Sean Ainsworth1, Linda Clerihew2, William McGuire3
Background - Methods - Results - Characteristics of Included Studies - References - Data Tables and Graphs
1Directorate of Women and Children's Health, NHS Fife (Acute Hospitals), Kirkcaldy, UK
2Tayside Institute of Child Health, Ninewells Hospital and Medical School, Dundee, UK
3Department of Paediatrics and Child Health, Australian National University Medical School, Canberra, Australia
Citation example: Ainsworth S, Clerihew L, McGuire W. Percutaneous central venous catheters versus peripheral cannulae for delivery of parenteral nutrition in neonates. Cochrane Database of Systematic Reviews 2007, Issue 3. Art. No.: CD004219. DOI: 10.1002/14651858.CD004219.pub3.
Directorate of Women and Children's Health
NHS Fife (Acute Hospitals)
Forth Park Hospital
Bennochy Road
Kirkcaldy
Fife
KY2 5RA
UK
E-mail: sean.ainsworth@faht.scot.nhs.uk
| Assessed as Up-to-date: | 08 July 2008 |
|---|---|
| Date of Search: | 31 May 2008 |
| Next Stage Expected: | 08 July 2010 |
| Protocol First Published: | Issue 2, 2003 |
| Review First Published: | Issue 2, 2004 |
| Last Citation Issue: | Issue 3, 2007 |
| Date / Event | Description |
|---|---|
| 02 June 2008 Updated | This updates the review "Percutaneous central venous catheters versus peripheral cannulae for delivery of parenteral nutrition in neonates" published in the Cochrane Database of Systematic Reviews, The Cochrane Library, Issue 2, 2004 (Ainsworth 2004). Conclusions not changed. |
| Date / Event | Description |
|---|---|
| 05 March 2007 New citation: conclusions changed | Substantive amendment |
Neonatal parenteral nutrition may be delivered via a peripheral cannula or a central venous catheter (either umbilical or percutaneous). Because of the complications associated with umbilical catheters, many neonatal units prefer to use percutaneous catheters following initial stabilisation. Although potentially more difficult to place, these catheters may be more stable than peripheral cannulae and need less frequent replacement. Both delivery methods may be associated with different risks of adverse events, including acquired systemic infection and extravasation injury.
To determine the effect of infusion via percutaneous central venous catheter vs. peripheral cannulae on nutrient input, growth and development, and complications in hospitalized neonates receiving parenteral nutrition.
The standard search strategy of the Cochrane Neonatal Review Group was used. This included searches of the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 2, 2008), MEDLINE (1966 - May 2008), EMBASE (1980 - May 2008), conference proceedings, and previous reviews.
Randomised controlled trials that compared delivery of intravenous fluids (primarily parenteral nutrition) via percutaneous central venous catheters vs. peripheral cannulae in hospitalized neonates.
Data were extracted using the standard methods of the Cochrane Neonatal Review Group, with separate evaluation of trial quality and data extraction by each author, and synthesis of data using relative risk (RR), risk difference (RD) and mean difference (MD).
Five trials recruiting a total of 432 infants were found. One study showed that use of a percutaneous central venous catheter was associated with a a smaller deficit between prescribed and actual intake during the trial period: Mean difference -7.1% (95% CI -11.02, -3.2). Infants in the percutaneous central venous catheter group needed significantly fewer catheters/cannulae: Mean difference -4.3 (95% CI -5.24, -3.43). Meta-analysis of data from all five trials did not find any evidence of an effect on the incidence of invasive infection: typical relative risk 0.93 (95% CI 0.69, 1.23); typical risk difference -0.02 (95% CI -0.10, 0.06).
Data from one small study suggest that using a percutaneous central venous catheter to deliver parenteral nutrition improves nutrient input. The significance of this in relation to long-term growth and developmental outcomes is unclear. Three studies suggested that the use of a percutaneous central venous catheter decreases the number of catheters/cannulae needed to deliver the nutrition. No evidence was found to suggest that percutaneous central venous catheter use increased the risk of adverse events, particularly systemic infection.
More trials are needed to determine whether delivering nutrition into superficial or deep veins is better for newborn infants. Preterm or sick newborn infants are often fed with a special nutrient solution that is delivered directly into the veins. The solutions can either be given into a superficial vein through a standard short (peripheral) cannula or into a large deep vein via a long (central) catheter. This review found limited data from five small randomised controlled trials that compared the effects of using these two different methods of delivering nutrient solutions. There is some evidence from one study that infants who received the solution into a deep vein received more nutrition. The use of central catheters has been thought to increase the risk of bloodstream infection in newborn infants, but this review did not find any evidence that this was the case. More trials are needed to determine which method is better at improving growth and development in newborn infants.
The appropriate methods of feeding neonates vary with gestational age and clinical state. Some neonates, particularly those that are preterm or sick, are slow to tolerate the introduction of enteral feeds because of delayed gastric emptying and intestinal peristalsis. Since early postnatal nutrition may have a significant effect on important outcomes, including long-term neurodevelopment, these infants often receive parenteral (intravenous) nutrition during the establishment of enteral nutrition (Wilson 1997). Parenteral nutrition may also be delivered during periods when enteral nutrition is not possible, such as when there is lack of maternal expressed breast milk or when feeding is specifically contraindicated because of significant gastrointestinal disease such as necrotising enterocolitis. Because modern perinatal care has resulted in improved the survival rates of preterm and sick newborn infants, there has been an increase in the number of infants who require prolonged parenteral nutrition.
