For newborns admitted to Neonatal Intensive Care Units (NICUs), one of the main risk factors for late-onset sepsis is the presence of a central venous catheter (CVC), which is often essential for the administration of medications and parenteral nutrition in this patient population. From a nosological perspective, sepsis associated with the presence of a venous catheter is defined by two acronyms: CRBSI (Catheter-Related Bloodstream Infection - a microbiological definition) and CLABSI (Central Line-Associated Bloodstream Infection - an epidemiological definition).

Among preventive strategies for CRBSI/CLABSI, antibiotic or antimicrobial catheter lock solutions - instilled in a volume equivalent to the catheter dead space and retained within the lumen until the next use - have demonstrated favorable efficacy in reducing infection risk.

Taurolidine 2% is considered a preferred agent due to its broad-spectrum antibacterial and antifungal activity and its lack of association with the development of antimicrobial resistance.

However, its prophylactic use in neonates remains largely investigational, with current evidence limited to small, retrospective observational studies involving catheters ≥3 Fr (e.g., femoral inserted central catheters - FICCs, centrally inserted central catheters - CICCs, and umbilical venous catheters).

In NICUs, epicutaneo-caval catheters (ECCs) are the most commonly used central venous access devices and represent a major source of catheter-related infections. Despite this, the use of antimicrobial lock prophylaxis in ECCs has been limited by concerns regarding catheter occlusion, given their smaller diameter (≤2 Fr). Nevertheless, available evidence indicates that short-duration locks, when combined with meticulous infusion line management, can be safely implemented without increasing the risk of catheter occlusion.

The aim of the study is to evaluate the efficacy of 2% taurolidine lock in the prevention of CLABSI/CRBSI in neonates with CVC (ECC, FICC, or CICC).

BACKGROUND

The diagnosis of sepsis remains, to this day, one of the most unfavorable for infants admitted to Neonatal Intensive Care Units (NICUs). Neonatal sepsis is classified as: early onset, when it occurs within 48-72 hours of life - typically due to vertical transmission; and late onset, when it occurs beyond 72 hours of life - presumably due to horizontal transmission.

Late-onset sepsis currently affects a clinically significant proportion of neonates, with rates exceeding 10% even in high-income countries, and is characterized by mortality rates reaching up to 35% in the most vulnerable patients, thus representing a major public health concern.

In NICUs, one of the main risk factors for late-onset sepsis is the presence of a central venous catheter (CVC), which is often essential for the administration of medications and parenteral nutrition in this patient population.

From a nosological perspective, sepsis associated with the presence of a venous catheter is defined by two acronyms: CRBSI (Catheter-Related Bloodstream Infection) and CLABSI (Central Line-Associated Bloodstream Infection).

CRBSIs - a microbiological definition - refer to bloodstream infections in which it is possible to document the presence of microorganisms originating from the vascular catheter. To establish this, it is necessary to obtain, simultaneously, a blood culture drawn directly from the catheter and a blood culture obtained via peripheral venipuncture.

CLABSIs - an epidemiological definition - include all bloodstream infections in which it is not possible to determine the source in patients with a vascular catheter in place.

Bloodstream infections in patients with vascular access may originate from extraluminal migration of microorganisms through the catheter insertion site (i.e., disruption of the skin barrier), from intraluminal entry of microorganisms due to manipulation of the catheter and the connected vascular devices, or from contamination of the infused solutions. Both routes of entry (extraluminal and intraluminal) contribute to the formation of a biofilm along the catheter - a stable structure resistant to antibiotic penetration that begins to develop within the first hours after catheter insertion. The biofilm may remain stable and clinically silent for the entire duration of catheter dwell time or may act as a source of bacteremia.

Several strategies have been proposed to prevent CRBSI/CLABSI in both adults and children. These include measures aimed at reducing infection risk during catheter insertion, proper care of the insertion site, and appropriate management of the infusion line.

