Comparison of ethanol lock and heparin lock solution as prevention of catheter-related bloodstream infection in hemodialysis patients: systematic review and meta-analysis of randomized controlled trials

Introduction: Catheter-related bloodstream infection (CRBSI) is a common complication of catheter use for vascular access in hemodialysis patients and a major cause of morbidity and mortality. Preventive measures, including antibiotic lock, are inadequate due to the risk of resistance and insufficient effect against bacterial biofilm. Ethanol, an antimicrobial substance, is a potential prophylactic lock-in preventing CRBSI. This study aims to assess ethanol lock’s effectiveness in preventing CRBSI in hemodialysis patients with a catheter as vascular access and its impact on catheter dysfunction. Methods: Researchers systematically searched online databases including Pubmed, Cochrane Library, and Science Direct for relevant randomized controlled trials (RCTs) published within 2011 until 2020. Relevant data were pooled in PICOs (Population, Intervention, Control, Outcomes) format and analyzed with Review Manager (version 5.3.5, Cochrane Collaboration, Denmark). Results: Seven RCTs involving 453 patients were assessed. The primary outcome indicates that prophylactic ethanol lock significantly reduces the incidence of CRBSI compared to that of heparin lock (RR=0.32, 95% CI 0.12-0.83, p=0.02, heterogeneity I 2 =68%). The secondary outcome suggests no significant difference in the incidence of catheter dysfunction in ethanol lock and heparin lock (RR=0.75, 95% CI 0.23-2.40, p=0.63, heterogeneity I 2 =68%). Conclusion: Ethanol is a potential prophylactic lock agent in preventing CRBSI in hemodialysis patients with catheter access. Further research is needed to synchronize the procedural use of ethanol lock and evaluate its long-term effect.


INTRODUCTION
Chronic Kidney Disease (CKD) remains a disease with high worldly prevalence. In late 2016, it is estimated that 3,7 million of the world population suffer from end-stage renal disease (ESRD), and this number continues to increase by 6% yearly. Around 3 million of them require hemodialysis (HD). 1 In Indonesia, according to The Indonesian Renal Registry (IRR), the number of CKD patients increased from 9.649 in 2010 to 30.831 in 2017, while the prevalence increased from 11.484 in 2010 to 77.892 in 2017. 2 Hemodialysis, particularly in Indonesia, remains the main treatment modality. There was a three-fold increase in the number of hemodialysis patients in 2011 than that in 2007. In the dialysis unit on Dr. Soetomo District Hospital Surabaya, the recorded number of CKD patients who undergo routine hemodialysis constantly increased for three years: 245 patients in 2013, 255 patients in 2014, and 303 patients in 2015. 3 Creating vascular access is a common procedure to aid in hemodialysis procedure. Several known types include catheter access, arteriovenous fistula (AVF), an arteriovenous graft (AVG). AVF and AVG, which are typically used for long-term use, require maturation for about six weeks, during which time patients are usually equipped with temporary catheter access for hemodialysis. The primary failure rate of AVF ranges between 30%-70%, and its 1-year patency ranges around 40%-70%. 4, 5 The complication of AVF and AVG, such as thrombosis or stenosis, requires repairing, catheter access is utilized to enable uninterrupted hemodialysis cycles. Furthermore, if possible physiologic sites for AVF or AVG are depleted due to recurrent failure, permanent catheter access may be considered. 4 Around 80% of patients initiate hemodialysis with a central venous catheter (CVC), which is used in continuation up to 90 days (68,5%) and one year in the form of a tunneled catheter (21%). 6,7 The long-term use of catheters carries risks, including catheter-related infection might occur on the exit site, inside the lumen, and bacteremia (catheter-related bloodstream infection/CRBSI). The incidence of vascular catheter-related bacteremia ranges around 48%-73%, responsible for 26% of catheter removal. 6 In 2011, The United States Centers for Disease Control and Prevention (CDC) reported 41000 cases of CRBSI which resulted in increased hospital admission rates. 8 Hemodialysis patients with CVC are at a 15-fold risk of CRBSI, with mortality rates ranging from 12% to 25%. The risk of sepsis with CVC is 8-fold greater than that with AVG and AVF. 7 Other complications include septic arthritis and epidural abscess. 9,10 CRBSI in hemodialysis patients with a long-term catheter is a nosocomial infection problem with high morbidity and mortality, resulting in higher cost and longer hospital stay. The main causative agent of CRBSI is Gram-positive coccus and Gram-negative basil and fungi that form biofilm on the catheter lumen. In non-permanent catheters, its removal may solve the problem, while in a permanent catheter, antimicrobial lock solutions might be of use. 11 Antimicrobial locks are further grouped into antibiotic-based and nonantibiotics. The current use of catheter locks commonly consists of heparin, which some studies suggest might aid in forming pathogenic biofilm inside the catheter lumen. 6 The use of antibiotics also raises concerns about antibiotic resistance. Several studies reported resistance in the use of high-dose gentamycin (4-27 mg/ ml) against Enterobacteriaceae sp. and Staphylococcus sp. and one case of death in the use of 1-4 mg/ml gentamycin lock. 6,7 Therefore, current use suggests low-dose antibiotics or replacement of heparin with other agents. However, lowdose antibiotics do not solve the concern for antibiotic resistance. 12 Other studies reported that antibiotics have inadequate biofilm penetration potency. Heparin replacement is deemed irrational due to its greater benefit of thrombus prevention than the risk of biofilm formation. 8 Due to the concern of antibiotic resistance with its use for locking agents, the authors evaluated the role of nonantibiotic antimicrobial lock solutions. The use of both low-and high-dose trisodium citrate result in inconsistent outcomes related to CRBSI. It is also not recommended due to a case of death and possible induced paresthesia and arrhythmia (by hypocalcemia or systemic embolism). Furthermore, its use requires a higher dose of thrombolytics to reduce catheter dysfunction. 6,7 Ethanol is a widely available, inexpensive antiseptic agent with bactericidal and fungicidal activity. It is a broad-spectrum agent that works through nonspecific protein denaturation and lipid dissolution. Several in vitro studies have reported its effectiveness against pathogenic biofilms. 8, 11 High-dose ethanol exposure has not been proved to induce resistance with a minimum side effect. 13 This study aims to evaluate the potency of ethanol lock in preventing CRBSI in hemodialysis patients in the form of metaanalysis from available clinical trials.