Parenteral nutrition usually consists of a glucose and electrolytes solution. More nutritionally complete formulations, "total parenteral nutrition", can also include an amino acid solution with minerals and vitamins in addition to fat as the principal non-protein energy source. The solutions are infused via either a short narrow-gauge peripheral venous cannula or a longer central venous catheter that extend into the larger veins such as the vena cavae (Shaw 1973; Trotter 1996).
Although it may be technically easier to site a peripheral cannula than to site a central venous catheter, peripheral cannulae are less stable and may require more frequent replacement. Once placed, a central venous catheter should remain in situ longer than a peripheral cannula (Shaw 1973), thus reducing the number of potentially painful procedures to which the infant is exposed. The need for frequent replacement of a peripheral cannula might also result in a significant cumulative period of interruption to the delivery of parenteral nutrition and in a nutrient deficit that which can have potential long-term consequences on growth and development (Embleton 2001). Another concern with the use of a peripheral cannula for delivering parenteral nutrition is the risk of extravasation injury. The subcutaneous infiltration of a hypertonic and irritant parenteral nutrition solution can cause local skin ulceration, secondary infection, and scarring.
In neonatal practice, a central venous catheter for the delivery of parenteral nutrition is usually placed percutaneously, although an umbilical venous or arterial catheter may also be used to deliver parenteral nutrition, particularly during the first week of life. Surgical placement, in which a deep vein is surgically exposed prior to cannulation and for which the infant may require a general anaesthetic, is less common. However, in infants in whom peripheral venous access is a major problem, often because of the frequency of previous cannulations, a surgically-placed central venous catheter may be the only option for the delivery of parenteral nutrition. This review will focus on the specific comparison of percutaneous central venous catheters vs. peripheral cannulae.
The most common complication associated with percutaneous central venous catheter use is nosocomial infection, which can include bacteraemia and invasive fungal infection. Micro-organisms can enter the bloodstream through the catheter entry site (Salzman 1995) or less commonly via the catheter hub (Salzman 1993). Catheter-associated thrombosis can act as a nidus for infection. It is often necessary to remove the catheter in order to clear the infection (Karlowicz 2002). The reported incidences of catheter-related sepsis in the neonatal intensive care unit vary from 5% (Neubauer 1995) to nearly 40% (Hruszkewycz 1991), depending on the criteria used to define catheter-related sepsis and the population studied. Extremely low birth weight infants (birth weight less than 1000 grams) are particularly at risk. Additional putative risk factors include prolonged use of parenteral nutrition and insertion of the catheter after the first week of life (Mahieu 2001). However, it is uncertain whether percutaneous central venous catheter use further increases the risk of infection in an "at-risk" population (Sohn 2001).
The use of a percutaneous central venous catheter to deliver parenteral nutrition may also be associated with iatrogenic injury, including embolism (air or thrombus) and infusate extravasation into tissue spaces. For example, there have been recent reports of cardiac tamponade following migration of the catheter tip to the pericardial space, and some authorities have recommended that central venous catheters should no longer be placed in the right atrium or ventricle because of this risk (Darling 2001).
Given that the choice of route for the delivery of parenteral nutrition may affect clinically important outcomes in neonates such as growth and development, the available evidence was systematically reviewed to determine if there are any implications for current practice or for future research.
To determine the effect of infusion of parenteral nutrition via a percutaneous central venous catheter vs. a peripheral catheter on nutrient input, growth and development and adverse consequences including bacteraemia or invasive fungal infection, cardiac tamponade, or other extravasation injuries in hospitalized newborn infants who require parenteral nutrition.
The following subgroup analyses were performed:
Neonates (newborn infants less than 28 days old at study entry) receiving parenteral nutrition, cared for in a hospital setting.
Trials comparing parenteral nutrition delivered via a percutaneously inserted central venous catheter vs. a peripheral cannula. Studies that compared delivery of parenteral nutrition via a surgically-placed central line (where the vein is surgically exposed prior to cannulation) or umbilical catheter are not be included. A minimum duration for trials was not specified.
Primary outcomes: Nutrient input, growth and development:
1. Average daily input of parenteral calories (kilocalories per kilogram per day) and protein (grams per kilogram per day) during trial period
2. Average daily proportion of prescribed parenteral calories and protein that were actually delivered during trial period
3. Short-term (prior to discharge from hospital) growth: weight gain (grams per day, or grams per kilogram per day), weight z-score at discharge, linear growth (millimetres per week), head growth (millimetres per week), skinfold thickness growth (millimetres per week)
4. Long-term (following discharge from hospital) growth: weight gain (grams per day, or grams per kilogram per day), linear growth (millimetres per week), head growth (millimetres per week), skinfold thickness growth (millimetres per week)
5. Neurodevelopmental outcomes during infancy and beyond using validated assessment tools such as Bayley Scales of Infant Development, and classifications of disability, including auditory and visual disability. Severe neurodevelopmental disability was defined as any one or combination of the following: non-ambulant cerebral palsy, developmental delay (developmental quotient less than 70), auditory and visual impairment.
