Evaluation of central line salvage for mucosal barrier injury laboratory-confirmed bloodstream infection (MBI-LCBI) management practices in patients with hematologic malignancies

Abstract

Patients with underlying hematologic malignancy (HM) and/or allogeneic hematopoietic stem cell transplant (HCT) recipients are at risk for mucosal barrier injury laboratory-confirmed bloodstream infection (MBI-LCBI) secondary to bacterial translocation. There is sparse data comparing MBI-LCBI management practices, in particular central venous catheter (CVC) salvage versus removal. We created a 22-item poll of Infectious Disease specialists at major US cancer centers on management controversies. Response rate was 44% (31/70). CVC salvage was a common practice among 87.5%. This was followed by a single center retrospective study (2017–2019) comparing outcomes related to CVC practices. We identified 115 patients, 52% (60/115) admitted for chemotherapy and 33% (38/115) for allogeneic HCT. The majority of patients (78%, 90/115) had their CVC removed. There was no difference in 72 h defervescence, microbiological clearance, in-hospital mortality, and 90-day recurrent infection between CVC salvage versus removal. CVC salvage is a safe approach in certain clinical scenarios.

Keywords:

• Bacteremia
• hematologic malignancies
• transplantation
• mucosal barrier injury

1. Introduction

Blood stream infection (BSI) is a major cause of neutropenic fever in patients with hematologic malignancy (HM) and those undergoing hematopoietic stem cell transplantation (HCT), ranging in incidence from 10% to 38% [1–4]. In 2013, the Centers for Disease Control and Prevention (CDC) and the National Healthcare Safety Network (NHSN) introduced a new definition of mucosal barrier injury laboratory-confirmed bloodstream infection (MBI-LCBI) to describe BSIs that occur in patients such as those with neutropenia or gastrointestinal graft versus host disease (GVHD), that are secondary to gut flora translocation rather than a direct catheter related infection [5,6]. This resulted in an at least 30% reduction of reporting of central line-associated bloodstream infection (CLABSI) rates among cancer centers [7].

Thus, while the epidemiology, risk factors and clinical outcomes of MBI-LCBI have been explored in the literature [4,8–11], there remains scarce data to guide management practices of MBI-LCBIs, including central venous catheters (CVC) salvage vs removal [8,9,12]. Practices vary widely by center. For example, in one study (31/80, 39%) of patients with any MBI-LCBI had their catheter removed [8], while in another center (54/55, 98%) of adult patients with Viridans group streptococci MBI-LCBI had their CVC retained [13]. Furthermore, management of CVCs in patients with MBI-LCBI can be additionally challenging as these patients are often thrombocytopenic with poor intravenous access. As it stands, U.S. and European guidelines recommend catheter removal in clinical situations associated with hemodynamic instability, thrombophlebitis, persistent bacteremia or presence of certain pathogens such as Pseudomonas spp, Stenotrophomonas maltophilia, and Acinetobacter spp, Candida spp and Mycobacteria [2,14] with no guidance otherwise.

In order to have a contemporary understanding of the impact of these practice differences, we created an anonymous poll to gauge MBI-LCBI management practices among Infectious Diseases physicians at U.S. cancer centers. This was followed by a retrospective study at our cancer center examining clinical outcomes associated with CVC management practices (removal vs salvage) in patients with MBI-LCBI.

2. Methods

2.1. Survey

We developed a 22-item online anonymous survey via Qualtrics software (Qualtrics, Provo, UT) to explore practitioners’ current knowledge, attitudes, and MBI-LCBI management practices among patients with HM and HCT recipients. In August 2018, the survey was distributed to 70 infectious disease consultants at 35 U.S. cancer centers and was open for 30 days. Participants agreed to participate in the survey without any compensation. The questionnaire consisted of three parts: respondent characteristics including geographic location and years of clinical experience, questions regarding their institution’s management of CVCs in patients with MBI-LCBI, and review of two clinical vignettes

2.2. Retrospective study

We retrospectively reviewed electronic medical records of consecutive adult patients (≥18 years of age) diagnosed with an underlying HM and/or admitted for allogeneic HCT to the Ohio State Comprehensive Cancer-James Cancer Hospital and Solove Research Institute between January 2017 to December 2019. Patients who met the definition for MBI-LCBI [5] with a qualifying central venous catheter were identified using the infection prevention BSI database. Data were collected from electronic medical records and managed using REDCap electronic data capture tools [15]. Patients were followed for 90 days from the day of MBI-LCBI diagnosis. Patients with thrombophlebitis, indwelling cardiac device, lacking CVC and prisoners were excluded. For patients with multiple episodes of bacteremia, only the first MBI-LCBI was included and analyzed during the study period.

