A Novel Nonantibiotic Nitroglycerin-Based Catheter Lock Solution for Prevention of Intraluminal Central Venous Catheter Infections in Cancer Patients

ABSTRACT

For long-term central lines (CL), the lumen is the major source of central line-associated bloodstream infections (CLABSI). The current standard of care for maintaining catheter patency includes flushing the CL with saline or heparin. Neither agent has any antimicrobial activity. Furthermore, heparin may enhance staphylococcal biofilm formation. We evaluated the safety and efficacy of a novel nonantibiotic catheter lock solution for the prevention of CLABSI. Between November 2015 and February 2016, we enrolled 60 patients with hematologic malignancies who had peripherally inserted central catheters (PICC) to receive the study lock solution. The study lock consisted of 15 or 30 μg/ml of nitroglycerin in combination with 4% sodium citrate and 22% ethanol. Each lumen was locked for at least 2 h once daily prior to being flushed. After enrollment of 10 patients at the lower nitroglycerin dose without evidence of toxicity, the dose was escalated to the higher dose (30 μg/ml). There were no serious related adverse events or episodes of hypotension with lock administration. Two patients experienced mild transient adverse events (one headache and one rash) possibly related to the lock and that resolved without residual effect. The CLABSI rate was 0 on lock days versus 1.6/1,000 catheter days (CD) off lock prophylaxis, compared with a rate of 1.9/1,000 CD at the institution in the same patient population. In conclusion, the nitroglycerin-based lock prophylaxis is safe and well tolerated. It may prevent CLABSI when given daily to cancer patients. Large, prospective, randomized clinical trials are needed to validate these findings. (This study has been registered at ClinicalTrials.gov under identifier NCT02577718.)

INTRODUCTION

Central venous catheters (CVC) are imperative for the treatment and management of cancer patients, particularly those with hematological malignancies who require a long-term access for the administration of chemotherapy, antibiotics, blood products, and fluids as well as for frequent blood draws. However, these lifelines are the leading source of bloodstream infections, causing at least 400,000 episodes of bloodstream infections yearly in cancer patients (1) and contributing to up to 62% of the bloodstream infections in patients with long-term CVC (2, 3). These catheter-associated infections constitute a major public health challenge, with an estimated mortality rate of 10 to 25% (4) and estimated cost of $45,000 per episode (5).

The Centers for Disease Control and Prevention (CDC) recognizes the gravity of central line-associated bloodstream infection (CLABSI) and have published guidelines for the prevention and management of this frequent and serious infection (6, 7).

For long-term CVC that remain in place for more than 10 days, the lumen of the CVC is the main source of colonization and subsequent bacteremia (8, 9). Solutions with antimicrobial, antibiofilm activity instilled in the lumen of the CVC may eradicate organisms and prevent biofilm formation, hence eliminating the source of CLABSI and preventing these serious infections (10).

The current standard of care for maintaining catheter patency includes flushing the lumen of the catheter with either a saline or heparin lock solution. Neither of these agents has any activity against organisms embedded in biofilms (11). Furthermore, heparin has been shown to enhance staphylococcal biofilm formation and could therefore predispose to CLABSI (12). The use of prophylactic antibiotic locks raises concerns about the emergence of antibiotic resistance (13). Lock solutions containing high concentrations of ethanol (70%) alone require prolonged exposure and have been associated with adverse events (AEs) (14). Ethanol concentrations higher than 28% may also cause protein precipitation (15) and have been implicated in compromising catheter polymers (16).

Recent in vitro studies of a nonantibiotic, heparin-free solution containing nitroglycerin (NTG) in combination with ethanol and citrate demonstrated a broad spectrum of antimicrobial efficacy and an ability to eradicate biofilm formation as well as an anticoagulant activity similar to that of heparin (17, 18).

We therefore conducted a pilot study to evaluate the safety and effectiveness of this novel nonantibiotic, nonheparin, nitroglycerin-based catheter lock solution used for the prevention of intraluminal central venous catheter infections in cancer patients.