Secondary outcomes: Adverse events:
1. Death (all cause) before 28 days
2. Death (all cause) before discharge from hospital
3. Confirmed invasive bacterial infection as determined by:
a. culture from a normally sterile site: cerebrospinal fluid, blood (from peripheral sites, not from indwelling catheters), urine (obtained by sterile urethral catheterisation or supra-pubic bladder tap), bone or joint, peritoneum, pleural space, or central venous line tip
b. findings on autopsy examination consistent with invasive bacterial infection
4. Confirmed invasive fungal infection as determined by:
a. culture from a normally sterile site: cerebrospinal fluid, blood (from peripheral sites, not from indwelling catheters), urine (obtained by sterile urethral catheterisation or supra-pubic bladder tap), bone or joint, peritoneum, pleural space, or central venous line tip
b. findings on autopsy examination consistent with invasive fungal infection
c. findings on ophthalmological examination consistent with fungal ophthalmitis or retinitis
d. pathognomonic findings on renal ultrasound examination: "renal fungal balls"
5. Extravasation injury as determined by:
a. subcutaneous injury resulting in skin ulceration
b. "deep" extravasation resulting in limb swelling
b. "central" extravasation-infusate in the pleural, peritoneal, or pericardial space
6. Number of cannulae/catheters per child used to administer parenteral nutrition to each infant during trial period
The standard search strategy of the Cochrane Neonatal Review Group was used, including electronic searches of The Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 2, 2008), MEDLINE (1966 - May 2008), EMBASE (1980 - May 2008). The search strategy included the following text words and MeSH subject headings: Infant-Newborn, infan$, neonat$, newborn, parenteral near3 nutrition, TPN, central near3 catheter, CVC, central near3 line, CVL. The search outputs will be limited with the relevant search filters for clinical trials. We did not apply any language restriction.
References in previous reviews and in studies identified as potentially relevant were examined. Abstracts presented at the Society for Pediatric Research, Pediatric Academic Society and the European Society for Pediatric Research or published in the journal Pediatric Research between 1980 and 2008 inclusive were hand searched. Trials reported only as abstracts were eligible if sufficient information was available from the report, or from contact with the authors to fulfil the inclusion criteria.
1. William McGuire (WM) and Linda Clerihew (LC) screened the title and abstract of all studies identified by the above search strategy. All review authors screened the full text of the report of each study identified as of potential relevance. Only the studies that met all of the pre-specified inclusion criteria were eligible for inclusion. The review authors resolved any disagreements by discussion until consensus was achieved.
2. The criteria and standard methods of the Cochrane Neonatal Review Group to assess the methodological quality of the included trials were used. The quality of the included trials in terms of allocation concealment, blinding of parents or caregivers and assessors to intervention, and completeness of assessment in all randomised individuals. Additional information from the authors of included trials to clarify methodology and results was requested as necessary.
3. A data collection form to aid extraction of relevant information and data from each included study was used. LC and Sean Ainsworth (SA) extracted the data separately, compared the data, and resolved differences by consensus and discussion with WM.
4. The standard methods of the Cochrane Neonatal Review Group to analyse and synthesize the data was used. Outcomes per infant, rather than per catheter, or per catheter-unit time were considered. Outcomes for infants needed to have been reported on an intention-to-treat basis. For example, if an infant was randomly allocated to parenteral nutrition via a percutaneous central venous catheter but subsequently received nutrition via a peripheral cannula because central access was not achieved, the outcome for this infant was still considered in the percutaneous central venous catheter group. The treatment effects of individual trials were estimated and heterogeneity was examined between trial results using the I2 test for dichotomous outcomes and ANOVA for continuous outcomes. A fixed effects model for meta-analyses was used. The effects seen for categorical data was expressed as relative risk, risk difference, and number needed to treat, with respective 95% confidence intervals. The effects seen for continuous data was expressed as weighted mean difference and 95% confidence interval.
Ten studies for potential inclusion were identified. Five of these studies, involving a total of 432 infants, fulfilled the inclusion criteria (Annibale 1995; Janes 2000; Ainsworth 2001; Wilson 2007; Barria 2007). Details of these trials are presented in the Characteristics of Included Studies Table. The other reports that were evaluated but not included were not randomised controlled trials (Cairns 1995; Parellada 1999; Liossis 2003), a trial that compared outcomes in neonates and children (Schwengel 2004) and one trial that did not compare percutaneous central venous catheters with peripheral cannulae (Childs 1995).
The included studies were undertaken in neonatal intensive care units in North America (Annibale 1995; Janes 2000; Wilson 2007), Chile (Barria 2007) and the United Kingdom (Ainsworth 2001) during the 1990s and early 2000s. The average gestational age of participating infants was 26 weeks (Janes 2000), 27 weeks (Wilson 2007), 28 weeks (Ainsworth 2001), 29 weeks (Annibale 1995) and 31 weeks (Barria 2007). In all of the trials, infants were recruited within the first week after birth and follow-up continued until the infant no longer required intravenous access for delivery of parenteral nutrition (when the infant tolerated enteral intake).
