Publication

• Title: 4% Tetrasodium EDTA to Prevent Central Venous Access Device–Associated Complications: A Randomized Clinical Trial
• Acronym: CLiCK — Control of Line Complications with KiteLock in the ICU
• Year: 2026
• Journal published in: JAMA
• Citation: Ornowska M, Wittmann J, Blitz S, Wong H, Vazquez-Grande G, Mitra AR, et al. 4% tetrasodium EDTA to prevent central venous access device–associated complications: a randomized clinical trial. JAMA. Published online May 18, 2026.

Context & Rationale

• Background

  • Central venous access devices (CVADs) are ubiquitous in critical care and are used for vasopressors, antibiotics, parenteral nutrition, haemodialysis, blood sampling, fluid resuscitation, and vesicant medications.
  • Important CVAD complications include central line–associated bloodstream infection (CLABSI), catheter occlusion, thrombolytic use, catheter removal due to occlusion, catheter colonisation, and catheter-associated venous thrombosis.
  • Biofilm formation and intraluminal thrombus are mechanistically linked: biofilm promotes microbial persistence and antimicrobial tolerance, while fibrin and thrombus provide surfaces for microbial adherence and line obstruction.
  • The protocol framed 4% tetrasodium EDTA (t-EDTA) as a non-antibiotic catheter lock with anticoagulant, antimicrobial, and antibiofilm properties, mediated through chelation of divalent cations required for coagulation, bacterial cell wall stability, and biofilm integrity.¹
  • A 2018 adult critical care systematic review reported CLABSI at 4.59 per 1000 catheter-days, catheter-associated venous thrombosis at 8.34 per 1000 catheter-days, and catheter occlusion affecting approximately 4% to 22% of CVADs, emphasising the clinical importance of line complications in ICU practice.²
  • Before CLiCK, ICU evidence for locking fluids was fragmented, with most data coming from haemodialysis, oncology, paediatric intestinal failure, home parenteral nutrition, and observational implementation studies rather than large blinded adult ICU trials.
  • A 2023 systematic review of locking fluids in adult critical care found limited high-quality evidence and highlighted that current ICU practice remained dominated by saline, heparin, and citrate rather than dual antibiofilm-anticoagulant locks.³
  • A 2025 EDTA meta-analysis suggested major reductions in catheter-related bloodstream infection or CLABSI and catheter colonisation, but it was based on small and heterogeneous studies across non-ICU and ICU-adjacent populations.⁴
• Research Question/Hypothesis

  • Does routine use of 4% t-EDTA locking fluid in unused CVAD lumens reduce CVAD-associated complications in adult ICU patients compared with usual control locking fluid?
  • The hypothesis was that a t-EDTA lock would reduce a composite event rate of CLABSI, catheter occlusion requiring alteplase, and catheter removal due to occlusion.
• Why This Matters

  • CVAD occlusion interrupts drug delivery, blood sampling, and renal replacement access, increasing nursing workload and often prompting alteplase use or line replacement.
  • CLABSI is uncommon in well-performing ICUs but has high morbidity, cost, and surveillance importance.
  • A safe non-antibiotic lock could reduce both mechanical and infectious line complications without the resistance concerns associated with antibiotic locks.
  • The trial addressed a practical bedside intervention that could be embedded into routine ICU nursing workflow if effective, safe, affordable, and acceptable.

