Original Investigation
Buttonhole Versus Rope-Ladder Cannulation of Arteriovenous
Fistulas for Hemodialysis: A Systematic Review
Ben Wong, MD,1 Maliha Muneer, BSc,2 Natasha Wiebe, MMath,2 Dale Storie, MA,2
Sabin Shurraw, MD,1 Neesh Pannu, MD,1 Scott Klarenbach, MD,1
Alexa Grudzinski, BSc,3 Gihad Nesrallah, MD,3 and Robert P. Pauly, MD1
Background: The buttonhole technique is an alternative method of cannulating the arteriovenous fistula
(AVF) in hemodialysis (HD), frequently used for home HD patients. However, the balance of risks and benefits
of the buttonhole compared with the rope-ladder technique is uncertain.
Study Design: A systematic review of randomized trials and observational studies (case reports, case
series, studies without a control group, non-English studies, and abstracts were excluded).
Setting & Population: HD patients (both in-center conventional HD and home HD) using an AVF for
vascular access.
Selection Criteria for Studies: We searched MEDLINE, EMBASE, EBM Reviews, and CINAHL from the
earliest date in the databases to March 2014 for studies comparing clinical outcomes of the buttonhole versus
rope-ladder technique.
Intervention: Buttonhole versus rope-ladder cannulation technique.
Outcomes: The primary outcomes of interest were patient-reported cannulation pain and rates of
AVF-related local and systemic infections. Secondary outcomes included access survival, intervention,
hospitalization, and mortality, as well as hematoma and aneurysm formation, time to hemostasis, and allcause hospitalization and mortality.
Results: Of 1,044 identified citations, 23 studies were selected for inclusion. There was equivocal evidence
with respect to cannulation pain: pooled observational studies yielded a statistical reduction in pain with
buttonhole cannulation (standardized mean difference, 20.76 [95% CI, 21.38 to 20.15] standard deviations),
but no difference in cannulation pain was found among randomized controlled trials (standardized mean difference, 0.34 [95% CI, 20.76 to 1.43] standard deviations). Buttonhole, as compared to rope-ladder,
technique appeared to be associated with increased risk of local and systemic infections.
Limitations: Overall poor quality and substantial heterogeneity among studies precluded pooling of most
outcomes.
Conclusions: Evidence does not support the preferential use of buttonhole over rope-ladder cannulation in
either facility-based conventional HD or home HD. This does not preclude buttonhole cannulation as being
appropriate for some patients with difficult-to-access AVFs.
Am J Kidney Dis. 64(6):918-936. ª 2014 by the National Kidney Foundation, Inc.
INDEX WORDS: Buttonhole; rope-ladder; cannulation technique; needling pain; arteriovenous fistula (AVF);
vascular access; hemodialysis (HD); end-stage renal disease (ESRD); access-related infection; systematic
review; chronic kidney disease (CKD).
T
he arteriovenous fistula (AVF) is the preferred
vascular access among hemodialysis (HD)
patients, owing to its lower infectious complications
and fewer failures compared with all other types of
access.1-5 Consistently successful cannulation of an
AVF is critical for achieving adequate HD and is
the lifeline for HD patients. The conventional method
of accessing the AVF for dialysis treatments is the
rope-ladder or rotating-site cannulation technique.
From the 1Division of Nephrology and Transplant Immunology,
University of Alberta; 2University of Alberta, Edmonton, Alberta;
and 3Western University, London, Ontario, Canada.
Received January 20, 2014. Accepted in revised form June 9,
2014. Originally published online August 7, 2014.
Address correspondence to Ben Wong, MD, University of Alberta,
Canada. E-mail: bcw@ualberta.net
2014 by the National Kidney Foundation, Inc.
0272-6386/$36.00
http://dx.doi.org/10.1053/j.ajkd.2014.06.018
918
This involves needle placement sites that are rotated
along the entire length of the fistula each time the
patient receives dialysis. An alternative method,
described by Twardowski and Kubara6 in 1979, is
called the buttonhole or constant-site technique. Here,
the HD needles are inserted into precisely the same
location, at an identical angle and depth for each
treatment. The result is a permanent, self-sealing,
fibrous tunnel through which the fistula can be
accessed repeatedly over years, and it initially was
advocated for use with fistulas having short useable
segments or for patients experiencing significant
cannulation discomfort.6
Although the buttonhole technique has not gained
widespread popularity among conventional in-center
HD patients, this technique has been used routinely
by many self-cannulating home HD patients. Purported benefits of the buttonhole technique include
less patient-perceived pain, more rapid hemostasis,
Am J Kidney Dis. 2014;64(6):918-936
Buttonhole Versus Rope-Ladder Cannulation
and fewer fistula complications, including local and
systemic infections, aneurysm formation, and AVF
failure.7,8 Though initially thought to be a safe
procedure, more recent literature raises concerns that
the buttonhole technique may be associated with
a higher incidence of access-related infections,
including life-threatening metastatic infectious
complications.
Given the absence of large high-quality studies of
patient-relevant outcomes of cannulation technique,
we sought to systematically review and summarize
the risks and benefits of buttonhole compared to ropeladder cannulation among HD patients using the AVF
for vascular access.
METHODS
This systematic review was reported in accordance with published guidelines.9,10 The review protocol was registered with the
National Institute for Health Research PROSPERO database
(registration number CRD42012002850).
Search Strategy
A health information specialist at the University of Alberta
(D.S.) was involved in development of the search strategy. Databases included MEDLINE (1946 to March 21, 2014), EMBASE
(1974 to week 12, 2014), Cumulative Index to Nursing and Allied
Health Literature (CINAHL; 1937 to March 21, 2014), and
Evidence-Based Medicine (EBM) Reviews (to March 21, 2014).
Detailed search strategies are in Item S1 (provided as online
supplementary material). Reference lists of included articles also
were reviewed for relevant citations. Pairs of reviewers (B.W.,
R.P.P., G.N., A.G., and S.S.) independently performed title and
abstract screening in duplicate, and any study considered potentially
relevant by at least 1 reviewer was retrieved for full-text review.
Selection Criteria
Two reviewers (B.W. and M.M.) independently assessed the
full text of each potentially relevant study for inclusion using
predetermined eligibility criteria. Studies of adults 18 years or
older using buttonhole cannulation in the home or in-center setting
to cannulate incident or prevalent fistulas were included if they
reported important clinical outcomes and included a rope-ladder
control group. We included randomized trials and observational
studies. We excluded abstracts, case reports, review articles, editorials without original data, and non-English publications. Disagreements were resolved by a third party (R.P.P.).
Data Extraction
All data were extracted in duplicate and included study characteristics (country, year, study design, sample size, and study
duration), patient characteristics (age, sex, modality, dialysis vintage
at baseline, and age of AVF at baseline), specifics of the buttonhole
cannulation technique, and patient outcomes. The primary outcomes
of interest were patient-reported cannulation pain and rates of AVFrelated local and systemic infections. Secondary outcomes included
access-related interventions, survival, hospitalization, and mortality,
as well as hematoma and aneurysm formation, time to hemostasis,
and all-cause hospitalization and mortality.
Risk of Bias Assessment
For randomized studies, we evaluated the risk of bias within
studies using criteria adapted from Higgins et al11; a risk of bias
assessment tool based on the Ottawa-Newcastle criteria was
applied to observational studies.12 These criteria include items of
Am J Kidney Dis. 2014;64(6):918-936
study design (selection of participants and matching for covariates
and outcome definitions), statistical analysis (calculation of sample
size and adjustment for potential confounding), and results (loss to
follow-up).
Data Synthesis and Analysis
We compiled characteristics of each included study in tabular
form and distilled this into individual outcomes tables according to
dialysis modality: in-center HD versus mixed (in-center HD and
home HD) or home HD. We analyzed data using Stata, version
13.1 (StataCorp LP). Median values were substituted for mean
values and missing standard deviations were imputed according to
Wiebe et al.13 We calculated and pooled the standardized mean
difference14 for cannulation pain; the rope-ladder cannulation
group mean minus the buttonhole cannulation group mean value
divided by their pooled standard deviation corrected for small
sample size bias using adjusted g of Hedges and Olkin.15 By
presenting the differences in mean values relative to standard deviation, we removed the heterogeneous effect of choice of pain
instrument. Due to other differences (eg, study design, patient
population, and variations in buttonhole cannulation technique)
expected between studies, we combined results using the KnappHartung random-effects model.16 Statistical heterogeneity was
quantified using the I2 statistic.17,18
RESULTS
Search Yield
The search yielded 1,379 records (Fig 1). In total,
171 citations were retrieved for detailed evaluation.
