Background
Central venous catheters (CVCs) are widely utilized in clinical practice, especially in intensive care units (ICUs) [
1]. These devices are inserted so as to enable the administration of fluids, blood products, medications, parenteral nutrition, and for the performance of dialysis and central venous pressure monitoring [
2,
3]. Currently, there are four types of CVCs: non-tunneled, tunneled, peripherally nserted central catheters (PICCs) and totally implantable venous access devices (TIVADs) [
4].
These central lines will remain in place for days or even weeks each time [
5]. Prolonged use may result in catheter occlusion, which may give rise to a requirement for the catheter to be treated, removed or replaced. Inserting a new central line creates latent threats, which could lead to disrupted treatment, increasing morbidity, and greater spending on health care [
5]. Generally, catheter obstruction can be defined as partial occlusion (inability to aspirate blood but ability to flush freely) or complete occlusion (inability to flush freely and withdraw blood). It is estimated that the occlusion rate is between 0% and 33% when using heparin saline (HS) solution [
6,
7]. Factors leading to catheter obstruction can be generally classified into three categories: mechanical causes, drug/mineral precipitates and clot formation, which is the most common reason overall [
8]. To avoid the risk of catheter occlusion, thrombosis and catheter-related bloodstream infection (CRBSI), proper catheter flushing and locking are always considered to be the primary intervention because of the effect of reducing blood reflux into the lumen [
8,
9].
Unfractionated heparin is well-known for its anticoagulant activity. Thus, heparin is widely used to maintain the patency of CVCs [
10]. Nonetheless, the efficacy of this practice has not been definitively shown. Moreover, the use of heparinised saline is associated with potential risks such as coagulation disorders, hypersensitivity reactions and heparin-induced thrombocytopenia (HIT) [
11,
12]. Researchers have been looking for a safe alternative to heparin, such as isotonic saline, vitamin C, lepirudin, sodium citrate or polygeline, to improve this situation [
13‐
16]. Especially important, if there was a suitable replacement for HS, that would be beneficial, especially for patients with contraindications to using HS.
There have been numerous publications in this field over the last few years, including a guideline [
17], several trials [
2,
10] and several reviews [
5,
9], including a Cochrane Review [
18]. Most of these studies indicate that normal saline (NS) is safe and efficacious in preventing catheter occlusion in adult populations with CVCs. The recent guideline concluded that routine flushing with NS is recommended. However, the Cochrane review showed that there is no clear evidence to indicate whether NS flushing is superior to flushing with HS solution.
CVC occlusion is a fairly common problem, but differences in methods of prophylaxis, diagnosis and treatment practices related to catheter lumen obstruction vary, perhaps as a result of a lack of appropriate clinical guidelines [
19]. An evidence-based, standardized flushing protocol is required for CVCs in adults. We conducted a systematic review and meta-analysis to evaluate the clinical efficacy (benefits and harms) of NS flushes and HS flushes for prevention of CVC lumen occlusion in adult patients.
Methods
Our systematic review and meta-analysis was conducted according to preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines (Additional file
1) [
20]. The PICO framework was applied to define the clinical question clearly (Additional file
2). The primary outcome was catheter occlusion. Secondary outcomes included: maneuver needed (patients who required catheter manipulation to maintain the patency of the lumen), HIT, haemorrhage, central venous thrombosis and CRBSI.
Search strategy and selection criteria
We systematically searched PubMed, Embase and the Cochrane library databases from the inception to 28 September 2016, using the following terms: “Sodium Chloride”, “Saline Solution, Hypertonic”, “NaCl”, “Heparin”, “Catheterization, Central Venous”, “Randomized Controlled Trial”, etc. (Additional file
3). There was no restriction on language. We also reviewed bibliographies in the retrieved articles to identify additional relevant studies. Only clinical randomized controlled trials (RCTs) of NS flushing vs flushing with HS solution in adults were included. Exclusion criteria were (1) age <18 years, and (2) case reports, letters, reviews, case-control studies and cohort studies, or non-human studies.
Data were independently extracted by three reviewers (ZL, YY and WX). The following information was abstracted from the included studies: Study ID, mean age (years), country origin, number of subjects (NS/HS), female (%), centre, ICUs, disease types, follow up (days), heparin concentration (IU/ml) and heparin volume (ml). Disagreements were resolved by consensus.
Assessment of study quality
The quality of the individual studies was assessed based on the Cochrane handbook for systemic reviews of interventions [
21].
Assessment of risk of bias
We performed sensitivity analysis to assess the influence of a single study on the pooled effects. Simultaneously, we used a funnel plot for assessment of publication bias [
22].
Statistical analysis
The pooled effects were analyzed by relative risk (RR) with 95% confidence interval (CI) for dichotomous outcomes. Statistical heterogeneity among trials was quantitatively assessed with the X
2 test,
P values and the
I
2
statistics [
23]. We pooled data using Mantel-Haenszel random-effects models, which are more conservative in their estimations [
24].
During the search process, there were four different kinds of subjects for analysis: patients (six studies), catheters (two studies), lumens (one study; multilumen CVCs) and line access (one study; flushing central lines before and after each use). Given this, our meta-analysis was to be analysed separately on the basis of definite features of units.
