Chester Clinic

 

Impregnated central venous catheters for prevention of bloodstream infection in children (the catch trial): a randomised controlled trial

Impregnated central venous catheters for prevention of
bloodstream infection in children (the CATCH trial):
a randomised controlled trial
Ruth E Gilbert, Quen Mok, Kerry Dwan, Katie Harron, Tracy Moitt, Mike Millar, Padmanabhan Ramnarayan, Shane M Tibby, Dyfrig Hughes, Carrol Gamble, for the CATCH trial investigators* Summary
Background Impregnated central venous catheters are recommended for adults to reduce bloodstream infections but Published Online
not for children because there is not enough evidence to prove they are eff ective. We aimed to assess the eff ectiveness March 3, 2016
of any type of impregnation (antibiotic or heparin) compared with standard central venous catheters to prevent http://dx.doi.org/10.1016/
bloodstream infections in children needing intensive care.
Methods We did a randomised controlled trial of children admitted to 14 English paediatric intensive care units. S0140-6736(16)00566-3
Children younger than 16 years were eligible if they were admitted or being prepared for admission to a participating *Members listed at end of paper
paediatric intensive care unit and were expected to need a central venous catheter for 3 or more days. Children were UCL Institute of Child Health,
randomly assigned (1:1:1) to receive a central venous catheter impregnated with antibiotics, a central venous catheter London, UK (Prof R E Gilbert MD,
impregnated with heparin, or a standard central venous catheter with computer generated randomisation in blocks of K Harron PhD); Paediatric and
Neonatal Intensive Care Unit
three and six, stratifi ed by method of consent, site, and envelope storage location within the site. The clinician (Q Mok MB) and Children's
responsible for inserting the central venous catheter was not masked to allocation, but allocation was concealed from Acute Transport Service
patients, their parents, and the paediatric intensive care unit personnel responsible for their care. The primary (P Ramnarayan MD), Great
outcome was time to fi rst bloodstream infection between 48 h after randomisation and 48 h after central venous Ormond Street Hospital for
Children, London, UK;
catheter removal with impregnated (antibiotic or heparin) versus standard central venous catheters, assessed in the Department of Biostatistics
intention-to-treat population. Safety analyses compared central venous catheter-related adverse events in the subset of (K Dwan PhD,
children for whom central venous catheter insertion was attempted (per-protocol population). This trial is registered Prof C Gamble PhD) and
with ISRCTN number, ISRCTN34884569.
Medicines for Children Clinical
Trials Unit University (T Moitt),
University of Liverpool,
Findings Between Nov 25, 2010, and Nov 30, 2012, 1485 children were recruited to this study. We randomly assigned Liverpool, UK; Department of
502 children to receive standard central venous catheters, 486 to receive antibiotic-impregnated catheters, and 497 to Infection, Barts Health NHS
receive heparin-impregnated catheters. Bloodstream infection occurred in 18 (4%) of those in the standard catheters Trust, London, UK
(M Millar PhD); Evelina
group, 7 (1%) in the antibiotic-impregnated group, and 17 (3%) assigned to heparin-impregnated catheters. Primary Children's Hospital,
analyses showed no eff ect of impregnated (antibiotic or heparin) catheters compared with standard central venous St Thomas's Hospital, London,
catheters (hazard ratio [HR] for time to fi rst bloodstream infection 0·71, 95% CI 0·37–1·34). Secondary analyses UK (S M Tibby MB); and Centre
showed that antibiotic central venous catheters were better than standard central venous catheters (HR 0·43, for Health Economics and
Medicines Evaluation, Bangor
0·20–0·96) and heparin central venous catheters (HR 0·42, 0·19–0·93), but heparin did not diff er from standard University, Bangor, Gwynedd,
central venous catheters (HR 1·04, 0·53–2·03). Clinically important and statistically signifi cant absolute risk Wales (Prof D Hughes PhD)
diff erences were identifi ed only for antibiotic-impregnated catheters versus standard catheters (–2·15%, 95% CI Correspondence to:
–4·09 to –0·20; number needed to treat [NNT] 47, 95% CI 25–500) and antibiotic-impregnated catheters versus Prof Ruth E Gilbert, UCL Institute
heparin-impregnated catheters (–1·98%, –3·90 to –0·06, NNT 51, 26–1667). Nine children (2%) in the standard of Child Health,
London WC1N 1EH, UK central venous catheter group, 14 (3%) in the antibiotic-impregnated group, and 8 (2%) in the heparin-impregnated [email protected]
group had catheter-related adverse events. 45 (8%) in the standard group, 35 (8%) antibiotic-impregnated group, and
29 (6%) in the heparin-impregnated group died during the study.
Interpretation Antibiotic-impregnated central venous catheters signifi cantly reduced the risk of bloodstream
infections compared with standard and heparin central venous catheters. Widespread use of antibiotic-impregnated
central venous catheters could help prevent bloodstream infections in paediatric intensive care units.
Funding National Institute for Health Research, UK.
frequent cause of bloodstream infections.1,2 US studies3–5 Bloodstream infections are important causes of adverse report the success of improved aseptic practices during clinical outcomes and costs to health services. Paediatric insertion and maintenance of central venous catheters intensive care units have one of the highest reported for reducing rates of catheter-related bloodstream rates of hospital-acquired bloodstream infections of any infections. The Department of Health in England clinical specialty, with central venous catheters being a invested in similar infection reduction initiatives, www.thelancet.com Published online March 3, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00340-8
Research in context
Evidence before this study
Added value of this study
We searched PubMed on Sept 1, 2014, initially for systematic To our knowledge, this is the fi rst trial to assess antibiotic and reviews or meta-analyses, using the clinical queries fi lter for heparin-impregnated central venous catheters in children and therapy studies or terms for meta-analysis and (catheter* OR in the context of low bloodstream infection rates associated central OR venous OR intravenous), (impregnated OR bonded OR with improved asepsis practices. We add new evidence of coated OR antibiotic OR heparin), and infection, with no language eff ectiveness of antibiotic central venous catheters for any restrictions. We identifi ed fi ve systematic reviews published since bloodstream infection, showing a 57% reduction compared 2008. The two most recent reviews were both published in the with standard central venous catheters in children. Cochrane Library. One included any type of central venous We confi rmed the eff ectiveness of antibiotic central venous catheter impregnation, but excluded children (consisting of catheters identifi ed in systematic reviews of trials in adults, with 56 randomised controlled trials, fi ve studies of antibiotic- a 75% reduction in the risk of catheter-related bloodstream impregnated catheters vs standard catheters, and 1 study of infections (HR 0·25, 0·07–0·90) compared with standard central heparin-impregnated catheters vs standard catheters). The other venous catheters, for the first time in children. We also report compared heparin bonded with standard central venous catheters for the first time that antibiotic central venous catheters are in children (two trials). All the trials assessed in these two reviews superior to heparin central venous catheters. These results are were included in an earlier systematic review and network based on secondary analyses so need to be interpreted with meta-analysis that comprised direct and indirect mixed treatment caution. Our results are consistent with previous studies comparisons of 45 randomised controlled trials assessing showing no effect of antibiotic impregnation on mortality or catheter-related bloodstream infections (consisting of six studies adverse effects.
