. Although individual circumstances vary, some institutions insert the CVL at the beginning of remission induction therapy, while others delay CVL placement until the patient has entered remission (typically 4 weeks from initiation of therapy). Although children with ALL are often neutropenic at diagnosis, early placement is often performed because of the added convenience and comfort of having a CVL in place during remission induction. A number of studies have addressed CVL complications in oncology patients (pediatric and adult), but data are sparse regarding optimal timing of CVL placement in pediatric ALL patients.1每15
The present study was undertaken to address the issues of optimal timing and type of CVL placement, relative to risks such as infection and thrombosis, in pediatric patients with lesser risk ALL. We hypothesized that early CVL placement and external CVLs are associated with higher complication rates compared to late CVL placement and internal CVLs.
PATIENTS AND METHODS
Patients
The treatment protocol was opened June 8, 1992 as a pilot study to assess feasibility. A second pilot study and the phase III study (Pediatric Oncology Group [POG] protocol 9201) enrolled a total of 697 patients through November 15, 1999. Eligibility criteria included patients aged 1 to 10 years with B-precursor ALL with an initial peripheral WBC lower than 50,000/mm3 and no overt CNS disease. In addition, patients' leukemia cells must have had simultaneous trisomy of chromosomes 4 and 10 (or, if genetic testing was uninformative, a DNA index of > 1.16) along with the absence of poor-risk lesions [t(1;19), t(9;22), and 11q23 rearrangements]. POG Protocol 9201 was approved by the Wake Forest University School of Medicine institutional review board, and informed consent was obtained for all patients. Similarly, institutional review board approval and informed consent was obtained by the treating institution for each patient registered onto the study. In addition, the present study was approved by the ALL Committee Chair of the POG (B. Camitta, personal communication, 2001).
Therapy
Patients on POG protocol 9201 were treated with a 4-week induction regimen that consisted of prednisone, vincristine (VCR), L-asparaginase, and intrathecal (IT) chemotherapy (initially methotrexate [MTX], hydrocortisone, and cytarabine; later changed to MTX only). This was followed by a 20-week consolidation phase (weeks 5 to 24) consisting of oral 6-mercaptopurine (6-MP), VCR, prednisone, and intravenous and IT MTX. Continuation therapy (weeks 25 to 130) also consisted of prednisone, VCR, MTX (intramuscular and IT), and 6-MP. The therapy for both pilot studies and the phase III study was essentially the same for all patients, with minor changes in IT scheduling and an adjustment of the 6-MP dose after the 33 patients in the first pilot were enrolled. To date, the therapy has been well tolerated with a 7-year event-free survival of more than 85%. (A.R.C., unpublished results).
Data Collection
We retrospectively examined data from patients on POG protocol 9201. Data were abstracted from patient-specific protocol flow sheets regarding demographics and blood counts (WBC, absolute neutrophil count [ANC], hemoglobin, and platelets) at the time of diagnosis and CVL placement, presence of CVL, CVL type (internal v external), and the dates of diagnosis, and CVL placement. CVL placement was defined as "early" if placed on or before treatment day 15, and "late" if placed on treatment day 16 or later. We chose this cutoff point between the early and late groups based on the distribution of known placement times, which was bimodal as reported on the flow sheets. In addition, occurrences of infectious and thrombotic complications were recorded. These data included a positive blood culture, infection of any type, thrombosis or occlusion associated with the CVL, and CVL removal. Data regarding these complications were recorded for weeks 1 through 25 of therapy, by which time the CVL complication rate for all groups had dropped dramatically. From all enrolled patients, a subset of patients was identified as having sufficient data for analysis.
Data Analysis
Statistical analysis was used to assess associations between the three response variables (positive blood culture, thrombosis, and CVL removal) and explanatory variables (age, sex, race, CVL type, early v late CVL placement, blood culture status, and CVL-associated thrombosis status) after accounting for dependence of observations within institutions. ANC at the time of placement was included as an additional explanatory variable to assess the effect of ANC and any change in other effects after adjusting for ANC. All analyses were performed using either simple or multiple logistic regression models as implemented in Stata Statistical Software 6.0 (Stata Corporation, College Station, TX). A backward stepwise approach was used to remove statistically nonsignificant effects of CVL type, CVL placement timing, blood culture status, and thrombosis status from the final multiple logistic regression models. Results are reported as unadjusted or adjusted odds ratios, 95% CIs, and P values. P .05 was considered statistically significant.
RESULTS
Patients
Flow sheets for all 697 patients were reviewed in detail through week 25 of therapy. Toxicity reporting, including infections and thromboses, was required by each treating institution. Reporting data on the presence or absence of CVL, type of CVL, and date of CVL placement was not required, but many patients' flow sheets did contain these data. Only the patients whose flow sheets contained data from these items were included in this report. Of 697 protocol-registered patients, 362 patients (52%) had sufficient data available for analysis (Fig 1). Patient characteristics are listed in Table 1.