The primary objective was to assess 90-day recurrence rate of BSI episodes between patients who had their CVC removed vs salvaged. The secondary objectives were to analyze the rates of defervescence within <72 h, microbiological clearance within <72 h, intensive care unit (ICU) transfer, in-hospital mortality, 90 day all-cause mortality, 90 day infection related mortality, and hospital length of stay. The study was approved by The Ohio State Cancer Institutional Review Board.

In our cancer center, neutropenic patients admitted for allogeneic HCT or induction chemotherapy do not routinely receive fluoroquinolone prophylaxis. Infection prevention practices include oral care and daily bathing with chlorhexidine gluconate soap. We also have a built-in order set for routine oral care that include sodium bicarbonate/saline mix swish and soft tooth brushing four times a day and as needed.

3. Study definitions

MBI-LCBI was defined as having bacteremia with one or more oral or gastrointestinal commensal bacteria in the setting of neutropenia (<500 cells/ul) and/or having gastrointestinal GVHD with ≥20 mL/kg diarrhea in a 24-hour period [5]. Only eligible organisms as outlined by the CDC were included in the analysis [5]. Recurrence was defined as occurrence of infection with the same organism regardless of site within 90 days of MBI-LCBI diagnosis. Baseline characteristics were collected at the time of MBI-LCBI including demographics, HM disease type, HCT characteristics, BSI related variables, CVC type, presence of GVHD or neutropenia (defined as absolute neutrophil count <500 cells/ul). We utilized Pitt bacteremia score (PBS) as a predictor of acute illness severity and early mortality as previously described [16]. PBS ranges from 0 to 14 points, with a score ≥4 commonly used as an indicator of critical illness and increased mortality. In-hospital all-cause mortality was defined as any death during the same admission for the bacteremia regardless of the cause.

4. Statistical analysis

The survey participant’s responses were summarized using descriptive statistics. The characteristics of patients and clinical outcomes were described using medians and ranges for continuous variables and frequencies and percentages for categorical variables. The comparison between the catheter salvage and catheter removal groups were conducted using Wilcoxon rank sum test for continuous data and Chi-square or Fisher’s exact test for categorical data. Time to bacteremia in transplant patients was calculated from the date of transplant to onset of bacteremia treating different types of bacteremia as competing risks and the cumulative incidence of gram negative, gram positive and polymicrobial were calculated respectively. Overall survival was defined from the date of bacteremia to the date of death censoring those who were alive at the date of last contact. Overall survival rates were estimated using the Kaplan–Meier method and compared between catheter salvage and catheter removal groups using the log rank test. All the statistical tests were two-sided and the significance level was set at 0.05. Stata 16 (College Station, Texas) was used for all the analyses.

5. Results

5.1. Survey analysis

A total of 70 Infectious Disease consultants at U.S. cancer centers received the anonymous poll in August 2018. The 30-day response rate was 44% (n = 31), of whom 77% (n = 24) fully completed all survey questions. Responders were from at least 13 U.S. cancer centers with wide geographic distribution (South 44%, Midwest 39%, West 13%, Northeast 4%). Half of the respondents had more than 10 years of clinical experience. Respondents’ comments on performing certain diagnostic and therapeutic practices for management of MBI-LCBI are illustrated in Figure 1. The majority of respondents (76%, 19/25) do not implicate the catheter as the source of MBI-LCBI. 87.5% of respondents salvage the CVC in cases of intestinal/oral flora MBI-LCBI not meeting guidelines for immediate catheter removal. Of those who salvage the catheter, 62.5% sometimes use adjunct lock therapy.

Figure 1. Survey results on diagnostic and therapeutic practice variations for mucosal barrier injury blood stream infections among cancer centers.

Based on the two clinical vignette responses (Escherichia coli and Streptococcus viridians uncomplicated MBI-LCBI in hemodynamically stable patients), 19% of respondents would utilize microbiological differentiation methods to determine the decision regarding CVC salvage vs removal in both scenarios. More than half of the respondents would salvage the CVC in an HCT recipient with transient uncomplicated pan-sensitive Escherichia coli (67%) or Streptococcus viridians (62%) MBI-LBI. The use of combination salvage and adjunct lock therapy was less common among respondents, 9% and 14% for Escherichia coli or Streptococcus viridians, respectively. Only 5% of respondents would remove the CVC in either scenario.

