We aimed to find out which was the better way to keep a newborn baby's intravenous line open and working -- either running a continuous amount of intravenous fluid through it (continuous infusion) or giving a small amount of fluid through it every few hours (intermittent flush) only. One study showed no difference between the two approaches for keeping a baby's intravenous line open and working and one study showed an advantage for intermittent flushes. The studies, however, had some problems in how the data were analysed and reported. Therefore, we are uncertain as to how reliable the results are and further research should be undertaken.
In the second study, only one of our primary outcomes was available: the mean (SD) number cannulas used per infant in the first 48 hours was less in the intermittent flush group with a mean difference of -0.76 cannulas (95% CI -1.37 to -0.15). No results were available for any of our other primary outcomes: in the published report, results were reported per catheter rather than per infant, a number of infants received more than one intravenous catheter (39 infants received an unknown number of catheters). The overall duration of cannula patency was significantly longer in the intermittent flush group with a mean duration of patency in the intermittent flush group of 2.1 days (SD 1.0) compared with the continuous infusion group where the mean duration of patency was 1.0 days (SD 0.5) - Student's t test P value 0.0003.
Insertion of peripheral intravenous cannulas in neonates can be a stressful experience for the baby, the new parent and the medical/nursing staff (Yeo 1998; Olds 2000; Cotton 1998). It is therefore desirable once the intravenous cannula is inserted that its patency be maintained for as long as possible. Each cannula should last as long as possible and for any given period of treatment a minimum number of cannulas should be used. To achieve this the cannula can be infused continuously with fluid at a low rate or flushed intermittently (usually every 4-8 hours) (Cotton 1998).
Continuous infusions require more nursing time and equipment and impede access of the mother to the infant but the cannula might last longer and have fewer complications such as extravasation or dislodgment. Intermittent flushing would decrease nursing time and equipment and allow greater access of the mother to her infant, but may decrease cannula life by blockage due to clotting. Currently there is little uniformity between neonatal nurseries as to which method is used and evidence is required to decide which method, if any, best maintains intravenous access.
There is limited evidence that the following factors may increase the risk of intravenous cannula failure: certain drugs (such as gentamicin or aminophylline) (Moclair 1995); size of cannula (Danek 1992).
A different question as to whether the continuous infusion fluid or the intermittent flush solution should contain heparin has been considered in a separate Cochrane review (Shah 2004). Shah et al concluded that "no conclusive evidence is available...to evaluate the effectiveness of heparin to prolong PIV [peripheral intravenous] cannula life in the neonatal population". Only eight eligible studies were identified and only five commented on duration of catheter use. Due to significant clinical heterogeneity and heterogeneity in treatment effect no recommendation could be made as to whether use of heparin was advisable or not without further research.
As with all invasive procedures, there is a significant risk of introducing infection into the newborn when inserting a peripheral intravenous catheter. Septicaemia, although an infrequent complication could potentially have a lethal outcome. It is therefore desirable that catheter patency be maintained as long as possible thus reducing the number of skin breakages made in the baby and ultimately decreasing the risk of infection.
Primary Objectives
The standard search strategy for the Cochrane Neonatal Review Group was used. See: Neonatal Review Group search strategy. We searched the following electronic databases: The Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 2, 2004), CINAHL (from 1982 - June 2004 ) and MEDLINE (from 1966 - June 2004).
Searches of the electronic databases were based on the following search terms:
MeSH terms: infusions, intravenous OR injections, intravenous;
OR text words: "IV treatment" OR "IV therapy" OR "IV drug" OR "IV medication"
OR "intravenous treatment" OR "intravenous therapy" OR "intravenous drug"
OR "intravenous medication" OR "intra-venous treatment" OR "intra-venous
therapy" OR "intra-venous drug" OR "intra-venous medication";
NOT MeSH term: Substance Abuse, Intravenous.
AND
MeSH term: infant, newborn OR text word "neonate"
AND
The highly sensitive search strategy developed by Kay Dickersin to identify RCTs (Dickersin 1994)
We also searched previous reviews including cross references. We also
used the results of searches of abstracts, conference and symposia proceedings
as done by the Cochrane Neonatal Review Group. No language restrictions were
applied. Published or un-published data were considered.
Criteria and methods used to assess the methodological quality of the trials: standard methods of the Cochrane Collaboration and its Neonatal Review Group were used. At least two of the reviewers worked independently to search for and assess trials for inclusion and methodological quality. Studies were assessed using the following key criteria: allocation concealment (blinding of randomisation), blinding of intervention, completeness of follow up and blinding of outcome measurement. The reviewers extracted data independently. Differences were resolved by discussion. An attempt was made to contact study investigators for additional information or data.
Data analysis:
For individual trials: for continuous variables such as duration of cannula patency, mean differences, and 95% confidence intervals were to be reported. For categorical outcomes such as mortality, the relative risks (RR) and 95% confidence intervals will be reported.