of antibiotic-impregnated catheters vs standard catheters—none By contrast with evidence from systematic reviews, we of these studies were in children; and three studies of identifi ed no signifi cant eff ect for heparin-bonded versus heparin-impregnated catheters vs standard catheters, two of standard central venous catheters. The lack of eff ectiveness of which were in children). For antibiotic (minocycline–rifampicin) heparin central venous catheters might relate to the low compared with standard central venous catheters, a pooled odds baseline event rate noted in this study, which was done after ratio (OR) for catheter-related bloodstream infection of 0·18 implementation of central venous catheter care bundles in (95% CI 0·08–0·34) was reported. We identifi ed one subsequent paediatric intensive care units to improve asepsis procedures randomised controlled trial which compared antibiotic during central venous catheters insertion and maintenance. (minocycline and rifampicin) and standard central venous Another potential explanation could be emergence of catheters for children undergoing heart surgery. The trial of resistance to benzalkonium chloride, the bonding agent used 288 participants was terminated early because of a low event rate for heparin, which is widely used in hand hygiene products.
(three catheter-associated bloodstream infections in each group). The mixed treatment comparison for heparin-bonded versus Implications of the available evidence
standard central venous catheters produced a pooled OR of 0·20 When combined with previous systematic reviews, our fi ndings (0·06–0·44), and for antibiotic-impregnated catheters compared establish the eff ectiveness of antibiotic-impregnated central with heparin central venous catheters (indirect comparisons venous catheters compared with standard central venous only), OR 1·18 (0·28–3·29). A previous cost-eff ectiveness analysis catheters and extend this evidence for paediatric use. based on trials in adults estimated that impregnated central Widespread use of antibiotic-impregnated central venous venous catheters would be cost eff ective even at baseline risks of catheters could help prevent bloodstream infections in bloodstream infection as low as 0·2%.
paediatric intensive care units. including the Saving Lives central venous catheter care of direct and indirect comparisons of impregnated bundle and the Matching Michigan scheme.6–8 and standard central venous catheters identifi ed Use of central venous catheters that are impregnated, that heparin-bonded or antibiotic-impregnated central for example with antibiotics, chlorhexidine, or heparin, venous catheters were the most eff ective options, with an has been recommended as part of these infection associated 70–80% reduction in the risk of catheter-reduction initiatives in the USA and the UK, but only for related bloodstream infections.
adults at high risk of bloodstream infections.7,9 Despite the large number of randomised controlled Impregnated central venous catheters have not been trials, uncertainty remains about the strength of the recommended for children.10 The evidence for reduced evidence for using impregnated central venous catheters, rates of catheter-related bloodstream infections with particularly for children. First, inherent biases exist in impregnated compared with standard central venous the use of catheter-related bloodstream infections (the catheters derives from trials predominantly of adults. primary outcome used in all previous trials) as a primary Systematic reviews11–15 draw on evidence from outcome because this could overestimate benefi ts of 56 randomised controlled trials. A network meta-analysis14 antibiotic impregnation.11,16 Criteria for catheter-related www.thelancet.com Published online March 3, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00340-8
bloodstream infection require positive cultures of the venous catheter insertion to receive a central venous same organism in the central venous catheter tip and in catheter impregnated with antibiotics, a central blood. This requirement might favour reduced catheter- venous catheter impregnated with heparin, or a standard related bloodstream infection in antibiotic-impregnated central venous catheter. The clinician or research nurse central venous catheters because antibiotics in the opened a pressure-sealed, sequentially numbered, opaque catheter tip might inhibit bacterial growth in culture envelope containing the central venous catheter media.17 Second, few studies6,7,18 have been done in the allocation. Randomisation sequences were computer context of the low infection rates associated with generated by an independent statistician in random improved asepsis programmes. Third, very few of these blocks of three and six, stratifi ed by method of consent, trials were in children.19–21 Compared with adults, children site, and envelope storage location within the site to help need narrower central venous catheters, which with easy access to envelopes (eg, for insertion in theatre thrombose more readily. Standard, non-impregnated and in paediatric intensive care unit).
central venous catheters are still used for most children The clinician responsible for inserting the central in UK paediatric intensive care units.10 However, there venous catheter was not masked to allocation (because of could be substantial benefi ts for children's health and the diff erent colour of strips for antibiotic and heparin health-care costs if impregnated central venous catheters central venous catheters), but because central venous could be shown to reduce rates of bloodstream infections.
catheters looked identical while in situ, allocation was We aimed to establish the eff ectiveness of any type concealed from patients, their parents, and the paediatric of impregnation (antibiotic or heparin) compared with intensive care unit personnel responsible for their care. standard central venous catheters for prevention of Labels identifying the type of central venous catheter bloodstream infections in children needing intensive received were held securely in a locked drawer in case care. A secondary aim was to establish which of the unmasking was needed. Participant inclusion in analyses three types of central venous catheter was most eff ective. and occurrence of outcome events were established We also investigated the eff ectiveness of type of central before release of the randomisation sequence for analysis venous catheter on catheter-related bloodstream and for the data monitoring committee.
infections, duration of care, and safety, including mortality and adverse events, such as antibiotic resistance.