5.2. Retrospective study results

5.2.1. Cohort Characteristics

A total of 115 patients were included. The median age of the cohort at bacteremia onset was 59, range 21–81 and 61% were male. The majority of patients (90/115, 78%) had their CVC removed, but median time to blood culture clearance did not differ between CVC removal and salvage groups (median 1, range; 0–6 vs median 1, range 0–3, respectively; p = 0.34). 96% (86/90) of patients in the CVC removal group, had their CVC removed within 3 days of BSI diagnosis, among which 11 patients (11/86, 13%) had the line removed on the same day (data not shown). The majority of patients (111/115, 97%) were neutropenic at the time of MBI-LCBI. Baseline characteristics are found in Table 1. The CVC removal group had a higher PBS range compared to the salvage group (median 1, range 0–4 vs median 0, range 0–2, p = 0.001). The pattern of positivity from both CVC and periphery was variable among both groups but only 22/90 (24%) were positive from catheter only in the CVC removal group compared to 9/25 (36%) in the salvage group (p < 0.001). Only 7 patients with salvaged CVC received adjunct lock therapy; range of therapy duration was 4–14 days. Transplant characteristics for those admitted for HCT are described in Supplementary Table S1.

Table 1. Baseline characteristics of patients at the time of mucosal barrier injury blood stream infection.

	Catheter removal n = 90		Catheter salvage n = 25		P-value
	N	%	N	%
Age, median (range)	59.5 (23–83)	–	63(22–74)	–	0.58
Age at bacteremia, median (range)	58.5 (22–81)	–	62 (21–73)	–	0.72
Gender
Men	52	57.6	18	72	0.25
Women	38	42.4	7	28
Malignancy type
Acute myeloid leukemia	53	58.9	13	52	0.09
Myelodysplastic syndrome (MDS/MPN/CML/CMML)	17	18.9	4	16
Acute lymphocytic leukemia	7	7.8	7	28
Non-Hodgkin’s lymphoma	8	8.9	0	0
Hodgkin’s lymphoma	2	2.2	1	4
Multiple Myeloma	3	3.3	0	0
Neutropenia < 500 cell/microL	87	96.7	24	96	0.99
Reason for admission
Chemotherapy	50	55.6	10	40	0.21
HCT^a and/or cellular therapy	26	28.9	12	48
Others	14	15.6	3	12
Type of bacteremia
Gram negative	54	60	14	56	0.84
Gram positive	28	31.1	8	32
Polymicrobial	8	8.9	3	12
Pitt bacteremia score, median (range)	1	0–4	0	0–2	0.001
Catheter type
Tunneled	54	60	16	64	0.39
Peripherally inserted central catheter	31	34.4	6	24
Port	5	5.6	3	12
Source of positive blood culture
Catheter and periphery	60	66.7	9	36	<0.001
Catheter only	22	24.4	9	36
Peripheral only	8	8.9	7	28
Days to blood culture clearance, median (range)	1	0–6	1	0–3	0.34

CMML: Chronic myelomonocytic leukemia; CML: Chronic myeloid leukemia; MDS: Myelodysplastic syndromes; MPN: Myeloproliferative neoplasms.

^aHCT: Hematopoietic stem cell transplantation.

5.2.2. Bacteremia episodes

About 59% of MBI-LCBI episodes were caused by gram negative bacteria and 9.6% were polymicrobial. Escherichia coli and Alpha-hemolytic Streptococcus were the most common pathogens reported (Table 2). 12% of patients in the CVC salvage group had hypotension (systolic blood pressure <100) at the time of BSI compared to 22% of patients in the CVC removal group, p = 0.40. Infectious disease consultation was obtained in (64/115, 55.6%) of all MBI-LCBI episodes with no difference between the CVC removal versus CVC salvage (54% vs 60%, p = 0.66). For patients undergoing allogeneic HCT, the median time from HCT to first bacteremia was 9 (1–15) days (Figure 2).

Figure 2. Cumulative incidence of mucosal barrier injury blood stream infection among allogeneic stem cell recipients (n = 38).

Table 2. Pathogens Causing Mucosal Barrier Injury Blood Stream Infection.

