Dr. E Ng
Literature search and identification of trials
Evaluation of methodological quality of trials
Data collection
Verification of data
Revision of the review
Dr. A Sinha
Literature search and identification of trials
Evaluation of methodological quality of trials
Data collection
Verification of data
Revision of the review
There were no changes to the conclusions of this review.
Heparin in intravenous fluids may reduce IV tube changes needed for newborn babies in neonatal intensive care, but more research is needed to determine its safety.
Babies in neonatal intensive care often need fluids intravenously (through a tube inserted into a vein). Sometimes the intravenous (IV) tube becomes blocked, as the blood clots and skin becomes swollen. Bacteria can also enter and cause serious infection. Regularly changing the tube (and which vein is used) can reduce some problems, but babies have few usable veins. The drug heparin used in the IV fluids could reduce blockages by thinning the blood, but it can have serious adverse effects. The review of trials found that more research is needed to determine whether heparin in IV fluids is advantageous for neonates without causing side effects.
Peripheral intravenous (PIV) catheters are widely used in modern medical practice. However, mechanical or infectious complications often necessitate their removal and/or replacement. Heparin has been shown to be effective in prolonging the patency of peripheral arterial catheters and central venous catheters, but may result in life threatening complications, especially in preterm neonates.
The primary objective was to determine the effectiveness of heparin versus placebo or no treatment on duration of PIV catheter patency, defined as number of hours of catheter use. The secondary objectives were to assess the effects of heparin on catheter blockage, phlebitis or thrombophlebitis, catheter related sepsis, and complications including abnormality of coagulation profile, allergic reactions to heparin, heparin induced thrombocytopenia, intraventricular/intracranial hemorrhage and mortality.
A literature search was performed using the following databases: MEDLINE (1966-February 2005), EMBASE (1980-February 2005), CINAHL (1982-February 2005), Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 1, 2005), and abstracts from the annual meetings of the Society for Pediatric Research, American Pediatric Society and Pediatric Academic Societies published in Pediatric Research (1991-2004). No language restrictions were applied.
Randomized or quasi-randomized trials of heparin administered as flush or infusion versus placebo or no treatment were included. Studies which included a neonatal population and reported on at least one of the outcomes were included.
The methodological quality of the studies was assessed using criteria for blinding of randomization, blinding of intervention, completeness of follow-up and blinding of outcome assessment. Data on relevant outcomes were extracted and the effect size was estimated by calculating WMD (weighted mean difference, 95%CI), RR (relative risk, 95% CI) and RD (risk difference, 95% CI).
Ten eligible studies were identified. Heparin was administered either as a flush solution, or as an additive to the total parenteral nutrition solution. Five studies reported data on the duration of use of the first catheter. Two of these studies found no statistically significant effect of heparin; two studies showed a statistically significant increase and one study showed a statistically significant decrease in the duration of PIV catheter use in the heparin group. The results were not combined for meta-analysis due to significant heterogeneity of the treatment effect (p < 0.01). In addition, there were marked differences between the studies in terms of the methodological quality, the dose, the timing, the route of administration of heparin and the outcomes reported. From a limited number of studies, there were no significant differences between the heparin and the placebo/no treatment groups in the risks of infiltration, phlebitis and intracranial hemorrhage.
Implications for practice: The effect of heparin on the duration of peripheral intravenous catheter use varied across the studies. Because of clinical heterogeneity and heterogeneity in treatment effect, recommendations for heparin use in neonates with PIV catheters cannot be made.
Implications for research: There are insufficient data concerning the effect of heparin for prolonging PIV catheter use in neonates. Further research on the effectiveness, the optimal dose, and the safety of heparin is required.
Intravenous infusion of drugs, fluids and nutrients has become an indispensable practice in present day medical care (Maki 1977; Lewis 1985; Tager 1983). Venous cannulation via peripheral intravenous (PIV) catheters is the simplest and most frequently used method for administration of an infusion. It is associated with inherent complications (Graham 1991) which can be mechanical or infectious (Lewis 1985; Tager 1983; Bossert 1994). Mechanical complications include thrombosis, dislodgement, extravasation, leakage, phlebitis and scar formation (Nieto-Rodrig 1992; Tomford 1984). Infectious complications include bacterial or fungal sepsis (Graham 1991). Thrombosis or phlebitis at the catheter site can act as a nidus for infection. Factors associated with the risk of these complications include duration of infusion, site of infusion, size and type of catheter, rate of flow through catheter, turbulence of fluid flow and characteristics of patients and infusate (Lewis 1985; Krafte-Jacobs 1995).
There have been various attempts to prevent or reduce PIV catheter related complications (Hecker 1992). These include designated intravenous therapy teams (Tomford 1984), frequent change in the site of catheters, placement of in-line filters (Roberts 1994), topical use of glyceryl trinitrate (Tighe 1995), hydrocortisone (Tighe 1995) or heparin (Vilardell 1999), intermittent heparinized-saline flush or lock and continuous infusion of heparin (Tanner 1980; Hanson 1976; Daniell 1973; Johnstone 1991).
In a neonate, frequent change in catheter site at predetermined timings is technically challenging. In addition, repeated attempts at cannulation break the skin barrier and predispose these patients to infection by commensals in the skin such as coagulase negative staphylococcus. Such infections can be life threatening in neonates. PIV catheters are often left in situ in neonates until complications arise (Moclair 1995). Therefore, methods that can prolong the duration of viability of these catheters may be beneficial in this population.
Heparin, an anticoagulant, has been administered as intermittent injection or continuous infusion to prevent thrombus formation and consequently prolong catheter patency (Moclair 1995; Bossert 1994). However, significant effects of heparin on coagulation profiles have been observed (O'Neill 1974), even at low dose. Other adverse effects associated with heparin use include allergic reactions, bleeding complications due to dosing error, intraventricular hemorrhage in preterm infants (Malloy 1995; Lesko 1986) and heparin induced thrombocytopenia (Potter 1992).