Procedures
Participation in the trial did not involve any changes to
standard clinical care or data collection apart from Study design and population
collection of an additional 0·5 mL of blood whenever a The CATCH trial was a pragmatic, three-group, blood culture sample was taken. The sample was sent for randomised controlled trial of children admitted to PCR testing for 16S ribosomal RNA (rRNA) of bacterial 14 paediatric intensive care units in England between ribosome protein to detect bacterial infection. All December, 2010, and November, 2012. Children younger randomly assigned and consented participants were than 16 years were eligible if they were admitted or being followed up until 48 h after central venous catheter prepared for admission to a participating paediatric removal or attempted central venous catheter insertion. intensive care unit and were expected to need a central The research nurse assessed routinely recorded daily venous catheter for 3 or more days. For children admitted hospital records up until 48 h after catheter removal to to paediatric intensive care units after elective surgery, we assess primary and secondary outcomes. To corroborate sought prospective written parental consent during data and to measure admissions, discharges, or death preoperative assessment. For children who needed a more than 48 h after catheter removal, we linked the central venous catheter as an emergency, we sought child's hospital administrative records, death certifi cation written parental consent after randomisation and records, and records from the national Paediatric Audit stabilisation (deferred consent) to avoid delaying treatment. Network (PICANet22) for 6 months after randomisation Parents consented to the use of their child's data for the to the child's trial record. We extracted the Paediatric trial, to follow-up using routinely recorded clinical data, Index of Mortality23 at admission to intensive care from and to an additional 0·5 mL of blood being collected for the PICANet database.
PCR testing whenever a blood culture was clinically Both types of impregnation involved internal and external needed (appendix). The Research Ethics Committee for surfaces. We used polyurethane central venous catheters See Online for appendix
South West England approved the study protocol (reference
manufactured by Cook Medical Incorporated, IN, USA. number 09/H0206/69). The protocol and the statistical Sizes used were French gauge 4 (double lumen), 5, or 7
analysis plan are available online and the full statistical (triple lumen). Cook reports a concentration of 503 μg/cm For the protocol see http://www.
analysis report is available on request from the authors.
minocycline and 480 μg/cm rifampicin for their antibiotic- nets.nihr.ac.uk/projects/impregnated central venous catheters, which reduces hta/081347 Randomisation and masking
biofi lm formation.24 Heparin bonding reduces thrombus Children were randomly assigned (1:1:1) at the bedside or and thereby biofi lm formation and uses benzalkonium in theatre (operating room) immediately before central chloride as an anti-infective bonding agent.16,25 www.thelancet.com Published online March 3, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00340-8
after randomisation and 48 h after central venous catheter The primary analysis for the trial compared any removal (or before death). All blood culture samples impregnated central venous catheters (antibiotic or included in the primary outcome were clinically heparin) with standard central venous catheters. Secondary indicated, defi ned by recorded evidence of infection analyses involved pairwise comparisons for the three types (one or more of: temperature instability, change in of central venous catheters for the primary outcome.
inotrope requirements, haemodynamic instability, or The primary outcome was time to the fi rst bloodstream poor perfusion) or removal of the central venous catheter infection based on blood cultures taken between 48 h because of suspected infection. Blood cultures were 1859 randomly assigned 1485 randomly assigned and 374 randomly assigned and deferred consent obtained consent not obtained* 501 prospective consent 180 not approached 984 deferred consent 17 no response consent 177 consent refused 502 allocated standard central 486 allocated antibiotic- 497 allocated heparin-impregnated 374 assigned but could no longer impregnated central venous central venous catheter participate in study 122 standard central venous 126 antibiotic-impregnated central venous catheter 126 heparin-impregnated central venous catheter 34 did not receive standard 49 did not receive antibiotic- 57 did not receive heparin- central venous catheter impregnated central impregnated central 1 received antibiotic- impregnated central 23 received standard 22 received standard central central venous catheter 12 received heparin- 5 received heparin- 2 received antibiotic- impregnated central impregnated central impregnated central 21 received no catheter 21 received no catheter 33 received no catheter 15 insertion attempted 14 insertion attempted 24 insertion attempted 468 received standard central 437 allocated central venous 440 allocated central venous catheter impregnated catheter impregnated 502 included in intention-to-treat 486 included in intention-to-treat 497 included in intention-to-treat Clinical indications for blood culture Clinical indications for blood culture Clinical indications for blood culture 213 had at least 1 blood culture 190 had at least 1 blood culture 190 had at least 1 blood culture sample taken for inclusion sample taken for inclusion sample taken for inclusion in sensitivity analysis in sensitivity analysis in sensitivity analysis 8 had no blood cultures 6 had no blood cultures 3 had no blood cultures 281 had no clinical indications 290 had no clinical indications 304 had no clinical indications Figure 1: Trial profi le
*Further details reported elsewhere.27 †Based on a clinically indicated blood culture sample taken at least 48 h after randomisation and less than 48 h after central
venous catheter removal. ‡Samples used in sensitivity analysis.
www.thelancet.com Published online March 3, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00340-8
Standard CVC Antibiotic-
Standard CVC Antibiotic-
impregnated impregnated
impregnated impregnated
CVC (n=486)
CVC (n=497)
CVC (n=486)
CVC (n=497)
Emergency (deferred (Continued from previous column) CVC inserted at same hospital† Elective (prospective Intensive care unit Operating theatre CVC inserted at other hospital†‡ Intensive care unit Operating theatre Prospective consent, Weight at admission (kg) CVC inserted at same hospital† Intensive care unit Operating theatre Triple lumen CVC† Admitted for surgery CVC inserted into PICU assessment (from 479 (95%) linked PICANet data) Primary reason for admission* Data are n (%). PICU=paediatric intensive care unit. PICANet=Paediatric Intensive Care Audit Network. CVC=central venous catheter. *n=479 for standard CVCs, n=456 for antibiotic-impregnated CVCs, and n=497 for heparin-impregnated CVCs. †n=481 for standard CVCs, n=465 for antibiotic-impregnated CVCs, and n=464 for heparin-impregnated CVCs. ‡CVCs were inserted by the retrieval team before to transfer to PICU.