RESULTS

We prospectively enrolled 60 high-risk patients with hematological malignancies and/or stem cell transplant to receive a nonantibiotic chelator-based lock solution that contained 15 or 30 μg/ml nitroglycerin in combination with 4% sodium citrate and 22% ethanol (NiCE) lock solution in their dual-lumen peripherally inserted central catheters (PICCs). The patients' median age was 67 years (range, 26 to 85 years), and 63% were male (Table 1). Of the 60 patients, 10% had received an allogeneic stem cell transplant, 7% experienced graft-versus-host disease, and 60% had a history of hypertension or were receiving antihypertensive medications (Table 1). The main usage for the PICC was administration of chemotherapy (95%), transfusions (88%), and antimicrobials (78%). For the antibiotic use, 68% of the patients received antibiotics on the days they received the locks and 68% received antibiotics on the days their PICCs were not locked. Similarly, there was no statistically significant difference in the percentage of patients receiving antifungals on days on versus days off lock (43% versus 48%; P = 0.58).

TABLE 1.

Baseline characteristics of the patients

Characteristic	Value for patients (n = 60)
Median age (range), yrs	67 (26–85)
Males, no. (%)	38 (63)
Type of cancer, no. (%)
Acute lymphatic leukemia	13 (22)
Acute myelogenous leukemia	39 (65)
Acute biphenotypic leukemia	1 (2)
Chronic lymphatic leukemia	1 (2)
Chronic monocytic leukemia	3 (5)
Lymphoma	1 (2)
Aplastic anemia	1 (2)
Myelofibrosis	1 (2)
HCT,a no. (%)	6 (10)
Graft-vs-host disease, no. (%)	4 (7)
Median days between HCT and study enrollment, no. (range)	443 (99–1,942)
Permanent foreign body (prosthetic valve), no. (%)	1 (2)
Hypertension or antihypertensive medications, no. (%)	36 (60)
Median SD of systolic blood pressure (range)	7.86 (1.15–41.48)

^aHCT, hematopoietic cell transplant.

The PICCs had been in place for a median duration of 37 days (range, 14 to 555 days) prior to enrollment. During the course of the study, 619 lock/flushes were administered to the 60 patients enrolled. The patients received a median of 9 days of lock prophylaxis (range, 1 to 24 days). Two patients did not want to continue to participate in the study and withdrew their consent after the first lock dose, but this was not due to any lock-related adverse event. In 47% of the patients both lumens were locked simultaneously, whereas in 40% both lumens were simultaneously locked on some days, and in 13% lumens were locked subsequently (Table 2).

TABLE 2.

Clinical data and catheter information

Variable	Value for patients (n = 60)
Use of catheters,a no. (%)
Chemotherapy	57 (95)
Transfusion	53 (88)
Antibiotics	47 (78)
Steroids	47 (78)
Interleukin	4 (7)
TPNb	2 (3)
Electrolytes	1 (2)
Median days of catheter use before lock prophylaxis, no. (range)	37 (14–555)
Dose of nitroglycerin administered in the lock, no (%)
15 μg/ml	10 (17)
30 μg/ml	50 (83)
Lumens locked, no. (%)
Simultaneously all the time	28 (47)
Not simultaneously all the time	8 (13)
Sometimes simultaneously and sometimes not	24 (40)
Median days of lock prophylaxis, no. (range)	9 (1c–24)
Catheter removal or exchange over guidewire, no. (%)	16 (27)
Median days between starting lock prophylaxis and catheter removal, no. (range)	18 (1–46)

^aMany patients used their catheters for multiple purposes.

^bTPN, total parenteral nutrition.

^cTwo patients did not want to continue to participate in the study and withdrew their consent after the first lock dose. Neither patient reported any lock-related adverse event.

The first 10 patients received the lower dose of nitroglycerin (15 μg/ml) without the occurrence of any episode of hypotension. All the subsequent 50 patients were enrolled at the higher dose (30 μg/ml) and continued to receive the higher dose according to the safety decision rules for dose escalation. None of the patients experienced drug-related hypotension as defined above.

The estimated mean readings of the three baseline blood pressure (BP) readings within 72 h prior to the first dose of lock administration were 138 mm Hg for systolic BP (SBP) and 76.2 mm Hg for the diastolic BP for all study patients.

