Effect of Buttonhole Cannulation With a Polycarbonate Peg on In-Center Hemodialysis Fistula Outcomes: A Randomized Controlled Trial

Original Investigation Effect of Buttonhole Cannulation With a Polycarbonate Peg on In-Center Hemodialysis Fistula Outcomes: A Randomized Controlled Trial Emma Vaux, MD, PhD, 1 Jennie King, RN, BSc(Hons), 1 Swee Lloyd, RN, 1 Jane Moore, RN, 1 Leo Bailey, RN, 1 Isabel Reading, MSc, PhD, 2 and Ramesh Naik, MD 1 Background: Quality improvement strategies to increase and maintain the numbers of arteriovenous fistulas (AVFs) are a critical drive in enhancing the quality of care of patients receiving treatment with hemodialysis. How the AVF is needled is an important consideration in AVF survival; the ideal cannulation technique has not been established to date. Study Design: Prospective randomized single-center trial. Setting & Participants: Patients on maintenance hemodialysis therapy (N 140). Intervention: A 1-year intervention of buttonhole (constant site) or usual-practice (different site) cannulation. Outcomes: Primary study outcome was AVF survival over 1 year, in which AVF failure was defined as an AVF no longer used for hemodialysis (also referred to as assisted patency). Secondary outcomes included primary patency, number of access interventions, bleeding time, infection rate, cannulation time and pain, and aneurysm formation. Results: Demographic data were similar for both groups. The primary outcome measure of AVF survival at 1 year was statistically significantly increased in the buttonhole group (100% vs 86% with usual practice; P 0.005, log-rank test). In the buttonhole group, there were fewer interventions (19% vs 39% in usual practice) and less existing aneurysm enlargement (23% vs 67% in usual practice). There were no bacteremia events in the buttonhole group and 2 in the usual-practice group (0.09/1,000 AVF days). There were no significant differences in bleeding times and lignocaine use between the 2 groups. Limitations: A single-center study, lack of blinding. Conclusions: In this study, AVF survival was significantly greater when using buttonhole cannulation. The buttonhole technique significantly decreased the need for access interventions and reduced existing aneurysm enlargement. Concerns of increased infection rates or prolonged bleeding times with the buttonhole technique were not seen in this study. The buttonhole technique should be considered the cannulation technique of choice for AVFs. Am J Kidney Dis. xx(x):xxx. 2013 by the National Kidney Foundation, Inc. INDEX WORDS: Buttonhole; cannulation; arteriovenous fistula; survival. Creating and maintaining vascular access has been described as the Achilles heel of hemodialysis (HD), exacerbated by the increasing numbers and advanced age of new patients with associated comorbid conditions, such as diabetes mellitus and peripheral vascular disease. 1 The arteriovenous fistula (AVF) is the preferred option for vascular access for the provision of maintenance HD treatment. Clinical opinion about the generally accepted method for access puncture is divided. 2 There are 3 recognized options for cannulation of an AVF 3 ; the rope ladder pattern, the area puncture (or area dilatation) technique, and the buttonhole technique. The original recommendation that puncture sites be changed for each dialysis session was made in 1966. 4 This was thought to provide good healing of the puncture wound and prevent problems such as hematoma, stenosis, infection, and pseudoaneurysm. 4 Based on this principle, the most common practice in most dialysis units has been to use a combination of rope ladder and area puncture techniques together in the same patient at different times and are referred to here, collectively, as usual-practice (different site, rotating) cannulation. However, with the rope ladder method, a fistula with long segments appropriate for cannulation is required, and the area puncture method weakens the fistula wall and leads to the development of (pseudo) aneurysms and strictures. 3 From the 1 Renal Unit, Royal Berkshire NHS Foundation Trust, Reading; and 2 NIHR Research Design Service South Central, University of Southampton, Southampton General Hospital, Southampton, United Kingdom. Received May 26, 2012. Accepted in revised form January 8, 2013. Trial registration: www.isrctn.org; study number: 27841616. Address correspondence to Emma Vaux, MD, PhD, Renal Unit, Royal Berkshire NHS Foundation Trust, Reading, United Kingdom RG1 5AN. E-mail: emma.vaux@royalberkshire.nhs.uk 2013 by the National Kidney Foundation, Inc. 0272-6386/$36.00 http://dx.doi.org/10.1053/j.ajkd.2013.01.011 Am J Kidney Dis. 2013;xx(x):xxx 1