Types of bacteria	Catheter Removal	Catheter Salvage
Gram Negative	n = 69 (%)	n = 17 (%)
Escherichia coli	25 (37)	7 (41)
Klebsiella  spp.	22 (32)	5 (29)
Enterobacter spp.	13 (19)	2 (11)
Citrobacter spp.	4 (5)	2 (11)
Serratia Marcescens	3 (4)	0 (0)
Proteus mirabilis	0 (0)	1 (6)
Capnocytophaga	1 (1)	0 (0)
Leptotrichia	1 (1)	0 (0)
Gram Positive	n = 51	n = 17
Alpha-hemolytic Streptococcus	20 (44)	5 (33)
Enterococcus spp	10^a (22)	2^b (13)
Fusebacterium	7 (15)	4 (27)
Streptococcus gallolyticus	2 (4)	0 (0)
Clostridium	4 (8)	2 (13)
Bifidobacterium	3 (6)	1 (6)
Granulicatella Adiacens	2 (4)	1 (6)
Rothia	1 (2)	0 (0)
Others	3^c (6)	2^d (13)
Total	N = 120	N = 34

^aE faecium (Vancomycin resistant), 4 E gallinarum, 1 E Hira.

^bPathogens are 1 E Hira, 1 E faecium (Vancomycin susceptible).

^cPathogens are 1 Finegoldia, 2 Gemella.

^dPathogens are Leuconstoc lactis and Rumminocus.

5.2.3. Clinical Outcomes

Clinical outcomes of CVC management are summarized in Table 3. None of the patients who had their CVC salvaged had recurrent bacteremia at 90 days with the same initial pathogen nor required ICU transfer. Approximately 58% (52/90) of patients who had their CVC removed underwent a new CVC placement within the same hospital stay. There was no difference in 72 h defervescence or microbiological clearance for patients who underwent CVC salvage versus removal (p= 0.30, p= 0.99, respectively). 90-day all-cause mortality rate was similar among both groups; 19% in the CVC removal group and 24% in the CVC salvage group, p= 0.58. Figure 3 illustrates 90-day survival of all cohort and HCT group based on CVC management. In-hospital all-cause mortality was numerically higher in the CVC removal group (5/90, 5.5%) compared to none in the salvage group but it did not reach statistical significance p = 0.58. There were no differences in hospital length of stay, 90 day recurrent infection at any site, infection related hospitalization, and 90 day infection related deaths.

Figure 3. Overall survival among patients who underwent catheter removal and salvage: (a) all cohort (n = 115). (b) patients admitted for allogeneic hematopoietic stem cell transplant (n = 38).

Table 3. Outcomes of patients with mucosal barrier injury blood stream infection based on catheter management.

Clinical Outcomes	Catheter Removal n = 90		Catheter Salvage n = 25		P-value
	N	%	N	%
Microbiological clearance within 72 h	88^a	97.8	25	100	0.99
Defervescence within 72 h of antibiotic initiation	77	85.6	24	96	0.30
Hospital Length of stay, days, median (range)	30	3–107	33	13–112	0.74
Intensive care unit Transfer	10^b	11.1	0	0	0.12
All-cause In-hospital mortality	5	5.5	0	0	0.58
All-cause mortality within 90 days	17	18.9	6	24	0.58
Recurrent bacteremia within 90 days	2^c	2.2	0	0	0.99
Infection related mortality within 90 days	1	1.1	1	4	0.39

^aThe two patients with delayed microbiological clearance; one cleared on day 4 (catheter removed on Day 2) and the other one cleared on day 6 (catheter removed on day 5).

^bAll ten patients had catheter removed, median days 1 (0–5). All but one had their bacteremia resolved by 72 h (one patient on day 6). None required pressor.

^cEvent occurred after 79 and 43 days.

6. Discussion

In this study, we explore the diverse management practices of MBI-LCBI among cancer centers in the US. The survey results emphasized that CVC salvage in MBI-LCBI cases is practiced by a majority of physicians and use of adjunct lock therapy is not routine. The aim of the retrospective analysis was to evaluate safety of CVC salvage when cultured organisms support gut translocation as the source of bacteremia. Our findings showed no in-hospital adverse events or recurrent bacteremia related to CVC salvage. When the catheter is deemed the source of infection, removal is essential due to risk of biofilm formation by certain pathogens and risk for recurrent infection. We identified a cohort of patients where catheter salvage was theoretically possible, but only 21.7% of these catheters were salvaged. As this population requires long term definitive intravenous access for chemotherapeutics, antimicrobials, and supportive blood products removing a catheter often results in rapid replacement with another one. Therefore, CVC salvage in selected clinical settings can provide a safe opportunity to reduce the number of procedures and associated risk, patient satisfaction and to a lesser extent overall cost and hospital length of stay [17].