The effectiveness of heparin for prolongation of PIV catheter life has been systemically reviewed previously. Goode 1991, Peterson 1991 and Randolph 1998 compared the efficacy of heparin versus normal saline in patients of all age groups. These reviews concluded that normal saline was as efficacious in prolonging catheter patency as heparin. However, only one study from these reviews studied neonates, so that the efficacy and safety of heparin was not ascertained specifically in neonates. Neonates are unique in their sensitivity and resistance to heparin (Vieira 1991) and in their higher propensity to develop intracranial hemorrhage (Lesko 1986).
The primary objective of this review was to determine the effectiveness of heparin administered as continuous infusion or intermittent injections, via PIV catheter, versus placebo or no treatment, on duration of catheter patency (defined as number of hours of catheter use) in neonates.
Secondary objectives were to determine the effects of heparin on:
A. Complications relating to PIV catheters:
1. Catheter blockage (defined as inability to infuse fluid or medication)
2. Phlebitis or thrombophlebitis developing at the site of or above the catheter (defined as pain, swelling, erythema or induration at the site of catheter with or without a palpable venous cord above the catheter site)
3. Catheter related sepsis (defined as signs and symptoms suggestive of sepsis with positive blood culture growing the same organism in samples obtained from venipuncture at a different sterile site and from the catheter or catheter tip)
4. Number of additional PIV catheter insertions
B. Complications relating to heparin:
1. Abnormality of coagulation profile
2. Incidence of allergic reactions to heparin
3. Heparin induced thrombocytopenia (development of thrombocytopenia {platelet count < 150,000 per cubic litre} after starting heparin in an infant with previously normal platelet count after exclusion of all other causes of thrombocytopenia and positive serotonin release test {Warkentin 1990})
4. Intraventricular/intracranial hemorrhage (development of recent onset of hemorrhage or extension of preexisting hemorrhage after starting heparin, according to the classification of Papile 1978)
C. Mortality
Subgroup analyses were planned a priori according to the method of heparin administration (continuous versus intermittent), and gestational age [preterm (less than 37 weeks gestation) versus term (more than or equal to 37 weeks gestation)]. A comparison of effects of different heparin doses was also planned if sufficient numbers of studies or trials using different dosage regimens were identified.
Randomized and quasi-randomized controlled trials in which heparin administration was compared to placebo or no treatment for the prevention of thrombosis or occlusion of PIV catheter.
Term or preterm infants who required PIV catheter as determined by the attending physicians during their stay in neonatal intensive care unit.
Heparin by infusion or intermittent injections via PIV catheter, versus placebo or no treatment. Heparin or placebo must have been administered during the entire duration of the catheter being in situ.
Studies that report on one or more of the following outcomes:
Primary outcome: Number of hours of catheter use, measured as time until catheter was removed.
Secondary outcomes:
1. Complications associated with PIV catheters including:
occlusion of the catheter (identified by inability to infuse fluids)
incidence of phlebitis or thrombophlebitis (one episode per patient)
catheter related sepsis (one episode per patient)
number of additional PIV catheters needed
2. Complications related to heparin
abnormal coagulation profile
allergic reactions
heparin induced thrombocytopenia
intracranial or intraventricular hemorrhage
3. Neonatal mortality
In this review, only data from the first catheter, retrieved either from the published studies or directly from the study authors, were used in reporting outcomes concerning the efficacy of heparin.
See: Cochrane Neonatal Collaborative Group search strategy
MEDLINE was searched (1966 to February 2005) using MeSH terms: heparin, infant, newborn, peripheral intravenous catheter, silastic catheter, indwelling catheter.
Other databases that were searched included: EMBASE (1980 to February 2005); CINAHL (1982 to February 2005); the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 1, 2005) and the reference lists of identified trials and abstracts from the annual meetings of the Society of Pediatric Research, American Pediatric Society and Pediatric Academic Societies published in Pediatric Research (1991-2004). No language restrictions were applied.
The following types of articles were excluded: letters, editorials/commentaries, reviews, lectures.
Standardized methods of the Cochrane Neonatal Collaborative Review Group were used to assess methodological quality of studies.
All published articles identified as potentially relevant by the literature search were assessed for inclusion in the review.
Quality of included trials was evaluated independently by all the reviewers, using the following criteria:
1. Masking of randomization
2. Masking of intervention
3. Complete follow-up
4. Masking of outcome assessment.
We attempted to obtain data from the primary author where published data provided inadequate information for the review, or the study population included all age groups and neonatal data could not be abstracted. No studies were identified which were reported only as abstracts. Retrieved articles were assessed and data were abstracted independently by all the reviewers. Discrepancies between reviewers were resolved by consensus.
Statistical methods included relative risk (RR), risk difference (RD), number needed to treat (NNT) and weighted mean difference (WMD) as appropriate. Ninety five percent confidence intervals were used for these estimates of treatment effects. A fixed effect model was used for meta-analyses.
Because a catheter may be changed owing to blockage, because of routine practice to avoid blockage, or because it is no longer needed, analysis of number of catheters used may not reflect an effect due to heparin use. Accordingly, we developed three potential approaches to statistical analysis. 1) We attempted to assess duration of the first catheter used by each patient and the reasons for discontinuation. Duration of patency before occlusion is a valid measure of effect. These data can be analyzed using mean difference in duration between the treated and untreated patients, or if hazard ratios are reported in the individual trials, these can be analyzed using the relative risk statistical procedure for meta-analysis. 2) If there are several catheters per patient, with different durations of patency for each and they are known to have become occluded, the mean duration of all catheters for each patient were to be obtained and analyzed as a mean of means. This procedure has the advantage of using more of each patient's data than simply relying on duration of the first catheter. 3) If trials report the number of occluded catheters for each patient, excluding those removed which were unoccluded, these data were to be analyzed using the mean number of catheters used.
Ten studies were included in this review (Alpan 1984; Golberg 1999; Heilskov 1998; Klenner 2003; Kotter 1996; Moclair 1995; Mudge 1998; Paisley 1997; Schultz 2002; Treas 1992). Clinical details concerning the participants, interventions and outcomes are given in the table, Characteristics of Included Studies.
