Hemodialysis International 2014; 18:481–487
Vascular access-related infection in nocturnal home hemodialysis Wesley N. HAYES,1 Karthik TENNANKORE,2 Marisa BATTISTELLA,2 Christopher T. CHAN2 1
Division of Nephrology, Hospital for Sick Children, Toronto, Canada; 2Division of Nephrology, Toronto General Hospital, Toronto, Canada
Abstract Frequent hemodialysis is associated with increased vascular access adverse events. We hypothesized that bacteremia would be more frequent in patients with central venous catheter (CVC) than arteriovenous fistula or arteriovenous graft (AVF/AVG) in nocturnal home hemodialysis (NHHD). We reviewed blood culture reports and concurrent clinical data for a cohort of one hundred eightyseven NHHD patients between January 1, 2006 and June 30, 2012. The primary outcome was time to first bacteremia, technique failure, or death after commencing NHHD. Types of bacteremia and clinical consequences were analyzed. Analyses were adjusted for a priori defined confounders. One hundred eighty-seven patients were included with a total follow up of six hundred five patient years. Initial vascular access was AVF in seventy-eight (42%) patients, AVG in eleven (6%) patients, and CVC in ninety-eight (52%) patients. A total of 79.3% of patients with a CVC reached the composite endpoint of bacteremia, technique failure, or death in the study period; 44.5% of patients with an AVF or AVG reached this composite endpoint. Adjusted time to first bacteremia, technique failure, or death was significantly shorter in patients with initial CVC access (hazard ratio 2.42, 95% confidence interval 1.50–3.90, p < 0.001). Risk factors for bacteremia were comorbid status quantified by the Charlson Comorbidity Index (p < 0.001) and diabetes (p < 0.001). Coagulase negative staphylococcus was the commonest organism cultured accounting for 51.4% bacteremias. The second commonest organism was staphylococcus aureus (20.3% bacteremias). Patients undergoing NHHD with a CVC have a shorter duration to first infection, technique failure, or death than those with permanent vascular access. Key words: Arteriovenous fistula, bacteremia, central venous catheter, nocturnal home hemodialysis, vascular access
BACKGROUND Vascular access complications in patients undergoing frequent nocturnal home hemodialysis (NHHD) are not well
Correspondence to: C. T. Chan, MD, FRCPC, Toronto General Hospital—University Health Network, 200 Elizabeth Street 8N room 842, Toronto, ON M5G 2C4, Canada. E-mail: [email protected]
characterized with conflicting reports from studies to date. The first randomized trial comparing conventional hemodialysis (CHD) to NHHD reported no significant difference in vascular access complications (bacteremic episodes, angiograms, and surgical interventions).1 In contrast, a trend for increased vascular access events (access repair, loss, or access-related hospitalization) among NHHD patients was noted in the Frequent Hemodialysis Network Nocturnal Trial, although this was not statistically significant.2,3
© 2014 International Society for Hemodialysis DOI:10.1111/hdi.12140
481
Hayes et al.
The primary objective of this report was to describe risk factors for bacteremia in a cohort of NHHD patients by exploring the effects of vascular access and patient comorbidities. Secondary objectives were to describe the overall rate of bacteremia and recurrent bacteremia. Given the increased risk of bacteremia with central venous catheters (CVCs) relative to permanent vascular access in CHD,4,5 we hypothesized that bacteremia risk would be greater in patients with CVC access than in those with permanent vascular access undergoing NHHD after adjustment for confounders.
METHODS
Index is a prospectively applicable method for classifying comorbid conditions and has been validated in the assessment of patients with end-stage renal disease.6,7
Microbiology Bacteremia implied microbiologically confirmed bloodstream infection with a single organism. Recurrent bacteremia was defined as two or more consecutive bacteremias with the same microorganism occurring within 28 days of the first treated episode. Each patient is offered retraining in vascular access technique after a bacteremia episode. Overall bacteremia rate was defined as the number of culture positive episodes per 1000 patient days.