Although median time to microbiological clearance in both groups was one day, CVC was removed in 64/90 (71%) patients within 48 h of BSI diagnosis. This group represents a missed opportunity for CVC salvage as immediate CVC removal were not indicated by the Infectious Diseases Society of America criteria [18] . BSI studies focused on CVC management in high-risk neutropenic patients and those with GVHD are limited. In two single cancer center studies, CVC was removed in 57% of BSI cases that did not meet criteria for CLABSI [19] while 91% (75/82) of patients with viridans group Streptococci bacteremia had CVC salvage, but no outcome data was reported in this study [13]. Another retrospective study found that CVC retention should be pursued in gram negative MBI-LCBI [20]. Similar to our study, there was no difference in 72 h defervescence or bacteremia resolution among patients with MBI [20].

The CVC removal group had a significantly higher PBS, possibly indicating that clinical factors influenced the decision for CVC removal. Only 10% of the CVC removal group required ICU transfer but none required pressor. Role of PBS in determining decision on CVC removal has not been evaluated previously. It will be interesting to see if outcomes remain the same when CVC salvage is pursued in a cohort of patients with uniform PBS score. Although microbiological tools such as time to positivity >120 min and quantitative colony count >3:1 [21,22] could assist with BSI source determination, broad availability and practicability remain challenging as reflected in our survey.

Other treatment techniques, such as antibiotic lock therapy are variable and not standardized. Non-antibiotic based lock therapies such ethanol or EDTA (Ethylenediaminetetraacetic acid) and nitroglycerin have been studied with conflicting results [23–26]. One small clinical trial showed minocycline, EDTA, and 25% Ethanol can successfully eradicate infection in salvaged CVC in cancer patients [27]. We did not observe recurrent BSI or any infection at any site in patients who had salvaged CVC without antibiotic lock therapy, about 61% (11/16). The small numbers preclude further recommendations regarding lock therapy without more data.

In contrast to CLABSI, there are no well-established preventive strategies to combat MBI-LCBI [6,28]. Some advocate that adherence to known infection prevention methods for CLABSI, use of less cytotoxic conditioning or chemotherapy, and optimizing oral hygiene measures can play a role in reducing MBI-LCBI [29–31]. A recent prospective study with a historic cohort showed that antiseptic chlorhexidine gluconate daily bathing can reduce incidence of BSI including MBI-LCBI among autologous and allogenic HCT without notable adverse effects [32].

Our study brings real world experience on CVC management practices through a national survey of US cancer centers and our single center experience. To the best of our knowledge, our study is first to focus on CVC practices among patients with HM who have both gram negative and gram positive MBI-LCBI. We selected clinically relevant outcomes that reflect the immediate effect of CVC management during the same hospital stay. We avoided focusing on remote outcomes that can be influenced by multiple confounders such as overall survival, all-cause mortality or relapse. We also utilized the PBS to predict the severity of illness at the time of diagnosis. Our study has several limitations. First, the survey does carry selection bias as it targeted oncology/transplant infectious disease physicians, a subspecialized group. However, the broad geographic representation and participant years of experience does reflect some diversity. Second, this is a single, center retrospective study with a small cohort with limited power to reflect the true failure rate of CVC salvage especially with differences in PBS among the two groups. Third, we did not collect information on antibiotic initiation timing as we have high adherence to standard febrile neutropenia practices. Time to antibiotic infusion as well as antibiotic therapy duration could have impacted clinical outcomes. Fourth, we did not report transplant related outcomes such as GVHD and transplant related mortality given the transient nature of this infection. Fifth, this is a study of MBI-LCBI and results should not be applied to blood stream infections with organisms such as Pseudomonas, multidrug resistant pathogens or cases with metastatic sources of infection. Finally, we excluded patients who had autologous transplant (as CVCs are usually removed) and those with indwelling foreign bodies which preclude generalization of our results. In conclusion, the survey and study provide practical insight on current MBI-LCBI management practices. Catheter salvage is a generally safe approach in certain clinical settings. Optimizing diagnostic and interventional practices when managing infections in this population is necessary to enhance overall clinical care while balancing risks and benefits. Further comparative clinical studies that focuses on common controversial clinical practices in this population are warranted.

Acknowledgements

ZEB and HC designed the study. ZEB, JK, ML collected the data. QZ designed and performed the statistical analysis. ZEB and HC drafted the manuscript. CL and MEL reviewed the written protocol and manuscript. All authors had access to the primary patient data, reviewed, analyzed, and interpreted the data. All authors reviewed the manuscript and agree with submission.