Design & Methods

• Research Question: In adult ICU patients with a CVAD and at least one unused lumen, does locking inactive lumens with 4% t-EDTA reduce CVAD-associated complication rates compared with control locking fluid?
• Study Type: Pragmatic, triple-blind, multicentre, cluster-randomised, crossover trial conducted in 6 Canadian hospitals: 3 community hospitals and 3 academic centres. Five ICUs were medical-surgical and one ICU also cared for cardiac surgical patients.
• Population:
  • Adults older than 18 years.
  • Admitted to a critical care unit.
  • Had a CVAD in place with at least one lumen not in use.
  • All CVAD types were eligible, including single-, double-, triple-, and quadruple-lumen central venous catheters; tunnelled and non-tunnelled catheters; dialysis lines; trialysis catheters; implanted vascular access devices; and peripherally inserted central catheters (PICCs).
  • Key exclusions were known sensitivity or allergy to t-EDTA, confirmed or suspected pregnancy, declining blood products, and refusal by the patient, treating physician, or temporary substitute decision-maker.
  • Waiver of consent procedures were approved because 4% t-EDTA was already approved by Health Canada for use as a CVAD locking fluid and because prospective consent was considered impracticable for routine line locking in ICU.
• Intervention:
  • Unused CVAD lumens were flushed with 20 mL 0.9% sodium chloride using a turbulent push-pause technique.
  • Prefilled, identical, masked syringes containing 2.5 mL of 4% t-EDTA were then instilled into inactive lumens.
  • For haemodialysis and trialysis lumens, the amount instilled was matched to the port volume, analogous to standard dialysis locking practice.
  • Locks were aspirated whenever possible before line reaccess or routine lock exchange.
  • Locking was performed by ICU and haemodialysis nursing staff, with daily research staff support for supply, protocol adherence, accountability, and outcome collection.
• Comparison:
  • Control syringes were identical and masked.
  • Control locking fluid was 0.9% sodium chloride for all catheter types except haemodialysis catheters.
  • Haemodialysis catheter control locking fluid was 4% citrate.
  • The comparator therefore reflected saline-based usual care for most CVC/PICC lumens and citrate-based usual care for dialysis lumens.
• Blinding: Participants, clinical staff, study staff, outcome adjudicators, and the primary data analyst were blinded. The randomisation schedule was generated by the study methodologist and disclosed only to the external study manager and pharmacists preparing the masked syringes.
• Statistics: The planned sample size was 1524 participants, giving 80% power at a two-sided α=.05 to detect a 30% relative reduction in the primary composite event rate, using simulation based on baseline complication rates from an observational period. The primary analysis was intention-to-treat, including all participants with at least one CVAD lock during ICU stay. Event rates were analysed with a negative binomial mixed-effects model using catheter-days as the offset, a random intercept for site, and adjustment for treatment period, age, sex, immunocompromised status, and baseline antibiotic, antifungal, or anticoagulant therapy.
• Follow-Up Period: Participants were assessed throughout ICU admission until ICU discharge, death, or the end of the treatment maintenance period, whichever occurred first. There was no planned post-ICU or post-hospital follow-up for the clinical outcome assessment.

Key Results

This trial was not stopped early. It completed the trial periods but enrolled slightly fewer than planned: 3124 ICU patients were assessed, 1574 were enrolled, 1470 were randomised, and 1468 were included in the intention-to-treat analysis after 2 post-randomisation exclusions.

• The primary result was positive for a composite event-rate endpoint, but the effect was driven almost entirely by fewer alteplase-treated occlusion events.
• The trial did not establish that t-EDTA reduces CLABSI, catheter-associated thrombosis, colonisation, ICU mortality, or length of stay.
• The most credible clinical signal is prevention of line occlusion in patients with longer-dwell CVADs or high protocol adherence, rather than broad prevention of all CVAD-associated harms.