Among these, promising results have been obtained with the use of antimicrobial solutions for catheter lock (i.e., infusion of a volume equal to the dead space of the infusion line followed by "closure" of the line until its next scheduled use), with the aim of preventing microbial colonization of the venous catheter and reducing the risk of bacteremia. In this context, 2% taurolidine is considered the agent of choice, as it exhibits antimicrobial activity against both bacteria and fungi and, not being classified as an antibiotic, does not induce antibiotic resistance [Sun Y, et al. PLoS One. 2020] The use of prophylactic catheter lock with 2% taurolidine in the neonatal setting remains limited to experimental and research contexts.

Currently available studies on this procedure in NICUs are observational, retrospective, involve small sample sizes, and have been conducted on vascular catheters with a diameter of at least 3 Fr (femorally inserted central catheters - FICCs, centrally inserted central catheters - CICCs, or umbilical venous catheters) [Savarese I, et al. J Hosp Infect. 2024].

The most commonly used catheters in neonatal care are epicutaneo-caval catheters (ECCs). These represent the first-line option when short-term CVC placement is required or when insertion into large vessels of the cervical or femoral regions is not feasible.

ECCs are the leading cause of catheter-related infections in NICUs. However, the use of antimicrobial locks for prophylactic purposes has not become widespread due to concerns about an increased risk of catheter occlusion, which is higher compared to FICC or CICC because of their smaller diameter (≤ 2 Fr). Careful management of the infusion line may allow the use of short-duration prophylactic locks (≤ 60 minutes) without increasing the risk of catheter occlusion, as demonstrated in recent feasibility studies [D'Andrea V, et al. J Hosp Infect. 2025].

STUDY OBJECTIVE

To evaluate the efficacy of 2% taurolidine lock in preventing the occurrence of CLABSI and CRBSI in neonates with a central venous catheter.

STUDY DESIGN

The study consists of a multicenter, triple-blind, randomized controlled trial (RCT) conducted in a population of neonates admitted to NICUs and carrying a central venous access (ECC, FICC, or CICC).

Blinding will be ensured by masking the study arm to which the patient has been allocated, using a placebo solution indistinguishable from the 2% taurolidine solution in terms of visual characteristics, smell, consistency, packaging appearance, and method of administration in the control arm, as well as by centralized data collection with masking of group allocation also for the biostatisticians responsible for data analysis.

The placebo solution must also ensure pharmacological inactivity, absence of clinical risk for the patient, contain known, non-allergenic excipients already used in clinical practice, and guarantee physiological neutrality and chemical-physical stability.

Enrolled neonates will be assigned to the experimental arm (intervention) or the control arm through a block randomization system designed to ensure, within each center, balance between the two arms among patients with gestational age < 28 weeks, between 28 and 32 weeks, and ≥ 32 weeks. Participation of centers in the study is contingent upon the presence of an internal protocol for surveillance, prevention, and treatment of infections, ensuring uniform management across both study arms.

Therefore, randomization will allow differentiation between the two arms solely on the basis of the intervention under study.

DATA COLLECTION

Patient participation in the study involves the collection of monitoring data at specific time points (enrollment: T-1; CVC lock: T-2, T-3, …; suspected catheter-related infection: T-event; CVC removal: T-CVC; discharge: T-end).

Specifically, enrollment (T-1) will coincide with the first prophylactic lock (with 2% taurolidine or placebo according to the randomization arm); T-2, T-3, … will correspond to subsequent locks; T-event will include all suspected episodes of catheter-related infection (catheter-related bloodstream infection - CRBSI, and central line-associated bloodstream infection - CLABSI) occurring more than 48 hours after catheter insertion or within 48 hours after its removal.

In the case of positive blood cultures, a diagnosis of CRBSI or CLABSI will be established according to the following definitions:

• CRBSI: clinically and microbiologically confirmed bacteremia, with isolation of the same microorganism from both the CVC and peripheral blood, together with at least one of the following criteria:

  1. quantitative ratio (colony-forming units, CFU) of catheter culture vs peripheral blood culture ≥ 3:1.
  2. differential time to positivity (DTP) between CVC culture and peripheral blood culture > 2 hours.