Vaux et al The buttonhole technique uses identical (constant) sites for cannulation. It was developed in the 1970s to help patients with limited puncture sites. 5 Subsequent observational studies have shown that constant-site insertion was easy and quick, cannulation was less painful, hematoma formation was decreased, there was less scar formation, and the infection rate was not significantly higher than with different-site needle insertion. 6-10 In addition, fistula 5-year survival rates using the buttonhole method were reported as 97.4% in home patients and 87% in outpatients. 11 However, the buttonhole method also has been associated with an increased risk of infection. 12-14 This and the requirement for the same puncture site have been the biggest barriers to acceptance of the buttonhole technique. With the recent introduction of a sterile polycarbonate peg product, a fixed puncture route can be established in a short period (2 weeks), making this technique a more feasible option. 15 The aim of this study was to determine, using a randomized controlled study design, the efficacy of the buttonhole cannulation method compared to usual practice with a primary outcome of AVF survival at 1 year and secondary outcomes of cannulation complications. METHODS Trial Design This was a single-center (3-site), randomized, controlled, parallelgroup study conducted in the United Kingdom with a 1-year follow-up. The study was designed to determine the efficacy of the buttonhole cannulation technique. Patients were randomly assigned in a 1:1 ratio to 1 of 2 groups. The study was given UK national ethical approval. Written informed consent was obtained from study participants. Participants The study took place at the Royal Berkshire National Health Service Foundation Trust from November 2007 to July 2010. A total of 270 patients were receiving treatment with HD. Eligible participants were all adults 18 years or older receiving treatment with HD and were already using or intending to use an AVF for vascular access. Both incident and prevalent patients were included and were representative of the general HD population. Exclusion criteria were presence of an arteriovenous graft, lack of capacity, living donor kidney transplantation date, and expected survival less than 12 months. Interventions Patients were randomly assigned to receive buttonhole cannulation or usual care. Usual care was cannulation of the AVF using the conventional different-site technique with a 25-mm length 15G sharp needle (Gambro). Buttonhole cannulation was a constantsite technique in which an individual needles the AVF at the same angle and depth of penetration every time. To establish the buttonhole tract, a dull-tipped thumbtack-shaped 5-mm long polycarbonate peg (BioHole Stick; Nipro Corp) 15 was used initially. The peg was placed at the site where the puncture needle had been removed and remained in place until the next HD session, when the peg was removed and a sharp puncture needle (25-mm length 15G; Gambro) was inserted along the track already formed by the peg left in place. These steps were repeated at each HD session over 14 days (a total of 6 times) to create a scar tissue tunnel. Thereafter, vascular access was achieved at each HD session by inserting the blunt needle (25-mm length 15G; Nipro Corp) through the established scar tissue tunnel track. Infection prevention measures are outlined in Item S1 (provided as online supplementary material). Outcomes All outcomes were recorded for each patient during 1 year after the start of the assigned cannulation technique. The primary outcome was AVF survival at 1 year, for which failure was defined as an AVF no longer used for successful HD (also known as assisted patency). Secondary outcomes identified primary patency, time to cannulation, and the complication risk with each cannulation technique, which included the number of access interventions (radiologic and/or surgical), bleeding time, infection rate, cannulation pain and local anesthetic use, new aneurysm formation, and existing aneurysm enlargement. Primary and primary assisted patency were defined as previously described. 