Of these, 23 primary articles were eligible for inclusion in this systematic review.6-8,19-38
Study and Participant Characteristics
There was significant heterogeneity among studies
(Table 1). Of the 23 studies included, 5 were randomized controlled trials (RCTs)7,19,21,23,29,38 (the
second study by MacRae et al38 was a long-term
follow-up of their original RCT19), while the rest
were observational studies of various designs: prospective cohort,8,28,37 retrospective cohort,20,30,34
prospective before-after,27,31,35 retrospective beforeafter,6,22,25,26,32,36 or cross-sectional.24,33 Eight of 23
studies did not have an explicit or even implicit prespecified primary outcome. Objectives of the various
investigations differed among the remaining studies
and the authors reported only on outcomes matching
their objectives; cannulation pain is not a necessary
outcome of buttonhole research so not all studies in
this systematic review addressed our primary outcome
(needling pain). Needling pain was reported as the
primary outcome in 5 studies, and infection rate, in
another 6. No study included only incident patients or
incident fistulas. Most (17 of 23) studies were restricted to in-center HD patients, whereas 3 were limited
to home HD patients and 3 included both modalities.
Sample size ranged from 14-447 patients. When reported, the mean age of patients ranged from 48.5-70.4
years, and most patients were men (range, 38%-82%).
Dialysis vintage was reported in only 10 studies, with
median and mean values ranging from 0.34-3.7 years.
919
Wong et al
Figure 1. Flow diagram documenting inclusion and exclusion of studies.
Age of the fistula at baseline was reported in only
13 studies and ranged from a minimum of 6 months to
a maximum mean of 5.2 years. Quality assessment is
reported in Tables S1 and S2.
Cannulation Technique
There were significant variations in (and reporting of)
cannulation technique among studies (Table S3, which
includes the 36 studies identified as relevant in Fig 1). In
terms of initial track formation, track maintenance,
initial disinfection, scab removal, second disinfection,
use of local anesthetic, needle insertion technique,
needle type and gauge, self- versus nurse cannulation,
antimicrobial topical prophylaxis, and postdialysis
routine, there was inconsistent (and frequently incomplete) reporting.
Cannulation Pain
Fourteen studies, including all 5 RCTs, reported on
patient-perceived pain (Table 2).7,8,19,21,23,24,26-29,31,33,35,37
One RCT29 reported a reduction in cannulation pain
920
associated with the buttonhole technique, though this
outcome was evaluated by a single item on a questionnaire 1 week after initiation of buttonhole puncturing and cannulation pain was not a primary
outcome. Three RCTs used objective instruments of
pain evaluation (visual analogue scales) and reported
no statistical differences in pain between the 2 cannulation methods.7,19,23 The other RCT21 reported a
nominally significant increase in pain with buttonhole
cannulation (P 5 0.05), although 8 of 58 patients
randomly assigned to buttonhole cannulation abandoned this technique due to discomfort and were not
included in the subsequent analysis. Results of observational studies were inconsistent, with 4 small studies
reporting statistically diminished pain with buttonhole
cannulation,26-28,35 whereas the largest prospective
cohort study reported increased pain with buttonhole
cannulation8; the other studies did not report a statistically meaningful difference.
Of the studies that enrolled only in-center HD
patients, pooling showed a statistical reduction in
Am J Kidney Dis. 2014;64(6):918-936
Population
Study
Study Type
Study Setting
Prespecified
Primary Outcome
Modality
Patient
Type
Mean
Age (y)
N
% Male
Dialysis
Vintage at
Baseline
(y)a
Age of AVF
at Baseline
(mo)a
Duration of
Follow-up
In-Center CHD
w34 patient-y
(BH), w22
patient-y (RL)
Poland
1969-1973
Not explicitly
stated
(multiple
outcomes
reported)
In-center HD
(2-63/wk)
Prevalent 6 20
incidentb
NR
NR
NR
NR
Toma29 (2003) RCT
Japan
Not explicitly
stated
(multiple
outcomes
reported)
In-center
CHD
Prevalent
80 (37 BH,
43 RL)
60 (BH),
64 (RL)
46 (BH),
37 (RL)
NR
42.6 (BH),
3 mo
45.1 (RL)
Marticorena27 Prospective
(2006)
before-after
Canada
2002-2003
Not explicitly
stated
(multiple
outcomes
reported)
In-center
CHD
Prevalent
14
48.5
43
NR
Minimum 12 12 mo
Figueiredo24
(2008)
Crosssectional
Brazil 2007
Needling pain
In-center
CHD
Prevalent
47 (19 BH,
28 RL)
57.3
38
NR
Median 14
NA
van Loon8
(2010)
Prospective
cohort
Netherlands
2007-2008
Not explicitly
stated
(multiple
outcomes
reported)
In-center
CHD
Prevalent
145 (75 BH,
70 RL)
67 (BH),
65 (RL)
59 (BH),
67 (RL)
NR
44 (BH),
31 (RL)
9 mo
USA
Birchenough22 Partially
(2010)
retrospective
2009-2010
before-after
Infection event
rate
In-center
CHD
Prevalent
NR
NR
NR
NR
NR
14 mo
Ludlow26
(2010)
Partially
Canada 2007
retrospective
before-after
Not explicitly
stated
(multiple
outcomes
reported)
In-center
CHD
Prevalent
29
65.9 (RL),
62.9 (BH)
62 (BH),
58 (RL)
3.2
32
3 mo
Struthers7
(2010)
RCT
Needling pain
In-center
CHD
Prevalent
56 (28 BH,
28 RL)
61 (BH),
60 (RL)
58
NR
28 (BH),
25 (RL)
6 mo
Labriola25
(2011)
Partially
Belgium
retrospective
2001-2010
before-after
Infection event
rate
In-center
CHD
Both
177
70.4
65.8
NR
NR
108 mo
Twardowski6
(1979)
Retrospective
before-after
UK
921
(Continued)
Buttonhole Versus Rope-Ladder Cannulation
Am J Kidney Dis. 2014;64(6):918-936
Table 1. Study and Patient Characteristics of Included Studies
922
Table 1 (Cont’d). Study and Patient Characteristics of Included Studies
Population
Study
Study Type
Study Setting
Prespecified
Primary Outcome
Modality
Patient
Type
N
Mean
Age (y)
% Male
Dialysis
Vintage at
Baseline
(y)a
Age of AVF
at Baseline
(mo)a
Duration of
Follow-up
Prospective
cohort
USA
Not explicitly
stated
(multiple
outcomes
reported)
In-center
CHD
Prevalent
45 (21 BH,
24 RL)
56.0 (BH),
66.5 (RL)
81 (BH),
83 (RL)
NR
Minimum 6
3 mo
MacRae19
(2012)
RCT
Canada
Needling pain
In-center
CHD
Prevalent
140 (70 BH,
70 RL)
70.3 (BH),
66.7 (RL)
72.9 (BH),
65.7 (RL)
Median
2.8 (BH),
3.0 (RL)
NR
2 mo for primary
outcome,
12 mo
for infectious
complications
Aitken33
(2013)
Crosssectional
UK
Prevalence
of AVFassociated
pain
In-center
CHD
Prevalent
447 (209 BH, 60.5
238 RL)
56.6
86.5%
.1y
62
NA
Kim35 (2013)
Prospective
before-after
Korea
2009-2010
Not explicitly
stated
(multiple
outcomes
reported)
In-center
CHD
Prevalent
32
62.4
63
NR
58.8
16 wk (BH),
8 wk (RL)
Smyth37
(2013)
Prospective
cohort
Australia
2011-2012
Signs/symptoms In-center
of infection,
CHD
hematoma
formation,
aneurysm
formation,
success of
cannulation
Prevalent
104 (41 BH,
63 RL)
60 (BH), 61
(RL)
34.1 (BH),
55.6 (RL)
NR
31.2
12 wk
NR
12 mo
Vaux21 (2013) RCT
UK 2007-2010 AVF survival
at 1 y
In-center
CHD
Both
127 (58 BH,
69 RL)
62 (BH),
64 (RL)
67 (BH),
63 (RL)
NR
MacRae38
(2014)
Canada
2006-2011
Access survival