Four of the studies did not report any data on the pre-specified primary outcomes for this review: nutritional input, growth, and development (Annibale 1995; Janes 2000; Wilson 2007; Barria 2007). The lead investigators for these trials were contacted in order to acquire the relevant data. Ainsworth 2001 reported the proportion of prescribed parenteral nutrition that was actually delivered during the trial period. All of the trials reported data on the incidence of bacteraemia or fungaemia. Three of the trials reported the number of insertion attempts and catheters required to maintain venous access, as well as the total duration of intravenous access (Annibale 1995; Janes 2000; Barria 2007). Wilson 2007 report the number of skin punctures but not the duration of venous access. Although not primary outcomes for any of the studies, data on neonatal deaths and deaths prior to hospital discharge were available for three of the trials (Ainsworth 2001; Janes 2000; Wilson 2007 ).
Three of the excluded studies (Cairns 1995; Parellada 1999; Liossis 2003) reported outcomes between cohorts of infants who received parenteral nutrition via either a percutaneous catheters or a peripheral cannula. Although the groups were matched in terms of gestation and birth weights, the potential for bias at selection for these groups was high. For example, clinicians may have been more likely to use a peripheral cannula in those infants thought to require parenteral nutrition for shorter durations. Schwengel 2004was excluded as this examined outcomes in paediatric surgical patients and enrolled all age ranges from neonates through to 14 year olds. No stratification had been used in the study and it was not possible to extract the neonatal data. The study by Childs 1995 looked at whether placing the tip of central venous catheters were better placed centrally or peripherally.
Quality assessments are detailed in the Characteristics of Included Studies Table. The methodology of the studies was generally good. Random allocation was concealed by using sealed opaque envelopes in four of the trials (Annibale 1995; Ainsworth 2001; Wilson 2007; Barria 2007). It is unclear if a satisfactory method of allocation concealment was used in the fifth trial (Janes 2000). The lead investigator was contacted in order to seek clarification. None of the trials was able to blind caregivers to the intervention. Follow-up appears to have been complete for the outcomes reported.
PERCUTANEOUS CENTRAL VENOUS CATHETERS VERSUS PERIPHERAL CANNULA (COMPARISON 1):
Primary outcomes: Nutrient input, growth and development
1. Average daily input of parenteral calories:
None of the studies reported the average daily input of parenteral calories (kilocalories per kilogram per day) and protein (grams per kilogram per day) during the trial period.
2. Average daily proportion of prescribed parenteral calories and protein that were actually delivered during trial period (Outcome 1.1):
Ainsworth 2001 reported a statistically significant difference in the deficit of delivered parenteral nutrition (from that actually prescribed) during the trial period: 3.2% (standard deviation 6.8%) in the percutaneous central venous catheter group vs. 10.3% (standard deviation 7.2%) in the peripheral cannula group: mean difference in the percentage of the prescribed nutritional intake actually received: -7.1% (95% confidence interval -11.02, -3.2).
3. Short-term growth parameters:
None of the studies reported any short-term growth parameters (prior to discharge from hospital).
4. Long-term growth parameters:
None of the studies reported any long-term growth parameters (following discharge from hospital).
5. Neurodevelopmental outcomes:
None of the studies reported any neurodevelopmental outcomes during infancy and beyond.
Secondary outcomes: Adverse events
1. Death (all causes) before 28 days (Outcome 1.2):
Three of the included studies report this outcome (Janes 2000; Ainsworth 2001; Wilson 2007). The authors of a fourth study (Barria 2007) confirmed that none of the infants entered into their study died either during the study period or prior to discharge. Seven of a total of 282 infants recruited to the four studies died. There was no statistically significant difference in the incidence of death before 28 days in either individual trial or on meta-analysis of the four trials: typical RR 1.31 (95% CI 0.36, 4.81); typical RD -0.01 (95% CI -0.03, 0.05). The lead investigator was contacted to inquire if the data on mortality are also available from the fifth trial (Annibale 1995).
2. Death (all cause) before discharge from hospital (Outcome 1.3):
Death before discharge from hospital: These data are identical to those for death before 28 days.
3. Confirmed invasive bacterial infection and 4. Confirmed systemic fungal infection (Outcome 1.4):
Reported as bloodstream infections (sepsis) in all of the trials.
Ainsworth 2001: 11 of 24 infants in the percutaneous central venous catheter group vs. 10 of 25 infants in the peripheral cannula group had an episode of bloodstream infection: RR 1.15 (95% CI 0.6, 21.9); RD 0.06 (95% CI -0.22, 0.34).
Annibale 1995: 22 of 75 infants in the percutaneous central venous catheter group vs. 27 of 75 infants in the peripheral cannula group had an episode of bloodstream infection: RR 0.81 (95% CI 0.51, 1.29); RD -0.07 (95% CI -0.22, 0.08).
Janes 2000 reported the total number of episodes of systemic infection in each group: 11 in the percutaneous central venous catheter group vs. 16 in the peripheral cannula group. However, some of the infants in each group had more than one episode of systemic infection during the trial period. The authors provided further information that 10 of 32 infants in the percutaneous central venous catheter group vs. 13 of 31 infants in the peripheral cannula group had an episode of bloodstream infection: RR 0.75 (95% CI 0.39, 1.44); RD -0.11 (95% CI -0.34, 0.13).