Patient cohort One hundred eighty-seven prevalent patients receiving NHHD in a single-center program between January 1, 2006 and June 30, 2012 were included. Patients received from four to seven overnight home hemodialysis treatments per week for the duration of the study period. Patients and carers were trained to follow standard protocols for accessing CVC and arteriovenous fistula or arteriovenous graft (AVF/AVG) following a standard handwashing technique. The CVC access protocol included the use of Tego connectors (ICU Medical Inc., San Clemente, CA, USA), which were exchanged every 7 days and cleaned with chlorhexidine prior to each access. Sodium citrate 4% was the standard CVC lock solution for all patients. Arteriovenous fistula or arteriovenous graft cannulation protocols included skin cleaning with chlorhexidine-based stick swabs. Patient choice determined the use of either the buttonhole or the rope-ladder technique. Patient demographic data and comorbidities were recorded. Results of all blood culture samples taken during the 6.5-year study period were retrospectively analyzed. Clinical data including details of hospital admissions and vascular access-related events were obtained from a prospective database with clinical chart review to clarify missing data where necessary. Vascular access was categorized into two groups, namely CVC and AVF/AVG. The small number of patients with AVG precluded their analysis as a separate group. We a priori specified eleven additional potential confounders (either baseline demographic or clinical variables). These factors were age, sex, body mass index (BMI), diabetes, congestive heart failure, stroke, peripheral arterial disease, hypertension, inflammatory disorder (vasculitis or other systemic autoimmune condition), smoking, and the Charlson Comorbidity Index. The Charlson Comorbidity
482
OUTCOME AND ANALYSES The primary outcome measure was time to the composite endpoint of first bacteremia, technique failure, or death. This was defined as time from entering the study to the first confirmed positive blood culture result, change in dialysis modality, or death. Secondary outcomes were time to first bacteremia, the overall bacteremia rate, and the recurrent bacteremia rate. Descriptive statistics for continuous variables were summarized using median ± interquartile range. Categorical variables were summarized using frequencies and proportions. Normally and non-normally distributed continuous variables were compared with the t test and Wilcoxon rank-sum test, respectively. Categorical variables were compared using Fisher’s exact test. In the primary analysis, time to first bacteremia was analyzed as a composite endpoint with dialysis technique failure and death given the informative nature of the latter two outcomes. Survival analysis was used with the failure event defined as the first blood culture proven bacteremia, technique failure, or death. Censoring occurred at transplantation and change of dialysis modality for reasons other than technique failure. Survival functions were compared using the log–rank test. A multivariable Cox proportional hazard model (including those confounders listed earlier) was used to explore independent associations with the composite endpoint. Secondary survival analyses on each individual outcome were also performed. The relationship of independent variables with bacteremia rate was explored with negative binomial regression. Statistical analyses were performed using Stata IC, version 12 (StataCorp, College Station, TX, USA). Ethics approval from Toronto General Hospital Research Ethics Board was obtained for this study.
Hemodialysis International 2014; 18:481–487
Bacteremia in home hemodialysis
Table 1 Demographics of nocturnal home hemodialysis patients Characteristic Age (yr) Male BMI (kg/m2) Type 1 diabetes Type 2 diabetes Congestive heart failure Stroke Peripheral arterial disease Hypertension Inflammatory disorder Charlson comorbidity index Smoking
All (n = 187)
AVF/AVG (n = 89)
CVC (n = 98)
P value*
44 (35–55) 115 (61.5) 24.4 (20.9–28.5) 6 (3.2) 41 (21.9) 23 (12.3) 15 (8.0) 10 (5.3) 166 (88.8) 14 (7.5) 3 (2–4) 27 (14.4)
44 (36–54) 59 (66.3) 23.9 (21.3–26.5) 0 (0.0) 9 (10.1) 7 (7.9) 8 (8.9) 3 (3.4) 83 (93.3) 4 (4.5) 2 (2–4) 18 (20.2)
44.5 (34–55) 56 (57.1) 25.0 (20.3–28.8) 6 (6.1) 32 (32.6) 16 (16.3) 7 (7.1) 7 (7.1) 83 (84.7) 10 (10.2) 3.5 (2–5) 9 (9.2)
