Changes in Central Line–Associated Bloodstream Infection Rates Among Children With Immune Compromised Conditions: An 11-Year Review

Abstract

This article describes changes in central line–associated bloodstream infection (CLABSI) rates among common causative organisms over an 11-year period on a pediatric inpatient unit prior to and following CLABSI reduction strategies. The setting for this descriptive cohort design study was a 32-bed inpatient unit in a tertiary pediatric hospital serving children with immune compromised conditions, including cancer and recipients of hematopoietic stem cell and solid organ transplants. Between January 2006 and December 2016, 265 CLABSIs involving 189 patients were reported. Data were organized into three time periods: 5-year preintervention baseline (2006-2010), implementation of maintenance care bundles (2011-2012), and addition of formalized supportive care practices to the maintenance care bundles (2013-2016). Organisms were categorized into four groups based on the National Health Safety Network organism list. Time-by-class Poisson regression models evaluated changes in CLABSI rates. Characteristics of patients who developed CLABSIs were unchanged. Infections occurred most frequently among patients with hematologic malignancies and neutropenia. Significant log rate decreases in CLABSI rates were observed with the implementation of maintenance care bundles plus enhanced supportive cares compared to preintervention baseline for the following organisms: (1) common commensal organisms (−1.05, p = .005), (2) mucosal barrier injury (MBI) organisms common to the mouth (−.708, p = .007), and (3) other noncommensal/non-MBI pathogens (−.77, p = .005). Rates were unchanged for MBI organisms common to the lower gastrointestinal tract. Central line maintenance care bundles and formalized supportive care practices resulted in sustained decreased CLABSI rates. Additional interventions are needed to reduce CLABSIs involving MBI-associated organisms common to the lower gastrointestinal tract.

Keywords

infection prevention central line–associated bloodstream infection pediatric oncology central line maintenance care bundles

Introduction

Although central line–associated bloodstream infections (CLABSIs) among children with cancer and other immune compromised conditions have decreased with the implementation of best practice bundles, these infections continue to occur (Bundy et al., 2014; Rinke et al., 2012). For children with cancer, the attributable cost of a CLABSI may be as much as $70,000 with an additional hospital length-of-stay of 21 days (Wilson, Rafferty, Deeter, Comito, Hollenbeak, 2014). Reported mortality rates for children with cancer and those undergoing hematopoietic stem cell transplant who develop bloodstream infections range from 1.5% to 10% (Bochennek et al., 2016; Celebi et al., 2013). Likewise, patients undergoing solid organ transplant face a heightened risk of bloodstream infection, with approximately 29% of adults undergoing liver transplantation experiencing a posttransplant infection (Santos, Hotchkiss, Chapman, & Olsen, 2016).

Mechanisms of Bloodstream Infection

Prevailing frameworks addressing causes of CLABSIs regard these events as breaches in care either during the insertion of the central venous catheter or during its maintenance care that result in the introduction of pathogens into the bloodstream (Agency for Healthcare Research and Quality, 2017; Centers for Disease Control and Prevention [CDC], 2017). The catheter insertion site as well as the catheter hub and needleless connector ports are regarded as potential sites for pathogen entry.

For patients with cancer and those undergoing hematopoietic stem cell transplant, translocation of pathogens across injured mucosal tissue represents an important mechanism of bloodstream infection (Gyarmati et al., 2015; Metzger et al., 2015; Simon et al., 2016; Steinberg, Robichaux, Tejedor, Reyes, & Jacob, 2013). Current pathophysiologic frameworks of mucositis indicate that injury is happening below the level of the epithelium before clinically evident signs and symptoms of injury are present (Sonis, 2009). Generation of reactive oxygen species, DNA strand breakage, and cytokine activation further contribute to tissue damage.

In 2013, the CDC added the mucosal barrier injury (MBI) designation to primary bloodstream infections that meet the clinical criteria of (1) presence of neutropenia and/or graft-versus-host disease and (2) infection involves an organism that is included on the CDC’s master list of organism as being associated with MBI (See et al., 2013). Although these infections include the MBI designation, they remain reportable as hospital-acquired infections.

The CDC revised their criteria for designating secondary versus primary bloodstream infections in 2015 (CDC, 2017). Included in these criteria was the requirement for radiographic evidence of gastrointestinal infection as a cause of secondary bacteremia. As a consequence of these revised definitions, many institutions experienced increases in CLABSI rates because events that previously would have been classified as secondary events were now classified as primary bloodstream infections (Grimes et al., 2016; Huber & Cycan, 2016).

Organisms and Infection Risk Factors

Among patients with cancer, those with hematologic malignancies are at heightened risk for bloodstream infection (Metzger et al., 2015; Rhee, Heung, Chen, & Chenoweth, 2015; Thurman, Abbott, Liu, & Larson, 2017). Approximately 20% of children with acute lymphoblastic leukemia and over 60% of children with acute myelogenous leukemia will experience at least one episode of bacteremia (Ammann et al., 2015; Sung, Lange, Gerbing, Alonzo, & Feusner, 2007). Gram-positive organisms, particularly coagulase-negative staphylococci and viridans group streptococci, predominate as leading causes of bacteremia (Gaur et al., 2013; Kelly et al., 2011; Lewis et al., 2014; Nielsen et al., 2016; Sung et al., 2007). For children with hematologic malignancies and those undergoing hematopoietic stem cell transplants, viridans group streptococcal infections most frequently occur during periods of neutropenia and are associated with significant morbidity, including septic shock and the need for intensive care services. Organisms associated with the lower gastrointestinal flora, specifically Escherichia coli, Enterococcus species, and Klebsiella species, are also important causative organisms. In the hematology/oncology/transplant population, these organisms are frequently associated with infections meeting MBI criteria (Metzger et al., 2015). Of note, vancomycin-resistant enterococci may be increasing as a source of CLABSIs (Kelly et al., 2011).

