The U chart demonstrates a sustained reduction in CCABSI rates after the best-practice CCABSI prevention bundle was implemented with ethanol lock prophylaxis. A indicates the CCABSI prevention bundle with ethanol lock prophylaxis offered to eligible patients with a history of 2 or more CCABSIs; and B, CCABSI prevention bundle with ethanol lock prophylaxis offered to all eligible patients.
Ardura MI, Lewis J, Tansmore JL, Harp PL, Dienhart MC, Balint JP. Central Catheter–Associated Bloodstream Infection Reduction With Ethanol Lock Prophylaxis in Pediatric Intestinal FailureBroadening Quality Improvement Initiatives From Hospital to Home. JAMA Pediatr. 2015;169(4):324-331. doi:10.1001/jamapediatrics.2014.3291
Children with intestinal failure are at high risk for developing central catheter–associated bloodstream infections (CCABSIs) owing to children’s chronic dependence on central venous catheters for parenteral nutrition.
To evaluate the effectiveness and safety of the addition of ethanol lock prophylaxis to a best-practice CCABSI prevention bundle on hospital and ambulatory CCABSI rates in children with intestinal failure.
Design, Setting, and Participants
Quality improvement and statistical process control analysis that took place at a tertiary care pediatric hospital and patient homes. Participants included children who were 18 years or younger with intestinal failure requiring a central venous catheter.
Central catheter–associated bloodstream infection prevention bundle that included daily ethanol lock prophylaxis.
Main Outcomes and Measures
Central catheter–associated bloodstream infection rates and safety outcomes (central catheter insertions, repairs, and hospitalizations) before (January 1, 2011-January 31, 2012) and after (February 1, 2012-December 31, 2013) ethanol lock prophylaxis bundle implementation.
Twenty-four children with intestinal failure received the ethanol lock prophylaxis CCABSI prevention bundle for a median of 266 days (range, 12-635 days). Rates of CCABSI decreased from 6.99 CCABSIs per 1000 catheter days at baseline to 0.42 CCABSI per 1000 catheter days after ethanol lock prophylaxis bundle implementation, despite an increase in the total number of catheter days. A subset of 14 children who received prolonged ethanol lock prophylaxis (≥3 months) had fewer median (range) central catheter insertions 0 (0-2) vs 3 (0-6); P = .001. The pre-ELP intervention CCABSI rates in this subset was 7.01 per 1000 catheter days vs 0.64 per 1000 catheter days for post-ELP intervention (P = .004). There were no significant differences in the total number of hospital admissions; however, there were fewer hospitalizations for fever and CCABSI (P = .003).
Conclusions and Relevance
A best-practice CCABSI prevention bundle that included ethanol lock prophylaxis in both the hospital and home was successfully implemented, well tolerated, and demonstrated a significant and sustained reduction in preventable harm in the form of CCABSIs in children with intestinal failure.