Wilson 2007: 15 of 36 infants in the percutaneous central venous catheter group vs. 13 of 50 infants in the peripheral cannula group had an episode of bloodstream infection: RR 1.25 (95% CI 0.67, 2.34); RD 0.07 (95% CI -0.12, 0.25).
Barria 2007; 1 of 37 infants in the percutaneous central venous catheter group vs. 2 of 37 infants in the peripheral cannula group had an episode of bloodstream infection: RR 0.50 (95% CI 0.05, 5.28); RD -0.03 (95% CI -0.12, 0.05).
Meta-analysis of the five trials did not detect any statistically significant difference in the incidence of systemic infection between the groups: typical RR 0.93 (95% CI 0.69, 1.23); typical RD -0.02 (95% CI -0.10, 0.06).
5. Extravasation injury (Outcome 1.5):
Three of the trials reported this outcome (Janes 2000; Ainsworth 2001; Wilson 2007). Barria 2007 reported the incidence of 'phlebitis' but this did not fulfil the criteria for extravasation injury. The lead investigator of the fifth trial was contacted to determine if the data were available (Annibale 1995).
Ainsworth 2001 reported that one infant in the percutaneous central venous catheter group vs. none in the peripheral cannula group had an extravasation injury (a self-limiting episode of groin swelling). Janes 2000 reported that there were no episodes of significant subcutaneous infiltration in either of the study groups. Wilson 2007 reported none of the infants in the percutaneous central venous catheter group vs. 5 of 50 in the peripheral cannula group had an extravasation injury. Meta-analysis of the three trials did not find any statistically significant difference: typical RR 0.36 (95% CI 0.07, 1.75); typical RD -0.04 (95% CI -0.09, 0.02).
6. Number of cannulae/catheters per infant used to administer parenteral nutrition during trial period (Outcome 1.6):
These data are not available from two of the trials (Ainsworth 2001; Wilson 2007). Janes 2000 reported that the mean number of cannulae/catheters used per infant was 4.8 (standard deviation 3.6) in the percutaneous central venous catheter group vs. 8.0 (standard deviation 4.2) in the peripheral cannula group. Annibale 1995 reported that the number of catheters per infant did not differ between the two groups (average 4 vs. 7). Barria 2007 reported that the mean number of cannulae/catheters used per infant was 8.6 (standard deviation 11.1) in the percutaneous central venous catheter group vs. 12.8 (standard deviation 13.8) in the peripheral cannula group.
Wilson 2007 reported median and ranges for the number of skin punctures; the median (range) of skin puncture in the percutaneous central venous catheter group was 9 (1, 74) vs. 14.5 (1, 111) in the peripheral cannula group. The authors of this study have been contacted for clarification about the means and standard deviations for these data.
Meta-analysis of the Annibale 1995, Janes 2000 and Barria 2007 studies shows a significant reduction in the number of number of cannulae/catheters in the percutaneous central venous catheter group; weighted mean difference -3.10 (-4.13, -2.06).
Sub-group analyses:
Most infants participating in the five included trials were preterm. No sub-group analysis by birth weight was done.
Five randomised controlled trials comparing the use of percutaneous central venous catheters vs. peripheral catheters for newborn infants who require parenteral nutrition were found. A total of 432 infants participated in the trials.
Limited data was found in the published reports on the pre-specified primary outcomes for this review. None of the trials assessed effects on growth or any longer term outcomes, nor did they plan to do so. Only one trial assessed nutrient input. Ainsworth 2001 found that infants randomly allocated to receive parenteral nutrition via peripheral cannulae had a statistically significantly higher nutritional deficit during the trial period than infants who received nutrition via a central venous catheter. The 7% difference in deficit seen would, if accumulated over the period of one week, result in the loss of one half of one day's nutrient requirements. The clinical significance of this finding is uncertain. Nutritional deficit during this very critical period of brain growth may have adverse consequences for long term neurodevelopmental outcomes. However, it may be that when enteral nutrition is introduced, catch-up growth can compensate for any deficiencies experienced during the early neonatal period.
All of the trials reported the incidence of systemic (bloodstream) bacterial or fungal infection. Meta-analysis did not demonstrate any statistically significant difference in the incidence of systemic infection between the groups. This finding should be interpreted with some caution as none of the trials blinded caregivers to the nature of the intervention and surveillance bias may have affected the results. Because percutaneous central venous catheters are often considered risk factors for sepsis, caregivers may have had a lower threshold for investigating or diagnosing systemic infection in the percutaneous central venous catheter group. This situation is further compounded by the difficulty in diagnosing true systemic infection vs. contamination of blood cultures with skin commensals. The reported systemic infection rates in the central venous catheter group might have been artificially high as caregivers, already more aware of the risks of sepsis might have, attributed symptoms to organisms isolated in contaminated cultures.