Of the 362 eligible patients, all achieved remission after 4 weeks of induction chemotherapy. One hundred eighty-two patients (50%) had a documented fever during the first 25 weeks of therapy. During this time period, 22 patients (6%) had a documented thrombosis associated with their CVL. Fifty patients (14%) had their CVL removed. Sixty patients (17%) had a positive blood culture, and 85 patients (23%) had a documented infection of any type (including positive blood cultures). All positive blood cultures were bacterial in etiology except in two patients who had fungal isolates: one of those two patients had an external CVL placed early, and the other had a port placed early. Two hundred eighty-five patients had data on both CVL type and timing of CVL placement. The majority of patients (53%) had an internal line (port) placed early. The primary outcomes assessed for this study were positive blood culture, CVL-associated thrombus, and CVL removal for any reason. There were no deaths associated with CVL placements or CVL complications.
Internal Versus External CVLs
Of the 362 patients known from flow sheets to have had a CVL, 308 patients had data available regarding internal versus external CVLs. Of these 308 patients, 245 patients (80%) had an internal line (port) and 63 patients (20%) had an external line. Of those with an external CVL, no information was reported as to whether the CVL was tunneled or nontunneled. Patients who had no known CVL or a CVL of unknown type (internal or external) were excluded from this analysis. External CVLs were more likely than internal CVLs to be associated with a positive blood culture (odds ratio, 3.1; 95% CI, 1.3 to 7.5; P = .01), thrombosis (odds ratio, 3.9; 95% CI, 1.5 to 10.3; P = .006), and CVL removal (odds ratio, 5.6; 95% CI, 2.7 to 11.6; P < .001), even when adjusted for age, sex, race, institution, and timing of placement (Table 2). These risks were also independent of ANC at the time of CVL placement (data not shown).
Timing of CVL Placement
Of the 362 patients, 339 patients had sufficient data regarding timing of CVL placement. Of these 339 patients, 234 patients (69%) had a CVL placed early, (defined as on or before day 15 of induction therapy; range, day 每3 to day 15); 105 patients (31%) had CVLs placed late (defined as day 16 or later of therapy; range, day 16 to day 133). By multivariate analysis, when compared to late line placement, early CVL placement was associated with an increased risk of having a positive blood culture (odds ratio, 2.2; 95% CI, 1.0 to 5.0; P = .05), even when adjusted for age, sex, race, institution, and type of CVL (internal v external; Table 2). Early CVL placement was not independently associated with thrombosis or CVL removal. The association of a positive culture with early CVL placement was also independent of ANC at the time of CVL placement (data not shown).
ANC
The mean ANC at diagnosis for all patients was 679/mm3 (standard deviation [SD], 1,053/mm3; range, 0/mm3 to 9,396/mm3); the mean ANC at time of CVL placement was 1,291/mm3 (SD, 1,999/mm3; range, 0/mm3 to 16,498/mm3); and the mean ANC at the time of remission was 3,162/mm3 (SD, 2,950/mm3; range, 35/mm3 to 16,544/mm3). The mean ANC at the time of CVL placement was significantly lower in the CVLs that were removed (mean, 601/mm3) than in those that were not removed (mean = 1,391/mm3; P = .004). Neutropenia was not associated with CVL thrombosis. Hemoglobin and platelet counts were not associated with CVL complications.
DISCUSSION
This study defines the independent contribution of CVL type to the risk of infectious and thrombotic complications among a large group of pediatric patients with lesser risk ALL, all of whom received essentially identical treatment. We found that, when compared with external CVLs, internal CVLs had a lower rate of infection, thrombosis, and need for removal. This finding was independent of the timing of CVL placement (early v late) and ANC at the time of CVL placement. We also found that CVLs placed later in induction therapy (on or after day 16 of induction) were associated with a lower rate of positive blood culture than those placed early (day 每3 to day 15).