Alpan 1984 performed a randomized controlled trial involving addition of heparin to total parenteral nutrition (TPN) solution in preterm infants. Preterm infants needing TPN were eligible for inclusion in the study. The treatment group received 1 IU/ml of heparin added to their TPN solution and the control group received no heparin in their TPN solution. All patients were cannulated using 22-gauge catheters. A total of 26 patients were enrolled, 13 infants in each group. These infants received 227 PIV catheters during the study period (105 in the heparin group and 122 in the control group) and all the catheters were managed according to the originally allocated policy. Baseline patient characteristics did not differ statistically significantly between the two groups. The duration of each catheter in situ was recorded. In addition, the incidences of phlebitis, local inflammation, intraventricular hemorrhage and septicemia were reported. Only the incidence of intraventricular hemorrhage was reported per patient and was used for this review; all other outcomes were reported per PIV catheter.
Golberg 1999 performed a randomized controlled trial comparing intermittent flush of heparinized normal saline (4 IU/ml) with normal saline via PIV catheters in neonates. A total of 60 neonates were randomized; however, 13 infants were removed from the analysis because of possible contamination. Of the remaining 47 neonates 23 were randomized to the heparinized normal saline and 24 to normal saline group. Baseline characteristics did not differ between the two groups. All patients received 24-gauge intravenous catheters. The duration of catheter patency and the incidence of phlebitis were the only outcomes used for this review.
Heilskov 1998 performed a randomized double-blind controlled 3-arm trial in which neonates requiring PIV catheters were enrolled. Neonates who were receiving anticoagulant medication via another site were excluded. Ninety neonates were randomized to receive intermittent flush of low dose heparinized saline at 2 IU/ml (n = 28), high dose heparinized saline at 10 IU/ml (n = 35) or normal saline (n = 27) via PIV catheters. Only one catheter per patient was included in the study. Baseline patient characteristics were not statistically significantly different among groups. The duration of catheter patency and the incidence of PIV infiltration were reported in the study and were used for the purpose of this review.
Klenner 2003 performed a randomized controlled trial involving addition of heparin to intravenous infusion fluids in neonates who were anticipated to need intravenous infusion for > 5 days. Neonates > 1 kg birth weight were eligible for inclusion in the study. The treatment group received 0.5 IU/ml of heparin added to their intravenous infusion solution and the control group received no heparin in their intravenous infusion solution. Patients were cannulated using 24 or 26 gauge catheters. A total of 296 patients were enrolled, 145 infants in heparin group and 151 patients in control group. These infants received 1257 PIV catheters during the study period (565 in the heparin group and 692 in the control group) and all the catheters were monitored carefully. The catheters were removed if there were signs of infection, phlebitis or extravasation. Baseline patient characteristics did not differ statistically significantly between the two groups. The duration of each catheter in situ was recorded. In addition, the incidences of phlebitis, infiltration/extravasation, thrombus formation and accidental dislodgement were reported. Patients were evaluated for heparin induced thrombocytopenia antibody assay at 1, 7, 14, 21 and 28 days (108 patients from heparin group and 105 patients from the control group) and for intraventricular hemorrhage at 1, 3, 7, 14, 28 and 42 days age if they were still in hospital. Outcomes were reported as per catheter. Additional outcomes on first catheter not reported were obtained from the study author.
Kotter 1996 performed a randomized study comparing intermittent flush of heparinized saline (10 IU/ml) and normal saline via PIV catheters in neonates. Eighty-four neonates were enrolled, and data were collected for a total of 227 catheters. However, 109 of these catheters were excluded from the analysis because they were used for continuous infusion for a period of time before converting to intermittent use (heparin was not used for continuous infusion, so data from these catheters cannot be used for analysis), or were removed while patent. The authors reported data on the remaining 118 catheters from 51 patients. Each catheter was managed according to the patient's original group allocation. The primary outcome reported was duration (hours) of catheter patency. Other outcomes reported were incidence of catheter occlusion, phlebitis and leakage. All outcomes were reported per catheter and thus were not used for this review.
Moclair 1995 performed a randomized dose comparison study of heparin in neonates comparing the addition of various doses of heparin or no heparin to peripheral TPN solution. Patients receiving a central venous catheter during the study period were withdrawn from the study (the actual number not reported in the study). Ninety-seven neonates were randomized, but seven patients were excluded from analysis due to death or transfer to another hospital. The remaining 90 neonates were randomly allocated to one of five groups; 0.1 IU/ml heparin group (n = 15 with 35 PIV catheters), 0.25 IU/ml heparin group (n = 16 with 47 PIV catheters), 0.5 IU/ml heparin group (n = 23 with 61 PIV catheters), 1 IU/ml heparin group (n = 16 with 30 PIV catheters) and a heparin free control group (n = 20 with 72 PIV catheters). Subsequent catheters in the neonates were managed according to the original allocation. All neonates received 24-gauge catheters during the study. Baseline patient characteristics were not reported. Primary outcome was survival of PIV catheters, as indicated by number of PIV catheters required per neonate, duration of patency (hours) of each catheter, and the "survival curve" for catheter patency for each group. The duration of catheter patency was reported per number of catheters, and the authors were not able to provide the data on the first catheters only, so none of the outcome data were used for this review.
Mudge 1998 performed a quasi-randomized blinded trial comparing intermittent flush of heparinized saline (10 IU/ml) and normal saline via PIV catheters. Both neonatal and pediatric patients were included. A total of 61 patients with 134 PIV catheters were enrolled. Fifty-six catheters were flushed with heparinized saline and 78 with saline. The frequency of flushing was not controlled. The primary outcome was catheter duration, defined as the time from catheter insertion or conversion from continuous to intermittent infusion to catheter removal or conversion from intermittent to continuous infusion. Other outcomes reported were incidence of catheter-related complications (erythema, edema, occlusion, leakage, and discomfort). The author when contacted provided data on the 15 neonates in the heparin group and the 25 neonates in the saline group. Duration of catheter patency for the first catheter, incidence of catheter occlusion and incidence of phlebitis were used for this review.