Although frequently blamed for increasing the risk of systemic bacterial or fungal infections, percutaneous central venous catheters are not the only piece of 'plastic' used in the high risk neonate that by-passes the body's innate defences. A high proportion of infants with percutaneous central venous catheters have or have had endotracheal tubes, nasogastric tubes, peripheral cannulae as well as having an immature gastrointestinal mucosa that predisposes them to infections. The results of this meta-analysis and three non-randomised cohort studies (Cairns 1995, Parellada 1999 and Liossis 2003) all suggest that there are no significant differences in systemic infection rates in infants receiving parenteral nutrition via a central venous catheter or via peripheral cannulae.
None of the included studies reported the routine use of prophylactic antibiotics in the prevention of sepsis. Use of antibiotic prophylaxis to prevent sepsis whilst umbilical arterial and venous catheters are in place has not been shown to be beneficial (Inglis 2004; Inglis 2005). Routine use of low dose vancomycin prophylaxis has been shown to reduce the incidence of sepsis in preterm infants with and without percutaneous central venous catheters, although there are concerns that this may increase the numbers of vancomycin-resistant strains of bacteria (Ocete 1998; Craft 2000; Sherer 2005).
Percutaneous central venous cannulae are thought to be more stable than peripheral cannulae and, therefore, need to be replaced less frequently. Some evidence for this advantage was found. Three studies (Annibale 1995, Janes 2000 and Barria 2007) found that the mean number of cannulae/catheters used was statistically significantly lower in the percutaneous central venous catheter group compared with the peripheral cannula group. On average, infants needed about three fewer cannulae/catheters during the trial period. Wilson 2007 reported the number of number of skin punctures required to maintain venous access. It is unclear whether these data were for the actual numbers of catheters or cannulae or included 'failed' attempts. The data were also presented as a median and range. Further data from the lead investigator need to be obtained for inclusion in future updates. The number of catheter/cannulae may indirectly reflect the number of painful procedures in these infants. However, it is not clear whether insertion of a catheter (usually through a larger bore needle and technically more difficult) is more painful for the infant than the insertion of several peripheral cannulae over time. No study has specifically addressed the issue of pain or used pain scores in this respect.
The numbers of infants in these studies were too small to make any conclusions regarding the effects of either the catheters or cannulae on serious clinical adverse effects such as extravasation injuries and cardiac tamponade. Cartwright 2004 reported on the use of 2186 catheters in one unit and found that, with careful adherence to policies for insertion, placement and aseptic precautions during times when the lines are accessed, percutaneous central venous catheters are a safe method of delivery parenteral nutrition in neonates.
Limited data from one small study suggest that the use of a percutaneous central venous catheter rather than a peripheral cannula is associated with a statistically significant smaller deficit in delivered parenteral nutrition. Percutaneous central venous catheter use resulted in fewer painful procedures (venepunctures) than peripheral cannula use. No evidence of an increased risk of adverse effects, particularly systemic infection was found. There were no data on long-term growth and developmental outcomes.
Further large, adequately powered, randomised controlled trials are needed to determine whether the use of a percutaneous central venous catheter rather than a peripheral cannula for delivering parenteral nutrition has an important benefits for newborn infants. Trials should examine the effects of this intervention on growth and neurodevelopmental outcomes, particularly in very preterm infants where early nutritional intake may play an important role.
We thank the lead investigators of all of the included trials for clarification of aspects of their studies.
Sean Ainsworth (SA) and William McGuire (WM) developed the protocol for this review. All review authors screened the title and abstract of all studies identified by the search strategy. Linda Clerihew (LC) and SA screened the full text of the report of each study identified as of potential relevance. LC and SA extracted the data separately, compared data, and resolved differences by consensus, and with discussion with WM. SA and WM completed the final review.
Sean Ainsworth is the lead investigator of one of the included trials (Ainsworth 2001).