• CLABSI: symptomatic bacteremia occurring in a patient with a central venous catheter in the absence of another identifiable source of infection; therefore, CLABSI can be diagnosed even in the absence of a blood culture drawn from the CVC. Consequently, CRBSI may be considered a subgroup of CLABSI.

If systemic antibiotic prophylaxis or therapy is initiated, the prophylactic lock protocol will be resumed 48 hours after antibiotic discontinuation.

In the presence of multiple CVCs, systemic antibiotic administration will lead to interruption of the lock protocol for all CVCs in that patient and will be resumed at least 48 hours after discontinuation.

If a T-event (CRBSI/CLABSI) occurs within 48 hours of CVC insertion, it will not be attributed to that CVC; therefore, T-event data for that specific CVC will not be collected. Data collection for any T-event will include the 48 hours following removal of the CVC.

T-CVC corresponds to CVC removal (for any reason), and T-end corresponds to patient discharge.

Data collection and management will be performed using dedicated software with secure access credentials for each center, access control systems, audit trails, and procedures ensuring data quality, integrity, and protection throughout all study phases.

SAMPLE SIZE CALCULATION

Considering previous studies, such as the feasibility study by Savarese et al., in which none of the 16 patients treated with prophylactic lock developed CRBSI, and a meta-analysis on the use of taurolidine locks in pediatric patients, in which a 77% reduction in CRBSI events was demonstrated, for our study a reduction in the risk of CRBSI equal to or greater than 75% (IRR = 0.25) was assumed.

The contribution of CRBSI to the total number of CLABSI in neonates admitted to NICUs is not known, since obtaining a blood culture sample from ECC is not always feasible. Therefore, it was assumed that CRBSI accounts for 80% of the events of the primary outcome and considered as the minimum clinically relevant threshold a 60% reduction in the risk of CLABSI (80% of 75%), obtaining IRR = 0.40.

For the estimation of the sample size, an incidence of 8.25 CLABSI per 1000 catheter-days and a mean at-risk time of 14 days per patient (CVC days net of antibiotic therapy) were assumed, data derived from the Coordinating Center (unpublished data, extrapolated from the INCAS trial [Cresi F, et al. Trials. 2024]). Homogeneous exposure among subjects was assumed; therefore, the analysis was conducted by modeling the number of events per neonate during the observation period. An iterative simulation process (7500 replications) was used, assuming a Poisson distribution for the number of events per patient and using a Poisson regression model with log link for the comparison between arms. A one-sided test with a significance level α = 0.05 (one-sided p-value < .05) and 1:1 allocation between the two arms were adopted; under these assumptions, the estimated power is approximately 80%, for a total sample size of 414 subjects. This value was increased by applying a correction factor of 1.25, in order to account for uncertainty in the parametric assumptions and potential heterogeneity among participating centers. An estimated 10% loss of neonates due to transfer or major protocol violations was also considered. The final planned sample size therefore results in 414 × 1.25 / 0.90 = 575, from which the total sample size is 576 neonates (288 per arm).

Simulations were performed using SAS 9.4 software (SAS Institute, Cary, NC, USA).

The sample size may be revised based on the characteristics of the Participating Centers and on the results obtained from the feasibility study (see below).

STATISTICAL ANALYSIS

Descriptive analysis will be performed by reporting categorical variables as absolute frequencies (percentages), and numerical variables as mean (standard deviation) or median (interquartile range), depending on their distribution.

Primary endpoint will be evaluated using Poisson regression with log link, in order to estimate the expected mean number of events during the observation period and the Incidence Rate Ratio (IRR) between the two treatment arms. If evidence of overdispersion emerges, robust standard errors or an appropriate scaling factor will be used. Time to first event will also be evaluated using a Cox regression model, with estimation of the Hazard Ratio (HR) and the corresponding survival curve according to Kaplan-Meier.

For the evaluation of secondary endpoints, appropriate models will be used for each of them.

Analyses on the entire sample (ITT-set) will be performed following the Intention-to-Treat principle, thus including all subjects assigned to the two arms regardless of adherence to the protocol. A second analysis will then be performed on the subpopulation for which no protocol deviations occurred (Per-Protocol set).