Table 1: Baseline characteristics
recorded as positive for the primary outcome if any organism was isolated that was not a skin commensal bacterium or if coagulase-negative staphylococci (or other Paediatric Index of Mortality (PIM2) score* skin commensal bacteria) were isolated and there were two or more positive cultures of the same organism within 48 h of each other. A clinical committee reviewed all primary outcomes involving positive cultures without knowledge of central venous catheter allocation status. A sensitivity analysis assumed that the primary outcome occurred for those with a record of clinical indication Clinical condition <72 h before randomisation but no blood culture taken in the primary outcome time window.
Secondary bloodstream infection-related outcomes Antibiotics received were catheter-related bloodstream infection based on the same organisms cultured from blood and the central venous catheter tip between 48 h after randomisation Clinical condition at randomisation and 48 h after central venous catheter removal, Infection suspected diff erential positivity of cultures from several central venous catheter lumens on two or more occasions, or exit site infection or central venous catheter removed for infection; rate of bloodstream infection per 1000 central Deferred consent, venous catheter-days based on one or more bloodstream infections between random isation and central venous (Table 1 continues in next column) catheter removal; and time to a composite measure of bloodstream infection comprising the primary outcome or a negative blood culture combined with a positive www.thelancet.com Published online March 3, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00340-8
compared the time to paediatric intensive care unit discharge, hospital discharge, and death within 30 days of randomisation between groups. Deaths were recorded by Bloodstream infection the research team or by linkage to death certifi cation data Median time to fi rst bloodstream 7·5 (4·5–11·2) 6·9 (6·0–8·0) 4·2 (3·1–8·4) ce of National Statistics. Cost-eff ectiveness analyses based on linked hospital resource data for 6 month follow-up will be reported elsewhere.
Safety analyses compared central venous catheter- related adverse events (including unexplained thrombo- Organism group† cytopenia after insertion of central venous catheter), mortality, and antibiotic resistance to minocycline (>0·5 μg/mL) or rifampicin (>1·0 μg/mL) based on etest strips (bioMérieux, USA) applied to organisms isolated from bloodstream infections. Incomplete laboratory Catheter-related bloodstream testing and reporting limited analyses of resistance in positive blood cultures and prevented analysis of Bloodstream infection rate per 8·24 (4·72–11·77) 3·31 (1·01–5·60) 8·78 (5·03–12·55) resistance in cultures from the central venous catheter 1000 central venous catheter-days tip (protocol).
Number of bloodstream infections Bloodstream infection or culture negative infection§ We based the sample size calculation for the primary analysis on relative risk (RR). We assumed that detection Median time to central venous 4·28 (2·30–6·97) 4·31 (2·13–7·0) 4·20 (2·24–6·97) of a RR of 0·5 in patients with a baseline risk of 10% catheter removal (days) would change policy. We assumed that the RR would Mortality ≤30 days after remain constant across baseline risks whereas the absolute risk diff erence would be more variable. 1200 children in a 2:1 ratio (impregnated:standard) were Median time to PICU discharge 5·1 (2·8–10·0) 4·4 (2·2–9·3) 4·9 (2·3–8·9) needed to achieve 80% power to detect an RR of 0·5 at a 5% level of signifi cance, based on an estimated Median time to hospital discharge 12·0 (6·4–25·6) 12·0 (6·7–22·7) 12·1 (6·4–22·5) bloodstream infection rate of 10% and allowing for 5% loss to follow-up. A lower than expected bloodstream Safety analyses¶ infection rate of 5% would have 62% power to detect a Central venous catheter-related RR of 0·5 or 80% power for a RR of 0·32.
Mortality ≤30 days after The independent data monitoring committee recommended continuation of the study until Nov 30, 2012, after reviewing the fi rst 209 children, an Data are n (%), median (IQR), or rate (95% CI), unless otherwise specifi ed. PICU=paediatric intensive care unit. *Includes one mixed bloodstream infection pathogen and skin organism. †Groups add to more than total number of organism interim analysis of 650 children using the Peto-Haybittle groups because multiple types of organisms were isolated on the same occasion in some patients. ‡Includes six stopping rule for the primary outcome had occurred, bloodstream infections due to coagulase negative staphylococci. §Composite measure of bloodstream infection recruitment had reached the original target of the including the primary outcome or a negative blood culture combined with a positive 16S PCR result for bacterial DNA, 1200 preschedule in June, 2012, and there were no safety removal of the central venous catheter because of suspected infection, or a start of antibiotics or change in type of antibiotics on the same or next day. ¶For the safety analysis, n=533 for the standard central venous catheter group, concerns. The recommendation for continuation aimed n=451 for the antibiotic-impregnated catheter group, and n=479 for heparin-impregnated catheter group.
to exhaust available funding.
Outcome data were analysed according to the intention- Table 2: Endpoint frequency according to central venous catheter allocation (intention-to-treat
analyses) and central venous catheter received (safety analyses)
to-treat principle. Safety analyses included the subset of children for whom central venous catheter insertion was For the protocol see http://www.