For pooled results: for continuous variables, weighted mean differences (WMD) and 95% confidence intervals will be reported. For categorical outcomes, the relative risks (RR) and 95% confidence intervals will be reported. For significant findings, the risk difference (RD) and number needed to treat (NNT) will also to be reported. Each treatment effect will be tested for heterogeneity to help determine suitability for pooling of results in a meta-analysis. The fixed effects model will be used for meta-analysis.
Two studies were identified for inclusion in this review (Kalyn 2000; Taylor 1989). Methods and clinical details including participants, interventions and outcomes are given in Table 1.
Included studies
Kalyn et al (Kalyn 2000) performed a multi-centre
randomised controlled trial. Infants were randomised either to a continuous
infusion of 0.5-1 mL/hr of 10% Dextrose or intermittent flushes with 0.5-1
mL every 6 hrs of 0.9% non heparinised saline. The infants were randomly
assigned to their groups using a computer generated random number. Blinding
of the interventions was obviously not done. Infants receiving intermittent
drug therapies via a peripheral intravenous catheter were eligible for inclusion
in the study but infants receiving continuous drug infusions or supplemental
fluids were not. A total of 95 neonates (42 to intermittent flush group and
53 to continuous infusion group) were cannulated with 238 catheters (84 to
intermittent flush group and 154 to the continuous infusion group) during
the study period. The primary diagnosis for all the infants was sepsis. Follow-up
continued until the intravenous catheter was either
removed because intravenous medication was stopped or the catheter was no
longer working; the infant was transferred elsewhere; or the infant required
maintenance intravenous fluids. The two groups were all well matched for
demographic and clinical variables, including birth weight, gestational age,
the site and size of the catheters and types of drugs being administered.
The duration of patency of each catheter was recorded as well as the reasons
for loss of patency: phlebitis, occlusion, leaking and infiltration. The
outcomes were reported per catheter used.
Taylor et al (Taylor 1989) did a randomised controlled trial in infants who were admitted to the 'intermediate care' nursery who either:
1. required intravenous medications but no additional intravenous fluids, or
2. had an umbilical arterial catheter in situ and required an intravenous cannula for medications.
The infants were randomised either to a continuous infusion group where
the intravenous line was kept patent by a continuous infusion of 10% dextrose
(without heparin) at a rate of 1.5 to 3.0 ml/hr; or intermittent flushes
where the intravenous cannula was kept patent with a heparin lock (0.5 ml
of heparinised saline) given every six hours or after injection of medications.
The method of randomisation was not stated. Blinding of the interventions
was obviously not done. One infant in the intermittent flush group was excluded
post randomisation because the "... catheter was needed only briefly, not
allowing for time for data collection". Data were collected on a total of
39 neonates (22 to intermittent flush group and 17 to continuous infusion
group) who were cannulated with an unknown number of catheters. The infants
were enrolled in the study until they no longer required intravenous access.
The two groups were all well matched for demographic and clinical variables,
including birth weight and gestational age. Outcome measures included the
number of days the infant was enrolled in the study; quantitation of parental
medications, blood products and intravenous fluids; number of line infiltrations;
duration of cannula patency (this seemed to have been measured for each cannula
regardless of the number used in each infant); and the number of times the
infant was removed from the incubator to be held by a nurse or parent.
Excluded Studies
A number of studies were found which examined the duration of peripheral
intravenous catheter patency with or without heparin intermittent flushes
(Danek 1992; Hanrahan 2000; LeDuc 1997; McMullen 1993; Nelson 1998).
However, none compared intermittent flushes with continuous infusions and
were therefore not included in the review. This is the subject of another
review by Shah et al (Shah 2004).
Kalyn's study (Kalyn 2000): the randomisation method was by a computer generated random number to assign infants by alternate sequential series; allocation was performed in a blind manner by using sealed opaque envelopes to assign each infant to either group; blinding of intervention was not done; follow-up was complete and outcome assessments were not blinded.
Taylor's study (Taylor 1989):
the randomisation method was not stated; allocation was performed in a blind
manner by using sealed opaque envelopes to assign each infant to either group;
blinding of intervention was not done; follow-up was complete and outcome
assessments were not blinded.
Two studies (Kalyn 2000; Taylor 1989) were eligible for inclusion.
Kalyn 2000
The lead author of Kalyn et al (Kalyn 2000)
was contacted for further information regarding this study: additional data
were available on the duration of patency for the first catheter used per
infant.