The literature comparing CVL type (internal v external) and the risk of infection is controversial,1,15每19 but most institutions currently favor ports over external CVLs. In our study, 17% of assessable patients had external CVLs placed. Our results support a lower rate of infection for internal CVLs. Although the exact reasons are unknown, infection risk is likely related to several characteristics of external CVLs: a permanent breach in the skin integrity, increased manipulation (both intentional and unintentional), and the practice of more frequent flushing. The issues of perioperative antibiotic prophylaxis14 and antibiotic-impregnated catheters20每22 have been raised but not well studied in this population. Other strategies to prevent CVL complications have recently been reviewed,23 as have conditions associated with CVL infection.24
Our data are in agreement with the findings of Ingram et al18 that external CVLs have more occlusion or thrombosis than internal CVLs. In that study, catheter diameter was suggested as one possibility for the observed difference between external and internal CVLs. Like infection risk, the amount of manipulation required of external CVLs may be a contributing factor to thrombosis risk. A recent study showed that CVL thrombosis risk is influenced by location of the CVL and insertion technique,6 neither of which were assessed in our study. The overall risk of thrombosis in children with ALL is exacerbated by L-asparaginase therapy.25 L-asparaginase therapy would explain any increased risk of thrombosis for CVLs placed early, but it would not explain the increased risk for external CVLs compared to internal CVLs. While CVL-related thrombosis has been well documented in children,12 primary thrombotic prophylaxis is not current standard practice in children with ALL, because no controlled trial has been published documenting safety and efficacy.
Evidence indicates that neutropenia is an independent risk factor for infection related to CVLs, and for sepsis of unknown origin.14,26 Shaul et al14 indicate that neutropenia and the failure to administer prophylactic antibiotics are risk factors for the development of early CVL infection in pediatric patients. To avoid early infection and possible premature CVL removal, the authors recommend that placement of permanent CVLs be postponed until the ANC is higher than 1,000/mm3. Perioperative antibiotic administration is also recommended.14 Our data support neutropenia as a contributing factor to the risk of CVL removal, though it is difficult to determine an independent effect of ANC because the patients' ANC generally improved during induction.
To our knowledge only one previous study19 has assessed the timing of CVL placement as a risk factor in a similar population. In that study, "early" was defined as during the first 3 weeks of induction therapy, and was not found to be a risk factor for infection. However, the study reported an unusually high rate of infections, was relatively small (n = 148), and was from a single institution.19 For some patients, for example, those with difficult intravenous access, the benefits of inserting a port early in induction therapy may outweigh the risks. The benefits of early CVL placement include reduced pain of phlebotomy and reliable access for medication administration (including at least one vesicant, vincristine) during the first 4 weeks of therapy. Certainly, the risks and benefits for individual patients must be considered. In general, however, the results of our study show that external CVLs and early placement of CVLs are associated with a higher risk of complications compared with internal CVLs and late placement of CVLs.
This study is limited by its retrospective nature and the fact that it did not include all patients on POG protocol 9201. Some patients had missing data regarding CVL type and timing of placement. Thus, it is possible that a selection bias exists in these data, with patients who had complications more likely to have data than those who did not have complications. It is also possible that thrombolytic agents such as urokinase or alteplase (t-PA) were used in some patients but not recorded on the data flow sheets. Thus, the rate of thrombosis may be underestimated in this study.
Data regarding the complications were recorded only through week 25 of therapy. However, it is our opinion that CVL complications beyond week 25 are unlikely to be related to the timing of insertion. This theory is supported by the rare occurrence of CVL complications at week 25 in our patients. Patients who had their CVLs placed early obviously had a longer time period to develop complications, and our results suggest that this extra time (during induction therapy) is the high-risk period. Therefore, we believe that CVLs placed early are at higher risk of complication due to the additional CVL exposure time during the first weeks of therapy. Although it could be argued that the distinction between early and late placement (day 15 of induction) is arbitrary, this cutoff was chosen based on the bimodal distribution of the data. Because it is generally hoped that a CVL will last throughout treatment (130 weeks in POG 9201), the results of this study should be considered when scheduling CVL placement in this patient population.
Another limitation of this study is that only lesser risk ALL patients were included. As remission rates are similar for standard and even high-risk ALL patients, it is possible that these results may be applicable to other pediatric patients with ALL. It should be emphasized, however, that patients receiving more intensive therapy were not included in this study.