Internal Validity

• Randomisation and Allocation: Randomisation was performed at the ICU cluster level, stratified by 3 hospital pairs based on study enrolment dates, with crossover after each treatment period. Allocation concealment was credible because the schedule was generated by the methodologist and shared only with an external study manager and pharmacists preparing the blinded syringes.
• Dropout and Exclusions: Of 1470 randomised patients, 2 were excluded post-randomisation: 1 withdrawal of consent and 1 co-enrolment in a competing study. No primary outcome data were missing among the 1468 patients included in the intention-to-treat analysis.
• Performance and Detection Bias: Triple blinding was a major strength. Participants, bedside staff, research staff, outcome adjudicators, and the primary analyst were blinded. Identical syringes and common lock/aspiration procedures reduced behavioural differences in line handling between groups.
• Protocol Adherence: Of 1468 participants, 1258 (85.7%) had at least 80% adherence to study locking protocol. The per-protocol population was 953 participants (64.9% of the intention-to-treat cohort), and per-protocol adherence varied markedly by site, from 42.3% to 98.6%.⁵
• Baseline Characteristics: Baseline characteristics were broadly similar. Mean APACHE IV score was 81.5 vs 82.1, SOFA score was 10.1 vs 10.4, and Charlson Comorbidity Index was 3.5 vs 3.5. The t-EDTA group had more systemic antibiotic use at entry (87.1% vs 79.0%), more systemic or subcutaneous anticoagulant use (68.8% vs 62.4%), and more total parenteral nutrition (2.2% vs 0.5%). Antibiotic, antifungal, and anticoagulant exposure were included in the adjusted model.
• Heterogeneity: Heterogeneity was substantial but appropriate for a pragmatic ICU device trial. The trial included multiple CVAD types, catheter insertion sites, medical-surgical and cardiac surgical critical care, community and academic hospitals, dialysis and non-dialysis lumens, and variable catheter dwell times. This improves real-world relevance but complicates interpretation of small subgroup signals.
• Timing: Patients were enrolled within the first 48 hours of ICU admission and followed until ICU discharge, death, or the end of the maintenance period. This timing was appropriate for ICU line management, but it excluded later post-ICU or ward complications among patients discharged with CVADs.
• Dose: The dose was pragmatic: 2.5 mL for most unused lumens and port-volume locking for haemodialysis/trialysis lumens. The intervention could only affect inactive lumens that were locked; continuously used lumens and short-dwell catheters had less opportunity to benefit.
• Separation of the Variable of Interest:
  • The experimental arm received 4% t-EDTA for inactive lumens.
  • The control arm received 0.9% sodium chloride for most CVC/PICC/implanted lumens and 4% citrate for haemodialysis catheters.
  • There were 19 993 recorded locking events: 9445 in the t-EDTA group and 10 548 in the control group.
  • Non-protocol locks and lower adherence likely diluted the treatment contrast and biased the intention-to-treat estimate towards the null.
• Key Delivery Aspects: The trial was embedded in bedside nursing workflow, with research staff maintaining daily ICU presence. This made the intervention clinically realistic, but also means implementation fidelity depended on nursing workload, staff turnover, and post-pandemic ICU pressures.
• Crossover: The cluster-crossover design was appropriate for a unit-level line-locking intervention and reduced between-hospital confounding. No participants enrolled during the first randomisation period remained in ICU after crossover, reducing individual carryover risk. Period effects remain possible because the study occurred across a changing post-COVID critical care environment.
• Adjunctive Therapy Use: Alteplase use was part of the outcome rather than an adjunctive rescue therapy independent of the endpoint. Microbiological testing, imaging, and alteplase prescribing were performed at clinician discretion, which preserves pragmatism but introduces local threshold effects.
• Outcome Assessment: The primary endpoint was a composite event rate, not a patient-level time-to-first event. CLABSI and catheter removal were clinically meaningful but rare; alteplase administration was common and measured at lumen level. This makes the primary outcome sensitive to occlusion practice and nursing thresholds for alteplase, although blinding substantially reduces differential bias.
• Statistical Rigor: The primary analysis matched the event-rate design using negative binomial mixed-effects modelling with catheter-days as offset. The trial narrowly missed its planned sample size, enrolling 1468 rather than 1524 participants, but achieved 96% of target. Secondary outcomes with fewer than 20 events were analysed with unadjusted models and should be interpreted cautiously. Subgroups were exploratory.

Conclusion on Internal Validity: Internal validity is strong for the conclusion that 4% t-EDTA reduced the composite ICU CVAD-complication event rate, principally by reducing alteplase-treated occlusions. Internal validity is limited for infection, thrombosis, colonisation, haemodialysis-catheter effects, and patient-level clinical outcomes because those events were uncommon, variably ascertained, or exploratory.