16S PCR result for bacterial DNA, removal of the central attempted, grouped by the subset in whom central venous venous catheter because of suspected infection, or a start catheters were actually received. A p value of 0·05 was of antibiotics or change in type of antibiotics on the same considered statistically signifi cant and 95% CIs were used or next day.
throughout. Absolute risk diff erences were calculated for Other secondary outcomes included time to central proportions. Time-to-event outcomes were analysed using venous catheter removal and time to central venous Kaplan-Meier curves and log-rank tests. Cox regression catheter thrombosis (defi ned by two episodes within was used to adjust primary and secondary analyses of 5 days of each other of diffi culty fl ushing the central time to fi rst bloodstream infection for prospective or venous catheter or drawing back blood from the deferred consent type and suspected infection at baseline. central venous catheter, one episode of swollen limb, In a post-hoc sensitivity analysis, we used cumulative central venous catheter removal because of thrombosis, incidence curves to assess competing risks from death for or a positive ultrasound indicating thrombosis). We also time to fi rst bloodstream infection. We applied Gray's test www.thelancet.com Published online March 3, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00340-8
Any impregnated vs standard
Antibiotic-impregnated vs standard
Heparin-impregnated vs standard
impregnated (secondary analysis)
Risk diff erence Hazard ratio Risk diff erence Risk diff erence Risk diff erence (–3·04 to 0·75) (–4·09 to –0·20) (–2·45 to 2·12) (–3·90 to –0·06) (–2·58 to 0·45) (0·25 to 1·21) (–3·28 to –0·27) (–2·20 to 1·44) (–2·81 to 0·02) bloodstream infection (–6·36 to 1·94) (0·40 to 1·34) (–9·14 to –0·73) (–4·60 to 5·70) (–9·89 to –1·08) infection per 1000 central venous catheter days (–5·90 to 2·98) (0·75 to 1·20) (–6·26 to 4·03) (–6·87 to 3·30) (–4·41 to 5·75) infection or culture-negative infection (–6·02 to 3·22) (–4·40 to 6·46) (–8·99 to 1·44) (–0·50 to 10·10) ≤30 days after Risk diff erence is based on the risk of the outcome during follow-up and displayed as percentage. PICU=paediatric intensive care unit. *Risk ratios. †Rate ratios. Table 3: Risk diff erences and relative eff ect measures by central venous catheter allocated (intention-to-treat analyses)
to detect whether there was a diff erence between author had full access to all the data in the study and had
impregnated and standard central venous catheters for fi nal responsibility to submit for publication.
the primary outcome.26 For secondary outcomes, binary
outcomes were analysed using the chi squared χ² test and Results
continuous outcomes analysed using the Mann-Whitney Between Nov 25, 2010, and Nov 30, 2012, 1859 children
U test. The rate of bloodstream infection (defi ned as the were randomly assigned (501 prospective,
total number of bloodstream infections per 1000 central 1358 emergency) to one of the central venous catheter
venous catheter-days occurring between randomisation groups. 984 (72%) of the emergency patients
and central venous catheter removal) was analysed using subsequently provided deferred consent, leaving
Poisson regression. All analyses were done with SAS 1485 participants for fi nal analysis (fi gure 1). Reasons for
software version 9.2. This trial is registered with not providing consent are reported in fi gure 1. Of the
ClinicalTrials.gov, number NCT01029717.
1485 randomly assigned participants, 75 (5%) did not receive a central venous catheter. In 53 of these Role of the funding source
participants, insertion was attempted but unsuccessful The manufacturer and the funder of the study had no and in 22, central venous catheter insertion was not role in study design, data collection, data analysis, data attempted (16 no longer required a central venous interpretation, or writing of the report. The corresponding catheter, fi ve reason unknown, and one patient died). www.thelancet.com Published online March 3, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00340-8
coagulase negative staphylococci (table 2). Details of all organisms isolated in the primary outcomes are given in the appendix. All outcomes are reported by central venous catheter type (table 2).
In the primary comparison, time to bloodstream infection did not diff er between impregnated central venous catheters (antibiotic or heparin combined) and standard central venous catheters (hazard ratio [HR] 0·71, favourable outcome 95% CI 0·37–1·34; table 3). In secondary, pairwise comparisons, antibiotic-impregnated central venous catheters reduced the risk of bloodstream infections compared with standard central venous catheters (HR 0·43, 0·20–0·96) and compared with heparin central venous catheters (HR 0·42, 0·19–0·93). Absolute risks of bloodstream infections diff ered signifi cantly at the 5% level only for antibiotic-impregnated central venous Number at risk
Time to first bloodstream infection (days) catheters compared with standard central venous catheters (–2·15%) and heparin-impregnated central venous catheters (–1·98%; table 3). The number needed to treat by using antibiotic-impregnated central venous was 47 Figure 2: Kaplan-Meier curve for time to fi rst bloodstream infection by central venous catheter allocation
(95% CI 25–500) compared with standard and 51 (26–1667) compared with heparin central venous catheters.
Of those who received a central venous catheter, more of Figure 2 shows the Kaplan-Meier curve for time to fi rst those randomly assigned to standard central venous bloodstream infection. No signifi cant diff erence was catheter received the allocated type of central venous observed in time to fi rst bloodstream infection catheter than those assigned to antibiotic-impregnated comparing any impregnated catheters with standard or heparin-impregnated catheter (fi gure 1). Most of the catheters (p=0·9) or heparin-impregnated catheters with central venous catheters received but not allocated standard catheters (p=0·9). The risk of bloodstream were standard central venous catheters (45 [69%] of 65; infection was reduced for antibiotic-impregnated fi gure 1). All randomly assigned and consented catheters compared with standard catheters (p=0·04) participants were followed up until 48 h after central and for antibiotic compared with heparin-impregnated venous catheter removal or attempted central venous catheters (p=0·03). The direction of these results was catheter insertion.
robust to the sensitivity analysis in which the 17 cases More than half (57%) of children were aged younger with clinical indicators but no blood culture taken were than 12 months at admission, with a third aged less than assumed to have a positive bloodstream infection 3 months (table 1). A third of children had surgery before (fi gure 1). The direction of results did not change in the admission to the paediatric intensive care unit and half regression analysis (appendix). Competing risks analyses had cardiovascular problems as their primary diagnosis using Gray's test indicated no diff erence between at admission. Central venous catheter insertion took impregnated and standard central venous catheters for place in the operating room for 437 (89%) of either competing risk (p=0·29 for bloodstream infection 493 participants in the prospective consent (elective) and p=0·89 for death; data not shown).28group, but in only 34 (4%) of 917 of the deferred consent No signifi cant diff erence was reported between any (emergency) group.