Study Characteristics
The study selection process is shown in Figure 1. All 7 studies that met the inclusion criteria are randomized controlled trials (RCTs  Sofroniadou et al. (2017), catheter dysfunction is defined as thrombosis. The inability to use the catheter at >200 ml/minute flow despite additional flushing and instillation of t-plasminogen activator as intraluminal thrombolysis. 19 Shresta & Raut (2015) defined catheter dysfunction as persistent <200 ml/minute flow after eliminating mechanical factors (kink or patient positioning). It evaluated the presence of thrombus by cutting the catheter following its removal. 17 Broom et al. (2012) defined catheter dysfunction as flow dysfunction (blockage or reduced flow) and mechanical dysfunction (ruptured catheter). 14 Vercaigne et al. (2015) described catheter dysfunction as two consecutive dialysis sessions with a flow of <300 ml/minute for at least 50% dialysis session. 18 Figure 2 presents the risk of bias of each indluded studies. All 7 studies did randomization, but 4 studies 13,16,18, 19 have low risk of bias in the "random sequence generation" parameter by elaborating the randomization method, and 3 studies 13,18, 19 have low risk of bias in the "allocation concealment" parameter. Five studies 13,14,16,18, 19 have high risk of bias on the "blinding of participants and personnel" parameter. Two studies 14, 16 have high risk of bias on the "blinding of outcome assessment" parameter. The high risk of the two aforementioned parameters  is due to the unique odor of ethanol that is easily recognizable by both healthcare workers and the patients, affecting the blinding process. On the "incomplete outcome data" parameter, 6 studies 13,14,16-19 have low risk of bias, while on the "selective reporting" parameter, 5 studies 13,16-19 have low risk of bias. The risk of bias assessment of the included studies is summarized in Figure 3.

The intervention of ethanol lock to the incidence of CRBSI in regular hemodialysis patients
The definition of CRBSI in each study is provided in Table 1 20 Vercaigne et al. (2015) reported 3 adverse events suspected to be related to the use of ethanol lock, including axillofemoral graft occlusion, gastrointestinal bleeding, and septic episode. However, the gastrointestinal bleeding occurred in an anemic patient with a benign ulcer in the stomach antrum. The patient with sepsis had preexisting pedal gangrene and up for amputation right before the study was conducted thus obscuring the effect of an ethanol lock to this adverse event. 18

DISCUSSION
Analysis of 7 studies found that the risk of CRBSI in ethanol lock was only 32% compared to heparin lock. Therefore ethanol lock significantly reduced the incidence of CRBSI in HD patients by 68%. These findings are similar and in line with the meta-analysis by Zhang P et al. (2018), with a decrease in the incidence of CRBSI about 34%. However, the population in the meta-analysis was diverse (not focused on only HD patients). 21 As the secondary outcome, the effect of ethanol lock on the incidence of catheter dysfunction has the same level as heparin lock, which is consistent with the meta-analysis by Zhang J et al. (2019), who also evaluated ethanol but in a more diverse population. 22 There are two ways for pathogen micro-organisms to enter the bloodstream and cause CRB. First, the extraluminal way describes organism contact on the