External Validity

• Population Representativeness: The cohort is representative of adult ICU patients with CVADs and at least one unused lumen in Canadian academic and community hospitals. It included a broad mix of CVCs, PICCs, haemodialysis catheters, trialysis lines, insertion sites, and medical-surgical ICU diagnoses.
• Settings Best Matched to the Trial: The findings apply best to high-income adult ICU settings with established CVAD maintenance protocols, access to prefilled locking syringes, nurse-led lumen locking, routine aspiration before access, and capacity to monitor adherence.
• Settings Less Well Matched: Generalisability is weaker for paediatric ICUs, pregnancy, wards, outpatient vascular access care, home parenteral nutrition, long-term tunnelled catheters, resource-limited settings, and hospitals using heparin rather than saline as the main non-dialysis control lock.
• Haemodialysis Catheters: Applicability to acute haemodialysis catheters is uncertain. The trial’s haemodialysis subgroup did not show benefit, had few events, and used 4% citrate as the control lock, which is an active comparator.
• Baseline Event Rates: ICUs with very low alteplase use or very low CLABSI rates may see less benefit. ICUs with high line-occlusion burden, longer CVAD dwell times, frequent PICC use, or high nursing workload may find the occlusion signal more clinically relevant.
• Implementation Requirements: Translation requires training, reliable supply of t-EDTA syringes, documentation of every lock, aspiration before use, and local evaluation of cost, nursing time, alteplase costs, and catheter replacement rates.
• Outcome Scope: The trial followed patients only during ICU stay, so it does not address complications after ICU discharge, transfer to ward, discharge with line in situ, or long-term vascular access outcomes.

Conclusion on External Validity: External validity is good for adult Canadian/North American–style ICUs using nurse-led CVAD locking protocols. It is limited for long-term vascular access populations, paediatrics, home parenteral nutrition, haemodialysis-focused practice, and systems where line maintenance, lock choice, alteplase thresholds, and catheter replacement decisions differ substantially.

Strengths & Limitations

• Strengths:
  • Large pragmatic adult ICU trial addressing a common bedside device-management problem.
  • Triple-blind design with identical masked syringes.
  • Cluster-crossover design appropriate for a unit-level nursing workflow intervention.
  • Near-complete randomised cohort retention with only 2 post-randomisation exclusions and no missing primary outcome data.
  • Inclusion of multiple CVAD types and both academic and community hospitals.
  • Pragmatic integration into routine ICU nursing practice.
  • Predefined statistical model for event rates and appropriate adjustment for site and treatment period.
  • No apparent safety signal, including no reported hypocalcaemia related to t-EDTA.
• Limitations:
  • The trial did not reach its planned sample size, although it reached 96% of target.
  • The primary composite was driven by alteplase-treated occlusion, not infection or thrombosis.
  • CLABSI, catheter removal due to occlusion, thrombosis, and colonisation events were too rare for precise conclusions.
  • The primary outcome mixed participant-level events with lumen-level alteplase events.
  • Alteplase use, microbiological testing, imaging, and line removal decisions were performed at clinician discretion.
  • Protocol adherence was imperfect; only 64.9% of participants were in the per-protocol subgroup.
  • The per-protocol and adherence analyses are biologically coherent but not protected by randomisation.
  • The study occurred across post-pandemic ICU operational changes and required a sixth ICU because enrolment was slower than expected.
  • Advanced biofilm analysis planned in the original protocol was not operationalised.
  • Follow-up ended at ICU discharge or treatment-period end, so later ward or outpatient CVAD complications were not measured.
  • The trial was supported by SterileCare Inc, which supplied the investigational product; the manuscript reports no funder role in design, conduct, analysis, interpretation, or publication decision.

Interpretation & Why It Matters

• Primary message

  CLiCK supports 4% t-EDTA as an effective ICU lock for reducing a composite CVAD-complication event rate, but the clinically demonstrated benefit is mainly fewer alteplase-treated occlusions.
• What changed

  The trial shifts t-EDTA from small, heterogeneous, and largely non-ICU evidence into a large blinded pragmatic adult ICU trial.
• What did not change

  It does not prove a reduction in CLABSI, catheter-related bloodstream infection, venous thrombosis, ICU mortality, or length of ICU stay.
• Practical significance

  For ICUs with frequent line occlusion or high alteplase use, t-EDTA may reduce medication cost, delays in therapy, failed access, and nursing time spent troubleshooting blocked lumens.
• Implementation caution

  Adoption should be targeted and audited: local baseline occlusion rates, alteplase thresholds, catheter dwell times, nursing workflow, syringe cost, and adherence are likely to determine real-world value.