16 Time to cannulation was defined as time from the start of the cannulation process to successful needle placement (patient connected to machine and successful start of bleed out achieved) for each HD session. Cannulation pain was assessed at the venous and arterial sites using a self-reported modified numeric pain rating scale (1 indicates pain free, and 10 unbearable pain; Royal Berkshire Hospital Pain Clinic, United Kingdom) and by the use of local anesthetic at each dialysis session. Bleeding time was measured at the end of each dialysis session. It was defined as the total time from removal of the first needle to achievement of hemostasis for both needles. Infection rate was defined by bacteremias and exit-site infections per 1,000 AVF days. Aneurysm formation description was assessed as a swelling of 0.5 cm or an increase in size 0.5 cm during the study. Photographs were obtained at 3-monthly intervals and assessed independently by 2 researchers blinded to allocation. Access interventions were defined as radiologic (fistuloplasty or thrombectomy) and surgical (revision and thrombectomy) interventions. Staff satisfaction with the cannulation technique was assessed by a voluntarily completed survey. Sample Size Assuming the control group has 70% AVF survival at 1 year, to detect an improvement to 90% AVF survival at 1 year requires 63 patients per year. 17 This assumes 80% power and 5% two-sided significance testing using the log-rank test. A total of 140 patients (70 per group) were recruited to allow for a drop-out rate of 10%. Randomization Randomization used minimization (administered by the Centre for Statistics in Medicine, University of Oxford, United Kingdom) to balance the 2 treatment groups for sex, age, site, age and duration of AVF use, any previous AVF intervention, presence of diabetes mellitus or peripheral vascular disease, and kidney transplantation listing. Blinding Patients and dialysis nurses were aware of the randomized treatment assignment and ascertainment of the primary outcome and many of the secondary outcomes were performed by individuals who were not blinded to the treatment assignment. Radiologic interventions, assessment of aneurysms, and recording of infections were performed by individuals who were not aware of 2 Am J Kidney Dis. 2013;xx(x):xxx

AVF Survival Using Buttonhole Technique Figure 1. Enrollment, randomization, and follow-up. Abbreviation: AVF, arteriovenous fistula. treatment assignment, and all investigators, staff, and participants were kept masked to the trial progress and outcome tallies until follow-up of all participants was completed and all data were entered into the database. Statistical Analysis Statistical analyses were performed using STATA, version 11.0 (StataCorp LP). All patients were followed up until transplantation, death, or the end of the trial (1 year of follow-up), with censoring of data at the time a patient underwent transplantation or died. All statistical tests were 2 tailed, and P 0.05 was considered to indicate statistical significance. Primary Outcome Analysis Intention-to-treat (ITT) survival analysis was used to compare time to AVF failure between the 2 cannulation groups, with log-rank test as the primary comparison. A Cox proportional hazards model also was fitted to estimate hazard ratios (HRs) with 95% confidence intervals (CIs). Secondary Outcome Analyses A secondary analysis of the primary outcome was performed to present findings based on actual treatment rather than ITT. Thus, for patients who switched treatments from their original assigned treatment during the course of the study, their data were censored at the point of treatment change. Informal ITT subgroup analyses were performed to explore whether the treatment effects on the primary outcome were different in patients with a new AVF versus those with an existing AVF. ITT survival analysis methods also were used to examine primary patency. Proportions of patients who had any anesthetic use and those who had anesthetic use in more than half their cannulation sessions were compared between treatment groups using Fisher exact test. Proportions of patients who had any access interventions, aneurysm formation, and aneurysm