Paisley 1997 performed a quasi-randomized study comparing heparinized saline flushes (0.6 ml of heparin solution containing 10 IU/ml of heparin) and 0.6 ml of normal saline via PIV catheters. Infants 32 weeks gestation or greater were eligible. Thirty-three infants received heparin and 54 patients received saline flushes. A total of 159 PIV catheters were studied. The data were reported per first PIV catheter and per all PIV catheters. Subsequent catheters in the same patient were managed according to the patient's original group allocation. The primary outcome was duration of catheter patency. Other outcomes reported were heparin induced thrombocytopenia, catheter blockade, phlebitis, infiltration and leaking. Duration of catheter patency and incidence of heparin induced thrombocytopenia were used for this review.
Schultz 2002 performed a randomized controlled trial involving heparinized flushes vs normal saline flushes. Neonates <30 days of age were eligible for inclusion in the study. Twenty patients received 0.5 ml of heparinized saline (2 IU/ml) flushes and twenty nine patients received 0.5 ml of normal saline flushes. All patients were cannulated using 24-gauge catheters. Only one catheter per patient were included in the study. Baseline patient characteristics did not differ statistically significantly between the two groups. The duration of each catheter in situ was recorded. In addition, the incidences of occlusion, phlebitis (redness and edema) and extravasation were reported and were used for this review.
Treas 1992 performed a quasi-randomized study comparing heparinized saline (0.5 IU/ml) and normal saline via PIV catheters in neonates. All neonates were cannulated using size 24-gauge catheters. Only the first three catheters used in a neonate were included in the study. In the treatment group, heparin was added to a continuous infusion or as intermittent flush solution. In the control group, saline was used as intermittent flush solution. Sixty-three neonates with 131 PIV catheters were in the heparin group, and 49 neonates with 122 PIV catheters were in the control group. Subsequent catheters in the same patient were managed according to the patient's original group allocation. Baseline characteristics were not different between the two groups. The primary outcome was duration of catheter patency. Other outcomes reported were incidence of phlebitis, catheter-related infection, and bleeding. Outcomes were reported per catheter. The author (personal communication) provided data on the duration of patency for the first catheter and this was the only outcome used for this review.
For each study included in the review, assessments of methodological quality are given in the table, Characteristics of Included Studies.
The methodological details for studies were extracted from the published information and personal contact with the authors.
Alpan 1984
Randomization was performed by a table of random numbers and was blinded.
Masking of intervention and outcome measures was ensured. Outcomes were reported
on all catheters and all neonates enrolled in the study.
Golberg 1999
Randomization was performed centrally in the pharmacy. The study medication
was centrally supplied as premixed vials, and the vial stoppers were all
punctured to ensure that the intervention was masked. Outcomes were reported
on 47 of the 60 neonates enrolled.
Heilskov 1998
Randomization was performed using a random allocation list. The study
solutions were coded to ensure masking of the intervention. Outcomes were
reported on all catheters and all neonates enrolled in the study.
Klenner 2003
Randomization was performed centrally in the pharmacy. The study solution
was centrally supplied in premixed form to ensure masking of intervention.
Outcomes were reported as per catheter, however, data on first catheter were
obtained from the author.
Kotter 1996
Randomization was performed by random selection of sealed envelopes.
Color-coded cards and data sheets were used to separate the two groups; however,
the treatment allocation was concealed, thus ensuring masking of intervention
and outcome measures. Outcomes were reported on 51 (118 catheters) of 84
(227 catheters) neonates enrolled.
Moclair 1995
Randomization was performed centrally by allocation of codes in the pharmacy.
TPN solutions were prepared with heparin added according to the code assigned.
Masking of intervention and outcome measures were ascertained. Outcomes
were reported on 90 of 97 neonates who completed the study.
Mudge 1998
Allocation was based on the month of entry. All patients enrolled in
a given month received the same flush solution (labelled solution A or B).
Masking of intervention and outcome measures were ascertained. Outcomes were
reported on all catheters of all patients studied.
Paisley 1997
Allocation was based on last digit of infant's hospital number. Masking
of allocation or outcome assessment was not performed. Outcomes were reported
separately on the first catheters and all catheters.
Schultz 2002
Randomization was performed centrally in the pharmacy. The study medication
was centrally supplied as premixed vials to ensure that the intervention
was masked. Outcomes were reported on all neonates enrolled. Only one catheter
per patient was included in the study.
Treas 1992
Allocation was based on sequential hospital number given by the admitting
office. Masking of intervention and outcome measures were not performed (personal
communication). Outcomes were reported on all enrolled neonates and their
first three catheters.
In many studies included in this review, the outcomes were reported per catheter rather than per patient. However, most patients received more than one PIV catheter during the study period. Attempts were made to contact the primary authors in order to obtain data of the first catheter of each patient. Where such data were not available only the long-term outcomes such as intracranial hemorrhage were included in the review.
After reviewing the studies, it was observed that a number of catheters were removed electively. An attempt was made to obtain individual patient data which was not successful.