Overview of Strategies to Reduce Infections

To date, strategies to reduce CLABSIs have emphasized preventing the introduction of outside pathogens. Similar to efforts to reduce CLABSIs in critically ill adults, best practice maintenance care bundles for central venous catheters have been implemented across multiple pediatric hematology/oncology/transplant settings (Duffy, Rodgers, Shever, & Hockenberry, 2015; Rinke et al., 2012). Results of a national quality improvement initiative among inpatient hematology/oncology/transplant centers found that implementation of best practice maintenance care bundles resulted in a 28% decrease in CLABSIs (Bundy et al., 2014). This study did not, however, report differences in CLABSI rates based on organism. The use of a central venous catheter care team with the goal of ensuring standardization of best care practices was deemed feasible but did not result in a significant decrease in CLABSIs (Secola et al., 2012).

The use of chlorhexidine (CHG) bathing to reduce CLABSIs in critically ill adults has had mixed results (Climo et al., 2013; Noto et al., 2015). Reported decreases in CLABSI rates have been largely driven by decreases in coagulase-negative staphylococci while decreases in CLABSI rates involving organisms common to the lower gastrointestinal tract have not been observed (Climo et al., 2013). Likewise, in critically ill children, per-protocol analyses identified significant decreases in bacteremia involving coagulase-negative staphylococci with the use of CHG bathing in children with and without central venous catheters (Milstone et al., 2013). No differences were observed in bacteremia events involving yeast or gram negative bacteria.

Among the pediatric hematology/oncology/transplant population, a single-site study comparing a 7-month baseline to a 7-month trial of CHG bathing on an inpatient pediatric oncology unit did not demonstrate differences in bloodstream infection rates (Raulji, Clay, Velasco, & Yu, 2015). Rates of hospital-acquired infections from other sources such as the urinary tract or respiratory tract also did not differ. This site reported a decrease in hospital-acquired infections among patients 12 years and older (p = .009) but not among other age-groups. Another pediatric hematology/oncology/transplant unit reported decreased CLABSI rates involving coagulase-negative staphylococci and viridans group streptococci following implementation of formalized supportive care protocols with increased attention to daily hygiene (Linder, Gerdy, Abouzelef, & Wilson, 2017).

To date, most institutions have reported changes in CLABSI rates over a relatively short period of time such as a 6-month intervention period. Additionally, evidence addressing sustained decreases in CLABSI rates over longer periods of time in both adult and pediatric settings is limited. Because CLABSIs are uncommon events, examination of a longer time range may provide a more meaningful perspective and analysis of progress in reducing events and provide direction for additional focused interventions.

Purpose

This article presents changes in CLABSI rates among common causative organisms over an 11-year period on a pediatric inpatient unit that provides care for children with immune compromised conditions. This time period includes a 5-year preintervention baseline (2006-2010) and 6 years following the introduction of intentional CLABSI reduction efforts. The intervention period is further divided into two time periods: (1) the implementation of the Children’s Hospital Association’s best practice central line maintenance care bundles (2011-2012) and (2) the addition of formalized supportive cares in addition to the best practice maintenance care bundles (2013-2016). The formalized supportive cares included daily soap-and-water bathing, twice daily oral care, and out-of-bed activity three times each day (Linder, Gerdy, Abouzelef, & Wilson, 2017).

Method

Design

This study used a descriptive cohort design. This study is guided by a practice-based evidence approach that emphasizes the use of data from routine clinical practice to identify the most effective and efficient practices in routine care (Horn et al., 2015; Horn & Gassaway, 2007). A practice-based evidence approach is particularly useful when best practices are not well established. Data can then be used to guide iterative decision making as part of quality improvement initiatives.

Setting

The setting was a 32-bed inpatient unit serving children with immune compromised conditions at a free-standing tertiary children’s hospital in the Intermountain West of the United States. The majority of the unit’s patient population includes children with cancer and hematologic conditions as well as those undergoing hematopoietic stem cell transplant. The unit also serves children undergoing kidney and liver transplantation. Over the past 11 years, the unit has averaged approximately 6,400 central line days each year. The unit averages approximately 1,300 admissions and 6,800 patient days per year. More than 85% of these days are attributed to patients with cancer or those undergoing hematopoietic stem cell transplant. Patients treated by the oncology service whose treatment results in prolonged and profound neutropenia (e.g., acute myelogenous leukemia) remain inpatient during the count recovery period. Antibacterial prophylaxis was not routinely prescribed for these patients during the time periods represented in these analyses.

Data Sources

CLABSIs

All positive blood culture events meeting National Health Safety Network (NHSN) criteria to be classified as CLABSIs from January 1, 2006, to December 31, 2016, were included. All events were reviewed by the hospital infection preventionist and were classified based on NHSN criteria in place at the time of the event. All CLABSIs, including those designated as MBI events from 2013 to 2016, are included.

Central Line Days

The unit’s central line days were calculated based on data reported by the hospital’s continuous improvement department. The hospital maintains a central line database that tracks central venous catheters from the time of placement to the time of removal. This database can be queried to provide a summary of each individual unit’s central line days. At the present time, the system is unable to provide a summary of central line days based on admitting service.

Patient Characteristics

Patient characteristics were extracted from review of the medical record to characterize the sample. Variables included the patient’s age at the time of the CLABSI, diagnosis, gender, admitting service, diagnosis, presence of relapsed disease, and neutropenia (absolute neutrophil count < 1.0/µL) at the time of the CLABSI.

Procedure

Ethical Considerations

Institutional review board approval with a waiver of informed consent was granted to support the review of individual CLABSI cases.