Children with intestinal failure (IF) often depend on a central venous catheter (CVC) to deliver parenteral nutrition (PN), which is essential for their growth and survival. Home PN infusions are being used with increasing frequency and have improved the quality of life of children with IF.1 However, long-term CVC use can result in infectious, mechanical, and thrombotic complications. Infectious complications, such as central catheter–associated bloodstream infections (CCABSIs) and recurrent sepsis episodes, remain a major burden, causing significant morbidity and directly influencing the long-term survival of children with IF.2 Despite adherence to central catheter insertion and maintenance care bundles, CCABSI rates in children with IF are unacceptably high, varying from 2 to 26.5 infections per 1000 catheter days. Further, children with IF have a higher frequency of repeated infections compared with children with other disorders who also require long-term CVCs.3- 6 Risk factors for CCABSIs are multifactorial and inherent to the IF diagnosis, including long-term CVC use, need for PN, age of 2 years or younger, underlying illness and residual small bowel length, associated liver disease, small-bowel bacterial overgrowth, and pathogen translocation across a dysmotile intestine.7- 9 Central catheter–associated bloodstream infections in this population can be caused by both bacteria and fungi owing to a single pathogen or polymicrobial.10
Central catheter–associated bloodstream infection prevention in children with IF remains a cornerstone of therapy and a continued challenge to improving the quality of patient care and clinical outcomes. Published data and national efforts have focused on CCABSI prevention in the hospital setting; however, there has been little progress on infection prevention associated with central catheter care in the home following hospital discharge. Published data regarding CCABSIs in other high-risk populations, such as children with malignancies with CVCs, confirm that community-associated CCABSIs occur at higher rates compared with in hospital settings and carry appreciable morbidity.11,12
There is no standard of care to guide clinicians in optimizing preventable harm in the form of CCABSIs in the home care setting. The use of ethanol locks as a prophylactic measure is a promising option for CCABSI prevention because ethanol readily penetrates biofilm, has anticoagulant and fibrinolytic properties, and is both bactericidal and fungicidal without promoting the emergence of antimicrobial resistance. Review of published data using 70% ethanol lock prophylaxis (ELP) in children with IF showed promise as a potentially effective and inexpensive prevention measure but the data are limited to small retrospective case series. In addition, these same data suggest possible but significant complications with ethanol lock use, including catheter thrombosis, which has precluded its more widespread use (Table 1).13- 18
In an attempt to reduce CCABSI infections in our patients and improve clinical outcomes, we performed a quality-improvement (QI) initiative intended to evaluate whether the addition of ELP to our CCABSI prevention bundle in both the hospital and home settings could safely reduce CCABSI rates in our pediatric patients with IF.
Nationwide Children’s Hospital (NCH) is a 427-bed academic tertiary care freestanding children’s hospital that serves as a primary referral center for central Ohio. The Intestinal Support Service (ISS) at NCH has followed up 296 patients with IF since its inception; 6% to 8% of patients with IF require long-term PN or fluids at any given time. We retrospectively (January 1, 2011-January 31, 2012) and prospectively (February 1, 2012-December 31, 2013) collected data on culture-proven bloodstream infections, CVC days, CVC replacements, CVC repairs, occlusions, and hospitalizations of pediatric patients with IF who required a CVC, long-term PN, or intravenous fluids and were followed up at NCH ISS.
The core improvement team for this QI project consisted of 2 pediatric gastroenterologists who provided clinical care to the patients with IF, a specialized CVC nurse who evaluated and performed all CVC care and teaching to parents about CVC care and the prevention bundle, an infectious disease physician, and a QI professional.
Potential patients were identified by the gastroenterologists and specialized CVC nurse during an ISS clinic visit or hospitalization. Families were offered the CCABSI prevention bundle with ELP if the patient weighed 5 kg or more, was clinically stable, had a functional silicone-based CVC that was needed for PN delivery for at least 1 month, was not allergic to alteplase, was not receiving citrate or metronidazole, and if the parents were willing and able to comply with the ELP procedure at home. Citrate and metronidazole were not allowed because of the respective risks of precipitation and theoretical concerns of disulfiram-like reactions if the ethanol had to be flushed through the CVC.
The Nationwide Children’s Hospital Institutional Review Board determined that this QI project did not meet the definition of human participants research and did not require Institutional Review Board review or consent.
At the start of the IF QI initiative in February 2012, the ISS service was already actively involved in the hospital wide QI health care–associated CCABSI prevention collaborative that began in March 2011 and included insertion, maintenance, and dressing change bundles. Using the Institute for Healthcare Improvement model, the QI team developed a key driver diagram with the aim of reducing combined health care–associated and community-acquired CCABSI rates in eligible patients with IF by 50% by April 30, 2012, and sustaining the improvement through December 31, 2013. The QI team created a clinical practice guideline that incorporated ELP into the CCABSI prevention bundle, which was accepted by gastroenterologists and infectious disease clinicians and approved by the NCH Pharmacy and Therapeutics Committee. The ELP CCABSI prevention bundle was offered in February 2012 to patients with IF with a history of having 2 or more CCABSIs in the previous year. The QI intervention was implemented and tested using rapid-cycle small tests of change (plan, do, study, and act). In October 2012, the ELP CCABSI prevention bundle was expanded to all patients with IF, regardless of CCABSI history, who had no other contraindication to receipt of ethanol. Collected data were presented at monthly infectious disease QI meetings and quarterly to the NCH QI leadership team.