For the 60 patients, there were 189 BP measurements during the first 2 weeks that they were enrolled in the study. The estimated means of SBP of the three readings were 137.4 mm Hg (SBP1, prior to lock instillation), 135.8 mm Hg (SBP2, 2 h after lock instillation but prior to flushing), and 136.9 mm Hg (SBP3, after flushing the lock) (Fig. 1). Analysis showed no statistical difference between any two of the three readings (P = 0.52). Analysis of diastolic BP changes showed similar results (P = 0.17). Overall, there were 280 BP measurements for the 60 patients during their 30 days of participation in the study. The analysis showed similar results, and there was no hypotension observed with any of the doses administered (15 μg/ml and 30 μg/ml). Two patients (3%) developed mild transient adverse events that might have been related to the lock: one patient who had a history of headaches experienced a mild headache that lasted only 1 min following flushing of the lock (after the 4th dose) and did not recur with subsequent administration of the study lock, and another patient reported a rash after receiving 23 daily doses of the medication, without incident. This rash resolved in 2 days. Both adverse events led to discontinuation of study drug per the patients' request.

FIG 1.

Lock administration and blood pressure (BP) monitoring during the first 2 weeks for all patients. Doses of nitroglycerin are in micrograms per milliliter. Systolic and diastolic BPs are in millimeters of mercury.

No CLABSI/catheter-related bloodstream infection (CRBSI) was observed during the 619 catheter days (CD) with lock prophylaxis, compared to 3 episodes of CLABSI/CRBSI during the 1,853 days without lock prophylaxis (P = 0.32). The incidence density for CLABSI reported per 1,000 CD) was 0/1,000 CD during lock days versus 1.6/1,000 CD during days without the lock (95% confidence interval [CI], 0.5 to 5.0 per 1,000 CD). The organisms causing CLABSI were Escherichia coli (2 cases) and Pseudomonas and Fusarium (polymicrobial in one case).

Compared to the historical control group of patients admitted to the same wards during the months of September and October 2015 that immediately preceded the study, infection control reported 8 episodes of CLABSI occurring with 4,285 CD which contributed to a CLABSI incidence density of 1.9/1,000 CD, compared to 0 in the lock group (95% CI, 0.9 to 3.7 per 1,000 CD) (Fig. 2).

FIG 2.

Comparison of CLABSI incidence density. Control 1 patients are the patients enrolled in this study during the time they were not receiving daily lock prophylaxis; control 2 patients consist of the historical similar patient population admitted to these hospital wards prior to the study (CLABSI/CRBSI rate reported by infection control).

The PICC was removed in 16 patients (27%) after a median duration of 18 days (range, 1 to 46 days).

PICC dysfunction was reported for one patient (2%) and consisted of complete occlusion of all lumens, resulting in removal of the catheter. However, this occurred when the patient was receiving heparin and was off NiCE lock prophylaxis.

Five patients (8%) died during the study period after a median of 26 days (range, 20 to 36 days), but their deaths were unrelated to the lock.

DISCUSSION

In this phase I/II open-label study, we demonstrated that NiCE lock/flush is safe, is associated with minimal and reversible adverse events (AEs), and has the potential to be efficacious in preventing CLABSI/CRBSI in high-risk cancer patients with long-term CVC.

In the 60-patient study group, the rate of CLABSI/CRBSI dropped from 1.6/1000 catheter days when the patients were not receiving the NiCE lock to 0 while patients were on the study lock. Only two transient AEs were reported during this study. These were associated with two NiCE lock administration in the setting of 619 NiCE lock/flush doses that were administered to the patients during the study.

Our group was the first to show that subpharmacological concentrations of nitroglycerin (NTG) are synergistic with 4% citrate and 22% ethanol. This triple-combination NiCE lock, which is a combination of subpharmacological concentrations of FDA-approved injectable agents, is capable of eradicating all resistant bacteria in addition to various Candida species in biofilms (17, 18).

All of these pathogens are well known to cause CLABSI (4, 7). Furthermore, this combination was found to be superior to taurolidine in eradicating resistant organisms such as vancomycin-intermediate Staphylococcus aureus (VISA) and Candida glabrata (17). Hence, the clinical results of this study demonstrating that NiCE completely prevented the occurrence of CLABSI are consistent with the high efficacy of NiCE demonstrated in our in vitro model in completely eradicating these organisms in a biofilm colonization model within 2 h (17).