enlargement were presented. Cannulation times, bleeding times, and pain scores were summarized within patient, and these summary measures then were compared between cannulation methods using Mann-Whitney U test. RESULTS Demographic Data Between November 2007 and July 2009, there were 140 participants randomly assigned to buttonhole cannulation (n 70) or usual practice (n 70; Fig 1). Patients were followed up until July 2010. In the buttonhole group, 12 patients never started for reasons unrelated to technique (Fig 1). In the usual-practice group, one patient never started because he or she withdrew consent. Five participants received kidney transplants and 13 died during the study period. A total of 58 participants in the buttonhole group and 69 in the usual-practice group were used in the analysis. The 2 groups were well matched with respect to all baseline characteristics (Table 1). Fourteen patients assigned to the buttonhole group changed to usual practice. The buttonhole could not be established or there were subsequent problems with needling in 4 patients, 2 patients transferred to a satellite unit where only usualpractice technique was used, and 8 patients asked to change to usual practice because of pain associated with the technique. Of these 14 patients, 12 maintained their Am J Kidney Dis. 2013;xx(x):xxx 3

Vaux et al Table 1. Demographics and Baseline Data Usual Practice (n 70) Buttonhole (n 70) Sex Male 44 (63) 47 (67) Female 26 (37) 23 (33) Age (y) 64 16.6 62 15.4 Cause of ESRD Diabetes mellitus 15 (21) 18 (26) Nephritis 16 (23) 14 (20) Congenital/hereditary diseases 12 (17) 13 (19) Renal vascular disease 3 (4) 8 (11) Multisystem diseases 7 (10) 6 (9) Obstructive uropathy 5 (7) 1 (1) Unknown 12 (17) 10 (14) Peripheral vascular disease No 66 (94) 65 (93) Yes 4 (6) 5 (7) Hemodialysis vintage 0-12 mo 21 (30) 24 (34) 1-5 y 43 (61) 40 (57) 5 y 6 (9) 6 (9) Ethnicity Asian 8 (11) 10 (14) Black 1 (1) 3 (4) White 61 (87) 57 (81) AVF characteristics New AVF 25 (36) 26 (37) Existing AVF 45 (64) 44 (63) Right-arm AVF 19 (27) 18 (25) Left-arm AVF 51 (73) 52 (74) Brachiocephalic 30 (43) 31 (44) Radiocephalic 40 (57) 39 (56) AVF vintage 6 mo 9 (13) 15 (21) 7-12 mo 9 (13) 9 (13) 13-24 mo 11 (16) 12 (17) 25-36 mo 4 (6) 1 (1) 37 mo 11 (16) 8 (11) Interventions to AVF before study start No 54 (77) 50 (71) Yes 16 (23) 20 (29) Previous accesses No 50 (71) 61 (87) Yes 20 (29) 9 (13) Previous catheter use No 22 (31) 21 (30) Yes 48 (69) 49 (70) Note: Values for categorical variables are given as number (percentage); age given as mean standard deviation. Abbreviations: AVF, arteriovenous fistula; ESRD, end-stage renal disease. AVFs for the rest of the study period, whereas the other 2 died. In the primary outcome analysis, these 14 patients were analyzed as being in the buttonhole group, with time to AVF survival being at the end of the study for 12 and at death for the other 2 (according to ITT principles). However, for the secondary as-treated analysis, time to AVF failure was censored at the date they changed cannulation practice. This was deemed to be a reasonable approach to an as-treated analysis given that the AVF had not actually failed at this point. Primary Outcome AVF survival (assisted primary patency) was statistically significantly greater in the buttonhole group than in usual practice (100% vs 86% in usual practice; P 0.005, log-rank test), with no AVF failures observed in the buttonhole group and 9 observed in the usualpractice group (Table 2). An HR from a Cox proportional hazards model could not be obtained because there were no AVF failures in one of the comparator groups; however, a Kaplan-Meier plot of AVF survival in the 2 groups was examined and shows the buttonhole group with clearly favorable results (Fig 2). Secondary Outcomes In the as-treated analysis of AVF survival, there was still statistically significantly greater survival in the buttonhole group than for usual practice (P 0.01, log-rank test), with no AVF failures observed in the buttonhole group and 9 observed in the usual-practice group. As in the primary analysis, an HR from a Cox proportional hazards model could not be obtained. The informal subgroup analysis indicated that the treatment effect of better survival in the buttonhole group was similar in patients with new and existing AVFs, but was statistically