impregnated catheter and standard central venous Clinical indicators of infection were recorded during catheters for the risk of catheter-related bloodstream the primary outcome time interval from 48 h after infections (p=0·13). The risk of catheter-related randomisation up to 48 h after central venous catheter bloodstream infection was signifi cantly lower for removal for 610 (41%) of 1485 participants, 593 (97%) of antibiotic-impregnated catheters versus standard whom had blood cultures taken (fi gure 1). Derivation of central venous catheters (p=0·03) and for antibiotic-the primary outcome and the number of bloodstream impregnated catheters versus heparin-impregnated infections excluded from the primary outcome is shown central venous catheters (p=0·03; table 3). The in the appendix. The primary outcome of bloodstream bloodstream infection rate per 1000 central venous infection was recorded for 42 children: 18 (4%) of catheter-days was lowest in the antibiotic group (table 2). 502 with standard central venous catheters, 7 (1%) No children had more than one bloodstream infection of 486 with antibiotic-impregnated catheters, and 17 (3%) while the trial central venous catheter was in place. The of 497 with heparin-impregnated catheters. Gram-positive association between bloodstream infection outcomes is organisms accounted for most bloodstream infections shown by time since randomisation (appendix). The (23 [55%] of 42), of which six (14%) were positive for composite measure of bloodstream infection or culture www.thelancet.com Published online March 3, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00340-8
negative infection did not diff er by central venous infection were recorded in the child, thereby recording catheter (table 2, 3). Indicators of infection contributing an outcome that clinicians would regard as potentially to the composite measure are shown in the appendix. serious and needing treatment. Restriction to clinically No other secondary outcomes were associated with type indicated blood cultures increased the clinical relevance of central venous catheter (table 3).
of the primary outcome, but, by contrast with routine The cohort for safety analyses (per protocol) was based blood culture sampling for all study participants, on children who had a central venous catheter insertion diminished the sensitivity of the study to detect attempted. These analyses comprised more children in bacteraemia. Only 41% of children had clinical indicators the standard group (n=533) than in the antibiotic (n=451) of bloodstream infection recorded during the primary or heparin groups (n=479; table 2; appendix p 2). No outcome interval but nearly all of these had a blood central venous catheter-related adverse events (31 events) culture taken. A third strength is the representativeness or mortality (148 events) were attributed to type of central of the study population in terms of children admitted to venous catheter received (table 2). Two children the 14 largest paediatric intensive care units (of a total of developed thrombocytopenia unrelated to the type of 24) across the country. We were able to enrol a similar central venous catheter. One was allocated to antibiotic proportion of emergency patients (two-thirds) as seen in and the other to heparin central venous catheter (full practice,29 enabled by the inclusion of retrieved children statistical analysis report available from the authors).
and the use of deferred consent.
Testing for antibiotic resistance varied by centre. Only Limitations of this study include the limited power of 12 of the 42 children with the primary outcome the study to detect diff erences in the primary outcome bloodstream infection had minocycline and rifampicin according to the type of central venous catheter. The resistance reported using etest strips; eight of 12 were trial was based on the best available evidence at the resistant to one or both antibiotics (three of fi ve in the time, which indicated large but equivalent benefi ts of standard group, two of two in the antibiotic-impregnated antibiotic and heparin central venous catheters group, and three of fi ve in the heparin-impregnated compared with standard.11–14 The key question, which group; appendix). Most resistance occurred in determined our primary analysis and sample size, was Gram-negative organisms (seven of nine organisms whether these benefi ts occurred in children. Secondary, cultured from eight bloodstream infection episodes; pairwise comparisons addressed which type of appendix). Resistance was detected in two bloodstream impregnated central venous catheter was best, but the infections that were positive for staphylococcal species: trial was not adequately powered to detect the anticipated one allocated to antibiotic and the other to heparin small diff erences between antibiotic and heparin central central venous catheters (appendix).
venous catheters. Power was further reduced by the low baseline rate of bloodstream infections.
Another limitation relates to fi nding that although In this fi rst trial to compare two types of impregnated antibiotic central venous catheters reduced bloodstream central venous catheters with standard central venous infections, we identifi ed no diff erences in secondary catheters in children our primary analysis showed no outcomes such as mortality, duration of central venous evidence of a statistically signifi cant diff erence between catheter insertion, or the composite measure of time to fi rst bloodstream infection for any impregnated bloodstream infections or culture negative infection. One central venous catheters (antibiotic and heparin potential reason is the complex and varied conditions and combined) compared with standard central venous disease processes aff ecting patients receiving intensive catheters. However, antibiotic impregnation reduced the care. Antibiotic central venous catheters might aff ect risk of bloodstream infection by 57% compared with bloodstream infection in these patients but not other standard central venous catheters, and by 58% compared outcomes. For example, none of the deaths were deemed with heparin-bonded central venous catheters. Antibiotic- to be directly attributable to bloodstream infections. A impregnated central venous catheters were associated second reason is the poor specifi city of the secondary with an absolute risk reduction of 2·15% compared with outcomes. Mortality and duration of central venous standard central venous catheters, meaning 47 children catheter placement are aff ected by a number of treatments, (95% CI 25–500) would need to be treated with an not just central venous catheter impregnation, thereby antibiotic-impregnated central venous catheter instead of biasing in favour of a null eff ect for these secondary a standard central venous catheter to prevent one case of outcomes. The reduction in the hazard ratio for antibiotic bloodstream infection.
versus standard central venous catheter was largest for The strengths of this study include the use of any catheter-related bloodstream infections (reduced by 75%), bloodstream infections as a clinically important primary less for bloodstream infections (reduced by 57%), and outcome thereby avoiding the biases inherent in small and not signifi cant for the composite measure of measuring catheter-related bloodstream infections. bloodstream infections or culture negative infection. Of A further strength was the restriction to positive blood these outcomes, catheter-related bloodstream infection is cultures that were clinically indicated, meaning signs of most specifi cally aff ected by antibiotic impregnation, www.thelancet.com Published online March 3, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00340-8
whereas the composite measure of bloodstream infection Contributors
All authors contributed to the design or conduct of the study. REG
is aff ected by other disease and treatment factors, thereby (Chief Investigator), QM, and CG conceived and designed the study. biasing towards the null eff ect.