0.997 0.23 0.34 0.03
Vascular access-related infection in nocturnal home hemodialysis Wesley N. HAYES,1 Karthik TENNANKORE,2 Marisa BATTISTELLA,2 Christopher T. CHAN2 1
Division of Nephrology, Hospital for Sick Children, Toronto, Canada; 2Division of Nephrology, Toronto General Hospital, Toronto, Canada
Abstract Frequent hemodialysis is associated with increased vascular access adverse events. We hypothesized that bacteremia would be more frequent in patients with central venous catheter (CVC) than arteriovenous fistula or arteriovenous graft (AVF/AVG) in nocturnal home hemodialysis (NHHD). We reviewed blood culture reports and concurrent clinical data for a cohort of one hundred eightyseven NHHD patients between January 1, 2006 and June 30, 2012. The primary outcome was time to first bacteremia, technique failure, or death after commencing NHHD. Types of bacteremia and clinical consequences were analyzed. Analyses were adjusted for a priori defined confounders. One hundred eighty-seven patients were included with a total follow up of six hundred five patient years. Initial vascular access was AVF in seventy-eight (42%) patients, AVG in eleven (6%) patients, and CVC in ninety-eight (52%) patients. A total of 79.3% of patients with a CVC reached the composite endpoint of bacteremia, technique failure, or death in the study period; 44.5% of patients with an AVF or AVG reached this composite endpoint. Adjusted time to first bacteremia, technique failure, or death was significantly shorter in patients with initial CVC access (hazard ratio 2.42, 95% confidence interval 1.50–3.90, p < 0.001). Risk factors for bacteremia were comorbid status quantified by the Charlson Comorbidity Index (p < 0.001) and diabetes (p < 0.001). Coagulase negative staphylococcus was the commonest organism cultured accounting for 51.4% bacteremias. The second commonest organism was staphylococcus aureus (20.3% bacteremias). Patients undergoing NHHD with a CVC have a shorter duration to first infection, technique failure, or death than those with permanent vascular access. Key words: Arteriovenous fistula, bacteremia, central venous catheter, nocturnal home hemodialysis, vascular access
BACKGROUND Vascular access complications in patients undergoing frequent nocturnal home hemodialysis (NHHD) are not well
Correspondence to: C. T. Chan, MD, FRCPC, Toronto General Hospital—University Health Network, 200 Elizabeth Street 8N room 842, Toronto, ON M5G 2C4, Canada. E-mail: [email protected]
characterized with conflicting reports from studies to date. The first randomized trial comparing conventional hemodialysis (CHD) to NHHD reported no significant difference in vascular access complications (bacteremic episodes, angiograms, and surgical interventions).1 In contrast, a trend for increased vascular access events (access repair, loss, or access-related hospitalization) among NHHD patients was noted in the Frequent Hemodialysis Network Nocturnal Trial, although this was not statistically significant.2,3
© 2014 International Society for Hemodialysis DOI:10.1111/hdi.12140
481
Hayes et al.
The primary objective of this report was to describe risk factors for bacteremia in a cohort of NHHD patients by exploring the effects of vascular access and patient comorbidities. Secondary objectives were to describe the overall rate of bacteremia and recurrent bacteremia. Given the increased risk of bacteremia with central venous catheters (CVCs) relative to permanent vascular access in CHD,4,5 we hypothesized that bacteremia risk would be greater in patients with CVC access than in those with permanent vascular access undergoing NHHD after adjustment for confounders.
METHODS
Index is a prospectively applicable method for classifying comorbid conditions and has been validated in the assessment of patients with end-stage renal disease.6,7
Microbiology Bacteremia implied microbiologically confirmed bloodstream infection with a single organism. Recurrent bacteremia was defined as two or more consecutive bacteremias with the same microorganism occurring within 28 days of the first treated episode. Each patient is offered retraining in vascular access technique after a bacteremia episode. Overall bacteremia rate was defined as the number of culture positive episodes per 1000 patient days.