significant only for patients with an existing AVF (P 0.03 [log-rank test Table 2. AVF Survival at 1 Year (Intention-to-Treat Analysis) Usual Practice (n 69) Buttonhole (n 58) Patient status Completed study 52 (75) 48 (83) Died 5 (7) 8 (14) Transplantation 3 (4) 2 (3) No. of AVF failures 9 (13) a 0 (0) Days to AVF failure Median [IQR] 268 [143-292] NA Minimum; maximum 7; 328 NA Note: Categorical variables given as number (percentage). Abbreviations: AVF, arteriovenous fistula; IQR, interquartile range; NA, not applicable. a Six AVFs thrombosed (not salvageable), 2 AVFs ligated (intractable infection, critically enlarging aneurysm), and 1 AVF abandoned (needling difficulties and surgical revision unsuccessful). 4 Am J Kidney Dis. 2013;xx(x):xxx

AVF Survival Using Buttonhole Technique Table 3. Comparison of Number of Interventions in Each Group Fistuloplasty for Stenosis Successful Thrombectomy Total No. of Interventions Usual practice, n 69 Buttonhole, n 58 25 (36) 2 (3) 27 (39) 10 (17) 1 (2) 11 (19) Figure 2. Arteriovenous fistula survival curves based on original randomization assignment (intent to treat). among existing AVFs] vs P 0.07 [log-rank test among new AVFs]) However, these analyses were based on only a small number of events; none in buttonhole group, 4 AVF failures in the usual-practice group with new AVFs, and 5 AVF failures in the usual-practice group with existing AVFs. Primary patency for AVFs using the buttonhole technique had a statistically significantly longer duration than that in the usual-practice group (P 0.01, log-rank test; Kaplin-Meier plot in Fig 3), ITT analysis. The HR of 0.46 to lose primary patency in the buttonhole group compared to usual practice (95% CI, 0.25-0.87) was observed (P 0.02). Primary patency at 1 year was 74% in the buttonhole group compared to 49% in usual practice. Both new and existing AVFs were included in the patency calculation. A decreased number of patients required access interventions in the buttonhole group (11 [19%]) compared to the usual-practice group (27 [39%]; Figure 3. One-year primary patency based on original randomization assignment (intent to treat); 1-year primary patency for arteriovenous fistulas using the buttonhole technique was 74%, and for usual practice, 49%. Table 3). To maintain functionality of the AVF, the buttonhole group required 0.2 interventions per patientyear, and patients with usual practice had 0.4 interventions per patient-year. Information on local anesthetic use was not available for 6 patients (1, buttonhole; and 5, usual practice). A similar proportion of patients required some use of local anesthetic (topical or injected) in the 2 treatment groups during at least one cannulation session (41% in the buttonhole group compared to 37% in usual practice) during the study (P 0.7, Fisher exact test). Similarly small proportions of patients required use of local anesthetic in more than half their cannulation sessions (9% in the buttonhole group vs 16% in the usual-practice group; P 0.3, Fisher exact test). All pain scores reported while the access was being cannulated according to the patient s assigned method (buttonhole or usual practice) were summarized within each patient by their average pain score over dialysis session and site (venous and arterial). Valid pain scores were not available for 6 patients (1, buttonhole; 5, usual practice). A borderline statistically significant difference in average pain score (1 no pain, 10 worst pain) was observed between the 2 groups (P 0.05, Mann-Whitney U test), with the median of the average pain scores being 1.3 (interquartile range [IQR], 1.2-1.9) with the buttonhole technique and 1.2 (IQR, 1.0-1.5) with usual practice. Eight patients in the buttonhole group abandoned buttonhole cannulation due to pain, and their subsequent pain scores (after they had switched to usual practice) have been excluded from this analysis. Analysis of pain scores therefore are likely to be biased due to these 8 patients withdrawing from the buttonhole group and should be viewed with caution. Patients pain scores were recorded a median of 98 (IQR, 50-126) times during the study. All cannulation times reported while the access was being cannulated according to the assigned method (buttonhole or usual practice) were summarized within each patient by the average time. Averages were used although cannulation times were not normally distributed in order to reflect the occasional extreme (long) cannulation times, as well as the more common shorter Am J Kidney Dis. 2013;xx(x):xxx 5