The intervention of ethanol lock to the incidence of catheter dysfunction in hemodialysis patients
The definition of catheter dysfunction in each study is provided in Table 1. From 7 included trials, only 4 (311 participants, skin with the catheter's outer surface, which migrates downward following the outer canal of the catheter to the tip where the hematogenous spread occurs. Second, the intraluminal way involves the transfer of organisms via the contact from an individual's (usually a healthcare worker) hand accessing the CVC or the patient's skin/surrounding clothing to the hub or catheter cap, resulting in contamination of the internal catheter surface. If the treatment of the exit site after insertion is done aseptically and properly, exit site infection can be minimized. However, the possibility of organism entry via the intraluminal route remains throughout catheter attachment. Intraluminal route infection begins within 24 hours of catheter insertion. Following adhesion, these micro-organisms form a structure called biofilms. A perfect biofilm consists of micro-organisms protected by a selfsecreted matrix of exopolysaccharides. Common organisms found in biofilms include Staphylococcus, Candida, and Pseudomonas. The presence of these biofilms facilitates ongoing and recurrent infection in patients with vascular catheters and the spread of hematogenous pathogenic micro-organisms. Therefore, the management of patients with CRBSI involves preventing micro-organisms attachment to the catheter and overcoming pathogenic biofilms by targeting the intraluminal catheter entry route. Moreover, heparin's current conventional lock is reported to play a role in forming pathogenic biofilms. 7, 10 Ethanol (30% to 70% concentration) possesses bactericidal properties and other advantages, namely inexpensive, ability to reduce biofilm formation, has no resistance to agents, has a broad spectrum of antimicrobial and antifungal properties. 6 Ethanol 70% works by damaging protein structures, unlike antibiotics, which may explain how ethanol reduces the risk of resistance. 23 The potency of ethanol as a catheter lock to reduce the incidence of systemic infection associated with vascular access catheter insertion has been evaluated from in vitro studies. A study by Alonso et al. in 2018 assessed the lowest effective concentration of ethanol in inhibiting biofilm formation in the catheter. The 40% ethanol reportedly does not affect the 60 IU heparin action, in which heparin is known as the standard vascular access catheter lock. A follow-up study was carried out on the pathological strain of CRBSI that was isolated directly from the patient from these results. Alonso et al. 2019, in their other study, evaluated a low concentration of ethanol lock of 40%-heparin 60 IU against clinical micro-organism strains (isolated from hospital CRBSI patients) with in vitro study models as a follow-up from previous studies using laboratory pathogenic strains. The result was that a low ethanol concentration of 40% for 72 hours reduces the metabolic activity of CRBSI organisms by 83% (no significant difference with that of 70% ethanol) and can be combined with heparin without reducing heparin activity which requires 24 hours of locking or more. In this study, it was also found that the composition of the ethanol solution was also efficient in reducing the biofilm mass in CRBSI by 50% of the clinical strain. However, unfavorable results were obtained in the regrowth inhibition test after 72 hours of locking. According to the study, microorganisms switch to a viable but nonculturable state (metabolically inactive) due to an unfavorable environment. 11 Apart from in vitro studies, the clinical use of ethanol has also been studied by several other studies. A review study on the use of 70% ethanol as a catheter locking in home parenteral nutrition (HPN) patients for the prevention of CRBSI showed good results, where several studies reported a decrease in the CRBSI rate from 8.3 to 2.7 per 1000 catheter days, and retrospective studies on 31 patients showed a decrease of 10.1 to 2.9 per 1000 catheter days. 23 A retrospective study by Kubiak et al. in 2014 on pediatric patients who received ethanol lock therapy with CVC catheter use stated that CVC in pediatric patients is often due to indications of chemotherapy and total parenteral nutrition. Thirty-five out of the 45 cases of CRBSI (78%) showed improvement in symptoms, and blood cultures became sterile after applying ethanol lock therapy for 4 to 12 hours daily for five days. 8 The advantages of ethanol lock in this retrospective clinical study in reducing CRBSI are in line with in vitro studies from Shrestha et al. at ≥30% concentration, in which ethanol is superior to several antibiotics (vancomycin, ciprofloxacin, minocycline, and rifampicin), with the ability of microbial eradication around 3,6 -3.9 log units at 2, 4, and 24 hours. At ≥35% concentration, ethanol has been shown to reduce the metabolic activity of matured Candida albicans biofilms by > 99%. The weakness of this retrospective study is the subjective observation of medical personnel treating patients with catheters and ethanol lock therapy. The result that there were no reports of side effects associated with ethanol lock can be explained by the minimal amount of ethanol that reached the systemic circulation in this study protocol. Most of it was aspirated after the dwelling period. Therefore, the ethanol lock solution administered in the CVC should be aspirated instead of flushed into the systemic circulation. 17 Aside from the need for parenteral nutrition and chemotherapy, long-term vascular access catheter in hemodialysis patients deserves attention due to the large population in need. Merikhi et al. (2019) evaluated the use of ethanol lock in pediatric hemodialysis patients suffering from CKD related to catheter-related infection compared to 10 mg clindamycin lock. Both regimens were combined with heparin. The antibiotic lock was administered 3x/week while ethanol was given 1x/week while the remaining sessions utilized heparin lock (2x/week). This resulted in decreased catheter site infection to only 12% in the ethanol lock group than the initial 44% in the antibiotic lock group. This study also observed that there was no protein precipitate sediment formation in the ethanol-heparin lock combination group. This indicates that the addition of heparin to ethanol as catheter lock can decrease the incidence of catheter thrombosis as a side effect of protein precipitation. 24 There are several possible side effects of ethanol lock, both to the catheter and to the body. Catheter-wise, ethanol is associated with catheter thrombosis and the mechanical integrity of the catheter. Catheter thrombosis is associated with protein precipitate that causes catheter occlusion, while ethanol is associated with reduced material integrity. Studies