Statistical analyses were done by KD and CG. Endpoint review for the Another factor likely to bias towards the null eff ect for primary outcome was done by QM, MM, and REG. REG, QM, KD, KH, secondary outcomes is the potential for rescue treatment and CG wrote the paper and all authors commented on the manuscript and approved the fi nal version.
in response to signs of bloodstream infections. Patients
in intensive care units are continuously monitored for Declaration of interests
We declare no competing interests.
changes in their condition and treated promptly. As a
result, signs of infection should be less likely to develop CATCH trial investigators
CATCH trial members: UK Q Mok, T Y Lee, S Riordan (Great Ormond into septic shock given good intensive care management. Street Hospital, London), I Macintosh, J McCorkell, K Stearn, R Mitchell Such responses introduce bias towards the null eff ect for (Southampton General Hospital, Southampton), S Tibby, J Harris, secondary outcomes such as mortality but are diffi cult to P Wellman (Evelina Children's Hospital, London), O Bagshaw, J Spry, measure adequately.
S Laker, N Holdback (Birmingham Children's Hospital, Birmingham), J Roche, S Cooper, D Hewett (Leeds General Infi rmary, Leeds), S Kerr, A further limitation was the fact that antibiotic F Haigh (Alder Hey Children's Hospital, Liverpool), M White, catheters had a brown line on the internal part of the M Schindler, C Traub, N Worrin (Bristol Royal Hospital for Children, catheter, but once inserted, the catheters were identical. Bristol), R Ramaiah, R Patel (Glenfi eld Hospital), D Macrae, S Bacon We found no evidence of diff erential blood culture (Royal Brompton Hospital, London), M Cooper, A Abdulla, A Brewer (St Mary's Hospital, London), R Agbeko, C Mackerness (Royal Victoria sampling by trial group (fi gure 1). The number of Infi rmary, Newcastle upon Tyne), P Davies, D Walsh, L Crate children who received their allocated central venous (Queens Medical Centre, Nottingham), R Agbeko, C Simmister catheter was slightly higher for those in the standard (Freeman Hospital, Newcastle upon Tyne), R Ramaiah, R Patel (Leicester Royal Infi rmary, Leicester). group, probably showing the fact that standard central Trial Steering Committee: UK R Tasker (chair, Addenbrooke's Hospital, venous catheters were the default catheters used in Cambridge), S Playfor (chair, Royal Manchester Children's Hospital, many units.10 Lastly, antibiotic resistance testing using Manchester), A Vail, D Roebuck, J Gray (Birmingham Children's Hospital, etest strips was not done for all positive blood cultures. Birmingham), H Greig-Midlane (parent representative, London).
Independent Data Safety and Monitoring Committee: UK P Ewings (chair), This represents local laboratory administration and M Sharland, N Modi.
processing, which centralised testing of positive cultures Trial management group: UK R Gilbert (chair), A Wade, (UCL Institute of could have mitigated. Where reported, resistance Child Health, London), C Gamble, K Dwan, T Moitt, R Breen (University occurred in all trial groups, predominantly in of Liverpool, Liverpool), D Hughes, C Ridyard (Bangor University, Bangor), Quen Mok, P Ramnarayan (Great Ormond Street Hospital, Gram-negative isolates, as expected. The low rates are London), S Tibby, J Harris (Evelina Children's Hospital, London), consistent with previous lack of evidence for the M Millar, L Draper (University of Leicester, Leicester), O Bagshaw emergence of resistance.30 (Birmingham Children's Hospital, Birmingham), B Muller-Pebody (Public The primary outcome, time to bloodstream infection, Health England, London), D Hewett (Leeds General Infi rmary, Leeds).
Quality assurance: UK M Blundell (University of Liverpool, Liverpool).
did not diff er between impregnated and standard Data management: UK S Howlin, L Finnetty (University of Liverpool, central venous catheters. However, secondary, pairwise Liverpool).
analyses of the type of central venous catheter, showed Administrative support: UK I Pribramska (UCL Institute of Child Health,
that only antibiotic-impregnated central venous London).
catheters reduced the risk of bloodstream infections Acknowledgments
compared with standard and with heparin-impregnated We thank the children and families who participated in the CATCH trial
and the principal investigators and research nurses at each study site. central venous catheters. The low rate of bloodstream We thank Roger Parslow and the Healthcare Quality Improvement infections in the standard and heparin groups and the Partnership for contributing data from the PICANet audit. We also multiple, pairwise comparisons, reduced the power of thank the Local Research Networks in England for supporting the trial our study. However, when combined with evidence implementation the Trial Steering Committee and the Independent Data Safety and Monitoring Committee for their oversight of the study. The from systematic reviews, our fi ndings establish the CATCH trial was funded by the National Institute for Health Research eff ectiveness of antibiotic-impregnated central venous Health Technology Assessment (NIHR HTA) programme (project catheters compared with standard central venous number 08/13/47). The views and opinions expressed therein are those catheters and extend this evidence for paediatric use. of the authors and do not necessarily refl ect those of the HTA programme, NIHR, National Health Service (NHS), or the Department For the fi rst time, we directly show the eff ectiveness of of Health. No funding was provided by the manufacturer (Cook) of the impregnation of central venous catheters with central venous catheters, although participating units could purchase antibiotics compared with heparin in this population, central venous catheters at a discount of 20% during recruitment to the even in the context of low rates of bloodstream study. Neither the funder nor the manufacturer had any involvement in the study design, interpretation of the results, or writing of the report.
infections.31 Widespread use of antibiotic-impregnated References
central venous catheters could help prevent bloodstream Abou Elella R, Najm H, Balkhy H, Bullard L, Kabbani M. Impact of infections in the paediatric intensive care unit. Whether bloodstream infection on the outcome of children undergoing these benefi ts outweigh the additional costs depends on cardiac surgery. Pediatr Cardiol 2010; 31: 483–89.
diff erential pricing of antibiotic and standard central 2 Elward AM, Hollenbeak CS, Warren DK, Fraser VJ. Attributable cost of nosocomial primary bloodstream infection in pediatric venous catheters by the manufacturer and the cost intensive care unit patients. Pediatrics 2005; 115: 868–72.
benefi ts of avoiding bloodstream infection.