Vaux et al durations. Valid cannulation times were not available for 8 patients (2, buttonhole group; 6, usual-practice group). A statistically significant difference in average cannulation time was observed between the 2 groups (P 0.001, Mann-Whitney U test), with a median of average cannulation times of 5.3 (IQR, 4.4-5.9) minutes with the buttonhole technique and 4.3 (IQR, 3.4-5.0) minutes with usual practice. The longest individual cannulation times observed were 80 minutes for a patient on one occasion in the buttonhole group, and 60 minutes for a patient on one occasion in usual practice. Patients cannulation times were recorded a median of 98 (IQR, 54-123) times during the study. There were 28 existing aneurysms at the start of the study (13, buttonhole; 15, usual practice). Significant enlargement of an existing aneurysm occurred less frequently in the buttonhole group (23%) compared to usual practice (67%; Table 4). At the start of the study, 99 patients did not have an aneurysm of the AVF. Fewer new aneurysms were observed in the buttonhole group (2 [4%] compared with 9 [17%] in usual practice; Table 4). All bleed times reported while the access was being cannulated according to the assigned method (buttonhole or usual practice) were summarized as for cannulation times and had the same patients with missing data. No difference in average bleeding times was observed between the 2 groups (P 0.3, Mann- Whitney U test; Fig 4), with a median of average bleed times of 7.9 (IQR, 6.3-10.4) minutes with the buttonhole technique and 9.1 (IQR, 6.9-11.3) minutes Table 4. Enlargement of Any Existing Aneurysm and New Aneurysm Formation Buttonhole Usual Practice Total no. with existing aneurysm 13 15 Change in aneurysm size Bigger 3 (23) 10 (67) Same size 6 (46) 4 (27) Smaller 1 (8) 1 (7) Unknown 3 (23) 0 (0) Total no. with no existing aneurysm 45 54 Development of new aneurysm Yes 2 (4) 9 (17) No 38 (84) 36 (67) Unknown 5 (11) 9 (17) Note: Values are given as number or number (percentage). Enlargement of an existing aneurysm was defined as an increase in size 0.5 cm during the 1-year study period. Missing data were from patients who died, underwent transplantation, or whose photograph was not obtained at appropriate time. New aneurysm formation was defined as development of an aneurysm 0.5 cm. Missing data were from patients who died, underwent transplantation, or whose photograph was not obtained at appropriate time. Figure 4. Bleeding time. Comparison of bleeding times after each hemodialysis session for the usual-practice and buttonhole cannulation techniques. Box indicates 25th and 75th percentiles (thick line is median value). Capped bars indicate minimum and maximum values, including outliers. The median (interquartile range) bleeding time for usual practice was 9.1 (6.9-11.3) minutes, and for the buttonhole technique, 7.9 (6.3-10.4) minutes. with usual practice. The longest individual bleed times observed were 120 minutes for a patient on one occasion in usual practice and 90 minutes for a patient on one occasion in the buttonhole group. Patients bleeding times were recorded a median of 84 (IQR, 45-122) times during the study. Infection rates were low. There was no bacteremia in the buttonhole group. There were 2 in the usualpractice group (0.09/1,000 AVF days), both caused by methicillin-resistant Staphylococcus aureus. There were 2 exit-site infections in the buttonhole group (both culture negative; 0.12/1,000 AVF days) and none in the usual-practice group. The response rate to the staff survey on satisfaction with the different methods of cannulation was 60% (18/30). Among respondents, 11 of the 18 nurses (61%) preferred using the buttonhole technique, 5 (28%) preferred usual practice, and 2 did not respond. DISCUSSION Previous research has focused on factors associated with fistula failure, such as surgical formation and radiologic procedures. Research describing critical review of needling methods is scarce. The results suggest that the buttonhole cannulation method offered significant advantages to the participants in this study in terms of fistula survival and cannulaassociated complications, such as existing aneurysm progression and need for access interventions. Concerns of increased infection rates and bleed times with the buttonhole technique were not seen in this study. Staff expressed a preference for the buttonhole technique despite the need to have to learn this as a new technique. The longer cannulation times for the 6 Am J Kidney Dis. 2013;xx(x):xxx