In-center
CHD
Prevalent
139 (70 BH,
69 RL)
70.2 (BH),
66.1 (RL)
73 (BH),
67(RL)
Median
Median
Median 19.2 mo
3.12 (BH),
26.2 (BH),
(BH), 17.2 mo
2.58 (RL)
31.9 (RL)
(RL)
USA
2004-2011
Primary AVF
patency
In-center
CHD
Prevalent
83 (45 BH,
38 RL)
60.9 (BH),
64.1 (RL)
64 (BH),
63 (RL)
0.36
RCT
Chan34 (2014) Retrospective
cohort
(Continued)
NR
Median 12 mo
Wong et al
Am J Kidney Dis. 2014;64(6):918-936
Pergolotti28
(2011)
Population
Study
Study Type
Study Setting
Prespecified
Primary Outcome
Modality
Patient
Type
Mean
Age (y)
N
Dialysis
Vintage at
Baseline
(y)a
% Male
Age of AVF
at Baseline
(mo)a
Duration of
Follow-up
Mixed and Home HD
Verhallen31
(2007)
Prospective
before-after
Netherlands
2004-2006
Not explicitly
stated
(multiple
outcomes
reported)
Home 3-53/
wk HD and
home NHD
Prevalent
33
49
73
2
NR
Mean 11 6 6 mo
Nesrallah32
(2010)
Retrospective
before-after
Canada
1998-2009
Staphylococcus
aureus
bacteremia
rate
Home NHD
Both
56
51.5
62
3.7
Minimum 6
286.9 patient-y
Van Eps30
(2010)
Retrospective
cohort
Australia
2003-2006
Rates of serious Home NHD
complications
and inrequiring
center
hospital
CHD
admissions
Prevalent
235 (63 NHD: 51.7 (NHD),
79.4 (NHD), 1.9 (NHD), NR
76% BH;
58.3 (CHD)
58.5 (CHD)
3.1 (CHD)
172 CHD:
92% RL)
Chow23 (2011) RCT
Australia
Needling pain
Home (unknown Both
prescriptions)
and in-center
CHD
69 (34 BH,
35 RL)
NR
70
NR
NR
O’Brien20
(2012)
Retrospective
cohort
Ireland
2004-2011
Infection event
rate
Home (unknown Prevalent
prescriptions)
and in-center
CHD
127 (74 BH,
53 RL)
Median
50 (BH),
68 (RL)
78 (BH),
55 (RL)
Median
3.2 (BH),
2.6 (RL)
56 (BH),
46 (RL)
NR
Muir36 (2014)
Retrospective
before-after
Australia
2003-2010
Coprimary:
systemic
AVF-related
infections,
fistula loss/
surgical
interventions
Home (overnight Both
or during
waking hours)
72 (BH),
30 (RL)
Median 52
73
0.34
NR
3,765 AVF-mo
(2,767 BH,
998 RL)
NHD: median
13.9 mo
(86.2 patient-y);
CHD: median
12 mo (142.7
patient-y)
6 mo
Note: Comparison group for all studies is RL cannulation, except for Nesrallah32 (in-center CHD with RL; home NHD prior to prophylactic mupirocin) and Van Eps30 (CHD [92% RL]).
Abbreviations: AVF, arteriovenous fistula; BH, buttonhole; CHD, conventional hemodialysis; HD, hemodialysis; NA, not applicable; NHD, nocturnal hemodialysis; NR, not reported; RCT,
randomized controlled trial; RL, rope-ladder; UK, United Kingdom; USA, United States of America.
a
Values are means unless otherwise indicated.
b
Whether or not the study also included some incident patients is unclear.
Buttonhole Versus Rope-Ladder Cannulation
Am J Kidney Dis. 2014;64(6):918-936
Table 1 (Cont’d). Study and Patient Characteristics of Included Studies
923
924
Table 2. Pain With Cannulation Outcome
Study
Study Type
RL Cannulation
BH Cannulation
Between-Group P
Notes
In-Center CHD
RCT
40.5% of pts “significant pain at the time of puncture at
baseline, but none with the BH technique”; 40.5% of pts
“pain with the BH technique was less than with the
conventional method”; 18.9% of pts “experienced the
same mild pain with either approach”
—
Assessed by a single item on questionnaire 1 wk after
initiation of BH puncturing
Struthers7 (2010)
RCT
1 at baseline; 1 at F/Ua
3 at baseline; 2.5 at
F/Ua
NR; NR
Measured by 10-point visual analogue scoring system
(not further described); 1/25 RL pts completing the
trial stopped using any local anesthetic vs 6/22 BH
patients (P , 0.01)
MacRae19 (2012)
RCT
1.5 (IQR, 0.4-3.2) at baseline; 1.2
(IQR, 0.4-2.4) at F/U
1.6 (IQR, 0.5-3.2) at
baseline; 1.5
(IQR, 0.5-3.4) at
F/U
P 5 0.8; P 5 0.6
Measured by 10-cm visual analogue scale; all pts
received a topical 5% lidocaine gel applied to AVF
for 5 min during the time of pain assessment;
respective within-group median pain scores did not
change over F/U of 8 wk
Vaux21 (2013)
RCT
1.2 (IQR, 1.0-1.5)
1.3 (IQR, 1.2-1.9)
P 5 0.05
Measured by self-reported modified numeric pain
rating scale (1 5 pain free, 10 5 unbearable pain); 8/
58 pts in BH group abandoned BH due to pain, and
subsequent pain scores excluded from the analysis;
37% RL pts vs 41% BH pts required topical or
injected local anesthetic use during at least 1
cannulation session during study (P 5 0.7)
Marticorena27 (2006)
Before-after
6.5 (IQR, 5.8-8.5) for venous
needle; 7.0 (IQR, 5.8-8.0) for
arterial needle
1.0 (IQR, 1.0-2.0)
for venous
needle; 1.0
(IQR, 1.0-2.0) for
arterial needle
P , 0.001; P , 0.001
Measured by 10-point visual analogue pain scale
(1 5 no pain, 10 5 extreme pain); daily HD pts
experienced less pain than those dialyzing 33/wk
throughout study (P , 0.01)
Figueiredo24 (2008)
Cross-sectional
3.08 6 2.28
2.4 6 1.7
P 5 0.1
van Loon8 (2010)
Cohort
1.0b
1.6b
P , 0.001
Ludlow26 (2010)
Before-after
2.6 6 1.4 for venous needle;
2.3 6 1.2 for arterial needle
1.9 6 1.1 for venous
needle; 1.7 6 0.8
for arterial needle
P 5 0.01; P 5 0.002
Measured by visual analogue scale, divided into 10
equal parts (0 5 lack of pain, 10 5 unbearable pain)
Measured by verbal rating 10-point scale (1 5 no pain,
10 5 extreme pain); use of local anesthetic cream
(lidocaine/prilocaine): 30% (RL) vs 8% (BH),
P , 0.001
Measured by 10-point visual analogue scale (1 5 no
pain; 10 5 severe pain)
Pergolotti28 (2011)
Cohort
Pts in BH group had average pain measurement 47.8%
less than pts in traditional group
(Continued)
P 5 0.005
Measured by visual analogue scale; no other data
reported
Wong et al
Am J Kidney Dis. 2014;64(6):918-936
Toma29 (2003)
Study
Study Type
RL Cannulation
BH Cannulation
Between-Group P
Notes
Aitken33 (2013)
Cross-sectional
3 (IQR, 2.5); .5 for 27.7%
3 (IQR, 2.4); .5 for
18.2%
P 5 0.4; P 5 0.09
Measured by 10-point visual analogue scale (0 5 “no
pain,” 10 5 “worst pain ever”); % of pts with severe
acute pain (score . 5); AVF pain unrelated to
cannulation per se (ie, chronic pain), was no different
between RL and BH
Measured by 10-cm visual analogue scale (0 5 “no
pain,” 10 5 “severe pain”)
Kim35 (2013)
Before-after
6.3 6 1.3 for venous needle;
6.1 6 1.2 for arterial needle
5.1 6 1.8 for venous
needle; 3.3 6 1.8
for arterial needle
P 5 0.001; P 5 0.001
Smyth37 (2013)
Cohort
1.82 6 0.934
1.90 6 1.069
P 5 0.6
Measured by 10-point visual rating scale (1 5 “no
pain,” 10 5 “terrible pain”); variable local
anesthetic use
Chow23 (2011)
RCT
0.81 (95% CI, 0.48-1.15) at
baseline; 0.71 (95% CI,
0.34-1.09) at F/U
0.81 (95% CI, 0.411.20) at baseline;
0.56 (95% CI,
0.13-0.99) at F/U
NS; NS
Mixed HD; measured by Wong-Baker Pain Rating
(visual analogue) Scale, from 0 (no pain) to 5 (worst
pain); use of lidocaine at last F/U: 77% of RL pts vs
44% of BH pts, P 5 0.01 (data unavailable for 17%
of pts)
Verhallen31 (2007)
Before-after
2.3 6 2.2
1.6 6 2.0
P 5 0.1
Home HD; measured by visual analogue scale; BH:
average pain score from 1.5-18 mo
Mixed and Home HD
Note: Unless otherwise indicated, values given in cannulation columns are presented as median (IQR), mean 6 standard deviation, or mean (95% CI).