Primary outcome: Number of hours of use of first catheter
Seven studies (Heilskov 1998; Paisley 1997; Mudge 1998; Treas 1992; Golberg 1999; Klenner 2003; Schultz 2002) reported on duration of use for the first catheter in each patient. Three studies found no statistically significant difference, three studies showed a statistically significant increase and one study showed a statistically significant decrease in the duration of catheter use in the heparin group. Heilskov 1998 found that the duration of use was not statistically different between the groups [mean (SD) catheter duration was 66.1 (32.1) hours in the heparin group vs 67.1 (33.1) hours in the control group]; Paisley 1997 also found that there was no statistically significant difference between the two groups [mean (SD) catheter duration was 66.2 (42.8) hours in the heparin group vs 76.4 (125.7) hours in the control group]. Mudge 1998 found that the duration of catheter use was statistically significantly increased in the heparin group [mean (SD) catheter duration was 50.8 (16.2) hours in the heparin group vs 38.0 (21.1) hours in the control group]; Treas 1992 also found that there was a statistically significant increase in the duration of catheter use in the treatment group [mean (SD) catheter duration was 62.8 (29.9) hours in the heparin group vs 27.3 (15.3) hours in the control group]. Golberg 1999 found that the duration of catheter use was statistically significantly shorter in the heparin group [mean (SD) catheter duration was 25.6 (22.7) hours in heparin group vs 56.5 (46.1) hours in control group]. Klenner 2003 found that the duration of catheter use was statistically significantly increased in the heparin group [mean (SD) catheter duration was 46.9 (27.9) hours in the heparin group vs 35.0 (17.8) hours in the control group]. Schultz 2002 found that the duration of use was not statistically different between the groups [mean (SD) catheter duration was 38.5 (33.3) hours in the heparin group vs 34.4 (27.3) hours in the control group]. However, there was strong evidence of heterogeneity for the treatment effect (p < 0.001) across these seven studies. In addition, there were important differences between the studies in terms of dosages of heparin, concentrations of heparin, timings of flushing of catheters, size of the catheters and methodology of the studies. After reviewing these differences among the studies, a decision was made not to report a summary measure of treatment effect.
Secondary outcomes:
1. Complications associated with PIV catheters:
Occlusion of the catheter (identified by inability to infuse fluids):
Four studies (Heilskov 1998; Mudge 1998; Klenner 2003; Schultz 2002) reported on this outcome for the first catheter placement. There was no statistically significant difference in the rate of catheter occlusion and subsequent infiltration (pooled RR 1.00 [95% CI 0.85, 1.16] and RD 0.00 [95% CI -0.09, 0.09]. There was evidence of heterogeneity for the treatment effect; p=0.01).
Incidence of phlebitis or thrombophlebitis:
Four studies (Golberg 1999; Mudge 1998; Klenner 2003; Schultz 2002 ) reported on this outcome for the first catheter placement. There was no statistically significant difference in the rate of phlebitis or thrombophlebitis (pooled RR 0.74 [95%CI 0.50, 1.10] and RD -0.06 [95% CI -0.13, 0.02]; p = 0.65).
Catheter related sepsis and number of additional PIV catheters needed:
None of the included studies reported on these outcomes.
2. Complications related to heparin:
Abnormal coagulation profile:
None of the studies reported on the incidence of abnormal coagulation profile. However, Alpan 1984 reported on the mean prothrombin time and mean partial thromboplastin time and there was no significant difference between the groups.
Allergic reactions and heparin induced thrombocytopenia:
Paisley 1997 and Klenner 2003 reported that none of the infants in their study developed heparin induced thrombocytopenia. No other study has reported on this outcome. Klenner 2003 prospectively evaluated 213 (those patients who received heparin for > 5 days) of the total 296 patients for heparin induced thrombocytopenia at specified intervals and did not observe any difference in mean optical density at 405 nm (enzyme linked immunosorbent assay) between groups (0.020 vs 0.019; p = 0.78).
Intracranial or intraventricular hemorrhage:
Two studies Alpan 1984 and Klenner 2003 reported on the incidence of intraventricular hemorrhage in their study patients. There was no statistically significant difference between the groups (pooled RR 0.37 [95% CI 0.12, 1.17] and RD -0.04 [95% CI -0.08, 0.00]).
3. Mortality:
No study reported on this outcome.
Included studies but no analyzable outcomes:
Kotter 1996 reported all the outcomes per number of catheters.
Moclair 1995 did not report any of the outcomes per number of patients. The results were expressed per number of catheters so were not included in the analysis. Survival curves for control group and all the heparin doses were calculated, based on all catheters, and the authors reported that addition of heparin prolongs the infusion site survival at concentration of 0.25 IU/ml. They also reported that the effect of heparin is sigmoidal with onset at 0.1 IU/ml and flattening at 0.5 IU/ml.
There is an extensive literature on the use of heparin to prolong the patency of PIV catheters. Randolph 1998 performed a systematic review on the use of heparin for PIV catheters in all age groups, and concluded that low dose heparin infusion was effective in prolonging patency of peripheral arterial catheters but not PIV catheters. The present review includes nine additional studies aiming to test the effectiveness of heparin use for PIV catheters in the neonatal population.
Two more studies were identified in this update in addition to eight studies identified in previous version (Shah 2002). Despite large number of studies identified, the total number of neonatal patients included in this systematic review was 721. The methodology of the systematic review process was hampered by many studies reporting outcomes per catheter rather than per patient; this practice leads to non-independence of multiple measures of the same outcome in the same patients, and thus creates potential bias in data analysis. Therefore, in performing this review, only data from the first catheters were considered.
This review was further limited by the significant variability amongst included studies in clinical details, methodological quality, and outcomes reported, thus precluding the reviewers from combining data on the primary outcome to estimate an overall effect size. The dose and method of administration of heparin varied widely, with concentrations varying from 0.1 to 10 IU/ml of heparin administered as either intermittent flush solution or as an additive in parenteral nutrition solutions, and frequency of flush was not always specified by the investigators. Catheter size was not uniform across the studies and may have affected the duration of catheter patency or the likelihood of the complications. The methodological quality of studies were variable, from randomized, double-blinded controlled trials to quasi-randomized studies, with some studies having inadequate or no blinding of intervention or outcome assessment, and incomplete followup. The outcomes reported were different among the studies. Although seven of the ten studies reported on the primary outcome of PIV catheter patency, few reported on secondary outcomes such as short and long term adverse effects of heparin, or complications related to PIV catheter insertions. It is noteworthy that none of the neonates developed heparin induced thrombocytopenia which was examined in two studies (n = 300).
It was not possible from the published reports to assess the effect of heparin among patients who had multiple catheter placements. Individual patient data, required for the planned analyses among patients having multiple catheters, were not available. Future studies should report the duration of catheter use separately for catheters removed because of occlusion, removed electively, or removed for other reasons. These data should be reported for the first catheter and for all catheters.