In light of the Ge and Schallom study [
2,
25], the duration of catheter placement is classified as short-term (less than 3–4 wks) and long-term (months to years). Simultaneously, subgroup analysis based on the length of indwelling time was carried out to characterize possible sources of heterogeneity(e.g. <30 and >30 days). All statistical tests were two-sided using an α level of 0.05. This meta-analysis was conducted using Stata 12.0 software (StataCorp, College Station, TX, USA) and Review Manager Version 5.3.5 (available from
http://tech.cochrane.org/revman/download).
Discussion
Our systematic review and meta-analysis did not demonstrate a general difference between use of NS or HS in adult populations. In subgroup analysis, stratified by the length of indwelling time (e.g. <30 and >30 days), there appear to be two conflicting conclusions (Fig.
6.) In the short run (<30 days), HS was slightly better than NS. A plausible reason may be that NS has no anticoagulation activity. However, a marginally significant association was observed between using NS vs HS and the incidence of catheter occlusion. Owing to the limited numbers of included studies and the effect sizes, we should treat such a result with caution. By comparison, NS could be equal, if not more effective, in the long run (>30 days). This has implications for patients in whom long-term catheter use may be necessary, for example, in patients undergoing cancer treatment or those requiring dialysis. From a long-term perspective, the use of NS in these patients has several advantages over HS solutions. To begin with, NS is an isotonic solution, which is in accordance with basic physiological needs. In addition, the use of NS will result in fewer side effects from heparin-related complications. Finally, as HS is several times more expensive than NS [
32], eliminating its use in flushing solutions has economic benefits.
To date, there have only been three relevant meta-analyses in this area (Additional file
4). The result of the first study (network meta-analysis) was no marked difference, when comparing adult patients with NS vs HS or other solutions in the flushing of CVCs [
33]. The second study, consistent with results of previous research, found that HS was not more effective than NS in reducing catheter occlusion when analysed in three different areas (participant, catheter and line access) [
18]. These findings challenged the continued use of HS in CVC flushing, as it is more expensive than saline solution. However, a recent study supported NS as a substitution for HS as a locking solution in CVCs in adult patients from the point of view of four different types of CVC [
34]. Data from these studies suggests that HS may not be required to maintain the patency of CVCs. In the absence of sufficient evidence to support the use of NS, the debate will be moot. For this reason, further study is needed in this field.
To our knowledge, our study might be the first meta-analysis from the viewpoint of four different CVC-related areas (patient, catheter, lumen and line access) and indwelling time (i.e. <30 and >30 days). Our results, in accordance with current studies, meta-analyses and reviews [
18,
27,
33], suggest that there is very little evidence to conclude that flushing with HS has more effect than NS flushing solution for CVC maintenance.
Only a few randomized controlled studies have compared NS with HS for maintenance of CVC lumen patency in adults. The Rabe study was the first RCT to compare the effects of NS versus HS and they determined that the use of a flush containing 5000 U/ml was more effective than NS [
13]. In particular, catheter survival rate was higher in the HS group than in NS group. In contrast to the results of the Rabe study, a large number of other studies suggest that the catheter lumen occlusion rate is not different between those with vs those without heparin [
10,
27‐
29]. NS flushing for CVCs has been applied in some American ICUs without supporting evidence [
35]. Parallel with the mainstream view, Schiffer et al. [
17] suggest that routine flushing of CVCs with NS to prevent occlusion is reasonable in this guideline. Morover, a report of a recent multicentre randomized trial [
27] argued that HS was not more effective than NS in reducing withdrawal or total occlusion. No statistical difference in catheter patency was observed when comparing NS to HS solution.
Due to the body’s physiologic response to the catheters, nearly 100% of CVCs will develop a “fibrin sheath”, which may increase the risk of catheter occlusion from 1 to 14 days after insertion of the indwelling catheter [
36,
37]. As the “fibrin sleeve” usually envelops the tip of the CVC, some argue that a heparin lock could not prevent thrombotic occlusion because of the difficulty in achieving an effective concentration on the outside of the catheter tip [
28]. Our findings support these suggestions.
Certainly, the occurrence of CVC occlusion is related to the catheter type, puncture site, heparin concentration, heparin volume, flush frequency, retaining time and the patient’s physical condition [
37,
38]. As there are indeterminate factors in this field, further studies, including well-designed trials, are warranted to assess these effects on clinical outcomes.
Various potential limitations should be taken into consideration. First, although the statistical heterogeneity was low, the clinical and methodological heterogeneity cannot be ignored. The latter two types of heterogeneity might be attributed to various types of participants, interventions, outcomes studied (partial or complete occlusion), study designs and study qualities. Second, the potential hazards might occur after long-term follow up, thus, some of these complications could be discarded due to the short duration of some included studies. Third, this meta-analysis was limited to studies conducted in Asia, Europe and North America, and thus, might not be generalizable to other parts of the world. Finally, there was a publication bias in our study as small studies with null results tend not to be published. Hence, uniform study design and multi-centre studies should be launched in different countries and regions to establish the best approach to long-term maintenance of CVCs.