Description of the condition

Peripherally inserted central catheters (PICCs), also commonly referred to as percutaneous central venous catheters (CVCs), provide a stable route for the intravenous delivery of drugs or fluids to preterm or sick newborn infants. Typically, PICCs are inserted via a superficial vein in the upper or lower limb, usually avoiding proximal sites such as the femoral or axillary veins, and are then advanced so that the tip of the catheter is sited at the vena caval junction with the right atrium of the heart (Jain 2013; Wrightson 2013). Peripherally inserted central catheters are less expensive and easier to insert than catheters placed directly or via a subcutaneous tunnel into central veins. Because they are more stable than short peripheral cannulae, their use reduces the risk of extravasation injury from hyperosmolar parenteral nutrition solutions and medications (Ainsworth 2015).

Bloodstream infection is the most common serious complication associated with the use of PICCs in newborn infants. The reported incidence ranges from about 2% to 30% depending on the precise diagnostic criteria and the demographics of the population studied (Chien 2002; Cartwright 2004; van der Zwet 2005; Garland 2008; Hoang 2008; Olsen 2009; O'Grady 2011; Ohki 2013). Very preterm infants are at the highest risk, but inter‐unit variation in the incidence of catheter‐associated bloodstream infection is not fully explained by case‐mix and may relate to care or infection control practices (Wong 2012). Newborn infants, particularly very preterm infants, with an acquired bloodstream infection have a higher risk of mortality and a range of important morbidities including bronchopulmonary dysplasia, necrotising enterocolitis, retinopathy of prematurity, hepatic dysfunction and prolonged hospitalisation (Saint 2000; Mahieu 2001a; Mahieu 2001b; Chapman 2003; Payne 2004; Adams‐Chapman 2006; Hermans 2007; Lahra 2009). Bloodstream infection is associated with higher rates of adverse neurodevelopmental outcomes including cognitive or sensory impairment and cerebral palsy (Stoll 2004; Shah 2008a; Bassler 2009).

The most common causes of PICC‐associated bloodstream infections in newborn infants are coagulase‐negative staphylococci, other Gram‐positive cocci (Staphylococcus aureus (S. aureus), enterococci), Gram‐negative bacilli and fungi (predominantly Candida species) (Makhoul 2002; O'Grady 2002; Stoll 2002; Isaacs 2003; Isaacs 2004; Gordon 2006; de Brito 2010). Micro‐organisms can gain access through the PICC entry site or via the catheter hub into the PICC lumen or tract. Pathogens adhere to the material of the PICC and secrete an intraluminal or extraluminal biofilm of extracellular polymeric substances (Machado 2009). Bacteria or fungi growing within the biofilm are relatively protected from the host's immune system and circulating antimicrobial agents thus enabling sustained colonisation (Ramirez de Arellano 1994; Stewart 2001). It is often necessary to remove the PICC in order to clear the infection (Benjamin 2001). A thrombus (blood clot) attached to the PICC may be an additional nidus (locus) for infection (Thornburg 2008). Some evidence exists that continuous heparin infusion can reduce the incidence of PICC‐occlusion in neonates, but the data are insufficient to determine the effect of this intervention on the risk of infection (Shah 2008b).

Several strategies to prevent PICC‐associated bloodstream infections have been developed and adopted, often as multifaceted packages of interventions ('care bundles'). These include strict aseptic precautions when inserting and accessing the PICC, use of needleless intravascular catheter systems and prompt removal when the PICC is no longer needed (O'Grady 2002; Yébenes 2004; Pronovost 2006; Borghesi 2008; Miller 2010; Sannoh 2010; Vanholder 2010; Wirtschafter 2010; Kaplan 2011; O'Grady 2011; Schulman 2011). Care bundles have been shown to reduce bloodstream infection rates in adult, paediatric and neonatal intensive care studies (Pronovost 2006; Miller 2010; Wirtschafter 2010; Kaplan 2011; Schulman 2011; Fisher 2013). Despite these strategies, however, PICC‐associated infections remain a major cause of morbidity and mortality in newborn infants and other interventions are required to reduce the infection rates further.

Description of the intervention

Although strong associations between PICC exposure, especially to deliver parenteral nutrition, and the risk of bloodstream infection have been described, observational studies have not provided consistent evidence about the effect of duration of PICC placement ('dwell time') on the risk of invasive infection in newborn infants. Some reports indicate an elevated risk of infection when the PICC has been in place for more than about 10 to 14 days, while others do not show any association until after a dwell time of four to six weeks (Smith 2008; Sengupta 2010; Wong 2012; Milstone 2013; Greenberg 2015). Furthermore, it is not certain to what extent PICC use is an independent risk factor for a bloodstream infection or whether observed associations exist because infants who are smaller, less mature, sicker and receiving more intensive and invasive support are also more likely to have a PICC in situ. A Cochrane review of randomised controlled trials of PICCs versus peripheral cannulae for delivering parenteral nutrition to neonates did not show an effect on invasive infection rates. Infants with a PICCin situ experienced fewer skin break procedures than infants with a peripheral cannula (or series of cannulae) and this may have balanced the overall risk for acquiring bloodstream infection (Ainsworth 2015).

This review examines the evidence from randomised controlled trials that early planned removal within a pre‐specified maximum dwell time versus allowing clinicians to determine when the PICC is removed or replaced affects the risk of bloodstream infection in newborn infants. Limited guidance about this issue is available. The 2011 US Centers for Disease Control and Prevention Healthcare Infection Control Practices Advisory Committee guidelines do not specify a recommended maximum dwell time and advise that PICCs should not be routinely replaced to prevent catheter‐related infections (O'Grady 2011). This uncertainty is reflected in surveys of practice that identify wide variation between neonatal care centres with regard to PICC insertion and maintenance (Sharpe 2013; Taylor 2014).

We define the maximum dwell time pragmatically as up to two weeks after insertion based on the typical duration of PICC use for infants in most neonatal units and the minimum time that observational studies have shown this to be associated with a rise in the risk of infection (Smith 2008; Sengupta 2010; Wong 2012; Milstone 2013). The intervention is the pre‐specified intent to remove or replace the PICC within this time period, and the control is either (i) any permissive approach that does not pre‐specify dwell time but that allows PICC removal or replacement based on clinical criteria (including suspected or confirmed bloodstream infection), or (ii) a longer pre‐specified dwell time than the intervention.

How the intervention might work

Pre‐specifying a fixed maximum dwell time, with planned removal rather than an expectant approach, may reduce the risk of a PICC being leftin situ when not actually in use or needed and reduce the overall length of PICC exposure. These effects could lower the risk of bloodstream infection and its associated complications if the PICC is an independent risk factor for infection. This intervention may also plausibly affect nutrient intake, either reducing receipt of parenteral nutrients, or prompting a more rapid progression to full enteral feeding, or both, with potential consequences for acute morbidity (principally the risk of acute necrotising enterocolitis), growth and development.

Why it is important to do this review

Given the potential for the planned duration of placement of PICCs to affect important outcomes for newborn infants, we undertook a systematic review to identify, appraise and synthesise the available evidence from randomised controlled trials.