Abbreviations: AVF, arteriovenous fistula; BH, buttonhole; CHD, conventional hemodialysis; CI, confidence interval; F/U, follow-up; HD, hemodialysis; IQR, interquartile range; RL, ropeladder; NR, not reported; NS, not significant; pts, patients; RCT, randomized controlled trial.
a
Values are medians.
b
Values are means.
Buttonhole Versus Rope-Ladder Cannulation
Am J Kidney Dis. 2014;64(6):918-936
Table 2 (Cont’d). Pain With Cannulation Outcome
925
Wong et al
cannulation pain with the buttonhole cannulation
technique among observational studies (standardized
mean difference, 20.76 [95% confidence interval
(CI), 21.38 to 20.15] standard deviations), but no
difference in cannulation pain among RCTs (standardized mean difference, 0.34 [95% CI, 20.76 to
1.43] standard deviations; Fig 2). Between-study heterogeneity was large (I2 5 89%). Two studies23,31
specifically addressed pain in home HD patients who
self-cannulated, and neither reported a significant difference between cannulation techniques. Local anesthetic use prior to cannulation was incompletely
reported (Table S3) and with mixed results (Table 2).
Local and Systemic Infection
The association between local and/or systemic
infection and cannulation technique was described in
15 studies (Table 3).6-8,19-23,25,29,30,32,34,36,37 Owing to
the heterogeneity of the case definition of infectious
complications, we have not pooled results. However,
all 10 studies (of which 4 were RCTs) including
exclusively conventional facility-based HD patients
reported a trend toward an increase in infectious events
among buttonhole cannulators. Furthermore, of the 5
studies that reported between-group statistics, 3
demonstrated significant differences.8,19,25 A similar
trend was observed for studies including selfcannulating home HD populations (1 was an
RCT).20,23,30,32,36 No study in this review had infectious complications as a primary end point and no RCT
was specifically powered to discern differences in
rates. Table 3 delineates culture results (when reported)
and shows a preponderance of Staphylococcus aureus
bacteremia with buttonhole cannulation, with complications including endocarditis, discitis, septic arthritis,
septic emboli, and death.
Three studies reported within-group improvement in
infectious complications when the buttonhole technique
was subjected to stricter procedures,22 a buttonhole
education workshop for nursing staff was introduced,25
or mupirocin prophylaxis cream was introduced.32
However, infectious event rates for buttonhole patients
were still higher than for the rope-ladder comparator
groups despite these measures to reduce infection risk.
Access Intervention
Seven studies evaluated the association between
need for access intervention and cannulation technique
(Table 4).8,21,26,27,34,36,38 A prospective cohort study
demonstrated a reduction in access interventions
(angioplasty) associated with buttonhole technique;
however, buttonhole patients were recruited from a
different HD unit than the rope-ladder control group.8
The RCT by Vaux et al21 showed a similar trend,
with rope-ladder patients requiring 0.4 intervention
(angioplasty, thrombectomy, or surgical revision) per
patient-year to maintain access functionality, and
buttonhole patients requiring only 0.2 intervention per
patient-year (buttonhole, n 5 58; rope-ladder, n 5 69;
significance level not reported). It is important to note
that a polycarbonate peg was used for initial buttonhole
track creation; patients assigned to the buttonhole
group also had fewer previous accesses at baseline
despite similar HD vintage, previous catheter use, and
number of interventions to the AVF prior to the start of
the study. A subsequent RCT by MacRae et al,38 in
long-term (.18 months) median follow-up, did not
demonstrate differences in intervention requirements,
including fistulogram, percutaneous transluminal angioplasty, and surgical intervention in an intention-totreat analysis, though an as-treated analysis suggested
that buttonhole technique required more fistulograms
(P , 0.001) and more percutaneous transluminal angioplasties (P 5 0.003). The other studies, consisting
of in-center HD26,27,34 or home HD patients,36 did not
demonstrate differences in access interventions.
Access Survival and Access-Related Hospitalization
and Mortality
Figure 2. Forest plot of cannulation pain among in-center hemodialysis (HD) patients (insufficient studies for creating forest
plot for home HD patients). Abbreviations: BH, buttonhole;
RCT, randomized controlled trial; RL, rope-ladder; SMD, standardized mean difference.
926
Three studies enrolling only in-center HD patients
evaluated AVF survival (Table 4). Two RCTs specifically evaluated AVF survival as the primary end
point.21,38 The RCT by Vaux et al,21 in which a
polycarbonate peg was used for initial buttonhole
track creation, reported 100% access survival for patients randomly assigned to buttonhole cannulation
compared to 86% in the rope-ladder group
(P 5 0.005). In contrast, in long-term follow-up of
the RCT by MacRae et al,38 similar median access
survival times were observed (rope-ladder vs buttonhole: 16.0 vs 18.4 months; P 5 0.2). Furthermore,
Chan et al34 found no difference in AVF patency at 3,
Am J Kidney Dis. 2014;64(6):918-936
Infection Rate
Study
Study Type
Micro-organisms Identified
Complications
RL Cannulation
BH Cannulation
Between-Group P
Notes
In-Center CHD
Toma29 (2003)
RCT
NR
NR
0 events/1,000
AVF-d
0.9 events/1000
AVF-d
NR
Local infectious event defined by
1 of the following: redness,
swelling, tenderness, exudate or
pus; limited to a 3-mo
observation period
Struthers7
(2010)
MacRae19
(2012)
RCT
NR
NR
0/28 (0%) pts
1/28 (3.6%) pts
NR
NR
RCT
BH: 1 S aureus
bacteremia; RL: no
positive cultures
NR
9.6 events/1,000
AVF-d
21.4 events/1,000
AVF-d
P 5 0.003
Vaux21 (2013)
RCT
NR
Bacteremias:
0.09 event/1,000
AVF-d; ESIs:
0/1,000 AVF-d
0.12 event/1,000
AVF-d
Bacteremias: 0/1,000
AVF-d; ESIs: 0.12
event/1,000 AVF d
NR; NR
Localized infection defined as
erythema, pus, or swelling at
fistula site during 8-wk study
period; after study completion
(within 12 mo) 2 BH pts
developed S aureus bacteremia
and 9 developed AVF abscess;
0 complications reported with RL
Infection rate defined by
bacteremias or ESIs
0.24 event/1,000
AVF-d
NR
Infectious events defined by
those requiring antibiotics
P 5 0.001
Access-related infection defined
by need for antibiotics; cannot
determine AVF-d of follow-up
BH: no positive
cultures; RL: 2
S aureus
bacteremias
Before-after
NR
NR
van Loon8
(2010)
Cohort
NR
0/70 (0%) pts
4/75 (5.3%) pts
Birchenough22
(2010)
Before-after NR
NR
0.13 event/1,000
AVF-d
Labriola25
(2011)
Before-after 22 S aureus, 11
S epidermidis
3 fatal endocarditis;
2 other fatal
bacteremias;
Event/1,000
AVF-d: period
1, 0.17 (95%
1.4 event/1,000 AVF-d; NR
Access infection defined by
after revised BH
temperature . 99 F, redness
policy/procedure
and swelling at cannulation
initiated: 0.82
sites, and positive blood
event/1,000 AVF-d
cultures if performed
Event/1,000 AVF-d:
P 5 0.003
Infectious event defined as nonallergic
period 3, 0.43
(periods 1 and
erythema, pain, or tenderness close
(95% CI, 0.29-0.61);
2 combined
to cannulation sites, necrotic scabs,
Twardowski6
(1979)
BH: 2 S aureus
bacteremias; 2 local
infections (S aureus
& Clostridium
perfringens); RL:
no positive cultures
(Continued)
Buttonhole Versus Rope-Ladder Cannulation
Am J Kidney Dis. 2014;64(6):918-936
Table 3. Infectious Outcomes
927
928
Table 3 (Cont’d). Infectious Outcomes
Infection Rate
Study
Study Type
Micro-organisms Identified
(RL vs BH
distribution NR)
Complications