AVF Survival Using Buttonhole Technique buttonhole technique may in part reflect this. Overall, time to cannulation was statistically significantly decreased in the usual-practice group, but this equated to only a 1-minute difference, which is unlikely to have any real implications in everyday practice. The polycarbonate peg was used to create the buttonhole track and enabled practical implementation by different nurses and the timely transition to blunt needles. Best results seemed achieved when the track was established by the same nurse until blunt needling was established, minimizing the risk of multitrack formation, difficulties inserting the needle, keeping the tunnel track narrow, and reducing the risk of an exit-site infection. This approach may have contributed to the superiority of the buttonhole technique for survival. In addition, the buttonhole approach is likely to minimize the risk of area dilatation and subsequent aneurysm and stenosis. Interestingly, pain scores were marginally higher in the buttonhole group. This may reflect their initial experience of establishing the buttonhole track, an interpretation supported by the 8 participants who transferred to usual practice because of pain (5 of 8 had been using usual-practice cannulation prior to the study). The buttonhole method has been associated with high risk of infection. 12-14 Bacteremia rates of the buttonhole technique have been reported to range from 0.15-0.6 per 1,000 patient-days, or 0.05-0.2 per patient-year. There was no bacteremia in the buttonhole group in this study. Overall AVF-years in this study are small and may not have been sufficient for an adequately powered comparison of bacteremia rates. Canadian experience has reported the efficacy of topical mupirocin prophylaxis to reduce the risk of patients on home care developing an S aureus bacteremia. 18 This practice is supported by some UK units, although there remains a concern for the emergence of mupirocin resistance. This practice has not been implemented in our unit. Zero tolerance to infection with rigorous implementation of basic hygiene has contributed to our unit s low infection rates, which are reinforced by a staff and patient education program on the care of vascular access. Diligent use of chlorhexidine, 2%, before and after scab removal prior to cannulation is essential. The use of the buttonhole method for AVF cannulation in this study was efficacious and safe. However, there are several limitations to our study that should be noted. Lack of blinding to treatment assignment can affect performance and ascertainment bias. Patients assigned to the buttonhole group had fewer previous accesses (9 vs 20) despite the same HD vintage, previous catheter use, and number of interventions to the AVF prior to the start of the study and might suggest some survivor bias to the buttonhole group. In addition, this was a single-center (3-site) study. As a result, the study potentially has limited external validity. Small randomized controlled trials also are more prone to random error and hence may overestimate treatment effects. There also may be unmeasured confounding prognostic factors that were not evenly distributed. One such example is colonization of a fistula with an organism prior to study enrollment. AVF survival observed in the usualpractice group was lower in this study than was estimated (the estimate on which the power calculation was based), partly as a result of both incident and prevalent AVFs being included (in both groups), rather than only new AVFs, and also because the estimate was based on our own unit experience, but also from other published studies. 