Primary outcomes
A significant difference (Chi squared test P < 0.001) was found when the two catheter groups were compared with respect to the removal or loss of patency of catheters. The results were presented by catheter rather than by infant although it was stated that results were also analysed by infant and similar results were obtained. The catheters in the intermittent flush group were less likely to infiltrate, leak or cause phlebitis (35.7%) than in the continuous infusion group. The continuous infusion group, however, were less likely to occlude (9.1%) than the intermittent flush group (25%). In the intermittent flush group 39% of catheters (compared with 24% in the continuous infusion group) did not have infiltration/phlebitis/leaking/occlusion because the catheter was either removed because intravenous medication was stopped, the infant was transferred elsewhere or the infant required maintenance intravenous fluids.
Taylor 1989
The lead author of Taylor et al (Taylor 1989)
was contacted for further information regarding this study: additional data
were available on post-randomisation exclusions, the number of cannulas used
during the first 48 hours in individual infants, and loss of cannula function.
Primary outcomes
Secondary outcomes
Even though one study showed no difference between the approaches for maintaining cannula patency and one showed an advantage for intermittent flushes, it is unfortunate that the way the data were analysed and reported in the two included studies makes the reliability of the results uncertain.
The reports (for both included studies) only included as their main outcome the average duration of cannula patency for all cannulas. That is, they reported duration of cannula patency for all cannulas in all infants in each treatment group averaged over the treatment course. Some infants would have had multiple cannulas and, therefore, the mean duration of cannula patency would include multiple measures of the duration of cannula patency in some infants with non-independence of those multiple measures. This would invalidate the assumptions necessary to ensure the reliability of the statistical tests used.
In the study by Kalyn et al (Kalyn 2000) the number of catheters per number of neonates was difficult to interpret and additional unpublished data are not available to clarify this. The reasons for re-cannulation of individual infants were not recorded in the study. However, the additional information that was provided by the study investigators for duration of patency for the first catheter used per infant did not show any significant difference between groups for this outcome.
Taylor et al found a longer duration of cannula patency (averaged over all catheters for that infant) with intermittent flushing and greater ease of handling of those infants. They also found that the mean number of cannulas used per infant in the first 48 hours was less for intermittent flushing. We had speculated that a baby requiring intermittent flushing of their intravenous line is far more portable than one with a continuous infusion line. We believed that this would make handling, feeding and caring for baby easier for mothers in neonatal nurseries and enable staff the opportunity to enhance maternal-infant bonding opportunities. These outcomes were assessed by Taylor et al (Taylor 1989) using a subjective nursing score (unblinded assessment) and they did find that the infants having intermittent flushes were easier to care for.
It should be noted that Taylor et al's study compared intermittent
flushing of the cannula with heparinised saline with a continuous infusion
that did not contain heparin. The use of heparinised saline for the intermittent
flush used to maintain cannula patency has been reviewed by Shah et al (Shah 2004).
Their systematic review found five studies that compared intermittent flush
with heparin and intermittent flush without heparin and reported the outcome
of duration of cannula patency for the first cannula used per infant. There
was no consistency of results for this outcome with two studies showing longer
duration with heparin, one study showing longer duration without heparin
and two studies showing no difference.
Nil
Study | Methods | Participants | Interventions | Outcomes | Notes | Allocation concealment |
Kalyn 2000 | Multi-centre randomised controlled trial with treatment group allocation determined by computer generated randomisation. Assessment of the following key criteria: | Neonates in the neonatal nursery who do not require IV fluids. Number of infants = 95 Number of catheters = 238 | Continuous infusion (CI) using 0.5-1 mL of 10% Dextrose; or intermittent flushing (IF) using 0.5-1 mL 0.9% normal saline flushed before and after every medication and every 6 hours. | Proportion of catheters with: 1. infiltration, phlebitis or leaking; 2. occlusion; 3. either removal because intravenous medication was stopped, the infant was transferred elsewhere or the infant required maintenance intravenous fluids. | The infants were randomised and allocated to either group, but the data were analysed and reported by catheter. The lead author of this study was contacted for further information regarding this study: additional data were available on the duration of patency for the first catheter used per infant. | A |
Taylor 1989 | Randomised controlled trial. The randomisation method was not stated. Assessment of the following key criteria: | Newborn
infants who were admitted to the 'intermediate care' nursery who either 1.
required intravenous medications but no additional intravenous fluids, or
2. had an umbilical arterial catheter in situ and required an intravenous
cannula for medications. Number of infants = 39 | Continuous infusion (CI) using 10% dextrose (without heparin) at a rate of 1.5 to 3.0 ml/hr; or intermittent flushing (IF) using a heparin lock (0.5 ml of heparinised saline) given every 6 hours or after injection of medications. | Outcome measures included: the number of days the infant was enrolled in the study; quantitation of parental medications, blood products and intravenous fluids; number of line infiltrations; duration of cannula patency (this seemed to have been measured for each cannula regardless of the number used in each infant); and the number of times the infant was removed from the incubator to be held by a nurse or parent. | A |
Study | Reason for exclusion |
Danek 1992 | Did not compare continuous infusion with intermittent flushes. |
Hanrahan 2000 | Historical cohort comparison only; did not compare continuous infusion with intermittent flushes. |
LeDuc 1997 | Did not compare continuous infusion with intermittent flushes. |
McMullen 1993 | Did not compare continuous infusion with intermittent flushes. |
Nelson 1998 | Did not compare continuous infusion with intermittent flushes. |
Kalyn A, Blatz S, J Pinelli. A comparison of continuous infusion and intermittent flushing methods in peripheral intravenous catheters in neonates. Journal of IV Nursing 2000;23:146-53.