Retrospective Review

A retrospective review of CLABSI events occurring between 2006 and 2010 was conducted prior to the initiation of intentional CLABSI reduction strategies to provide a baseline for comparison.

Continuing Review

As part of the unit’s ongoing efforts to reduce CLABSIs, individual cases are reviewed with the hospital infection preventionist and summarized as they occur.

Data Preparation and Analysis

Data were entered into Excel spreadsheets in preparation for analysis. Data were organized into three key time periods associated with the implementation of unit-based CLABSI-reduction strategies. These included (a) 2006 to 2010: a 5-year preintervention baseline for comparison; (b) 2011 to 2012: implementation of the Children’s Hospital Association best practice maintenance care bundles; and (c) 2013 to 2016: implementation of formalized supportive care practices in addition to the maintenance care bundles.

CLABSI rates per 1,000 central line days were calculated for each of the three time periods as were rates for individual organisms. Exact binomial 95% confidence intervals were calculated for each rate.

Organisms were grouped based on the source of the organism using the NHSN master organism list (NHSN, 2017). Organisms listed on the mucosal barrier injury (MBI) tab of the master organism list were further divided into those organisms common to the mouth and those common to the lower gastrointestinal tract. The four groups of organism included in the analyses were (a) common commensal organisms (e.g., coagulase-negative staphylococci), (b) MBI organisms common to the mouth (e.g., viridans group streptococci), (c) MBI organisms common to the lower gastrointestinal tract (e.g., Klebsiella spp.), and (d) other pathogens that are not designated as common commensal or MBI-associated organisms (e.g., Pseudomonas spp.).

Patient characteristics were assessed for comparability across the three time periods (preintervention [2006-2010], bundles [2011-2012], and bundles + formalized supportive cares [2013-2016]) using chi-square tests. Poisson regression (with log [catheter line time] as offset) was used to compare CLABSI rates for the individual organisms and organism groups using post hoc Tukey HSD (honest significant difference) tests to compare across the three time periods.

All analyses were performed using R software (3.4.3) with results considered statistically significant if p < .05.

Results

Patient Characteristics

Table 1 summarizes demographic and clinical characteristics of patients who developed CLABSIs during each time period. Between 2006 and 2016, 265 CLABSI events were identified in 189 individual patients. Twenty of these events involved two or more organisms, and 289 total organisms were identified. No significant differences in patient characteristics were present across each of the three periods. CLABSIs most frequently occurred in children with hematologic malignancies who were being treated for a primary diagnosis of cancer by the oncology service. Fifty-two patients experienced two or more CLABSIs (range two to five). Within this subgroup of patients with repeat CLABSIs, 44 (85%) had a diagnosis of leukemia. Of note, no patients with sarcoma developed a CLABSI following the introduction of best practice bundles.

Table 1.

Patient Characteristics.

	Preintervention (2006-2010), n = 156; n (%)	Bundles (2011-2012), n = 44; n (%)	Bundles + formalized supportive cares (2013-2016), n = 65; n (%)	χ², Significance
Gender				0.32, p = .85
Male	88 (56)	24 (55)	34 (51)
Female	68 (44)	20 (45)	31 (49)
Age-group (years)				8.52, p = .58
0-4	74 (47)	23 (52)	28 (43)
5-9	20 (13)	5 (11)	12 (18)
10-14	32 (21)	6 (14)	15 (23)
15-19	26 (17)	8 (15)	7 (11)
20+	4 (3)	2 (5)	3 (5)
Diagnosis				14.02, p = .17
Leukemia
ALL	37 (24)	13 (30)	21 (32)
AML	65 (42)	20 (45)	19 (29)
Other leukemia	5 (3)	3 (7)	3 (5)
Lymphoma
Hodgkin lymphoma	1 (1)	0 (0)	0 (0)
Non-Hodgkin lymphoma	6 (4)	0 (0)	1 (2)
Solid tumor
Sarcoma	6 (4)	0 (0)	0 (0)
Brain tumor	12 (8)	2 (5)	5 (8)
Other solid tumor	8 (5)	2 (5)	7 (11)
Benign hematologic condition	10 (6)	0 (0)	3 (5)
Solid organ transplant	3 (2)	1 (2)	4 (6)
Other	3 (2)	3 (7)	2 (3)
Relapsed disease				4.69, p = .32
No	110 (70)	29 (66)	37 (57)
Yes	34 (22)	12 (27)	19 (29)
N/A	12 (8)	3 (7)	9 (14)
Medical team				8.63, p = .20
Hematology/oncology	122 (78)	35 (80)	48 (74)
BMT	28 (18)	7 (16)	13 (20)
Solid organ transplant	1 (1)	1 (2)	4 (6)
Other	5 (3)	1 (2)	0 (0)
Neutropenia (ANC < 1.0/µL)				0.13, p = .94
No	40 (26)	11 (25)	18 (28)
Yes	116 (74)	33 (75)	47 (72)

Note. ALL = acute lymphocytic leukemia; AML = acute myeloid leukemia; BMT = bone marrow transplant; ANC = absolute neutrophil count.

Overall CLABSI Rates

Table 2 presents a comparison of overall CLABSI events and rates across each of the three time periods. Overall, CLABSI rates decreased by an expected log rate of −1.48 (p < .01), that is, the CLABSI rate was e^−1.48 = 0.23 times less following implementation of the bundles plus supportive cares compared with the preintervention period. Post hoc Tukey tests identified significant differences in rates only between the 5-year preintervention period and following implementation of enhanced basic supportive cares in addition to best practice bundles (z = −3.43, p < .01).

Table 2.

Comparison of CLABSI Events and Rates.