In summary, these data show that for pediatric patients with lesser risk ALL, CVLs placed early and external CVLs are associated with higher complication rates. We recommend that internal CVLs (ports) should be the line of choice. To minimize infectious and thrombotic complications, consideration should be given to delaying port placement until the end of remission induction therapy. Prospective studies are needed to further clarify optimal timing for CVL placement in these and other ALL patients.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
NOTES
Presented in part at the 15th Annual Meeting of the American Society of Pediatric Hematology/Oncology and Pediatric Academic Society, Baltimore, MD, May 2-5, 2002.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Wiener ES, McGuire P, Stolar CJ, et al: The CCSG prospective study of venous access devices: An analysis of insertions and causes for removal. J Pediatr Surg 27:155每163, 1992
Johnson EM, Saltzman DA, Suh G, et al: Complications and risks of central venous catheter placement in children. Surgery 124:911每916, 1998
van Hoff J, Berg AT, Seashore JH: The effect of right atrial catheters on infectious complications of chemotherapy in children. J Clin Oncol 8:1255每1262, 1990
Cairo MS, Spooner S, Sowden L, et al: Long-term use of indwelling multipurpose silastic catheters in pediatric cancer patients treated with aggressive chemotherapy. J Clin Oncol 4:784每788, 1986
Rackoff WR, Ge J, Sather HN, et al: Central venous catheter use and the risk of infection in children with acute lymphoblastic leukemia: A report from the Children's Cancer Group. J Pediatr Hematol Oncol 21:260每267, 1999
Male C, Chait P, Andrew M, et al: Central venous line-related thrombosis in children: Association with central venous line location and insertion technique. Blood 101:4273每4278, 2003
Dawson S, Pai MK, Smith S, et al: Right atrial catheters in children with cancer: A decade of experience in the use of tunnelled, exteriorized devices at a single institution. Am J Pediatr Hematol Oncol 13:126每129, 1991
Kappers-Klunne MC, Degener JE, Stijnen T, et al: Complications from long-term indwelling central venous catheters in hematologic patients with special reference to infection. Cancer 64:1747每1752, 1989
Abrahm JL, Mullen JL: A prospective study of prolonged central venous access in leukemia. JAMA 248:2868每2873, 1982
Reilly JJ Jr, Steed DL, Ritter PS: Indwelling venous access catheters in patients with acute leukemia. Cancer 53:219每223, 1984
Gorelick MH, Owen WC, Seibel NL, et al: Lack of association between neutropenia and the incidence of bacteremia associated with indwelling central venous catheters in febrile pediatric cancer patients. Pediatr Infect Dis J 10:506每510, 1991
Massicotte MP, Dix D, Monagle P, et al: Central venous catheter related thrombosis in children: Analysis of the Canadian Registry of Venous Thromboembolic Complications. J Pediatr 133:770每776, 1998
Landoy Z, Rotstein C, Lucey J, et al: Hickman-Broviac catheter use in cancer patients. J Surg Oncol 26:215每218, 1984
Shaul DB, Scheer B, Rokhsar S, et al: Risk factors for early infection of central venous catheters in pediatric patients. J Am Coll Surg 186:654每658, 1998
Wurzel CL, Halom K, Feldman JG, et al: Infection rates of Broviac-Hickman catheters and implantable venous devices. Am J Dis Child 142:536每540, 1988
Mueller BU, Skelton J, Callender DP, et al: A prospective randomized trial comparing the infectious and noninfectious complications of an externalized catheter versus a subcutaneously implanted device in cancer patients. J Clin Oncol 10:1943每1948, 1992
Decker MD, Edwards KM: Central venous catheter infections. Pediatr Clin North Am 35:579每612, 1988
Ingram J, Weitzman S, Greenberg ML, et al: Complications of indwelling venous access lines in the pediatric hematology patient: A prospective comparison of external venous catheters and subcutaneous ports. Am J Pediatr Hematol Oncol 13:130每136, 1991
Abbas AA, Fryer CJ, Paltiel C, et al: Factors influencing central line infections in children with acute lymphoblastic leukemia: Results of a single institutional study. Pediatr Blood Cancer 42:325每331, 2004
Raad I, Darouiche R, Dupuis J, et al: Central venous catheters coated with minocycline and rifampin for the prevention of catheter-related colonization and bloodstream infections: A randomized, double-blind trial〞The Texas Medical Center Catheter Study Group. Ann Intern Med 127:267每274, 1997
Maki DG, Stolz SM, Wheeler S, et al: Prevention of central venous catheter-related bloodstream infection by use of an antiseptic-impregnated catheter: A randomized, controlled trial. Ann Intern Med 127:257每266, 1997
Hanna H, Benjamin R, Chatzinikolaou I, et al: Long-term silicone central venous catheters impregnated with minocycline and rifampin decrease rates of catheter-related bloodstream infection in cancer patients: A prospective randomized clinical trial. J Clin Oncol 22:3163每3171, 2004
McGee DC, Gould MK: Preventing complications of central venous catheterization. N Engl J Med 348:1123每1133, 2003
Castagnola E, Molinari AC, Fratino G, et al: Conditions associated with infections of indwelling central venous catheters in cancer patients: A summary. Br J Haematol 121:233每239, 2003
Mitchell LG, Andrew M, Hanna K, et al: A prospective cohort study determining the prevalence of thrombotic events in children with acute lymphoblastic leukemia and a central venous line who are treated with L-asparaginase: Results of the Prophylactic Antithrombin Replacement in Kids with Acute Lymphoblastic Leukemia Treated with Asparaginase (PARKAA) Study. Cancer 97:508每516, 2003
Howell PB, Walters PE, Donowitz GR, et al: Risk factors for infection of adult patients with cancer who have tunnelled central venous catheters. Cancer 75:1367每1375, 1995(Thomas W. McLean, Christe)