19 Implementation of the buttonhole technique was within our busy dialysis units with no extra personnel or resources potentially enhancing applicability. The main strength of this study is to allow a definitive larger multicenter study to be planned and powered appropriately for efficacy outcomes and rigorous safety assessment of infection rates. ACKNOWLEDGEMENTS The authors acknowledge the assistance of the nurses on the dialysis units at the Royal Berkshire National Health Service Foundation Trust. Aspects of this work were presented in abstract form at the British Renal Society/Renal Association conference in Birmingham, England, June 6-8, 2011. Support: The BioHole pegs and an educational and travel grant of 13,000 were provided by Nipro Corp to the Renal Unit, Royal Berkshire National Health Service Foundation Trust. The grant was directed for use of dissemination of any study findings. Nipro Corp had no role in study design; collection, analysis, and interpretation of data; writing the report; or the decision to submit the report for publication. Financial Disclosure: The authors declare that they have no other relevant financial interests. SUPPLEMENTARY MATERIAL Item S1: Infection prevention measures. Note: The supplementary material accompanying this article (http://dx.doi.org/10.1053/j.ajkd.2013.01.011) is available at www.ajkd.org. REFERENCES 1. Konner, K. A primer on the AV fistula Achilles heel, but also Cinderella of hemodialysis. Nephrol Dial Transplant. 1999; 14(9):2094-2098. 2. National Kidney Foundation. KDOQI vascular access guidelines. http://www.kidney.org/professionals/kdoqi/guideline_uphd_ PD_VA/hd_guide1.html. Accessed November 30, 2006. 3. Kronung G. Plastic deformation of Cimio fistula by repeated puncture. Nephrol Dial Transplant. 1994;13(6):35-38. 4. Brescia MJ, Cimino JE, Appel K, Hurwich BJ. Chronic hemodialysis using venepuncture and a surgically created arteriovenous fistula. N Engl J Med. 1966;275(20):1089-1092. Am J Kidney Dis. 2013;xx(x):xxx 7

Vaux et al 5. Twardowski Z, Kubara H. Different sites versus constant sites of needle insertion into arteriovenous fistula for treatment by repeated dialysis. Dial Transplant. 1979;8(10):978-980. 6. Scribner B. The overriding importance of vascular access. Dial Transplant. 1984;13(10):625. 7. Goovaerts T. Analysis of the buttonhole puncture technique. Am J Nephrol. 2004;24(6):140-146. 8. Twardowski ZJ. Constant site (buttonhole) method of needle insertion for hemodialysis. Dial Transplant. 1995;24:559-576. 9. Struthers J, Allan A, Peel RK, et al. Buttonhole needling of arteriovenous fistulae: a randomised controlled trial. ASAIO J. 2010;56(4):319-322. 10. Pergolotti A, Rich E, Lock K. The effect of the buttonhole method versus the traditional method of AV fistula cannulation on hemostasis, needle stick pain, pre-needle stick anxiety, and presence of aneurysms in ambulatory patients on hemodialysis. Nephrol Nurs J. 2011;38(4):333-336. 11. Sucker A, Schatzle-Schuler G, Pankiewicz T, et al. Constant site needle insertion versus different site needle insertion into the Cimino Brescia fistula. Proc EDTNA. 1984;13:59-66. 12. Verhallen AM, Kooistra MP, van Jaarsfeld BC. Cannulating in hemodialysis: rope-ladder or buttonhole technique? Nephrol Dial Transplant. 2007;22(9):2601-2604. 13. Van Loon M, Goovaerts T, Kessel A, et al. Buttonhole needling of hemodialysis arteriovenous fistulae results in less complications and interventions compared to the rope-ladder technique. Nephrol Dial Transplant. 2010;25(1):225-230. 14. Chow J, Rayment G, San Miguel S, et al. A randomised controlled trial of buttonhole cannulation for the prevention of fistula access complications. J Renal Care. 2011;37(2):85-93. 15. Toma S, Shinzato T, Fukui H, et al. A timesaving method to create a fixed puncture route for the buttonhole technique. EDTA- ERA. 2003;18(10):2118-2121. 16. Sidawy AN, Gray R, Besarab A, et al. Recommended standards for reports dealing with arteriovenous hemodialysis accesses. J Vasc Surg. 2002;35(3):603-610. 17. Barthel FMS, Royston P, Babiker A. A menu-driven facility for complex sample size calculation in randomized controlled trials with a survival or a binary outcome: update. Stata J. 2005;5(1):123-129. 18. Nesrallah G. Cuerden M, Wong JHS. Staphylococcus aureus bacteremia and buttonhole cannulation: long-term safety and efficacy of mupirocin prophylaxis. Clin J Am Soc Nephrol. 2010; 5(6):1047-1053. 19. Turmel-Rodrigues L, Pengloan J, Rodrigue H, et al. Treatment of failed native arteriovenous fistulae for hemodialysis by interventional radiology. Kidney Int. 2000;57(3):1124-1140. 8 Am J Kidney Dis. 2013;xx(x):xxx