	Preintervention baseline (2006-2010)	Best practice maintenance bundles (2011-2012)	Best practice bundles + formalized supportive cares (2013-2016)	Log rate change, significance (2-tailed)
Total central line days	32,229	13,364	25,159
Combined CLABSIs
Total CLABSIs	156	44	65
CLABSI rate [95% CI]	4.84 [4.11, 5.66]	3.29 [2.39, 4.42]	2.58 [1.99, 3.29]^a	−1.48, p < .01
Common commensal organisms
Coagulase-negative staphylococci events	28	6	6
Coagulase-negative staphylococci rates [95% CI]	0.87 [0.58, 1.26]	0.45 [0.16, 0.98]	0.24 [0.09, 0.52]^a	−1.29, p < .01
Other common commensal organism events	5	1	3
Other common commensal organism rates [95% CI]	0.16 [0.05, 0.36]	0.07 [0.002, 0.42]	0.12 [0.02, 0.35]	−0.26, p = .72
Mucosal barrier injury organisms common to the mouth
Viridans group streptococci events	36	10	14
Viridans group streptococci rates [95% CI]	1.12 [0.78, 1.55]	0.75 [0.36, 1.38]	0.56 [0.30, 0.93]	−0.70, p = .027
Candida spp. events	11	2	2
Candida spp. rates [95% CI]	0.34 [0.17, 0.61]	0.15 [0.02, 0.54]	0.08 [0.01, 0.29]	−1.46, p = .058
Other MBI organisms common to the mouth	5	4	4
Other MBI organisms common to the mouth rates [95% CI]	0.16 [0.05, 0.36]	0.07 [0.08, 0.77]	0.16 [0.04, 0.41]	0.025, p = .97
Mucosal barrier injury organisms common to the lower gastrointestinal tract (GI)
Enterococcus spp. events	19	4	10
Enterococcus spp. rates [95% CI]	0.59 [0.35, 0.92]	0.30 [0.08, 0.77]	0.40 [0.19, 0.73]	−0.39, p = .31
Klebsiella spp. events	14	4	11
Klebsiella spp. rates [95% CI]	0.43 [0.24, 0.73]	0.30 [0.08, 0.77]	0.44 [0.22, 0.78]	0.01, p = .99
Escherichia coli events	11	4	5
Escherichia coli rates [95% CI]	0.34 [0.17, 0.61]	0.30 [0.08, 0.77]	0.20 [0.06, 0.46]	−0.54, p = .32
Enterobacter spp. events	9	4	6
Enterobacter spp. rates [95% CI]	0.28 [0.13, 0.53]	0.30 [0.08, 0.77]	0.23 [0.09, 0.52]	−0.16, p = .77
Other MBI organisms common to the lower GI tract events	2	2	5
Other MBI organisms common to the lower GI tract rates [95% CI]	0.06 [0.008, 0.22]	0.15 [0.02, 0.54]	0.20 [0.06, 0.46]	1.16, p = .16
Noncommon commensal and non-MBI pathogens
Pseudomonas spp. events	11	3	3
Pseudomonas spp. rates [95% CI]	0.34 [0.17, 0.61]	0.22 [0.05, 0.66]	0.12 [0.02, 0.35]	−1.05, p = .11
Methicillin sensitive staphylococcus aureus events	6	1	4
Methicillin sensitive staphylococcus aureus rates [95% CI]	0.19 [0.07, 0.41]	0.07 [0.002, 0.42]	0.16 [0.04, 0.41]	−0.16, p = .81
Other noncommon commensal and non-MBI pathogens events	8	4	2
Other noncommon commensal and non-MBI pathogens rates [95% CI]	0.25 [0.11, 0.49]	0.30 [0.08, 0.77]	0.08 [0.01, 0.29]	−1.14, p = .15

Note. CLABSI = central line–associated bloodstream infection; MBI = mucosal barrier injury.

Rate for 2013 to 2016 is significantly less than that for 2006 to 2010 (p < .01).

Individual Organisms

Among individual organisms, significant expected log rate decreases were present only for coagulase-negative staphylococci and viridans group streptococci. For coagulase-negative staphylococci, the expected log rate decrease was −1.29 (p ≤ .01), that is, the CLABSI rate was e^−1.29 = 0.28 times less following implementation of the bundles plus supportive cares compared the preintervention period. For viridans group streptococci the expected log rate decrease was −.70 (p = .03), meaning the CLABSI rate was e^−.7 = 0.497 times less following implementation of the bundles plus supportive cares compared the preintervention period. Of note, the expected log rate decrease for Candida species approached significance with an expected log rate decrease of −1.46 (p = .058), that is, the CLABSI rate was e^−1.46 = 0.23 times less following implementation of the bundles plus supportive cares compared the preintervention period. Post hoc Tukey tests identified significant differences in rates of infection with coagulase-negative staphylococci between the preintervention period and following implementation of enhanced supportive cares (z = −2.87, p = .011). Although decreases in rates of infection involving Candida and Pseudomonas species did not reach statistical significance, the observed decreases were clinically important.

Organism Groups

Table 3 presents a comparison of CLABSI events and rates based on the four groups of organism. Expected log rate decreases were significant for three of the groups: common commensal organisms, MBI organisms common to the mouth, and other pathogens that are not considered common commensal or MBI-associated organisms (p < .01). Additionally, CLABSI rates were significantly less during 2013 to 2016 (bundles + formalized supportive cares) compared with 2006 to 2010 (preintervention comparison) for both common commensal organisms (z = −2.80, p = .014) and MBI organisms common to the mouth (z = −2.69, p = .019). CLABSI rates were unchanged for MBI organisms common to the lower gastrointestinal tract.

Table 3.

Log Rate Changes by Organism Group.