Because of heterogeneity among included studies, and despite the best effort of data collection and analysis, the question of the effectiveness and safety of heparin use in PIV catheters in neonates was unable to be definitively answered by this review.
No conclusive evidence is available from this systematic review on which to evaluate the effectiveness of heparin to prolong PIV catheter life in the neonatal population. Although limited data did not suggest significant adverse effects from heparin use, the safety of such practice has not been thoroughly investigated.
A randomized, controlled trials of sufficient power is needed to evaluate the efficacy of heparin in neonates with PIV catheters. Studies to identify the minimal effective dose of heparin and to evaluate the safety of heparin administration in this setting will also be necessary before routine use can be recommended in neonates, particularly in the preterm population. A detailed history of all catheters should be recorded, including the reason for catheter removal (blocked, suspected or proven infection, routine or elective replacement, no longer needed). Time until catheter blockage should be assessed by "survival" function analysis.
We would like to thank the primary authors (Moclair 1995, Mudge 1998, Treas 1992, Hanrahan 2000, LeDuc 1997, Taylor 1989, Klenner 2003) for providing additional data from their studies. We are also grateful for Dr. Arne Ohlsson's valuable input in the revision of this manuscript.
None.
| Study | Methods | Participants | Interventions | Outcomes | Notes | Allocation concealment |
| Alpan 1984 | Randomization - yes Single centre study Blinding of randomization - yes Blinding of intervention - yes Blinding of outcome assessment - yes Complete follow up - yes | Preterm infants admitted to NICU Group 1: n=13, Mean (SD) GA 32.0 (2.2) weeks Mean (SD) BW 1397 (543) g Group 2: n=13, Mean (SD) GA 31.2 (3.1) weeks Mean (SD) BW 1361 (627) g | Group 1: heparin 1 IU/ml in TPN infusate Group 2: no heparin | Duration of catheter use Phlebitis Septicemia Coagulation parameters (PT and PTT) | Only long-term outcome such as intraventricular hemorrhage reported per number of patients while others are reported per number of catheters | A |
| Golberg 1999 | Randomization - yes Single centre study Blinding of randomization - yes ( by pharmacist) Blinding of intervention - yes (multidose vial) Blinding of outcome assessment - yes Complete follow up - no | All patients in the NICU who required PIV catheters Group 1: n=24 Group 2: n=23 Mean (SD) GA of the cohort 35.07 (5.4) weeks Mean (SD) BW of the cohort 2664 (1175.6) g | Group 1: heparin 4 IU/ml Group 2: normal saline Dose 1 ml flush at least every 6 hours | Duration of catheter use Phlebitis | Positive pressure technique for flushing | A |
| Heilskov 1998 | Randomization - yes Single centre study Blinding of randomization - yes Blinding of intervention - yes (treatment solution with the code number) Blinding of outcome assessment - yes Complete follow up - yes | All neonates admitted to NICU or intermediate care nursery who required PIV catheter Exclusion criteria: Neonates who had received anticoagulant within the last 48 hours, BW <800 grams, Parents - non English speaking Group 1: n=27, Mean (SD) BW 2323 (816) g Group 2: n=28, Mean (SD) BW 2521 (891) g Group 3: n=35, Mean (SD) BW 2677 (998) g | Group 1: heparin 2 IU/ml Group 2: heparin 10 IU/ml Group 3: normal saline Catheter flushed at least every 6 hours | Duration of catheter use Infiltration | Positive pressure technique for flushing Many canula had IV fluid running through prior to locking for intermittent use | A |
| Klenner 2003 | Randomization - yes Single centre study Blinding of randomization - yes Blinding of intervention - yes Blinding of outcome assessment - yes Complete follow up - yes | Infants with anticipated need for intravenous infusion for >5 days and birth weight >1000g and postnatal age <28 days Group 1: n=145, Mean (SD) GA 35.3 (3.5) weeks Mean (SD) BW 2400 (890) g Group 2: n=151, Mean (SD) GA 35.6 (3.3) weeks Mean (SD) BW 2416 (737) g | Group 1: heparin 0.5 IU/ml in intravenous infusion Group 2: saline | Duration of catheter use Extravasation/infiltration Phlebitis Thrombosis Heparin induced thrombocytopenia | A | |
| Kotter 1996 | Randomization - yes (pharmacist) Single center study Blinding of intervention - yes Blinding of outcome assessment - not stated Complete follow up - no | All NICU patients under one month of age who required intermittent PIV catheter locks Group 1: n= 27 (75 catheters) Group 2: n=24 (43 catheters) Mean GA of the cohort 36.2 weeks Mean BW of the cohort 2581 g | Group 1: heparin 10 IU/ml Group 2: normal saline Dose: 0.5 ml after each infusion and every 4 hours | Duration of catheter use Infiltration Phlebitis Occlusion Leaking | The outcomes were reported per number of catheters | A |
| Moclair 1995 | Randomization - yes (pharmacist) Single centre study Blinding of randomization - yes Blinding of intervention - yes Blinding of outcome assessment - not stated Complete follow up - no (97 entered the study, 7 dropped due to death or transfer) | All NICU patients receiving TPN through peripheral intravenous catheter Group 1: n=20 (72 catheters) Group 2: n=15 (35 catheters) Group 3: n=16 (47 catheters) Group 4: n=23 (61 catheters) Group 5: n=16 (30 catheters) | Group 1: heparin 0.1 IU/ml Group 2: heparin 0.25 IU/ml Group 3: heparin 0.5 IU/ml Group 4: heparin 1 IU/ml Group 5: no heparin added to TPN infusate | Duration of catheter use Extravasation of catheters | A | |