The ELP CCABSI prevention bundle consisted of interventions that were already performed in the hospital but added the 70% ethanol lock as prophylaxis to be performed both in the hospital and home (Box). Central venous catheter use and troubleshooting of the ELP CCABSI prevention bundle were assessed and tracked by a core team of 3 people (P.L.H., M.C.D., and J.P.B.). The ethanol solution was prepared in a sterile fashion by the patient’s home care or NCH pharmacy and was provided in individual syringes each week for home use with a 9-day expiration date.
Prevention bundle includes:
Daily dressing/site assessment and weekly sterile dressing changes by a specialized CVC nurse or parent
In hospitalized patients only, daily chlorhexidine wipes (if ≥2 mo of age and no known allergy)
Use of two 15-s alcohol scrub/dry to the central catheter hub with each catheter entry
Use of alcohol-impregnated disinfection caps (Curos)
Caps changed daily in hospitalized patients
ELP was prepared by diluting a 98% ethanol ampule with sterile water for a final concentration of 70%
First-time ELP procedure:
The specialized CVC nurse performed the initial ELP instillation in the GI clinic or inpatient floor
Type of CVC material was confirmed as silicone
Catheter was deemed functional by CVC nurse (able to draw and flush easily)
Alteplase was instilled into CVC lumen for minimum of 2 h and withdrawn in its entirety
CVC was flushed with 5 to 10 mL of normal saline
CVC nurse determined CVC volume by drawing back until blood first entered the syringe attached to a catheter with a cap for a Broviac and an extension piece and cap for a PICC
CVC nurse determined the total volume of ethanol to be instilled for ELP by measuring the CVC volume on each patient and adding a correction factor:
In children weighing <15 kg, 0.1 mL plus CVC volume
In children weighing ≥15 kg, 0.2 mL plus CVC volume
Maximum volume of ethanol to be used for ELP was 3 mL
Parents were taught the ELP procedure and demonstrated understanding and competency
Families received a detailed instruction sheet as a reference to be presented to any health care professional who would need to access the CVC
Subsequent ELP procedures:
Patients with double-lumen CVCs alternated the lumen of the ELP daily
Heparin was deleted from all PN and medication orders both at home and in the hospital and could not be used for flushing
When used in the hospital setting, ethanol was removed from the CVC prior to the patient leaving the floor for any procedure
CVC flushed with 5 to 10 mL of normal saline
Instillation of the predetermined volume of 70% ethanol into the CVC
The 70% ethanol volume was allowed to dwell for a minimum of 2 h and a maximum of 24 h daily, depending on the patient’s scheduled use of the CVC for medication and PN, while still allowing for the fewest number of catheter entries throughout the day
At the end of ethanol lock dwell period, twice the amount of the predetermined ethanol volume was withdrawn from the CVC
CVC was flushed with 5 to 10 mL of normal saline before administering any medications or PN
Abbreviations: CVC, central venous catheter; ELP, ethanol lock prophylaxis; GI, gastrointestinal; PICC, peripherally inserted central catheter; PN, parenteral nutrition.
Intestinal failure was defined as residual intestinal function inadequate to sustain growth, hydration, or electrolyte homeostasis and included children with short-bowel syndrome from any cause, intestinal motility disorders, and mucosal enteropathies. A CCABSI was considered health care associated if it met the National Healthcare Safety Network guidelines for CCABSIs.19 Laboratory evaluation for a CCABSI consisted of drawing blood from the CVC for bacterial aerobic, bacterial anaerobic, and isolator quantitative blood cultures. Laboratory-confirmed bloodstream infections associated with clinical symptoms that were not related to an infection at another site and occurred within 48 hours of hospitalization were considered community acquired. For CCABSI with coagulase-negative Staphylococcus species, patients had to have either 2 or more days of bacteremia with coagulase-negative Staphylococcus or fever and isolation of coagulase-negative Staphylococcus from 2 distinct blood culture bottles drawn at the same time not related to another site of infection, as well as have completed appropriate antibiotic therapy as deemed necessary by a clinician. Central catheter–associated bloodstream infection rates were defined as CCABSIs per 1000 catheter days. The number of new central catheter insertions included replacement for any reason (malfunction, occlusion, or CCABSI). The number of hospitalizations included hospitalization for any reason and hospitalization specifically for fever and CCABSI.