	Preintervention baseline (2006-2010)	Best practice bundles (2011-2012)	Best practice bundles + formalized supportive cares (2013-2016)	Log rate change, significance (2-tailed)
Common commensal organisms
Events	33	7	9
Rate	1.02	0.52	0.36^a	−1.05, p = .005
Mucosal barrier injury organisms common to the mouth
Events	52	16	20
Rate	1.56	1.20	0.79^a	−0.708, p = .007
Mucosal barrier injury organisms common to the lower gastrointestinal tract
Events	55	18	37
Rate	1.66	1.35	1.47	−0.149, p = .484
Noncommon commensal and nonmucosal barrier injury pathogens
Events	25	8	9
Rate	0.78	0.60	0.36	−0.77, p = .005

Rate for 2013 to 2016 is significantly less than that for 2006 to 2010 (p < .05).

Discussion

CLABSIs remain an important subset of infections for children with cancer and other conditions resulting in immune compromise. To date, interventions have emphasized the prevailing model of CLABSIs as resulting from organisms entering the bloodstream as a consequence of breaches in care. Despite these efforts, CLABSIs continue to occur. The addition of formalized supportive cares in addition to the best practice maintenance care bundles contributed to sustained significant decreases in the overall CLABSI rate at this institution; however, rates of infection involving MBI organisms common to the lower gastrointestinal tract did not decrease. The examination of data over an 11-year period provided important insights into the characteristics of patients who develop CLABSIs as well as organisms involved in CLABSI events.

Patient Characteristics

The clinical characteristics of patients who acquired CLABSIs over the 11-year period remained unchanged. Patients with leukemia continued to comprise the largest diagnostic group who developed CLABSIs, specifically during periods of prolonged and profound neutropenia. These patients also more frequently developed subsequent infections relative to other diagnostic groups. This project also identified subsets of patients at particularly low risk for infection, specifically, patients with sarcoma.

Organism Characteristics

Although overall CLABSI rates decreased over the 11-year period, the only two organisms for which significant rates of decrease were observed were coagulase-negative staphylococci and viridans group streptococci. Because these two groups of organisms have historically been the most common organisms involved in CLABSIs, decreasing the rates of infection involving these organisms has driven the decrease in the overall CLABSI rate. Despite overall decreases in the rate of viridans group streptococci, rates of infection involving this organism group remained the highest relative to other organism species.

Grouping the organisms into classes based on the NHSN’s master organism list supported additional examination of organisms involved in CLABSIs based on their potential source of introduction into the bloodstream. Decreases in rates of CLABSIs resulting from common commensal organisms, MBI organisms common to the mouth, and non-MBI/noncommon commensal organisms were observed. These decreases suggest current successes of interventions targeted at reducing the introduction of organisms through enhanced maintenance care of the central venous catheter and patient hygiene efforts targeted at reducing the concentration of bacteria on the skin and in the mouth.

Consistent with previously reported studies (Climo et al., 2013; Milstone et al., 2013), the rate of infection involving MBI organisms common to the lower gastrointestinal tract remained unchanged across the 11-year period. This lack of responsiveness to interventions to date supports current evidence that CLABSIs involving these organisms in patients with neutropenia likely represent a distinct subset of infection associated with translocation of organisms across injured mucosal tissue (Metzger et al., 2015; See et al., 2013). As such, these infections will require additional intervention strategies.

Clinical Implications

Results of this study provide guidance in directing interventions for ongoing efforts to reduce CLABSIs in pediatric hematology/oncology/transplant patients. Although past attention has been given to standardizing care across individual units and institutions, emphasizing more targeted interventions to selected subgroups of patients, specifically patients with leukemia, may provide a more meaningful approach. A recent study has demonstrated the safety and efficacy of administering levofloxacin to patients with leukemia during the induction period as antibacterial prophylaxis (Wolf et al., 2017). Such an approach may offer potential benefit to oncology service patients who remain hospitalized for count recovery during periods of profound and prolonged neutropenia and warrants investigation. Likewise, patients who are at particularly low risk for an infection may be able to avoid additional interventions such as CHG bathing and prophylactic antibiotics from which they are less likely to benefit.

The use of practice-based evidence techniques that emphasize a more rigorous description of unit-based data can support institutions in refining interventions targeted at the hematology/oncology/transplant settings. Treating CLABSI-MBI events as a distinct subset of infection with a distinct pathophysiology can help shift the paradigm of regarding hematology/oncology/transplant units as underperforming units. Reporting CLABSI rates with MBI events removed in addition to the overall CLABSI rates can further help to provide a more meaningful comparison across other units within and across institutions.

Limitations

Limitations of this project include data from a single institution. The scope of the patient population served by this institution may not be reflective of other settings, therefore, the results may have limited generalizability. Additionally, variations in institutional supportive care practices across institutions such as hospitalization during the count recovery period and the use of antibacterial prophylaxis during periods of prolonged and profound neutropenia may further contribute to variation in reported CLABSI rates across institutions. In 2016, this center adopted more stringent blood culture practices ensuring that two blood culture samples were obtained from patients developing new fevers and increasing the volume of blood obtained with each specimen. This practice change increased the potential for more CLABSIs to be identified, particularly those involving common commensal organisms and viridans group streptococci in which the same organism grew from both cultures.

Institutional mechanisms for reporting central venous catheter days during the time period reported in this article do not support further delineation of central line days based on patient diagnosis or admitting service. This limits the ability to complete more refined analyses of CLABSI rates among specific patient populations served by the unit.

A potential limitation of the analyses presented in this article is the additional delineation of MBI organisms. These organisms were attributed to either the oral cavity or lower gastrointestinal tract based on review of its primary location as part of the human host flora. For additional clarity in comparing trends over time, CLABSI rates for each major organism genus were provided to support comparisons.