| Mudge 1998 | Quasi-randomized Single center study Blinding of randomization - no Blinding of intervention - yes Blinding of outcome assessment - yes Complete follow up - no (40 out of 42 infants) | Patients admitted to NICU or general pediatric units Group 1: Total of 78 patient randomized to control group of which 25 were neonates Group 2: Total of 56 patients randomized to heparin group of which 15 were neonates Mean BW of the neonatal cohort was 2164.8 g | Group 1: heparin 10 IU/ ml Group 2: normal saline Dose: 1 ml flush, frequency not specified | Duration of catheter use Occlusion of catheter Phlebitis | Allocation was based on the month of the study Data on neonates was provided by primary author | C |
| Paisley 1997 | Quasi-randomized Single center study Blinding of randomization - no Blinding of intervention - no Blinding of outcome assessment - no Complete follow up - yes | Infants > 32 weeks gestation admitted to NICU and continuing care nursery Group 1: 33 infants Group 2: 54 infants Mean (SD) gestational age of the group 38.46 (2.48) weeks | Group 1: heparin 10 IU/ml Group 2: normal saline flushes Dose: 0.6 ml flush every 4 hours | Duration of catheter use Heparin induced thrombocytopenia Occlusion Leaking Phlebitis Infiltration | Allocation was based on the last digit of the hospital number All the outcomes apart from duration of catheter patency were reported in the combined group (treatment and control) | C |
| Schultz 2002 | Randomization - yes Single centre study Blinding of randomization - yes Blinding of intervention - yes Blinding of outcome assessment - yes Complete follow up - yes | All neonates who required intermittent intravenous catheter locks Group 1: n=20, Mean (SD) GA 33.2 (3.7) weeks Mean (SD) BW 2020 (830) g Group 2: n=29, Mean (SD) GA 33.7 (4.0) weeks Mean (SD) BW 2170 (910) g | Group 1: heparin 2 IU/ml Group 2: normal saline flushes Dose: 0.5 ml flush every 3 hours | Duration of catheter use Occlusion Extravasation Phlebitis | A | |
| Treas 1992 | Quasi-randomized Single center study Blinding of randomization - no Blinding of intervention - no Blinding of outcome assessment - no Complete follow up - yes | Infants admitted to NICU Group 1: n=63 (catheter 131) Mean (SD) GA 33.8 (3.0) weeks Mean (SD) BW 1977 (595) g Group 2: n=49 (catheters 122) Mean (SD) GA 34.7 (3.2) weeks Mean (SD) BW 2300 (806)g | Group 1: infusion containing 0.5 IU/ml of heparin or flush containing 0.5 IU/ml of heparin Group 2: normal saline Dose: Volume of flush not specified | Duration of catheter use Infection Infiltration Leaking Phlebitis | Allocation was based on hospital number All the outcomes were reported per number of catheters Data for the first catheters were provided by the author | C |
| Study | Reason for exclusion |
| Danek 1992 | Data on neonates not available |
| Hanrahan 2000 | Non-randomised |
| LeDuc 1997 | Data on neonates not available |
| McMullen 1993 | Data on neonates not available |
| Nelson 1998 | Data on neonates not available |
| Taylor 1989 | Compared heparin flush with heparin infusion |
Alpan G, Fabian E, Springer C, Glick B, Goder K, Armon J. Heparinization of alimental solutions administered through peripheral veins in premature infants: a controlled study. Pediatrics 1984;74:375-8.
Golberg 1999 {published data only}
Golberg M, Sankaran R, Givelichan L, Sankaran K. Maintaining patency of peripheral intermittent infusion devices with heparinized saline and saline. Neonatal Intensive Care 1999;12:18-22.
Heilskov 1998 {published and unpublished data}
Heilskov J, Kleiber C, Johnson K, Miller J. A randomized trial of heparin and saline for maintaining intravenous locks in neonates. Journal of the Society of Pediatric Nursing 1998;3:111-6.
Klenner 2003 {published and unpublished data}
Klenner AF, Fusch C, Rakow A, Kadow I, Beyersdorff E, Eichler P, Wander K, Lietz T, Greinacher A. Benefit and risk of heparin for maintaining peripheral venous catheters in neonates: a placebo-controlled trial. Journal of Pediatrics 2003;143:741-5.
Kotter 1996 {published data only}
Kotter RW. Heparin vs saline for intermittent intravenous device maintenance in neonates. Neonatal Network 1996;15:43-7.
Moclair 1995 {published and unpublished data}
Moclair A, Bates I. The efficacy of heparin in maintaining peripheral infusions in neonates. European Journal of Pediatrics 1995;154:567-70.
Mudge 1998 {published and unpublished data}
Mudge B, Forcier D, Slattery MJ. Patency of 24-gauge peripheral intermittent infusion devices: a comparison of heparin and saline flush solutions. Pediatric Nursing 1998;24:142-5.
Paisley 1997 {published data only}
Paisley MK, Stamper M, Brown J, Brown N, Ganong LH. The use of heparin and normal saline flushes in neonatal intravenous catheters. Pediatric Nursing 1997;23:521-7.
Schultz 2002 {published data only}
Schultz AA, Drew D, Hewitt H. Comparison of normal saline and heparinized saline for patency of IV locks in neonates. Applied Nursing Research 2002;15:28-34.
Treas 1992 {published and unpublished data}
Treas LS, Latinis-Bridges B. Efficacy of heparin in peripheral venous infusion in neonates. Journal of Obstetric, Gynecologic, and Neonatal Nursing 1992;21:214-9.
Danek GD, Noris EM. Pediatric IV catheters: efficacy of saline flush. Pediatric Nursing 1992;18:111-3.
Hanrahan 2000 {published and unpublished data}
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 and unpublished data}
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.
Taylor 1989 {published data only}
Taylor J, Shanon R, Kilbride H. Heparin lock intravenous line use in newborn infants: a controlled trial. Clinical Pediatrics 1989;28:237-40.
* indicates the primary reference for the study
Bossert E, Beecroft PC. Peripheral intravenous lock irrigation in children: current practice. Pediatric Nursing 1994;20:346-9,355.
Daniell HW. Heparin in the prevention of infusion phlebitis: a double-blind controlled study. JAMA : The Journal of the American Medical Association 1973;226:1317-21.
Goode CJ, Titler M, Rakel B, Ones DS, Kleiber C, Small S, Triolo PK. A meta-analysis of effects of heparin flush and saline flush: quality and cost implications. Nursing Research 1991;40:324-30.