1 nonfatal
endocarditis;
1 metastatic
costovertebral
arthritis
RL Cannulation
BH Cannulation
CI, 0.0860.31); period 2,
0.11 (95% CI,
0.0014-0.63)
period 4, 0.34
(95% CI, 0.19-0.55)
Between-Group P
compared with
period 3)
Notes
or drainage from cannulation site(s)
and/or bacteremia caused by a skin
micro-organism even without local
signs; period 1: all pts using RL
technique, period 2: progressive
switch to BH method, period 3: all
pts using BH method before
educational workshops, period 4: all
pts using BH method after
educational workshops
Signs and symptoms of infection
(redness, swelling, heat,
discharge, pain) assessed by
clinical observation, and
pathology (wound swab,
blood culture) when applicable
Smyth37 (2013) Cohort
BH: NR; RL: 1
S aureus local
infection
NR
0.76 event/1,000
AVF-d
0.87 event/1,000
AVF-d
P 5 0.9
Chan34 (2014)
Cohort
NR
NR
5/45 (8%) pts
3/38 (11%) pts
P 5 0.6
Bacteremia rates defined as at
least 1 positive blood culture
with definite or probable association
with infection secondary to the AVF,
treated with antibiotics
Chow23 (2011)
RCT
BH: 1 Klebsiella
pneumoniae and
coagulase-negative
staphylococcus;
RL: no positive
cultures
Presented with
1/35 (2.9%) pts
features of local
infection and fever
4/34 (11.8%) pts
P 5 0.1
Mixed HD; exit-site infection 6
bacteremia; the only pt in RL
group with infection was
dialyzing at home and had
self-initiated BH cannulation;
after the infection, the pt
resumed RL cannulation
without incident; cannot
determine AVF-d of follow-up
Nesrallah32
(2010)
2 septic pulmonary
Before-after BH: 11 S aureus,
emboli; 1 cervical
8 coagulase-negative
spine abscess
staphylococcus,
resulting in
1 Enterococcus faecalis,
quadriplegia;
1 Enterobacter cloacae,
1 septic joint;
1 hemolytic
6 local infections
streptococcus; RL: NR
with fever
Event/1,000 AVF-d:
before mupirocin
prophylaxis, 0.23;
after mupirocin
prophylaxis, 0.03
NR
Home HD; primary outcome
event defined as S aureus
bacteremia rate
Mixed and Home HD
(Continued)
Wong et al
Am J Kidney Dis. 2014;64(6):918-936
0.005 event/1,000
AVF-d
Infection Rate
Study
Study Type
Micro-organisms Identified
Complications
RL Cannulation
BH Cannulation
Between-Group P
Notes
Van Eps30
(2010)
Cohort
Among NHD pts (76%
using BH), 71.4% of
positive cultures
isolated Gram-positive
organisms (60% of
these are S aureus);
Gram-negative
organisms isolated
in 7.1%
NR
0.09 (95% CI,
0.04-0.16)
event/pt-y; ie,
0.25 event/1,000
AVF-d
0.22 (95% CI,
0.12-0.38)
events/pt-y; ie,
0.60 event/1,000
AVF-d
P 5 0.04 for
IRR of 3.0
(1.04-8.66)
Mixed HD (CHD pts: 92% RL; NHD
pts: 76% BH); septic events,
defined as admission to hospital
with local changes at access site
suggesting infection (eg, erythema,
warmth, pain, discharge from
cannulation, or CVC exit sites)
and/or fever or rigors, which required
intravenous antibiotic treatment
documented or suspected sepsis
when the dialysis access was
deemed by the treating clinicians
to be the most probable portal of
entry for the infection and no other
cause was identified
O’Brien20
(2012)
Cohort
BH: 8 S aureus, 1
S epidermidis; RL:
no positive cultures
4 endocarditis
(1 fatality);
1 cervical discitis
0 event/1,000
AVF-d
0.073 event/1,000
AVF-d
P 5 0.02
Muir36 (2014)
Before-after NR
NR
Event/1,000 AVF-d: Event/1,000 AVF-d:
measure 1,
measure 1, 0.20;
measure 2, 0.18;
0.03; measure 2,
0.07; measure 3,
measure 3, 0.39
0.10
Mixed HD; AVF-associated
infection defined by a positive
blood culture yielding typical
pathogens with cellulitis overlying
the AVF
Home HD; measure 1: local AVF
infections defined as erythema,
pain, or swelling of fistula site
requiring treatment with oral
antibiotics with negative blood
cultures in the absence of
systemic symptoms; measure
2: systemic AVF-related infection
events (blood culture–positive
sepsis or complicated infection
[eg, endocarditis] in absence of
identifiable non-AVF source);
measure 3: total AVF-related
infection events (systemic
infections 1 local infections)
NS; NS;
P # 0.05
Abbreviations: AVF, arteriovenous fistula; BH, buttonhole; CI, confidence interval; CVC, central venous catheter; ESI, exit-site infection; HD, hemodialysis; CHD, conventional hemodialysis;
IRR, incidence rate ratio; NHD, nocturnal hemodialysis; NR, not reported; NS, not significant; pt, patient; RCT, randomized controlled trial; RL, rope-ladder; S aureus, Staphylococcus aureus;
S epidermidis, Staphylococcus epidermidis.
Buttonhole Versus Rope-Ladder Cannulation
Am J Kidney Dis. 2014;64(6):918-936
Table 3 (Cont’d). Infectious Outcomes
929
930
Table 4. Secondary Outcomes
Study
RL Cannulation
BH Cannulation
Between-Group P
Study Type
Dialysis Modality
Note
Vaux21 (2013)
RCT
In-center
27/69 (39%) pts; 0.4
intervention/pt-y
11/58 (19%) pts; 0.2
intervention/pt-y
NR
MacRae38 (2014)
RCT
In-center
Event/pt-y: fistulogram, 0.99;
percutaneous transluminal
angioplasty, 0.90; surgical
intervention, 0.09
0%
P 5 0.2; P 5 0.4; Rates adjusted for age, coronary artery
P 5 0.4
disease, and AVF location
Marticorena27 (2006) Before-after In-center
Event/pt-y: fistulogram, 0.75;
percutaneous transluminal
angioplasty, 0.72; surgical
intervention, 0.11
0%
NS
Percent pts requiring interventions to
maintain access patency
van Loon8 (2010)
Cohort
41/70 (58.6%)
10/75 (13.3%)
P 5 0.001
Access interventions (angioplasty,
thrombectomy, surgical revision,
antibiotics for access infections); 41
interventions in RL pts in 21 pts; 10
interventions in BH pts in 6 pts; RL vs BH
pts recruited from different HD units
Ludlow26 (2010)
Before-after In-center
Surgical repair of access: 6
Surgical repair of access: 2
NS; NS; NS; NS
(6.3%); venogram: 14 (14.7%);
(6.9%); venogram: 1 (3.4%);
venogram & angioplasty: 13
venogram & angioplasty: 1
(13.7%); CVC insertion: 8
(3.4%); CVC insertion: 2 (6.9%)
(8.4%)
Chan34 (2014)
Cohort
27/38 (71%) pts
29/45 (64%) pts
P 5 0.5
Referrals for intervention include 1 or a
combination of the following: physical
examination abnormalities, poor blood
flows, elevated venous/arterial pressures,
increased bleeding, inability to cannulate,
decrease in Transonic access flows, or
decrease in dialysis adequacy measured
by Kt/V or URR
Muir36 (2014)
Before-after Home
0.27 event/1,000 AVF-d
0.29 event/1,000 AVF-d
NS
Fistula loss/surgical intervention
requirement events (defined using
composite end point of any surgical
revision or episode requiring permanent
abandonment of AVF and/or creation of
new AVF); radiologic interventions were
excluded
Access Intervention
In-center
In-center
—
Wong et al
Am J Kidney Dis. 2014;64(6):918-936
(Continued)
Radiologic (fistuloplasty or thrombectomy) or
surgical (revision or thrombectomy)
interventions; pts assigned to BH group
had fewer previous accesses at baseline
(9 vs 20) despite same HD vintage,
previous catheter use, and no. of
interventions to AVF prior to start of study
Study
RL Cannulation
BH Cannulation
Between-Group P
Study Type
Dialysis Modality
Vaux21 (2013)
RCT
In-center
9/69 (13%)
MacRae38 (2014)
RCT
In-center
16.0 (IQR, 10.6-29.3) mo
Chan34 (2014)
Cohort
In-center
3 mo: 89%, 6 mo: 71%, 12 mo:
60%
Toma29 (2003)
RCT
In-center
,5 min in 27.9%; 5-10 min in
58.1%; .10 min in 14.0%
,5 min in 54.1%; 5-10 min in
40.5%; .10 min in 5.4%
NR
Hemostasis defined as time until bleeding
stopped at puncture site after withdrawal
of needle at end of treatment
Struthers7 (2010)
RCT
In-center
6.7 6 0.55 min (baseline);
6.7 6 0.52 min (6 mo)
5.7 6 0.60 min (baseline);
5.0 6 0.43 min (6 mo)
NS
—
MacRae19 (2012)
RCT
In-center
Rate of no bleeding post-HD:
23.6/1,000 dialysis sessions;
rate of any bleeding post-HD:
97.6/1,000 dialysis sessions
Rate of no bleeding post-HD:
28.3/1,000 dialysis sessions;
rate of any bleeding post-HD:
97.2/1,000 dialysis sessions
P 5 0.4; P 5 0.9
F/U of 8 wk
Vaux21 (2013)
RCT
In-center
9.1 (IQR, 6.9-11.3) min
7.9 (IQR, 6.3-10.4) min
P 5 0.3
Hemostasis defined as total time from
removal of first needle to achievement of
hemostasis for both needles
Marticorena27 (2006) Before-after In-center
20 (IQR, 15-40) min
13 (IQR, 9-20) min
P , 0.001
Hemostasis determined as follows: 5 min
after removing each needle (venous first),
pressure on site was released; if blood
appeared through folded gauze, pressure
was reapplied immediately in same
manner; this process was repeated every
5 min until there was no sign of fresh blood
in needle entrance site
Ludlow26 (2010)
Before-after In-center
14.08 6 3.31 min
13.72 6 3.99 min
NS
Hemostasis not defined
Pergolotti28 (2011)
Cohort
Pts in BH group had average hemostasis time 23.7% shorter than pts P 5 0.007
in RL group
Kim35 (2013)
Before-after In-center
Access Survival
0/58 (0%)
Note
P 5 0.005
Access failure defined as AVF no longer
used for successful HD at 1 y
18.4 (IQR, 10.9-32.7) mo
P 5 0.2
3 mo: 86%, 6 mo: 72%, 12 mo:
57%
P 5 0.6
Access survival defined as time from study
start until time of access abandonment,
death, modality transfer, or end of F/U
Primary patency evaluated from initial date
of BH or matched RL to date of referral for
access dysfunction
Hemostasis
In-center
14/64 (21.9%) evaluations
34/64 (53.1%) evaluations
931
(Continued)
P 5 0.001
Hemostasis defined as cessation of
puncture bleeding at arterial site as
assessed at 1-min and then 30-sec
intervals following needle removal
Proportion of pts achieving complete
hemostasis within 15 min after needle
removal
Buttonhole Versus Rope-Ladder Cannulation
Am J Kidney Dis. 2014;64(6):918-936
Table 4 (Cont’d). Secondary Outcomes
932
Table 4 (Cont’d). Secondary Outcomes
Study
Study Type
Dialysis Modality
RL Cannulation
BH Cannulation
Between-Group P
Note
Chow23 (2011)
Verhallen31 (2007)
RCT
Mixed
Before-after Home
8.7 6 3.6 min
Struthers7 (2010)
RCT
In-center
Hematoma Formation
27 reported episodes over 6 mo 19 reported episodes over 6 mo
NR
Hematoma defined as peripuncture
subcutaneous infiltration
MacRae19 (2012)
RCT
In-center
436/1,000 dialysis sessions
295/1,000 dialysis sessions
P 5 0.03
F/U of 8 wk; hematoma categorized as 04.9, 5-10, .10 cm (not further defined)
Twardowski6 (1979)
Before-after In-center
12.5%
0.1%
NR
Hematoma formation not defined;
presumably expressed as percentage of
no. of HD sessions
van Loon8 (2010)
Cohort
14.0 6 15.6 (no. per pt over 9 mo) 2.0 6 3.7 (no. per pt over 9 mo)
P , 0.0001
Hematoma defined as abnormal infiltration
of blood
Ludlow26 (2010)
Smyth37 (2013)
Before-after In-center
Cohort
In-center
33.7% of pts over 3 mo
31.7% of pts over 12 wk
37.9% of pts over 3 mo
26.8% of pts over 12 wk
NS
P 5 0.8
Hematoma defined as needle infiltration
Hematoma defined as abnormal localized
infiltration of blood caused by needle
cannulation
Chow23 (2011)
RCT
Mixed
0% of pts over 6 mo
11.8% of pts over 6 mo
P 5 0.03
Hematoma formation not defined
Aneurysm Formation
30% 6 7% increase in diameter of 1% 6 22% increase in diameter of P , 0.01
AVF over 6 mo
AVF over 6 mo
In-center
No difference in hemostasis time
7.6 6 4.0 min
P 5 0.004
Details NR
Hemostasis defined as average
compression time after cannula removal;
BH value based on times collected from
1.5-18 mo of F/U
RCT
In-center
Vaux21 (2013)
RCT
In-center
17% of pts develop new aneurysm 4% of pts develop new aneurysm NR; NR
over 12 mo; 67% of pts with
over 12 mo; 23% of pts with
enlargement of existing
enlargement of existing
aneurysm over 12 mo
aneurysm over 12 mo
van Loon8 (2009)
Cohort
In-center
67% of pts over 9 mo
1% of pts over 9 mo
P , 0.001
Aneurysm formation defined as localized
dilatation of vessel
Smyth37 (2013)
Cohort
In-center
28.6% of pts over 12 wk
7.3% of pts over 12 wk
P 5 0.02
Aneurysm defined as localized dilatation of
vessel on visual inspection
Fistulas from both groups were
photographed and maximum transverse
diameter measured prerandomization and
at study termination
Aneurysm defined as new
dilatation . 0.5 cm; enlargement of
existing aneurysm defined as
dilatation . 0.5 cm
Note: In cannulation columns, values given with IQRs are medians; values given before and after 6 are mean and standard deviation, respectively.
Abbreviations: AVF, arteriovenous fistula; BH, buttonhole; CVC, central venous catheter; F/U, follow-up; HD, hemodialysis; IQR, interquartile range; NR, not reported; NS, not significant; pt,
patient; RCT, randomized controlled trial; RL, rope-ladder; URR, urea reduction ratio.
Wong et al
Am J Kidney Dis. 2014;64(6):918-936
Struthers7 (2010)
Buttonhole Versus Rope-Ladder Cannulation
6, and 9 months between the rope-ladder and
buttonhole cannulation techniques. None of the
studies was powered to assess access-related hospitalization and mortality rates, and follow-up generally
was too short to yield meaningful results (median or
mean, #14 months in 15 of 23 studies).
Other Secondary Outcomes
Ten studies, including all 5 RCTs, reported some
measure of hemostasis after needle removal; however,
the evaluation of hemostasis was inconsistent between studies (Table 4).7,19,21,23,26-29,31,35 While 4 of
5 observational studies reported a statistically significant shortened time to hemostasis, 4 RCTs (of both
in-center HD patients and a mixture of in-center HD
and home HD patients) concluded that buttonhole
cannulation did not affect hemostasis,7,19,21,23 and one
suggested a benefit but did not report summary statistics to substantiate that conclusion.29
There also was considerable variability in the
definition of hematoma formation and how this was
quantified (Table 4). Qualitatively, 2 studies reported
statistically reduced rates of hematoma with buttonhole cannulation,8,19 2 further studies reported large
differences favoring buttonhole cannulation without
reporting summary statistics,6,7 and 2 studies demonstrated similar hematoma rates between the ropeladder and buttonhole techniques.26,37 One RCT
(evaluating a mixture of in-center HD and home HD
patients) showed an increased frequency of hematoma
formation with the buttonhole technique, but this
finding was attributed to lack of buttonhole experience by the nursing staff.23
All 4 studies (of only in-center HD patients) reporting on aneurysm formation suggested a benefit
with buttonhole versus rope-ladder cannulation, though
for only 3 were these results statistically significant
(Table 4).7,8,37
One study assessed all-cause hospitalization rate.30
However, it compared home HD patients versus incenter HD patients without directly comparing difference in cannulation technique. All-cause mortality
was not assessed systematically in any study.
DISCUSSION
Cannulation dogma suggests that compared to
traditional rope-ladder cannulation, buttonhole cannulation causes less pain, facilitates ease of cannulation, reduces hematoma formation from needle
infiltration, results in more rapid hemostasis, and decreases the incidence of aneurysm formation. In this
review, we used rigorous methods to summarize the
available literature by describing the benefits and
risks of buttonhole cannulation. For our primary
outcome of cannulation pain, we found no evidence in
pain score reduction with buttonhole cannulation
Am J Kidney Dis. 2014;64(6):918-936
irrespective of whether patients were self-cannulating
home HD or in-center HD patients. Compared to the
rope-ladder technique, buttonhole cannulation also
appeared to be associated with increased risk of local
and systemic infections.