| Methods | Blinding of randomisation: Yes Blinding of intervention: No Complete follow-up: Yes Blinding of outcome measurement: Yes |
|---|---|
| Participants | 49 infants cared for a regional neonatal intensive care unit who, in the opinion of the attending clinician were likely to need parenteral nutrition for greater than five days. Median gestation of recruited infants was 28 weeks. |
| Interventions | Delivery of parenteral nutrition either via a percutaneous central venous catheter (N = 24), or a peripheral cannula (N = 25). |
| Outcomes | 1. Episodes of "sepsis"- bacteraemia or fungaemia. |
| Notes | Infants in the two arms were of similar gestation and birth weight and age at randomisation. The infants in the percutaneous central venous catheter group were statistically significantly less likely to have had an umbilical line in situ and to have received parenteral antibiotics prior to randomisation. |
| Item | Judgement | Description |
|---|---|---|
| Allocation concealment? | Yes | A - Adequate |
| Methods | Blinding of randomisation: Yes Blinding of intervention: No Complete follow-up: Yes Blinding of outcome measurement: Can't tell. |
|---|---|
| Participants | 150 neonates (less than 6 days old) cared for in a large neonatal unit, who were thought likely to require intra-venous access for at least 3 days. Exclusion criteria not stated. |
| Interventions | Delivery of parenteral nutrition either via a percutaneous central venous catheter (N = 75), or a peripheral cannula (N = 75). |
| Outcomes | 1. Incidence of bacterial or fungal sepsis. |
| Notes | Trial reported in abstract form only. The lead investigator kindly provided further details of the trial methodology. |
| Item | Judgement | Description |
|---|---|---|
| Allocation concealment? | Yes | A - Adequate |
| Methods | Blinding of randomisation: Yes Blinding of intervention: No Complete follow-up: Yes Blinding of outcome measurement: Can't tell |
|---|---|
| Participants | 74 neonates cared for in a single regional neonatal intensive care unit who were likely to require intravenous fluids for more than 5 days. Median gestation of recruited infants was 31 weeks. |
| Interventions | Delivery of intravenous fluids (including parenteral nutrition) either via a percutaneous central venous catheter (N = 37), or a peripheral cannula (N = 37). |
| Outcomes | 1. Length of stay |
| Notes | Infants were of similar gestation, birth weights and age at randomisation. Catheters / cannulae were used for intravenous ('clear') fluids as well as parenteral nutrition (proportion of infants receiving parenteral nutrition slightly higher in central venous catheter arm) Further details on study methodology and results were available from the investigators. |
| Item | Judgement | Description |
|---|---|---|
| Allocation concealment? | Yes | A - Adequate |
| Methods | Blinding of randomisation: Can't tell Blinding of intervention: No Complete follow-up: Yes Blinding of outcome measurement: Can't tell |
|---|---|
| Participants | 63 infants of birth weights less than 1251 grams, cared for in a neonatal intensive care unit, and likely to require intra-venous maintenance fluids or total parenteral nutrition at one week of age or when their umbilical venous catheter was removed. |
| Interventions | Delivery of parenteral nutrition via a percutaneous central venous catheter (N = 32), vs. a peripheral cannula (N = 31). |
| Outcomes | 1. Incidence of bacterial or fungal sepsis |
| Notes | Random allocation was achieved using a "computer-generated block random number table". |
| Item | Judgement | Description |
|---|---|---|
| Allocation concealment? | Unclear | B - Unclear |
| Methods | Blinding of randomisation: Yes Blinding of intervention: No Complete follow-up: Yes Blinding of outcome measurement: Can't tell |
|---|---|
| Participants | 96 infants of birth weights less than 1251 grams or gestation less than 32 weeks at birth, cared for in a neonatal intensive care unit, and likely to require intra-venous maintenance fluids or total parenteral nutrition for at least 5 days of age. |
| Interventions | Delivery of parenteral nutrition via a percutaneous central venous catheter (N = 46), vs. a peripheral cannula (N = 50). |
| Outcomes | 1. Incidence of bacterial or fungal sepsis or death. |
| Notes | Infants were of similar gestation, birth weights and age at randomisation. |
| Item | Judgement | Description |
|---|---|---|
| Allocation concealment? | Yes | A - Adequate |
| Reason for exclusion | Not a randomised controlled trial. Matched historic cohorts. |
|---|
| Reason for exclusion | Comparison of percutaneous central venous catheters where the tip is sited either in a peripheral vein or a central vein (not percutaneous central venous catheters vs. peripheral cannulae). |
|---|
| Reason for exclusion | Not a randomised controlled trial. Matched historic cohorts. |
|---|
| Reason for exclusion | Non-randomised comparison of infants with percutaneous central venous catheters and matched controls who had peripheral cannulae |
|---|
| Reason for exclusion | Randomised trial of percutaneous central venous catheters vs. peripheral cannulae in neonates and children (ages ranging from neonatal through 14 years) without stratification by age group. We are seeking relevant subgroup data from the trial authors. |
|---|
Ainsworth SB, Furness J, Fenton AC. Randomized comparative trial between percutaneous longlines and peripheral cannulae in the delivery of neonatal parenteral nutrition. Acta Paediatrica 2001;20:1016-20.
Annibale DJ, Bissinger RL, Husley TC, Headdon P, Ohning BL. Early percutaneous central venous catheterization (PCVC) in neonates. Pediatric Research 1995;37:1148.
Barría RM, Lorca P, Muñoz S. Randomized controlled trial of vascular access in newborns in the neonatal intensive care unit. Journal of Obstetric, Gynecologic and Neonatal Nursing 2007;36:450-6.
Cairns PA, Wilson DC, McClure BG, Halliday HL, McReid M. Percutaneous central venous catheter use in the very low birth weight neonate. European Journal of Pediatrics 1995;154:145-7.
Childs AM, Murdoch Eaton DG, Standring P, Puntis JW. A prospective comparison of central and peripheral vein access for parenteral nutrition in the newborn. Clinical Nutrition 1995;14:303-5.
Liossis G, Bardin C, Papageorgiou A. Comparison of risks from percutaneous central venous catheters and peripheral lines in infants of extremely low birth weight: a cohort controlled study of infants < 1000 g. The Journal of Maternal-Fetal & Neonatal Medicine 2003;13:171-4.
Cartwright DW. Central venous lines in neonates: a study of 2186 catheters. Archives of Disease in Childhood Fetal & Neonatal Edition 2004;89:F504-8.
Craft AP, Finer NN, Barrington KJ. Vancomycin for prophylaxis against sepsis in preterm neonates. Cochrane Database of Systematic Reviews 2000, Issue 1. Art. No.: CD001971. DOI: 10.1002/14651858.CD001971.