No thrombotic occlusion was reported when the patients were receiving the NiCE component compared to heparin. The NiCE antithrombotic activity was clinically equivalent to that of heparin in this study. We did not observe any case of complete occlusion of all the lumens necessitating removal of the PICC during lock prophylaxis. Only one patient had an occlusion of the PICC, and this occurred while he was on heparin and off the NiCE lock.

These data are consistent with our previous data showing anticoagulant activity of NiCE equivalent to that of heparin, with the distinction of better antimicrobial activity with NiCE (17). This activity is related to the role that each component of NiCE plays.

This lock combination contains subpharmacological doses of nitroglycerin. Nitroglycerin inhibits platelet activation and hence contributes to the anticoagulant activity of the NiCE lock (19, 20). In addition, nitroglycerin has been used in vitro extensively and as a flush at a higher concentration and was previously shown to be well tolerated. Nitroglycerin has long been recognized to have therapeutic benefit in treating hypertension and in wound healing (21). Nitroglycerin has an extremely short half-life in blood (a few minutes) and is very rapidly metabolized and excreted (21). Clinical studies have demonstrated the safe use of bolus intravenous (i.v.) nitroglycerin for uterine relaxation in complicated obstetric deliveries (22) and in emergency treatment of congestive heart failure at concentrations and dose levels that were 8- to 67-fold higher than the 15- to 30-μg/ml concentrations used in this study (23, 24). Two trials report that bolus doses of 2 mg of nitroglycerin given up to 10 times in 3- to 5-min intervals were tolerated in emergency treatment of heart failure (23, 24). Boluses of 250 to 500 μg were well tolerated by both fetus and mother in separate clinical trials with pregnant females (25). Furthermore, nitroglycerin is known to be metabolized by nitrate reductase enzymes to nitric oxide (26), which has been shown to exert broad-spectrum antimicrobial activity against viruses, bacteria, fungi, and parasites (27–29).

In addition, this lock contains 4% citrate. This dose has been approved by the FDA and has been extensively used in blood transfusions and hemodialysis. The mechanism of anticoagulation leading to this antithrombotic activity is the fact that citrate is a chelator that binds to factor IV and has intrinsic anticoagulant activity (30, 31). Citrate was shown in a large multicenter trial to provide antithrombotic activity comparable to that of heparin against catheter occlusion (32). Citrate at 4% has been shown to have outcomes with regard to flow-related catheter exchange or tissue plasminogen activator (TPA) use equivalent to or better than heparin locking (33, 34). Moreover, citrate also avoids heparin-associated bleeding complications and provides an effective alternative for patients with suspected or confirmed heparin-induced thrombocytopenia (34). Moreover, as a chelator of iron, calcium, and magnesium, citrate has been shown to be a microbial biofilm disruptor, since these metallic ions are necessary components of the biofilm matrix (35).

Furthermore, this lock contains only 22% ethanol to avoid any adverse event related to alcohol and to prevent any damage to the catheter polymer. Ethanol has broad-spectrum antimicrobial activity. However, as a single agent, ethanol has not been shown to be clinically effective as a lock solution even when concentrations higher than 50% were used (14). The use of high concentrations of ethanol as a catheter lock solution may be associated with side effects such as headache, dizziness, nausea, light-headedness, and dyspnea (14). In addition, at concentrations of >28%, ethanol has been shown to cause protein precipitation (15), and concentrations of ≥60% have been reported to cause mechanical damage to the integrity of the polymer (16). The ethanol concentration used in this lock, 22%, was well within the safe range.

Hence, the safety profile of the subpharmacological concentration of the three components of NiCE as outlined above provides the biologic explanation as to why our patients reported only two mild and transient adverse events during the course of the study. The first one was a headache that lasted a minute and did not recur with subsequent administration of the study lock. The second adverse event was a rash that appeared after 23 doses and resolved in 2 days. This transient rash could be related to the lock or to other medications that the patient was receiving concomitantly or a transfusion received earlier the same day. There is only one case report in the literature of a rash that was associated with i.v. administration of nitroglycerin, and it occurred after the patient developed a rash in response to dermal nitroglycerin.