Patient cohort One hundred eighty-seven prevalent patients receiving NHHD in a single-center program between January 1, 2006 and June 30, 2012 were included. Patients received from four to seven overnight home hemodialysis treatments per week for the duration of the study period. Patients and carers were trained to follow standard protocols for accessing CVC and arteriovenous fistula or arteriovenous graft (AVF/AVG) following a standard handwashing technique. The CVC access protocol included the use of Tego connectors (ICU Medical Inc., San Clemente, CA, USA), which were exchanged every 7 days and cleaned with chlorhexidine prior to each access. Sodium citrate 4% was the standard CVC lock solution for all patients. Arteriovenous fistula or arteriovenous graft cannulation protocols included skin cleaning with chlorhexidine-based stick swabs. Patient choice determined the use of either the buttonhole or the rope-ladder technique. Patient demographic data and comorbidities were recorded. Results of all blood culture samples taken during the 6.5-year study period were retrospectively analyzed. Clinical data including details of hospital admissions and vascular access-related events were obtained from a prospective database with clinical chart review to clarify missing data where necessary. Vascular access was categorized into two groups, namely CVC and AVF/AVG. The small number of patients with AVG precluded their analysis as a separate group. We a priori specified eleven additional potential confounders (either baseline demographic or clinical variables). These factors were age, sex, body mass index (BMI), diabetes, congestive heart failure, stroke, peripheral arterial disease, hypertension, inflammatory disorder (vasculitis or other systemic autoimmune condition), smoking, and the Charlson Comorbidity Index. The Charlson Comorbidity
482
OUTCOME AND ANALYSES The primary outcome measure was time to the composite endpoint of first bacteremia, technique failure, or death. This was defined as time from entering the study to the first confirmed positive blood culture result, change in dialysis modality, or death. Secondary outcomes were time to first bacteremia, the overall bacteremia rate, and the recurrent bacteremia rate. Descriptive statistics for continuous variables were summarized using median ± interquartile range. Categorical variables were summarized using frequencies and proportions. Normally and non-normally distributed continuous variables were compared with the t test and Wilcoxon rank-sum test, respectively. Categorical variables were compared using Fisher’s exact test. In the primary analysis, time to first bacteremia was analyzed as a composite endpoint with dialysis technique failure and death given the informative nature of the latter two outcomes. Survival analysis was used with the failure event defined as the first blood culture proven bacteremia, technique failure, or death. Censoring occurred at transplantation and change of dialysis modality for reasons other than technique failure. Survival functions were compared using the log–rank test. A multivariable Cox proportional hazard model (including those confounders listed earlier) was used to explore independent associations with the composite endpoint. Secondary survival analyses on each individual outcome were also performed. The relationship of independent variables with bacteremia rate was explored with negative binomial regression. Statistical analyses were performed using Stata IC, version 12 (StataCorp, College Station, TX, USA). Ethics approval from Toronto General Hospital Research Ethics Board was obtained for this study.
Hemodialysis International 2014; 18:481–487
Bacteremia in home hemodialysis
Table 1 Demographics of nocturnal home hemodialysis patients Characteristic Age (yr) Male BMI (kg/m2) Type 1 diabetes Type 2 diabetes Congestive heart failure Stroke Peripheral arterial disease Hypertension Inflammatory disorder Charlson comorbidity index Smoking
All (n = 187)
AVF/AVG (n = 89)
CVC (n = 98)
P value*
44 (35–55) 115 (61.5) 24.4 (20.9–28.5) 6 (3.2) 41 (21.9) 23 (12.3) 15 (8.0) 10 (5.3) 166 (88.8) 14 (7.5) 3 (2–4) 27 (14.4)
44 (36–54) 59 (66.3) 23.9 (21.3–26.5) 0 (0.0) 9 (10.1) 7 (7.9) 8 (8.9) 3 (3.4) 83 (93.3) 4 (4.5) 2 (2–4) 18 (20.2)
44.5 (34–55) 56 (57.1) 25.0 (20.3–28.8) 6 (6.1) 32 (32.6) 16 (16.3) 7 (7.1) 7 (7.1) 83 (84.7) 10 (10.2) 3.5 (2–5) 9 (9.2)
0.997 0.23 0.34 0.03