ORIGINAL ARTICLE
revealed that protein precipitate starts forming in ethanol concentration of 28% or higher, but no precipitate was observed in heparin lock. This leads to the combined use of ethanol and heparin to minimize serum protein precipitate in catheters. 25 The use of a higher concentration of ethanol (70%) is associated with its effect on mechanical integrity. There was no significant difference in catheter integrity (stress/strain test and elastic modulus) in 70% ethanol and heparin after 26 weeks in either silicon or carbothane catheter group. 12 However, it is worth noting that this was an in vitro study using catheters from one producer. Therefore, ethanol is advisably used only with alcohol (ethanol)compatible catheters made of carbothane or silicon. Presumed systemic effects of ethanol including abnormal liver function test, mainly elevated transaminases, reports of headache, nausea, dizziness, and fatigue. These reports are more commonly found in studies that flushed, instead of aspirated, ethanol lock. 7,26 From here, several means to minimize the adverse effect of ethanol lock can be inferred: a) Only use ethanol as a lock for producers-approved catheters b) Use the minimum effective concentration (30%-40% is suggested) with minimum dwell time to preserve the integrity of the catheter, and use other additional agents (such as heparin) to prevent intraluminal obstruction. Ethanol is advisably aspirated instead of flushed following the end of dwell time. Observe for signs and symptoms of alcohol intoxication. 6 CDC and Infectious Diseases Society of America (IDSA) do not recommend routine use of antimicrobial lockin hemodialysis patients with CVC. Antimicrobial locks are recommended for patients with a history of recurrent CRBSI due to concerns of potential antimicrobial resistance. 7, 12 The European Best Practices Report, however, concluded that the effectiveness of antimicrobial lock to reduce CRBSI overcomes the risk and adverse effects, thus commending its use as a prophylactic in all ESRD patients with CVC. 12 Meanwhile, ethanol may be an alternative as an additional catheter locking agent due to its potency in reducing catheter-related systemic infection and near-zero risk of resistance up to this day.
A shortcoming of this study is that there are varied methods and interventions, notably the concentration of the ethanol and ethanol alone or in combination with other substances in each included study, posing as a potential bias of the analytic result. However, through the results of this study, ethanol as a nonantibiotic antimicrobial agent will be more convincing as a means to reduce CRBSI and is deemed safe for clinical practice.

CONCLUSION
Ethanol, both as an additional regimen to the standard heparin lock or as an individual catheter locking solution, is a potential prophylactic agent in preventing and reducing the incidence of catheterrelated bloodstream infection (CRBSI) in hemodialysis patients with vascular access in the form of either temporary or permanent tunneled catheter. Ethanol application as catheter lock effect on catheter dysfunction has the same level as standard heparin locks in general, therefore its effect on catheter dysfunction can be ignored. The key application of ethanol as therapy in CRBSI cases is clinically concerned with several factors, namely the concentration of ethanol, the method of locking (flushing or aspiration), the frequency of locking, the length of time of locking, and the combination with heparin. The authors assume that a higher concentration of ethanol used as a catheter lock will result in shorter required locking/dwelling time and less frequent usage, and vice versa. Therefore, further clinical studies are needed to determine the relationship between concentration, locking time, and usage frequency of ethanol lock.