Surveillance definitions for CLABSIs continue to be refined by the CDC. As a consequence, an event meeting criteria to be defined as a CLABSI in a given year may or may not have been classified as a CLABSI in a previous or subsequent year.

A limitation of the analytic approach is that it assumes immediate change as a result of the intervention and does fully address changes in variation and correlation following implementation of the intervention. Emerging statistical approaches, such as a robust interrupted time series model (Cruz, Bender, & Ombao, 2017) are designed to address these limitations and can be implemented in future longitudinal analyses.

Directions for Future Research

Directions for future research include ongoing attention to understanding individual patient characteristics. Within this study, nearly one third of patients experienced more than one CLABSI and 85% of patients with repeat infections had leukemia. Exploration of both disease-related and patient-specific characteristics may further support stratification of patients based on risk and guide more targeted interventions.

The lack of change in CLABSI rates involving MBI organisms common to the lower gastrointestinal tract emphasizes the crucial need for additional research and the development of interventions targeted at reducing this subgroup of infections. Regardless of whether these events are viewed as hospital-acquired, they represent an important source of morbidity for pediatric hematology/oncology/transplant patients. Areas of current interest include gastrointestinal dysbiosis, which can occur as a consequence of the illness and nutritional deficits, as well as exposure to chemotherapy, radiation, and antibiotic therapy (Galloway-Pena et al., 2016; Taur & Pamer, 2016). Failure of systemic immune responses, such as in the context of prolonged and profound neutropenia further contribute to infection risk. Polymorphisms in individual genes may further convey an increased risk of bacterial translocation in the context of an altered microbiome (Rayes et al., 2016).

Interventions targeted at reducing injury to the oral mucosa are also needed. Despite current improvements, MBI organisms common to the mouth remain an important source of infection, particularly in the context of profound and prolonged neutropenia. Low level laser light therapy has been proposed as means to promote the integrity of the oral mucosa, which may be protective against infection and has been reported to be tolerated by children (Eduardo et al., 2015; Migliorati et al., 2013).

As future studies are reported, attention needs to be given not only to reporting changes in overall CLABSI rates but also to changes in rates of infection involving specific organisms. Decreases in more common causative organisms such as coagulase-negative staphylococci can mask a lack of change in CLABSI rates involving other organisms (Climo et al., 2013). Such additional detail in reporting can provide a more meaningful evaluation of organisms for which a given intervention is or is not effective. Additional studies reporting the sustained effectiveness of interventions across time are also needed.

Conclusion

Results of this study emphasize ongoing challenges in reducing CLABSIs in pediatric hematology/oncology/transplant patients. Patients with hematologic malignancy, specifically leukemia, are most frequently affected by CLABSIs. While strategies to date resulted in decreases in some groups of organisms, rates of infection involving MBI organisms common to the lower gastrointestinal tract remained unchanged over an 11-year period. Study results emphasize the need for more specific, targeted interventions that address the underlying pathophysiology of these infections.

Footnotes

Authors’ Note

Andrew Wilson is also affiliated to PAREXEL International Corporation.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The pre-intervention retrospective review of CLABSIs between 2006-10 was supported by a grant from the Utah Organization of Nurse Leaders.

ORCID ID

Lauri A. Linder

Author Biographies

Lauri A. Linder, PhD, APRN, CPON^®, holds a joint appointment as an Associate Professor at the University of Utah College of Nursing and as a Pediatric Oncology Clinical Nurse Specialist at Primary Children’s Hospital. Her research interests include supportive care and symptom management for children, adolescents, and young adults with cancer. with attention to novel uses of technology to support patients in communicating their experiences

Cheryl Gerdy, MSN, RN, CPON^®, is the service line director for the Cancer, Blood, and Transplant Service at Primary Children’s Hospital.

Yeonjung Jo, BS, is finishing her MStat in Mathematics at the University of Utah and working as a graduate research assistant at the College of Nursing since 2017. She is planning to pursue PhD in Biostatistics. She is interested in clinical data, missing data problems, causal inference and longitudinal data analysis.

Andrew Wilson, PhD, MStat, is the Pharmacoepidemiology and Statistics Director in the Real World Data Services department at PAREXEL. He is also Adjunct Assistant Professor at the University of Utah College of Nursing and Department of Family and Preventive Medicine in the School of Medicine. His research interests are in graphical methods, including causal modeling and network analysis, predictive analytics, and reproducibility and transparency in research, all within a real-world context.

References

Agency for Healthcare Research and Quality. (2017). Topic: Central line-associated bloodstream infections (CLABSI). Retrieved from https://www.ahrq.gov/topics/central-line-associated-bloodstream-infections-clabsi.html

Ammann

R. A.

Laws

H. J.

Schrey

Ehlert

O. Moser

Dilloo

. . . Simon

(2015). Bloodstream infection in paediatric cancer centres: Leukaemia and relapsed malignancies are independent risk factors. European Journal of Pediatrics, 174, 675-686. doi:10.1007/s00431-015-2525-5

Bochennek

Hassler

Perner

Gilfert

Schöning

Klingebiel

. . . Lehrnbecher

(2004). Infectious complications in children with acute myeloid leukemia: Decreased mortality in multicenter trial AML-BFM 2004. Blood Cancer Journal, 6, e382.