Controversies & Subsequent Evidence

• The composite endpoint is the central interpretive issue: The primary endpoint was positive, but almost all effect came from reduced alteplase-treated occlusion. This is clinically useful, but it is not equivalent to demonstrating fewer bloodstream infections, fewer thromboses, or fewer line replacements.
• CLABSI was too rare for conclusions: There were only 5 CLABSI events: 2 with t-EDTA and 3 with control. The confidence interval for CLABSI was extremely wide, so CLiCK should not be described as proving infection prevention.
• CLABSI is a surveillance endpoint, not the most specific infection endpoint: Catheter-related bloodstream infection requires more rigorous evidence that the catheter is the source, such as quantitative culture or differential time-to-positivity. The trial used CLABSI because it enabled surveillance-style pragmatic ascertainment, but that choice trades specificity for feasibility.⁸
• Event-rate analysis is not the same as patient-level benefit: The primary endpoint was analysed as an incidence rate. Alteplase use was assessed at lumen level, whereas several other components were participant level. This makes the result relevant to workload and catheter function, but less straightforward for counselling about individual patient risk.
• Guidelines still emphasise line-care fundamentals: Contemporary infusion and intravascular catheter guidance continues to prioritise education, aseptic insertion and maintenance, appropriate flushing and locking technique, daily review of line necessity, and selective rather than universal antimicrobial-lock strategies.⁶⁷
• Haemodialysis catheters remain unsettled: A previous multicentre randomised haemodialysis study found that t-EDTA reduced tunnelled catheter colonisation but did not show a statistically significant CRBSI reduction, while CLiCK’s haemodialysis subgroup showed no benefit signal. A 2024 haemodialysis centre comparison suggested reduced alteplase use with 4% EDTA versus 4% citrate, but it was retrospective and operational rather than definitive.⁹¹⁰
• Evidence outside ICU is favourable but not directly transferable: Home parenteral nutrition and patient-acceptability studies suggest t-EDTA can be feasible and may reduce catheter complications in long-term access settings, but those populations differ from ICU patients in dwell time, lock frequency, catheter type, infection ecology, and patient/caregiver handling.¹¹¹²
• Network evidence may favour other locks for infection prevention: A 2026 Bayesian network meta-analysis ranked taurolidine-based locks highest for CLABSI prevention, with EDTA-based locks also showing benefit versus saline or heparin. CLiCK adds strong ICU evidence for occlusion reduction, but it does not settle the comparative infection-prevention hierarchy across lock solutions.¹³
• The biological signal is plausible but incomplete: t-EDTA’s chelation mechanism fits occlusion prevention and early biofilm disruption. The absence of a clear infection or thrombosis signal may reflect low event rates, short ICU dwell times, insufficient exposure in continuously used lumens, or the fact that infection prevention requires bundle-level interventions beyond lock solution alone.
• Cost-effectiveness is unresolved: The protocol planned an analysis of costs related to alteplase use, but this was deferred to a future health economic evaluation. Routine adoption should therefore consider local costs of t-EDTA syringes, alteplase, catheter replacement, nursing time, and CLABSI-prevention infrastructure.

Summary

• CLiCK was a pragmatic, triple-blind, cluster-randomised crossover trial in 1468 adult ICU patients with CVADs and at least one unused lumen.
• 4% t-EDTA reduced the primary composite event rate: 13.09 vs 19.94 events per 1000 catheter-days; rate ratio 0.68; 95% CI 0.47 to 0.96; P=.03.
• The benefit was driven by fewer catheter occlusions requiring alteplase: 66 vs 112 events; rate ratio 0.66; 95% CI 0.46 to 0.96.
• The trial did not demonstrate clear reductions in CLABSI, catheter-associated thrombosis, colonisation, ICU mortality, or ICU length of stay.
• No adverse events, serious adverse events, or hypocalcaemia attributable to t-EDTA were reported.