Our study has several limitations. We acknowledge that this was a single-center, nonrandomized, open-label study with a small sample size. We did not have enough power to show efficacy in preventing CLABSI. The patients were able to receive the NiCE lock on a daily basis only as inpatients. Our results must be confirmed in a large multicenter randomized control trial.

In conclusion, our novel NiCE triple-combination, nitroglycerin-based antibiotic-free lock solution was well tolerated and may prevent CLABSI. This NiCE lock could be useful in cancer patients, transplant recipients, hemodialysis patients, and patients with short-bowel syndrome who require long-term PICCs. Because this lock does not contain any antibiotic, there is less concern regarding antimicrobial resistance. It has the potential to be highly efficacious and needs to be tested in a phase III trial.

MATERIALS AND METHODS

Patient population.

We enrolled 60 patients in a prospective open-label, single-institution, two-dose pilot study at the University of Texas MD Anderson Cancer Center (MDACC) in Houston, TX, between November 2015 and February 2016. This study has been registered at ClinicalTrials.gov under registration no. NCT02577718.

Eligible patients were patients who were hospitalized at the MD Anderson Cancer Center and were 18 years of age or older. We targeted cancer patients with hematological malignancies who were at higher risk for catheter infections and who had a 5.0 French size, dual-lumen PICC that had been in place for at least 14 days. The PICC was anticipated to remain in place for at least 30 days (Fig. 3).

FIG 3.

Patient timeline. NiCE, nitroglycerin, citrate, and ethanol lock; PICC, peripherally inserted central catheter.

Patients were ineligible if any of the following was true: they had minocycline-rifampin-coated CVC, were hypotensive with a systolic blood pressure (SBP) reading of <110 mm Hg at any time over the 3 days prior to study entry, were not awake, were not alert, could not express pain or discomfort related to the catheter locks, had an existing local or systemic infection, or had signs and symptoms of localized catheter-related infection (tenderness and/or pain, erythema, swelling, or purulent exudates within 2 cm of the entry site), or had an occluded catheter, defined as a catheter in which one could neither withdraw blood nor instill 3 ml of fluid without resistance through any lumen. We excluded patients with multiple coexisting central venous catheters at the time of enrollment, who had a known history of allergic reaction to ethanol, nitroglycerin (NTG), or citrate, who were on disulfiram or metronidazole, who were dependent on alcohol determined by a single affirmative response to the CAGE questionnaire (36), or who were receiving phosphodiesterase type 5 (PDE-5) inhibitors (such as sildenafil, tadalafil, and vardenafil).

Intervention.

This was a 2-dose trial of a nonantibiotic chelator-based lock solution that contained 15 or 30 μg/ml of nitroglycerin in combination with 4% sodium citrate and 22% ethanol (NiCE) used to prevent CLABSI.

After providing informed consent, the first 10 patients were enrolled at the lower dose of NTG (15 μg/ml). Subsequent patients were enrolled at a higher dose of NTG (30 μg/ml) according to safety-driven decision rules for dose escalation.

Patients initially had their PICCs flushed with normal saline. BP, heart rate (HR), and symptoms and adverse events were recorded prior to instillation of the lock solution and served as the baseline. After the instillation of the lock solution, BP, HR, and symptoms and adverse events associated with the lock prophylaxis (including local catheter site pain or irritation, headache, facial flushing, alcohol taste, dyspnea, palpitation, hypotension, peripheral edema, syncope, dizziness, and catheter dysfunction leading to removal of the PICC, including total occlusion or rupture of the catheter) were recorded. BP and HR were checked 2 h after instillation, both before and after flushing of the lock solutions, and were compared to baseline.

Patients received a standard catheter flush (saline) followed by the study NiCE catheter lock solution on a daily basis as long as they were inpatients, for at least 2 days and up to 30 ± 7 days. After discharge, patients may have continued to receive the study lock solution on an outpatient basis at MDACC once a week for up to 30 ± 7 days.

Each catheter lumen was filled with a volume of 0.8 to 1 ml of NiCE lock solution and the solution was allowed to remain in the catheter for at least 2 h once daily, prior to being flushed with 10 ml of normal saline. We preferentially locked all the lumens at the same time. However, if this was not feasible, lumens were locked subsequently at different times.