Graham DR, Keldermans MM, Klemm LW, Semenza NJ, Shafer ML. Infectious complications among patients receiving home intravenous therapy with peripheral, central, or peripherally placed central venous cathters. The American Journal of Medicine 1991;91:95s-100s.
Hanson RL, Grant AM, Majors KR. Heparin-lock maintenance with ten units of sodium heparin in one milliliter of normal saline solution. Surgery, Gynecology & Obstetrics 1976;142:373-6.
Hecker JF. Potential for extending survival of peripheral intravenous infusions. BMJ 1992;304:619-24.
Johnstone JM. Heparinised saline - is the heparin really necessary? The Australian Nurses' Journal 1991;20:31-2.
Krafte-Jacobs B, Sivit CJ, Mejia R, Pollack MM. Catheter-related thrombosis in critically ill children: comparision of catheters with and without heparin bonding. Journal of Pediatrics 1995;126:50-4.
Lesko SM, Mitchell AA, Epstein MF, Louik C, Giacoia GP, Shapiro S. Heparin use as a risk factor for intraventricular hemorrhage in low-birth-weight infants. New England Journal of Medicine 1986;314:1156-60.
Lewis GBH, Hecker JF. Infusion thrombophlebitis. British Journal of Anaesthesia 1985;57:220-33.
Maki DG. Preventing infection in intravenous therapy. Anesthesia and Analgesia 1977;56:141-53.
Malloy MH, Cutter GR. The association of heparin exposure with intraventricular hemorrhage among very low birth weight infants. Journal of Perinatology 1995;15:185-91.
Nieto-Rodriguez JA, Garcia-Martin MA, Barreda-Hernandez MD, Hervas MJ, Cano-Real O. Heparin and infusion phlebitis: a prospective study. The Annals of Pharmacotherapy 1992;26:1211-4.
O'Neill TJ, Tierney LM Jr, Proulx RJ. Heparin lock-induced alterations in the activated partial thromboplastin time. JAMA : The Journal of the American Medical Association 1974;227:1297-8.
Papile L,Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weight less than 1500 grams. Journal of Pediatrics 1978;92:529-34.
Peterson FY, Kirchhoff KT. Analysis of the research about heparinized versus nonheparinized intravascular lines. Heart Lung 1991;20:631-40.
Potter C, Gill JC, Scott JP, McFarland JG. Heparin-induced thrombocytopenia in a child. Journal of Pediatrics 1992;121:135-8.
Randolph AG, Cook DJ, Gonzales CA, Andrew M. Benefit of heparin in peripheral venous and arterial catheters: systematic review and meta-analysis of randomised controlled trials. BMJ 1998;316:969-75.
Roberts GW, Holmes MD, Staugas REM, Day RA, Finlay CF, Pitcher A. Peripheral intravenous line survival and phlebitis prevention in patients receiving intravenous antibiotics: heparin/hydrocortisone versus in-line filters. The Annals of Pharmacotherapy 1994;28:11-6.
Tager IB, Ginsberg MB, Ellis SE, Walsh NE, Dupont I, Simchen E, Faich GA. An epidemiologic study of the risks associated with peripheral intravenous catheters. American Journal of Epidemiology 1983;118:839-51.
Tanner WA, Delaney PV, Hennessy TP. The influence of heparin on intravenous infusions: a prospective study. The British Journal of Surgery 1980;67:311-2.
Tighe MJ, Wong C, Martin IG, McMahon MJ. Do heparin, hydrocortisone, and glyceryl trinitrate influence thrombophlebitis during full intravenous nutrition via a peripheral vein? JPEN. Journal of Parenteral and Enteral Nutrition 1995;19:507-9.
Tomford JW, Hershey CO, McLaren CE, Porter DK, Cohen DI. Intravenous therapy team and peripheral venous catheter-associated complications. a prospective controlled study. Archives of Internal Medicine 1984;144:1191-4.
Vieira A, Berry L, Ofosu F, Andrew M. Heparin sensitivity and resistance in the neonate: an explanation. Thrombosis Research 1991;63:85-98.
Vilardell M, Sabat D, Arnaiz JA, Bleda MJ, Castel JM, Laporte JR, Vallve C. Topical heparin for the treatment of acute superficial phlebitis secondary to indwelling intravenous catheter: a double-blind, randomized, placebo-controlled trial. European Journal of Clinical Pharmacology 1999;54:917-21.
Warkentin TE, Kelton JG. Heparin and platelets. Hematology/oncology Clinics of North America 1990;4:243-64.
Shah P, Ng E, Sinha AK. Heparin for prolonging peripheral intravenous catheter use in neonates. In: The Cochrane Database of Systematic Reviews, Issue 4, 2002.
| Comparison or outcome | Studies | Participants | Statistical method | Effect size |
|---|---|---|---|---|
| 01 Heparin versus control (normal saline or no treatment) | ||||
| 01 Duration of catheter patency (hours), first catheter | WMD (fixed), 95% CI | No total | ||
| 02 Catheter occlusion, first catheter | 4 | 475 | RR (fixed), 95% CI | 1.00 [0.85, 1.16] |
| 03 Phlebitis or thrombophlebitis, first catheter | 4 | 432 | RR (fixed), 95% CI | 0.74 [0.50, 1.10] |
| 04 Intracranial hemorrhage | 2 | 322 | RR (fixed), 95% CI | 0.37 [0.12, 1.17] |
| 05 Heparin induced thrombocytopenia | 0 | 0 | RR (fixed), 95% CI | No numeric data |
Dr Ajay K Sinha
Clinical Lecturer
Department of Child Health
Queen Mary Hospital
Elizabeth Ward, Garden House
Royal London Hospital
London
UK
E1 1BB
Telephone 1: 0044 20 7377 7712
Telephone 2: 0044 1708 473 721
E-mail: ajaysinha@mcmail.com
Secondary address (home):
6 Squirrels Heath Lane
Romford
UK
RM2 6DS
| 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. |