Darling JC, Newell SJ, Mohamdee O, Uzun O, Cullinane CJ, Dear PR. Central venous catheter tip in the right atrium: a risk factor for neonatal cardiac tamponade. Journal of Perinatology 2001;21:461-4.
Embleton NE, Pang N, Cooke RJ. Postnatal malnutrition and growth retardation: an inevitable consequence of current recommendations in preterm infants? Pediatrics 2001;107:270-3.
Hruszkewycz V, Holtrop PC, Batton DG, Morden RS, Gibson P, Band JD. Complications associated with central venous catheters inserted in critically ill neonates. Infection Control and Hospital Epidemiology 1991;12:544-8.
Inglis GDT, Davies MW. Prophylactic antibiotics to reduce morbidity and mortality in neonates with umbilical artery catheters. Cochrane Database of Systematic Reviews 2004, Issue 3. Art. No.: CD004697. DOI: 10.1002/14651858.CD004697.pub2..
Inglis GDT, Davies MW. Prophylactic antibiotics to reduce morbidity and mortality in neonates with umbilical venous catheters. Cochrane Database of Systematic Reviews 2005, Issue 4. Art. No.: CD005251. DOI: 10.1002/14651858.CD005251.pub2.
Karlowicz MG, Furigay PJ, Croitoru DP, Buescher ES. Central venous catheter removal versus in situ treatment in neonates with coagulase-negative staphylococcal bacteremia. The Pediatric Infectious Disease Journal 2002;21:22-7.
Mahieu LM, De Muynck AO, Ieven MM, De Dooy JJ, Goossens HJ, Van Reempts PJ. Risk factors for central vascular catheter-associated bloodstream infections among patients in a neonatal intensive care unit. The Journal of Hospital Infection 2001;48:108-16.
Nakamura KT, Sato Y, Erenberg A. Evaluation of a percutaneously placed 27-gauge central venous catheter in neonates weighing less than 1200 grams. Journal of Parenteral and Enteral Nutrition 1990;14:295-9.
Neubauer AP. Percutaneous central i.v. access in the neonate: experience with 535 silastic catheters. Acta Paediatrica 1995;84:756-60.
Ocete E, Ruiz-Extremera A, Goicoechea A, Lozano E, Robles C, Rey ML, Salmeron J. Low-dosage prophylactic vancomycin in central-venous catheters for neonates. Early Human Development 1998;53 (suppl):S181-6.
Salzman MB, Isenberg HD, Shapiro JF, Lipsitz PJ, Rubin LG. A prospective study of the catheter hub as the portal of entry for microorganisms causing catheter-related sepsis in neonates. The Journal of Infectious Diseases 1993;167:487-90.
Salzman MB, Rubin LG. Intravenous catheter-related infections. Advances in Pediatric Infectious Diseases 1995;10:337-68.
Shaw JCL. Parenteral nutrition in the management of sick low birth weight infants. Pediatric Clinics of North America 1973;20:333-58.
Sherer CR, Sprague BM, Campos JM, Nambiar S, Temple R, Short B, Singh N. Characterizing vancomycin-resistant enterococci in neonatal intensive care. Emerging Infectious Diseases 2005;11:1470-2.
Sohn AH, Garrett DO, Sinkowitz-Cochran RL, Grohskopf LA, Levine GL, Stover BH, Siegel JD, Jarvis WR. The Pediatric Prevention Network. Prevalence of nosocomial infections in neonatal intensive care unit patients: Results from the first national point-prevalence survey. Journal of Pediatrics 2001;139:821-7.
Thureen PJ, Hay WW Jr. Early aggressive nutrition in preterm infants. Seminars in Neonatology 2001;6:403-15.
| Outcome or Subgroup | Studies | Participants | Statistical Method | Effect Estimate |
|---|
| 1.1 Percentage deficit in nutrient delivery per infant | 1 | 49 | Mean Difference (IV, Fixed, 95% CI) | -7.10 [-11.02, -3.18] |
| 1.2 Death before 28 days (all causes) | 4 | 282 | Risk Ratio (M-H, Fixed, 95% CI) | 1.31 [0.36, 4.81] |
| 1.3 Death before hospital discharge (all causes) | 4 | 282 | Risk Ratio (M-H, Fixed, 95% CI) | 1.31 [0.36, 4.81] |
| 1.4 Invasive bacterial and fungal infection | 5 | 432 | Risk Ratio (M-H, Fixed, 95% CI) | 0.93 [0.69, 1.23] |
| 1.5 Extravasation injury | 3 | 208 | Risk Ratio (M-H, Fixed, 95% CI) | 0.36 [0.07, 1.75] |
| 1.6 Number of cannulae/catheters per infant | 3 | 287 | Mean Difference (IV, Fixed, 95% CI) | -3.10 [-4.13, -2.06] |
This review is published as a Cochrane review in The Cochrane Library, Issue 4, 2008 (see http://www.thecochranelibrary.com for information). Cochrane reviews are regularly updated as new evidence emerges and in response to feedback. The Cochrane Library should be consulted for the most recent version of the review. |