Analyses will follow the intention-to-treat principle. The sample will be described using appropriate summary statistics: mean and standard deviation or median and interquartile range for continuous variables depending on distribution; frequencies and percentages for categorical variables.

The interim analysis, planned at 50% of events, will use an O'Brien-Fleming boundary for efficacy decision; the final analysis will take into account the same sequential rules.

The primary endpoint, namely the incidence of CLABSI per 1000 catheter-days, will be analyzed using count models through Poisson regression or, in the presence of overdispersion, negative binomial regression, with offset for catheter-days. The treatment effect will be expressed as rate ratio (incidence rate ratio, IRR) with 95% confidence interval.

If a patient contributes with more than one CVC, robust variance estimates will be used to account for clustering at the patient level; CVC removal or end of follow-up will result in censoring of the observation.

Secondary endpoints will be evaluated using two-sided tests: for time-dependent outcomes, Kaplan-Meier curves for the two arms will be presented and compared using the log-rank test; binary outcomes will be analyzed with χ² test or Fisher's exact test; continuous outcomes with t-test or Wilcoxon test, depending on data distribution.

The level of significance will be set at p < 0.05 (two-sided); for the primary endpoint, the formal decision will follow the thresholds of the sequential design.

A per-protocol analysis will also be performed, defining as protocol deviation the omission of more than 20% of the planned locks.

Analyses will be performed primarily in R; for descriptive tables and graphs, Jamovi or GraphPad Prism may also be used.

FEASIBILITY SUB-STUDY (INTERNAL PILOT)

The multicenter study includes an initial feasibility phase conducted at the Coordinating Center only. For this purpose, a total of 100 neonates will be recruited and randomly allocated 1:1 into the two arms.

In this phase, the following objectives will be evaluated:

• to estimate the incidence of CLABSI/CRBSI.
• to assess the feasibility of the protocol and operational procedures, including maintenance of blinding.
• to assess treatment safety.
• to estimate the frequency of drop-out and data completeness.
• to estimate the at-risk time (CVC days net of antibiotic therapy).

Feasibility endpoints will include:

• time required to enroll 100 neonates.
• duration of follow-up.
• percentage of drop-out.
• percentage of major protocol deviations.
• frequency of adverse events, distinguished as major and minor.
• estimation of at-risk time. If the baseline assumptions are correct, out of a total of 100 neonates allocated 1:1 between the arms, approximately 8 events are expected to be observed.

Data collected in this phase will be used to verify the plausibility of the assumptions used for planning (incidence and at-risk time) and to optimize operational procedures. If the estimates are substantially discordant from the assumptions, the sample size may be re-evaluated and, if necessary, updated through a protocol amendment before completion of enrollment. No formal efficacy analyses are planned in this phase. The analysis will be mainly descriptive, and parameter estimates will be reported with 80% confidence intervals.

Data collected in this phase will be included in the final analysis of the study, as they are obtained with a design, eligibility criteria, intervention, outcomes, and data collection methods overlapping with those planned for the multicenter phase.

DISCUSSION

This study will evaluate the efficacy of 2% taurolidine in reducing the incidence of catheter-related infections (CLABSI, CRBSI) in the population of neonates admitted to NICUs, an effect that is currently not adequately documented. If the results are favorable and clinically relevant, and no safety-related signals emerge, prophylactic CVC lock with 2% taurolidine could be considered for inclusion in standard clinical practice also in the neonatal population.

The size and methodology of the study are favorable premises to robustly demonstrate the clinical effects and safety of the prophylactic lock protocol with 2% taurolidine for neonates admitted to NICUs.

If the results demonstrate a clinically significant reduction in CRBSI and CLABSI, in the absence of adverse effects, prophylactic CVC lock with 2% taurolidine could become part of the standard of care also in the neonatal population.

The results of the study will be disclosed within 12 months from the conclusion of the trial.

Inclusion Criteria:

• presence of a CVC (ECC, CICC, or FICC) for at least 48 hours
• absence of systemic antibiotic therapy in the 48 hours preceding enrollment

Exclusion Criteria:

• presence of an umbilical venous catheter
• suspected thrombophilic disorder