Taylor 1989 {published data only}
Taylor J, Shannon R, Kilbride HW. Heparin lock intravenous line: use in newborn infants. Clinical Pediatrics 1989;28:237-40.
Danek GD, Noris EM. Pediatric i.v. catheters: efficacy of saline flush. Pediatric Nursing 1992;18:111-3.
Hanrahan 2000 {published data only}
Hanrahan KS, Kleiber C, Berends S. Saline for peripheral intravenous locks in neonates: evaluating change in practice. Neonatal Network 2000;19:19-24.
LeDuc 1997 {published data only}
LeDuc K. Efficacy of normal saline solution versus heparin solution for maintaining patency of peripheral intravenous catheters in children. Journal of Emergency Nursing 1997;23:306-9.
McMullen 1993 {published data only}
McMullen A, Fioravanti ID, Pollack V, Rideout K, Sciera M. Heparinized saline or normal saline as a flush solution in intermittent intravenous lines in infants and children. MCN. The American Journal of Maternal Child Nursing 1993;18:78-85.
Nelson 1998 {published data only}
Nelson JJ, Graves SM. 0.9% Sodium chloride injection with and without heparin for maintaining peripheral indwelling intermittent-infusion devices in infants. American Journal of Health-System Pharmacy 1998;55:570-3.
* indicates the primary reference for the study
Cotton MC, Turner BS, Miller-Bell M. Pharmacology in neonatal care. In: Merenstein GB, Gardener SL, editor(s). Handbook of Intensive Care. 5th edition. St Louis: Mosby, 1998:148-62.
Danek GD, Noris EM. Pediatric i.v. catheters: efficacy of saline flush. Pediatric Nursing 1992;18:111-3.
Dickersin K, Scherer R, Lefebvre C. Identifying relevant studies for systematic reviews. British Medical Journal 1994;309:1286-91.
Malcom IL, Tudehope DI, Thearle MJ. Parent-infant attachment and support for parents experiencing perinatal loss. In: Malcom IL, Tudehope DI, Thearle MJ, editor(s). Essentials of neonatal medicine. 3rd edition. Oxford: Blackwell Science, 2000:297-303.
Moclair A, Bates I. The efficacy of heparin in maintaining peripheral infusions in neonates. European Journal of Pediatrics 1995;154:567-70.
Olds SB, London ML, Wieland Ladewig PA. The newborn at risk: conditions present at birth. In: Olds SB, London ML, Wieland Ladewig PA, editor(s). Maternal-newborn nursing: a family and community-based approach. 6th edition. New Jersey: Prentice Hall Health, 2000:804-903.
Shah PS, Ng E, Sinha AK. Heparin for prolonging peripheral intravenous catheter use in neonates. In: The Cochrane Database of Systematic Reviews, Issue 3, 2004.
Ward R M. Neonatal pharmocology. In: Kenner C, Brueggemeyer A, Gunderson LP, editor(s). Comprehensive neonatal nursing. 1st edition. Philadelphia: W.B.Saunders Company, 1993:926-939.
Anonymous. Evidence for the ten steps to successful breastfeeding. Revised edition. Geneva: World Health Organisation, 1998.
Yeo H. Stress in the neonatal unit. In: Yeo H, editor(s). Nursing the Neonate. Oxford: Blackwell Science, 1998:278-92.
Comparison or outcome | Studies | Participants | Statistical method | Effect size |
---|---|---|---|---|
01 Continuous infusion versus intermittent flushing | ||||
01 Duration of cannula patency for the first cannula used per infant (hours) | WMD (fixed), 95% CI | No total | ||
02 Number of cannulas used per infant in the first 48 hours | WMD (fixed), 95% CI | No total |
Ms Anndrea Flint
Clinical Nurse Consultant
Grantley Stable Neonatal Unit
Royal Women's Hospital
Butterfield St
Herston
Brisbane
Queensland AUSTRALIA
4029
Telephone 1: +61 7 3636 8918
Facsimile: +61 7 3636 5259
The review is published as a Cochrane review in The
Cochrane Library, Issue 4, 2005 (see http://www.thecochranelibrary.com for
information). Cochrane reviews are regularly updated as new evidence emerges
and in response to comments and criticisms, and The Cochrane Library should
be consulted for the most recent version of the Review. |