Bundy

D. G.

Gaur

A. H.

Billett

A. L.

Colantuoni

E. A.

Miller

M. R.

, Children’s Hospital Association Hematology/Oncology CLABSI Collaborative. (2014). Preventing CLABSIs among pediatric hematology/oncology inpatients: National collaborative results. Pediatrics, 134, e1678-e1685. doi:10.1542/peds.2014-0582

Celebi

Sezgin

M. E.

Cakir

Baytan

Demirkaya

Sevinir

… Hacimustafaoglu

(2013). Catheter-associated bloodstream infections in pediaric hematology-oncology patients. Pediatric Hematology and Oncology, 30, 187-194. doi: 10.3109/08880018.2013.772683

Centers for Disease Control and Prevention. (2017). Central line-associated bloodstream infection (CLABSI). Retrieved from https://www.cdc.gov/hai/bsi/bsi.html

Climo

M. W.

Yokoe

D. S.

Warren

D. K.

Perl

T. M.

Bolon

Herwaldt

L. A.

. . . Wong

E. S.

(2013). Effect of daily chlorhexidine bathing on hospital-acquired infection. New England Journal of Medicine, 368, 533-542. doi:10.1056/NEJMoa1113849

Cruz

Bender

Ombao

(2017). A robust interrupted time series model for analyzing complex health care intervention data. Statistics in Medicine, 36, 4660-4676. doi:10.1002/sim.7443

Duffy

E. A.

Rodgers

C. C.

Shever

L. L.

Hockenberry

M. J.

(2015). Implementing a daily maintenance care bundle to prevent central line-associated bloodstream infections in pediatric oncology patients. Journal of Pediatric Oncology Nursing, 32, 394-400. doi:10.1177/1043454214563756

10.

Eduardo

F. P.

Bezinelli

L. M.

de Carvalho

D. L. C.

Lopes

R. M. G.

Feranades

J. F.

Brumatti

. . . Correa

(2015). Oral mucositis in pediatric patients undergoing hematopoietic stem cell transplantation: Clinical outcomes in a context of specialized oral care using low-level laser therapy. Pediatric Transplantation, 19, 316-325. doi:10.1111/petr.12440

11.

Galloway-Pena

Smith

D. P.

Sahasrabhojane

Ajami

N. J.

Wadsworth

W. D.

Daver

N. G.

. . . Shelburne

S. A.

(2016). The role of the gastrointestinal microbiome in infectious complications during induction therapy for acute myelogenous leukemia. Cancer, 122, 2186-2196. doi:10.1002/cncr.30039

12.

Gaur

A. H.

Bundy

D. G.

Gao

Werner

E. J.

Billett

A. L.

Hord

J. D.

. . . Miller

M. R.

(2013). Surveillance of hospital-acquired central line-associated bloodstream infections in pediatric hematology-oncology patients: Lessons learned, challenges ahead. Infection Control & Hospital Epidemiology, 34, 316-320. doi:10.1086/669513

13.

Grimes

McMullen

K. M.

Leone

Goris

Mueller

Carroll

. . . Babcock

H. M.

(2016). Impact of 2015 National Healthcare Safety Network (NHSN) definition changes on intensive care unit (ICU) central line-associated bloodstream infection (CLABSI) at a large healthcare system. Open Forum Infectious Diseases, 3(Suppl. 1), 946. doi:10.1093/ofid/ofw194.86

14.

Gyarmati

Kjellander

Aust

Kalin

Ohrmalm

Giske

C. G.

(2015). Bacterial landscape of bloodstream infections in neutropenic patients via high throughput sequencing. PLoS One, 10, e0135756. doi:10.1371/journal.pone.0135756

15.

Horn

S. D.

Corrigan

J. D.

Bogner

Hammond

F. M.

Steel.

R. T.

Smout

R. J.

. . . Whiteneck

G. G.

(2015). Traumatic brain injury: Practice-based evidence study: Design and patients, centers, treatments, and outcomes. Archives of Physical Medicine and Rehabilitation, 96 (8 Suppl.), S178-S196. doi:10.1016/j.apmr.2014.09.042

16.

Horn

S. D.

Gassaway

(2007). Practice-based evidence study design for comparative effectiveness research. Medical Care, 45(Suppl. 2), S50-S57.

17.

Huber

Cycan

(2016). The impact of NHSN definition changes on the attribution of secondary blood stream infections. Open Forum Infectious Diseases, 3(Suppl. 1), 1489. doi:10.1093/ofid/ofw172.1191

18.

Kelly

Conway

Wirth

Potter-Bynoe

Billett

A. L.

Sandora

T. J.

(2011). Moving CLABSI prevention beyond the ICU: Risk factors in pediatric oncology patients. Infection Control & Hospital Epidemiology, 32, 1079-1085. doi:10.1086/662376

19.

Lewis

Yanofsky

Mitchell

Dix

Ethier

M. C.

Gillmeister

. . . Sung

(2014). Predictors and outcomes of viridans group streptococcal infections in pediatric acute myeloid leukemia: From the Canadian infections in AML research group. Pediatric Infectious Diseases Journal, 33, 126-129. doi:10.1097/INF.0000000000000058

20.

Linder

L. A.

Gerdy

Abouzelef

Wilson

(2017). Using practice-based evidence to improve supportive care practices to reduce central line–associated bloodstream infections in a pediatric oncology unit. Journal of Pediatric Oncology Nursing, 34, 185-195. doi:10.1177/1043454216676838

21.

Metzger

K. E.

Rucker

Callaghan

Churchill

Jovanovic

B. D.

Zembower

T. R.

Bolon

M. K.

(2015). The burden of mucosal barrier injury laboratory-confirmed bloodstream infection among hematology, oncology, and stem cell transplant patients. Infection Control & Hospital Epidemiology, 36, 119-124. doi:10.1017/ice.2014.3

22.

Migliorati

Hewson

Lalla

R. V.

Antunes

H. S.

Estilo

C. L.

Hodgson

. . . Elad

(2013). Systematic review of laser and other light therapy for the management of oral mucositis in cancer patients. Supportive Care in Cancer, 21, 333-341. doi:10.1007/s00520-012-16-5-6

23.