Descriptive statistics were used for demographic characteristics of the cohort. The primary outcome measure was CCABSIs per 1000 catheter days. Monthly CCABSI rates were plotted as a function of catheter days and, using statistical process control methods, a U chart was used to represent the results of CCABSI implementation, with upper and lower control limits determined using standard statistical process control parameters (±3 SD). The U-chart denominator was the total number of catheter days in patients who received ELP both pre- and post-ELP implementation; the same 24 patients were analyzed across time. To assess safety, a subset of patients who received the ELP CCABSI prevention bundle for a prolonged period (≥3 months) was analyzed for differences in outcomes pre- and post-ELP implementation. The measured outcomes included CCABSI absolute numbers and rates per 1000 catheter days, hospitalizations, catheter adverse events including need for repairs, alteplase or cysteine, and new CVC insertion for any reason. The number of hospitalizations, catheter insertions, and use of alteplase or cysteine were also confirmed with data derived from the NCH electronic data warehouse and the pharmacy dispensing registry. Wilcoxon signed rank tests were used to compare differences between skewed distributions where patients served as their own controls and the timeframe before and after the bundle implementation was the same per individual patient to reduce bias. Pre- and post-ELP implementation CCABSI rates in this subset of patients were also compared using a 2-sample Poisson rate exact test (Minitab, version 22.214.171.124). A P value of less than .05 was considered statistically significant.
Nationwide Children’s Hospital complies with the Centers for Disease Control and Prevention–based best-practice central catheter care insertion and maintenance bundles, with a hospital initiative that started in March 2011, and was reflected in a hospital wide CCABSI rate of 0.95 per 1000 catheter days. In this study, patients with IF had received the standard NCH BSI bundle for almost 1 year before adding ELP to the CCABSI prevention bundle in February 2012. Compliance with the hospital CCABSI maintenance bundle in the IF population was 97%, with a rate of 1.2 CCABSIs per 1000 catheter days in the gastroenterology ward from January 1, 2011, to February 1, 2012. The overall combined (community-acquired and health care–associated) baseline CCABSI rate in the IF population was 6.99 CCABSIs per 1000 catheter days at the start of the QI initiative.
Twenty-four of 26 eligible children with IF (92%) received the ELP CCABSI prevention bundle for a median of 266 catheter days (range, 12-635 days; Table 2). Of the remaining eligible patients, 1 declined and the other was receiving a different prophylaxis regimen started at another institution; these patients were not included in the analysis. Eighteen children (75%) were receiving oral antibiotics (amoxicillin/clavulanate, trimethoprim/sulfamethoxazole, or ciprofloxacin most commonly) for small-bowel bacterial overgrowth or motility and all had gastrostomy tubes and/or enterostomies in place. No patients were receiving metronidazole at the start of ELP; however, as outpatients, 10 children (42%) received metronidazole concomitantly with ELP with no adverse events noted.
Seven months after implementing the ELP CCABSI prevention bundle, overall CCABSI rates decreased from 6.99 to 3.83 CCABSIs per 1000 catheter days (Figure). Given this improvement, the ELP CCABSI prevention bundle was then offered to all patients with IF regardless of history of CCABSI or age at the CVC. During the next 14 months, CCABSI rates decreased to 0.42 CCABSIs per 1000 catheter days after ELP implementation in all eligible patients, despite an increase in total number of catheter days (Figure).