www.thelancet.com Published online March 3, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00340-8
Bhutta A, Gilliam C, Honeycutt M, et al. Reduction of bloodstream 18 Harron K, Parslow R, Mok Q, et al. Monitoring quality of care through infections associated with catheters in paediatric intensive care unit: linkage of administrative data: national trends in bloodstream infection stepwise approach. BMJ 2007; 334: 362–65.
in UK paediatric intensive care units 2003–2012. Crit Care Med 2015; Pronovost P, Needham D, Berenholtz S, et al. An intervention to 43: 1070–78.
decrease catheter-related bloodstream infections in the ICU. 19 Pierce CM, Wade A, Mok Q. Heparin-bonded central venous lines N Engl J Med 2006; 355: 2725–32.
reduce thrombotic and infective complications in critically ill Miller MR, Griswold M, Harris JM 2nd, et al. Decreasing PICU children. Intensive Care Med 2000; 26: 967–72.
catheter-associated bloodstream infections: NACHRI's quality 20 Cox EG, Knoderer CA, Jennings A, et al. A randomized, controlled transformation eff orts. Pediatrics 2010; 125: 206–13.
trial of catheter-related infectious event rates using Bion J, Richardson A, Hibbert P, et al. ‘Matching Michigan': antibiotic-impregnated catheters versus conventional catheters in a 2-year stepped interventional programme to minimise central pediatric cardiovascular surgery patients. J Pediatr Infect Dis 2013; venous catheter-blood stream infections in intensive care units in 2: 67–70.
England. BMJ Qual Saf 2012; 22: 110–23.
21 Abdelkefi A, Achour W, Ben Othman T, et al. Use of heparin-coated Department of Health. Saving Lives: reducing infection, delivering central venous lines to prevent catheter-related bloodstream clean and safe care: London: Department of Health, 2007. infection. J Support Oncol 2007; 5: 273–78.
HM Treasury. Pre-budget report and comprehensive spending 22 Universities of Leeds and Leicester. Paediatric Intensive Care Audit review. London: Stationery Offi Network national report 2011–2013. Leeds: University of Leeds, 2013. O'Grady NP, Alexander M, Burns LA, et al. Guidelines for the 23 Slater A, Shann F, Pearson G. PIM2: a revised version of the prevention of intravascular catheter-related infections. Paediatric Index of Mortality. Intensive Care Med 2003; 29: 278–85.
Clin Infect Dis 2011; 52: e162–93.
24 Cook Incorporated. Cook Spectrum® Central Venous Catheter 10 Harron K, Ramachandra G, Mok Q, Gilbert R. Consistency between Minocycline/Rifampin ANtibiotic Impregnated Power Injectable. guidelines and reported practice for reducing the risk of Instructions for use. https://www.cookmedical.com/data/IFU_ catheter-related infection in British paediatric intensive care units. PDF/C_T_CTULMABRM_REV3.PDF (accessed Jan 23, 2016). Intensive Care Med 2011; 37: 1641–47.
25 Raad I, Hanna H, Maki D. Intravascular catheter-related infections: 11 Gilbert R, Harden M. Eff ectiveness of impregnated central venous advances in diagnosis, prevention, and management. catheters for catheter related blood stream infection: a systematic Lancet Infect Dis 2007; 7: 645–57.
review. Curr Opin Infect Dis 2008; 21: 235–45.
26 Gray RJ. A class of K-sample tests for comparing the cumulative 12 Hockenhull J, Dwan K, Boland A, et al. The clinical eff ectiveness incidence of a competing risk. Ann Stat 1988; 16: 1141–54.
and cost-eff ectiveness of central venous catheters treated with 27 Harron K, Woolfall K, Gamble C, Mok Q, Ramnarayan P, Gilbert R. anti-infective agents in preventing bloodstream infections: Deferred consent for randomised controlled trials in emergency a systematic review and economic evaluation. Health Technol Asses paediatric care settings: evidence for trial design and ethics 2008; 12: 1–154.
committees. Pediatrics 2015; 136: e1316–22.
13 Lai NM, Chaiyakunapruk N, Lai NA, O'Riordan E, Pau WSC, Saint S. 28 Harron K, Mok Q, Dwan K, et al. Catheter infections in children Catheter impregnation, coating or bonding for reducing central (CATCH): a randomised controlled trial and economic evaluation venous catheter-related infections in adults. Cochrane Database Syst Rev comparing impregnated and standard central venous catheters in 2013; 6: CD007878.
children.Health Technol Assess (in press).
14 Wang H, Huang T, Jing J, et al. Eff ectiveness of diff erent central 29 Harron K, Mok Q, Parslow R, Muller-Pebody B, Gilbert R, venous catheters for catheter-related infections: a network Ramnarayan P. Risk of bloodstream infection in children admitted meta-analysis. J Hosp Infect 2010; 76: 1–11.
to paediatric intensive care units in England and Wales following 15 Shah PS, Shah N. Heparin-bonded catheters for prolonging the emergency inter-hospital transfer. Intensive Care Med 2014; patency of central venous catheters in children. 40: 1916–23.
Cochrane Database Syst Rev 2014; 17: CD005983.
30 Falagas ME, Fragoulis K, Bliziotis IA, Chatzinikolaou I. 16 Geff ers C, Zuschneid I, Eckmanns T, Rüden H, Gastmeier P. Rifampicin-impregnated central venous catheters: a meta-analysis of The relationship between methodological trial quality and the eff ects randomized controlled trials. J Antimicrob Chemother 2007; of impregnated central venous catheters. Intensive Care Med 2003; 59: 359–69 .
29: 403–09.
31 Harron K, Wade A, Muller-Pebody B, et al. Risk-adjusted 17 Mermel LA, Stolz SM, Maki DG. Surface antimicrobial activity of monitoring of blood-stream infection in paediatric intensive care: heparin-bonded and antiseptic-impregnated vascular catheters. a data linkage study. Intensive Care Med 2013; 39: 1080–87.
J Infect Dis 1993; 167: 920–24.
www.thelancet.com Published online March 3, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00340-8

Source: http://catchtrial.org.uk/Documents/The%20LANCET%20-%20CATCH%20paper.pdf