Milstone

A. M.

Elward

Song

Zerr

D. M.

Orscheln

Speck

. . . Perl

T. M.

(2013). Daily chlorhexidine bathing to reduce bacteraemia in critically ill children: A multicenter, cluster-randomised, crossover trial. Lancet, 381, 1099-1106. doi:10.1016/S0140-6736(12)61687-0

24.

National Health Safety Network. (2017). NHSN master organism list. Retrieved from https://www.cdc.gov/nhsn/xls/master-organism-com-commensals-lists.xlsx

25.

Nielsen

M. J.

Claxton

Pizer

Lane

Cooke

R. P. D.

Paulus

Carrol

E. D.

(2016). Viridans group streptococcal infections in children after chemotherapy or stem cell transplantation: A 10-year review from a tertiary pediatric hospital. Medicine, 95, e2952. doi:10.1097/MD.0000000000002952

26.

Noto

M. J.

Domenico

H. J.

Byrne

D. W.

Talbot

Rice

T. W.

Bernard

G. R.

Wheeler

A. P.

(2015). Chlorhexidine bathing and healthcare-associated infections: A randomized clinical trial. JAMA Journal of the American Medical Association, 313, 369-378. doi:10.1001/jama.2014.18400

27.

Raulji

C. M.

Clay

Velasco

L. C.

(2015). Daily bathing with chlorhexidine and its effects on nosocomial infection rates in pediatric oncology patients. Pediatric Hematology and Oncology, 32, 315-321. doi:10.3109/08880018.2015.1013588

28.

Rayes

Morrow

A. L.

Payton

L. R.

Lake

K. E.

Lane

Davies

S. M.

(2016). A genetic modifier of the gut microbiome influences the risk of graft-versus-host disease and bacteremia after hematopoietic stem cell transplantation. Biology of Blood and Marrow Transplantation, 22, 418-422. doi:10.1016/j.bbmt.2015.11.017

29.

Rhee

Heung

Chen

Chenoweth

C. E.

(2015). Central line-associated bloodstream infections in non-ICU inpatient wards: A 2-year analysis. Infection Control & Hospital Epidemiology, 36, 424-430. doi:10.1017/ice.2014.86

30.

Rinke

M. L.

Chen

A. R.

Bundy

D. G.

Colantuoni

Fratino

Drucis

K. M.

. . . Miller

M. R.

(2012). Implementation of a central line maintenance care bundle in hospitalized pediatric oncology patients. Pediatrics, 130, e996-e1004. doi:10.1542/peds.2012-0295

31.

Santos

A. Q. C.

Hotchkiss

R. S.

Chapman

W. C.

Olsen

M. A.

(2016). Epidemiology of bloodstream infections in a multicenter retrospective cohort of liver transplant recipients. Transplant Direct, 2, e67.

32.

Secola

Azen

Lewis

M. A.

Pike

Needleman

Sposto

Doering

(2012). A crossover randomized prospective pilot study evaluating a central venous catheter team in reducing catheter-related bloodstream infections in pediatric oncology patients. Journal of Pediatric Oncology Nursing, 29, 307-315. doi:10.1177/1043454212461714

33.

See

Iwamoto

Allen-Bridson

K. T.

Horan

Magill

S. S.

Thompson

N. D.

(2013). Mucosal barrier injury laboratory-confirmed bloodstream infection: Results from a field test of a new national healthcare safety network definition. Infection Control & Hospital Epidemiology, 34, 769-776. doi:10.1086/671281

34.

Simon

Furtwängler

Graf

Laws

H. J.

Voight

Piening

. . . Ammann

R. A.

(2016). Surveillance of bloodstream infections in pediatric cancer centers: What have we learned and how do we move on? German Medical Science Hygiene and Infection Control, 11. doi:10.3205/dgkh000271

35.

Sonis

S. T.

(2009). Mucositis: The impact, biology, and therapeutic opportunities of oral mucositis. Oral Oncology, 45, 1015-1020. doi: 10.1016/j.oraloncology.2009.08.006

36.

Steinberg

J. P.

Robichaux

Tejedor

S. C.

Reyes

M. D.

Jacob

J. T.

(2013). Distribution of pathogens in central line–associated bloodstream infections among patients with and without neutropenia following chemotherapy: Evidence for a proposed modification to the current surveillance definition. Infection Control & Hospital Epidemiology, 34, 171-175. doi:10.1086/669082

37.

Sung

Lange

B. J.

Gerbing

R. B.

Alonzo

R. A.

Feusner

(2007). Microbiologically documented infections and infection-related mortality in children with acute myeloid leukemia. Blood, 110, 3532-3539. doi:10.1182/blood-2007-05-091942

38.

Taur

Pamer

E. G.

(2016). Microbiome mediation of infections in the cancer setting. Genome Medicine, 8, 40. doi:10.1186/s13073-016-0306-z

39.

Thurman

C. B.

Abbott

Liu

Larson

(2017). Risk for health care-associated bloodstream infections in pediatric oncology patients with various malignancies. Journal of Pediatric Oncology Nursing, 34, 196-202. doi:10.1177/1043454216680596

40.

Wilson

M. Z.

Rafferty

Deeter

Comito

M. A.

Hollenbeak

C. S.

(2014). Attributable costs of central line–associated bloodstream infections in a pediatric hematology/oncology population. American Journal of Infection Control, 42, 1157-1160.

41.

Wolf

Tang

Flynn

P. M.

Pui

C.-H.

Gaur

A. H.

Sun

. . . Jeha

(2017). Levofloxacin prophylaxis during induction therapy for pediatric acute lymphoblastic leukemia. Clinical Infectious Diseases, 65, 1790-1798. doi:10.1093/cid/cix644