Control.

In this study, every patient served as his own control while he was not receiving daily study lock prophylaxis but was receiving a heparin or saline lock. In addition, the rate of CLABSI infections was compared to the rate reported by infection control for the same patient population admitted to these hospital wards prior to the study activation during the same fiscal year.

Follow-up and outcome.

Blood pressure measurements were checked at baseline prior to instillation of NICE and then rechecked 2 h later both before and after flushing the NiCE lock solution with 10 ml of saline (Fig. 1). Blood pressure measurements related to the lock were done twice weekly while the patient was hospitalized and once weekly in the outpatient setting.

During the study period, patients were monitored very closely for development of adverse events, catheter function, and the development of CLABSI.

We monitored the adverse events associated with the lock prophylaxis (including local catheter site pain or irritation, headache, facial flushing, alcohol taste, dyspnea, palpitation, hypotension, peripheral edema, syncope, dizziness, and catheter dysfunction, including total occlusion or rupture of the catheter requiring removal of the catheter) and the impact of adverse events on possible study withdrawal or PICC removal. These lock-related AEs were evaluated within 24 h after the instillation of the NiCE lock on the first and last days.

The CLABSI/CRBSI rate for the enrolled patients while receiving daily lock prophylaxis was compared to the CLABSI/CRBSI rate for the same patients enrolled in this study while they were not receiving daily lock prophylaxis, as well as to the CLABSI/CRBSI rate reported by infection control for the same patient population admitted to these hospital wards prior to the study activation.

Documentation of a CLABSI resulted in early completion for that patient; however, patients who received at least one dose of the study lock were included in the safety outcome analysis. Patients were monitored for safety for an additional 30 ± 7 days following the last dose of the study lock solution.

Study data were collected and managed using Research Electronic Data Capture (REDCap) tools hosted at the MD Anderson Cancer Center (37). REDCap is a secure, web-based application designed to support data capture for research studies, providing (i) an intuitive interface for validated data entry, (ii) audit trails for tracking data manipulation and export procedures, (iii) automated export procedures for seamless data downloads to common statistical packages, and (iv) procedures for importing data from external sources.

Definitions.

Drug-related hypotension was defined as a drop in SBP that exceeded the normal BP variability of a patient by 30%, that was associated with clinical signs and symptoms (dizziness and syncope), and that was unexplained by factors other than the NTG lock (such as other antihypertensive drugs, sepsis, and bleeding).

The BP variability was calculated based on the standard deviation (SD) of a patient's blood pressure measured in the 3 days preceding participation in this trial.

PICC-associated bloodstream infections were defined as either CRBSI as defined by the Infectious Diseases Society of America (IDSA), particularly in patients with mucosal barrier injury (MBI), such as neutropenic patients, or intraluminal CLABSI, as defined by the Centers for Disease Control and Prevention (CDC), particularly in non-MBI patients. Patients with evidence of CLABSI/CRBSI that occurred on days on which the NiCE lock was not administered were not considered as treatment failure.

Statistical analysis.

Descriptive statistics were used to summarize patients' demographic, clinical, catheter use, and lock therapy information. They were also used to analyze patients' lock therapy-related hypotension, adverse events, and CLABSI/CRBSI data. This was a two-dose safety and efficacy trial. The decision rules about safety monitoring and dose escalation were determined using modified toxicity probability interval methods (mTPI) (38, 39). Linear regression analysis with repeated measures was performed to compare patients' blood pressure measured at the three different time points: prior to instillation of daily lock solution in the catheter lumen, 2 h after instillation of the lock but prior to flushing, and after flushing of the lock. The Poisson rates of CLABSI/CRBSI per 1,000 catheter days were estimated and compared between same patients with and without lock therapy during the study period. The CLABSI/CRBSI rates were also compared between patients when receiving lock therapy and a historical control group coming from the same patient population. All the tests were 2-sided tests with a significance level of 0.05. The stopping rules were generated in the mTPI Excel macro. The operating characteristics were calculated in the mTPI R package. The data analyses were performed using SAS version 9.3 (SAS Institute Inc., Cary, NC).