Pre-ELP implementation, CCABSIs were primarily community acquired and the microbiology was heterogeneous, with 35% of CCABSIs being polymicrobial. No individual patient had an increase in CCABSIs while receiving ELP. Post-ELP intervention, 2 patients had 3 breakthrough CCABSIs (2 CCABSIs from coagulase-negative staphylococci in 1 patient and 1 polymicrobial bacterial CCABSI in another); the polymicrobial CCABSI was health care associated.
Four patients no longer required PN and their CVCs were removed; thus, the bundle was stopped after a median of 56 days of ELP (range, 20-79 days). Two patients had mechanical difficulties with their CVC at 34 and 310 days after starting ELP. One patient had complete occlusion of the CVC requiring replacement and the other had difficulty with flushing the CVC but the CVC was retained; both these families chose not to continue with ELP. No other patients reported adverse effects with ELP at the doses provided.
A subset of 14 patients in whom ELP was continuously used for 3 months or more (median, 415 days [range, 104-635 days]) was further analyzed to capture possible complications associated with prolonged ethanol use (Table 3 and Table 4). The median time of analysis before and with ELP intervention per individual patient was 12 months (range, 3-17 months). The pre-ELP intervention CCABSI rates in this subset was 7.01 CCABSIs per 1000 catheter days compared with post-ELP intervention rates of 0.64 (P = .004). There were also significantly fewer central catheter insertions in these patients after ELP was implemented in the prevention bundle (P = .001). There were no significant differences in the total number of hospital admissions. There was a significant difference in the number of hospital admissions for fever and CCABSI after implementation of the ELP CCABSI prevention bundle (P = .003). There were no significant differences in CVC-related complications (defined as central catheter insertions for any reason, use of alteplase, central catheter repairs, and catheter occlusions; Table 4) pre- and post-implementation of the ELP CCABSI prevention bundle.
Long-term PN via a CVC is an integral part of the treatment for patients with IF. The number and type of CCABSIs, recurrent sepsis episodes, and their contribution to worsening cholestatic liver disease account for most of the morbidity and mortality in children with IF.20 Despite successful hospital interventions and strict adherence to the Centers for Disease Control and Prevention recommendations for central catheter maintenance care bundles,21,22 overall CCABSI rates in our patients with IF remained unacceptably high, demonstrating an area for improvement. Guidelines from the Infectious Diseases Society of America state that the use of an ethanol lock can be considered for therapy of mixed bacterial infections but there remains insufficient data for infectious disease consultants to recommend ethanol locks for either standard therapy or prevention of CCABSIs.23 It is highly unlikely that a multi-institutional clinical trial will be able to be performed that will provide the adequate power needed to confirm the effectiveness and acceptable adverse effect profile of ELP in children with IF, nor be able to take into account the varying ethanol concentrations, frequencies (daily vs weekly), and optimal dwell times in distinct catheter types. A meta-analysis by Oliveira et al24 identified 4 retrospective studies in 53 pediatric patients with IF and found that the use of ELP decreased CCABSI rates by 81% compared with traditional heparin locks. Our work in patients with IF shows a significant and sustained reduction in CCABSI rates associated with the implementation of an ELP CCABSI prevention bundle.