Overall Takeaway

CLiCK is an important practice-informing ICU device trial. It supports 4% t-EDTA as a safe, plausible, and effective strategy to reduce alteplase-treated CVAD occlusion, especially where adherence is high and catheter dwell time is longer, but it should not be interpreted as definitive evidence of CLABSI or thrombosis prevention.

Bibliography

• 1.Ornowska M, Wong H, Ouyang Y, White A, Mitra A, Willems S, et al. Control of Line Complications with KiteLock (CLiCK) in the critical care unit: study protocol for a multi-center, cluster-randomized, double-blinded, crossover trial investigating the effect of a novel locking fluid on central line complications in the critical care population. Trials. 2022;23(1):719.
• 2.Takashima M, Schults J, Mihala G, Corley A, Ullman A. Complication and failures of central vascular access device in adult critical care settings. Crit Care Med. 2018;46(12):1998-2009.
• 3.Ornowska M, Smithman J, Reynolds S. Locking solutions for prevention of central venous access device complications in the adult critical care population: a systematic review. PLoS One. 2023;18(10):e0289938.
• 4.Chakraborty S, Bhattacharya S, Ghosh A, Chatterji S. Role of EDTA lock solution in preventing catheter-related bloodstream infections or central line-associated bloodstream infections: a systematic review and meta-analysis. Infect Control Hosp Epidemiol. 2025;46(10):1035-1042.
• 5.Ornowska M, Wittmann J, Blitz S, et al. Supplemental content to: 4% tetrasodium EDTA to prevent central venous access device–associated complications: a randomized clinical trial. JAMA. Published online May 18, 2026.
• 6.Nickel B, Gorski L, Kleidon T, Kyes A, DeVries M, Keogh S, et al. Infusion therapy standards of practice, 9th edition. J Infus Nurs. 2024;47(1S Suppl 1):S1-S285.
• 7.O’Grady NP, Alexander M, Burns LA, Dellinger EP, Garland J, Heard SO, et al. Guidelines for the prevention of intravascular catheter-related infections. Clin Infect Dis. 2011;52(9):e162-e193.
• 8.Mermel LA, Allon M, Bouza E, Craven DE, Flynn P, O’Grady NP, et al. Clinical practice guidelines for the diagnosis and management of intravascular catheter-related infection: 2009 update by the Infectious Diseases Society of America. Clin Infect Dis. 2009;49(1):1-45.
• 9.Kanaa M, Wright MJ, Akbani H, Laboi P, Bhandari S, Sandoe JAT. Cathasept line lock and microbial colonization of tunneled hemodialysis catheters: a multicenter randomized controlled trial. Am J Kidney Dis. 2015;66(6):1015-1023.
• 10.Gage B, Shalansky K, Lau W, Harris C, Kiaii M. Comparison of tetrasodium EDTA 4% with sodium citrate 4% as line-locking solutions at 2 tertiary hemodialysis centres. Can J Hosp Pharm. 2024;77(1):e3447.
• 11.Hill J, Garner R. Efficacy of 4% tetrasodium ethylenediaminetetraacetic acid (T-EDTA) catheter lock solution in home parenteral nutrition patients: a quality improvement evaluation. J Vasc Access. 2021;22(4):533-539.
• 12.Saucedo L, Ghorbani Y, Heusser M, Chagas G, Arca C, Schwenger KJP, et al. Tetrasodium EDTA central venous catheter lock solution in home parenteral nutrition patients: ease of use and patient satisfaction, a prospective study. J Vasc Access. Published online September 17, 2025.
• 13.Liu J, Lin J, Liu H, Soh KG, Soh KL, Dong X. Efficacy of catheter lock solutions on central-line associated bloodstream infections (CLABSI) prevention: a systematic review and Bayesian network meta-analysis. J Vasc Access. Published online February 8, 2026.