Most CCABSIs in our patients with IF were community acquired. This provided a unique opportunity for a multidisciplinary QI initiative to evaluate the efficacy of adding ELP to a CCABSI prevention bundle in children with IF in both the home and hospital settings. Our results demonstrated a significant reduction in CCABSI rates from a baseline of 6.99 to 0.42 CCABSIs per 1000 catheter days and a decreased need for CVC replacement for any reason after the introduction of ELP to the CCABSI prevention bundle in patients with IF. This is important because the loss of vascular access sites owing to repeated catheter removal from infectious complications jeopardizes the already limited venous access sites for CVC insertion for nutritional delivery. Our cohort of 24 patients did not change during the course of the intervention but had an increase in overall catheter retention and decrease in new central catheter insertion, which was similar to results from other studies.14,15,18
There are case reports in the literature regarding catheter-related complications associated with ethanol lock use.17 In our cohort, we limited the use of ELP to patients with silicone-based CVCs based on concerns from prior studies of possible mechanical complications with the use of ethanol locks in polyurethane catheters. However, there is a paucity of data regarding catheter integrity with use of long-term ethanol as would be used in the setting of prophylaxis. Crnich et al25 found that there were no significant differences in catheter integrity in silicone vs polyurethane catheters that had been locked with 70% ethanol for a 10-week period. Guenu and colleagues26 performed both electron microscopy and gas chromatography/mass spectrometry to confirm that silicone dialysis catheters exposed to 60% ethanol for 15 days had neither luminal surface damage nor evidence of structural degradation. Catheters exposed to 95% ethanol for 15 days did not demonstrate changes in their luminal surface; however, there was evidence of release of cyclic polydimethylsiloxane polymers from the catheter surface, most prominent in the first 4 hours after the initial immersion in ethanol. Our patients received daily 70% ethanol locks for a prolonged period (median, 266 days [range, 12-635 days]) but did not demonstrate an increase in catheter repairs or complications compared with preintervention numbers. A pediatric case series also confirmed the lack of adverse CVC-related events on follow-up despite prolonged ethanol lock use.18 One patient in our cohort had an adverse event, specifically, a CVC-related occlusion leading to CVC replacement, which was thought to be secondary to ethanol. However, we were unable to prove that this event was directly caused by the ELP, which the patient had been receiving for 310 days. No patients experienced adverse clinical symptoms attributable to the ethanol (eg, dizziness or mental status changes). This may be owing to the low volumes of ethanol that were tailored to each individual’s catheter size and the fact that we did not routinely flush the ethanol into the CVC after dwell times. Although we have confirmation of outpatient ELP delivery, we were unable to measure compliance with the ELP CCABSI bundle at home. Thus, it is difficult to parse how much of the reduction in overall CCABSI rates was owing to improvement with bundle compliance vs institution of ELP therapy alone.
Despite these limitations, our study is unique in that it is the largest study, to our knowledge, evaluating the long-term use of a standardized CCABSI bundle with ELP in both hospital and home settings for the prevention of CCABSIs in pediatric patients with IF. There was a statistically significant reduction in CCABSI rates despite an increase in overall catheter days. Further, our study has demonstrated the greatest reduction in CCABSI rates, to our knowledge, with more than 90% total reduction in this high-risk population, despite an already lower CCABSI rate compared with contemporary cohorts (Table 1). Use of QI methods have also allowed us to quantify and confirm the effect and sustainability of the intervention over a prolonged time.
Our data support the efficacy of ELP as part of a best-practice bundle for CCABSI prevention in pediatric patients with IF. Our ELP CCABSI prevention bundle can be used on a routine basis in clinical practice to decrease preventable harm in the form of CCABSI, both in the hospital and at home, without an increase in adverse events.
Corresponding Author: Monica I. Ardura, DO, MSCS, Nationwide Children’s Hospital, Infectious Diseases and Immunology, 700 Children’s Dr, C5C.J5428, Columbus, OH 43205 (email@example.com).
Accepted for Publication: October 30, 2014.
Published Online: February 2, 2015. doi:10.1001/jamapediatrics.2014.3291.
Author Contributions: Drs Ardura and Balint had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Ardura, Lewis, Harp, Dienhart, Balint.
Acquisition, analysis, or interpretation of data: Ardura, Lewis, Tansmore, Balint.
Drafting of the manuscript: Ardura, Lewis, Harp.
Critical revision of the manuscript for important intellectual content: Ardura, Lewis, Tansmore, Dienhart, Balint.
Statistical analysis: Ardura, Lewis.
Administrative, technical, or material support: Tansmore, Harp.
Study supervision: Lewis.
Conflict of Interest Disclosures: None reported.
Additional Contributions: We thank Tami Kelly, RN, BSN, CIC, an epidemiology nurse specialist at Nationwide Children’s Hospital, and Richard Brilli, MD, and Wallace Crandall, MD, Nationwide Children’s Hospital and The Ohio State University College of Medicine, for their critical review of the manuscript. They did not receive financial compensation for their contributions.