Pain with cannulation is the most comprehensively
studied outcome among those included in this review
and the primary end point of 3 RCTs.7,19,23 Though
often cited as a key benefit of buttonhole cannulation,
the literature does not support the contention that pain
is reduced with this technique in either facility-based
HD or self-cannulating home HD patients. The trial of
MacRae et al,19 in particular, demonstrates median
pain scores of 1.2 (interquartile range, 0.4-2.4) versus
1.5 (range, 0.5-3.4; P 5 0.6) on a 10-point visual
analogue scale for rope-ladder versus buttonhole
cannulation at 8 weeks in the setting of an adequately
powered and well-designed RCT in facility-based
HD. While not obviously generalizable to selfcannulating home HD patients, cannulation pain in
this latter population has been the subject of only 2
studies, neither of which report statistically significant
improvements in pain.23,31 However, both these studies
are small and potentially underpowered, raising the
possibility of failing to detect a meaningful difference
between cannulation techniques when one may exist in
reality (type 2 error); a larger study is needed to address
this definitively.
A key risk with buttonhole cannulation is local and/
or systemic infection, and a consistent signal for increased infectious complications (with major morbidity) compared with rope-ladder emerges. However, a
granular comparison of infectious outcomes between
buttonhole and rope-ladder techniques is not possible
because of the disparate and often imprecise definition of those outcomes. These range from “exit-site
infection,”23 in whatever way that is interpreted by
individual clinicians, to local infection defined by
“one of the following: redness, swelling, tenderness,
exudate or pus,”29 to nominally defined specific events
such as S aureus bacteremia confirmed by blood
culture.32 A second potential source of variation in
outcomes is the patient population, whether selfcannulating at home or nurse-cannulated in facilitybased HD. Although home HD patients tend to be
younger and healthier than the overall dialysisdependent population, they may have potentially
higher infectious complications because: (1) they
typically dialyze more frequently and thus access their
AVFs more often, (3) there is infrequent surveillance
by dialysis professionals to intervene at early signs of
an infection, and (3) lapses in strict adherence to
protocol are possibly more likely to occur in an unsupervised home environment. This is particularly
relevant because a disproportionate number of home
HD patients use buttonhole cannulation compared
933
Wong et al
with facility-based comparator groups, which tend to
use the rope-ladder technique.20,30,32 Thus, it is not
surprising that in all studies including home HD patients, the infectious event rate is higher for buttonhole than rope-ladder cannulation even if not always
statistically significant.20,23,30,32,36 Finally, a worrisome signal appears to emerge that the complications
of infectious events associated with buttonhole cannulation include more S aureus bacteremias, leading to
sepsis, endocarditis, discitis, septic arthritis, septic
emboli, and death. This important observation requires
more definitive study.
There is a paucity of literature concerning the
secondary outcomes under consideration in this review. Whether considering access survival, access
interventions, access-related hospitalizations or mortality, hemostasis, or all-cause hospitalizations or
mortality, there are no data favoring one cannulation
technique over the other. Only hematoma and aneurysm formation appear to favor buttonhole users.
Although much has been written about buttonhole
cannulation, many of these studies were simply
descriptive (without an appropriate control group) and
were not included in this systematic review.39-51 Even
in the studies that informed this review, there are a
number of critical limitations to temper definitive
conclusions. First, most studies were observational in
nature, often with poorly defined or inappropriate
control groups. Quasi-experimental designs, which
dominate this literature, are subject to all the same
limitations to internal validity as other observational
studies (selection bias, information bias, and confounding), but also bias arising from temporal changes
in natural history and clinical care that are unrelated to
the intervention under investigation.52 Even among
the RCTs, there is an incomplete description of the
randomization process, including allocation concealment, in 4 of 5 trials, and the inherent nature of the
intervention that precludes blinding (Table S2). Second, there were small sample sizes, accordant with
limited power to detect meaningful differences where
they may exist in reality; most studies had far fewer
than 50 participants per intervention arm. Third,
follow-up duration was short: most studies had 12 or
fewer months of follow-up. This is particularly relevant for infectious complications because it is likely
that serious adverse infectious events are late complications of buttonhole cannulation. Fourth, there
was inconsistent description of outcomes: pooling
point estimates (whether of infection rates, aneurysm
formation rates, etc) requires reasonably homogeneous
definitions of those outcomes in order to yield a
meaningful estimate. For most outcomes, pooling data
from the studies included in this review would be
misleading, as described. Finally, there was a paucity
of details describing creation and maintenance of the
934
buttonhole: review of the literature suggests significant
variability in how buttonholes initially are developed
and the precise stepwise execution of the procedure
once in use (Table S3). As an illustrative example, one
can reasonably speculate that cannulation by variably
experienced staff can easily result in recurrent injury to
the tunnel track, resulting in pain, needle infiltration,
and perhaps infection, all complications that potentially
are avoidable if cannulation is restricted to highly
specialized nurses or the patients themselves. Comparison of the 2 recent RCTs showed significant differences in buttonhole cannulation comfort levels
between nurses in both studies.19,21 Most cannulators
in the study by Vaux et al21 preferred the buttonhole
technique, whereas the opposite was true in the study
by MacRae et al,19 and this comfort level (perhaps
reflecting experience) is a critical and modifiable
determinant of adverse events.
Of the outcomes under consideration in this systematic review, the best available data are for cannulation pain and infectious complications and
would suggest that buttonhole cannulation does not
significantly alleviate cannulation pain and likely
increases the risk of potentially serious infectious
complications. Thus, the evidence does not support the
preferential routine use of buttonhole over rope-ladder
cannulation in facility-based or home HD. The use of
buttonhole cannulation may be appropriate for patients
with specific indications such as tortuous, deep, or
short AVFs, as well as patients with a history of
difficult cannulations (though there is no specific evidence to guide this practice). However, because of the
overall poor quality of the literature, larger RCTs
with longer follow-up and well-defined clinical outcomes are needed to definitively identify the risks
and benefits of buttonhole in comparison to ropeladder cannulation.
ACKNOWLEDGEMENTS
Support: None.
Financial Disclosure: The authors declare that they have no
relevant financial interests.
Contributions: Research idea and study design: BW, RPP, SK,
NP; literature search: DS; title/abstract review: BW, RP, GN, AG,
SS; full-text review: BW, MM; data extraction: BW, RPP; risk of
bias assessment: BW, RPP; data synthesis/analysis: BW, NW,
RPP; supervision or mentorship: RPP. Each author contributed
important intellectual content during manuscript drafting or revision and accepts accountability for the overall work by ensuring
that questions pertaining to the accuracy or integrity of any portion
of the work are appropriately investigated and resolved. BW takes
responsibility that this study has been reported honestly, accurately, and transparently; that no important aspects of the study
have been omitted; and that any discrepancies from the study as
planned (and registered) have been explained.
SUPPLEMENTARY MATERIAL
Table S1: Quality assessment of observational studies.
Table S2: Quality assessment of RCTs.
Am J Kidney Dis. 2014;64(6):918-936
Buttonhole Versus Rope-Ladder Cannulation
Table S3: Variations of buttonhole creation and cannulation.
Item S1: Search strategy summary.
Note: The supplementary material accompanying this article
(http://dx.doi.org/10.1053/j.ajkd.2014.06.018) is available at
www.ajkd.org
REFERENCES
1. Dhingra RK, Young EW, Hulbert-Shearon TE, Leavey SF,
Port FK. Type of vascular access and mortality in U.S. hemodialysis patients. Kidney Int. 2001;60(4):1443-1451.
2. Hoen B, Paul-Dauphin A, Hestin D, Kessler M. EPIBACDIAL: a multicenter prospective study of risk factors for
bacteremia in chronic hemodialysis patients. J Am Soc Nephrol.
1998;9(5):869-876.
3. Pastan S, Soucie JM, McClellan WM. Vascular access and
increased risk of death among hemodialysis patients. Kidney Int.
2002;62(2):620-626.
4. Huber TS, Carter JW, Carter RL, Seeger JM. Patency of
autogenous and polytetrafluoroethylene upper extremity arteriovenous hemodialysis accesses: a systematic review. J Vasc Surg.
2003;38(5):1005-1011.
5. Perera GB, Mueller MP, Kubaska SM, Wilson SE,
Lawrence PF, Fujitani RM. Superiority of autogenous arteriovenous hemodialysis access: maintenance of function with fewer
secondary interventions. Ann Vasc Surg. 2004;18(1):66-73.
6. Twardowski Z, Kubara H. Different sites versus constant
sites of needle insertion into arteriovenous fistulas for treatment by
repeated dialysis. Dial Transplant. 1979;8(10):978.
7. Struthers J, Allan A, Peel RK, Lambie SH. Buttonhole
needling of arteriovenous fistulae: a randomized controlled trial.
ASAIO J. 2010;56(4):319-322.
8. van Loon MM, Goovaerts T, Kessels AG, van der
Sande FM, Tordoir JH. Buttonhole needling of haemodialysis
arteriovenous fistulae results in less complications and interventions compared to the rope-ladder technique. Nephrol Dial
Transplant. 2010;25(1):225-230.
9. Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA
Group. Preferred Reporting Items for Systematic Reviews and
Meta-analyses: the PRISMA statement. PLoS Med. 2009;6(7):
e1000097.
10. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of
observational studies in epidemiology: a proposal for reporting.
Meta-analysis of Observational Studies in Epidemiology
(MOOSE) Group. JAMA. 2000;283(15):2008-2012.
11. Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane
Collaboration’s tool for assessing risk of bias in randomised trials.
BMJ. 2011;343:d5928.
12. Wells GA, Shea B, O’Connell D, et al. The NewcastleOttawa Scale (NOS) for assessing the quality of nonrandomized
studies in meta-analysis. http://www.ohri.ca/programs/clinical_
epidemiology/oxford.asp. Accessed May 4, 2014.
13. Wiebe N, Vandermeer B, Platt RW, Klassen TP, Moher D,
Barrowman NJ. A systematic review identifies a lack of standardization in methods for handling missing variance data. J Clin
Epidemiol. 2006;59(4):342-353.
14. Hedges L, Olkin I. Estimation of a single effect size:
parametric and nonparametric methods. In: Statistical Methods for
Meta-analysis. Toronto, Canada: Academic Press; 1985:75-106.
15. Hedges L, Olkin I. Statistical Methods for Meta-analysis.
San Diego, CA: Academic Press; 1985: Chapter 5.
16. Cornell JE, Mulrow CD, Localio R, et al. Random-effects
meta-analysis of inconsistent effects: a time for change. Ann Intern
Med. 2014;160(4):267-270.
Am J Kidney Dis. 2014;64(6):918-936
17. Higgins JP, Thompson SG. Quantifying heterogeneity in a
meta-analysis. Stat Med. 2002;21(11):1539-1558.
18. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring
inconsistency in meta-analyses. BMJ. 2003;327(7414):557-560.
19. MacRae JM, Ahmed SB, Atkar R, Hemmelgarn BR.
A randomized trial comparing buttonhole with rope ladder
needling in conventional hemodialysis patients. Clin J Am Soc
Nephrol. 2012;7(10):1632-1638.
20. O’Brien F, Kok H, O’Kane C, et al. Arterio-venous fistula
buttonhole cannulation technique: a retrospective analysis of infectious complications. Clin Kidney J. 2012;5:526-529.
21. Vaux E, King J, Lloyd S, et al. Effect of buttonhole cannulation with a polycarbonate peg on in-center hemodialysis fistula outcomes: a randomized controlled trial. Am J Kidney Dis.
2013;62(1):81-88.
22. Birchenough E, Moore C, Stevens K, Stewart S. Buttonhole
cannulation in adult patients on hemodialysis: an increased risk of
infection? Nephrol Nurs J. 2010;37(5):491-498. 555; quiz 499.
23. Chow J, Rayment G, San Miguel S, Gilbert M. A randomised controlled trial of buttonhole cannulation for the prevention
of fistula access complications. J Ren Care. 2011;37(2):85-93.
24. Figueiredo AE, Viegas A, Monteiro M, Poli-deFigueiredo CE. Research into pain perception with arteriovenous
fistula (AVF) cannulation. J Ren Care. 2008;34(4):169-172.
25. Labriola L, Crott R, Desmet C, Andre G, Jadoul M.
Infectious complications following conversion to buttonhole cannulation of native arteriovenous fistulas: a quality improvement
report. Am J Kidney Dis. 2011;57(3):442-448.
26. Ludlow V. Buttonhole cannulation in hemodialysis: improved
outcomes and increased expense—is it worth it? CANNT J. 2010;
20(1):29-37.
27. Marticorena RM, Hunter J, Macleod S, et al. The salvage of
aneurysmal fistulae utilizing a modified buttonhole cannulation
technique and multiple cannulators. Hemodial Int. 2006;10(2):
193-200.
28. Pergolotti A, Rich E, Lock K. The effect of the buttonhole
method vs. 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.
29. Toma S, Shinzato T, Fukui H, et al. A timesaving method
to create a fixed puncture route for the buttonhole technique.
Nephrol Dial Transplant. 2003;18(10):2118-2121.
30. Van Eps CL, Jones M, Ng T, et al. The impact of extendedhours home hemodialysis and buttonhole cannulation technique on
hospitalization rates for septic events related to dialysis access.
Hemodial Int. 2010;14(4):451-463.
31. Verhallen AM, Kooistra MP, van Jaarsveld BC. Cannulating in haemodialysis: rope-ladder or buttonhole technique?
Nephrol Dial Transplant. 2007;22(9):2601-2604.
32. Nesrallah GE, Cuerden M, Wong JH, Pierratos A. Staphylococcus aureus bacteremia and buttonhole cannulation: longterm safety and efficacy of mupirocin prophylaxis. Clin J Am
Soc Nephrol. 2010;5(6):1047-1053.
33. Aitken E, McLellan A, Glen J, Serpell M, Mactier R,
Clancy M. Pain resulting from arteriovenous fistulae: prevalence
and impact. Clin Nephrol. 2013;80(5):328-333.
34. Chan MR, Shobande O, Vats H, et al. The effect of
buttonhole cannulation vs. rope-ladder technique on hemodialysis
access patency. Semin Dial. 2014;27(2):210-216.
35. Kim MK, Kim HS. Clinical effects of buttonhole cannulation method on hemodialysis patients. Hemodial Int. 2013;17(2):
294-299.
935
Wong et al
36. Muir CA, Kotwal SS, Hawley CM, et al. Buttonhole cannulation and clinical outcomes in a home hemodialysis cohort and
systematic review. Clin J Am Soc Nephrol. 2014;9(1):110-119.
37. Smyth W, Hartig V, Manickam V. Outcomes of buttonhole
and rope-ladder cannulation techniques in a tropical renal service.
J Ren Care. 2013;39(3):157-165.
38. MacRae JM, Ahmed SB, Hemmelgarn BR; Alberta Kidney
Disease Network. Arteriovenous fistula survival and needling
technique: long-term results from a randomized buttonhole trial.
Am J Kidney Dis. 2014;63(4):636-642.
39. Ball LK, Treat L, Riffle V, Scherting D, Swift L. A multicenter perspective of the buttonhole technique in the Pacific
Northwest. Nephrol Nurs J. 2007;34(2):234-241.
40. Castro MC, Silva Cde F, Souza JM, et al. Arteriovenous
fistula cannulation by buttonhole technique using dull needle.
J Bras Nefrol. 2010;32(3):281-285.
41. Doss S, Schiller B, Moran J. Buttonhole cannulation—an
unexpected outcome. Nephrol Nurs J. 2008;35(4):417-419.
42. Hartig V, Smyth W. “Everyone should buttonhole”: a novel
technique for a regional Australian renal service. J Ren Care.
2009;35(3):114-119.
43. Hashmi A, Cheema MQ, Moss AH. Hemodialysis patients’
experience with and attitudes toward the buttonhole technique
for arteriovenous fistula cannulation. Clin Nephrol. 2010;74(5):
346-350.
44. Marticorena RM, Hunter J, Cook R, et al. A simple method
to create buttonhole cannulation tracks in a busy hemodialysis
unit. Hemodial Int. 2009;13(3):316-321.
936
45. Marticorena RM, Hunter J, Macleod S, et al. Use of the
BioHole device for the creation of tunnel tracks for buttonhole
cannulation of fistula for hemodialysis. Hemodial Int. 2011;15(2):
243-249.
46. Pipkin M, Craft V, Spencer M, Lockridge RS Jr. Six years
of experience with nightly home hemodialysis access. Hemodial
Int. 2004;8(4):349-353.
47. Silva Gdos S, Silva RA, Nicolino AM, et al. Initial experience with the buttonhole technique in a Brazilian hemodialysis
center. J Bras Nefrol. 2010;32(3):257-262.
48. Ward J, Shaw K, Davenport A. Patients’ perspectives of
constant-site (buttonhole) cannulation for haemodialysis access.
Nephron Clin Pract. 2010;116(2):c123-c127.
49. Sukthinthai N, Sittipraneet A, Tummanittayangkoon B,
Vasuvattakul S, Chanchairujira T. Buttonhole technique better
than area puncture technique on hemostasis and pain associated
with needle cannulation. J Med Assoc Thai. 2012;95(suppl 2):
S208-S212.
50. Hill AA, Vasudevan T, Young NP, et al. Use of an
implantable needle guide to access difficult or impossible to cannulate arteriovenous fistulae using the buttonhole technique.
J Vasc Access. 2013;14(2):164-169.
51. Kandil H, Collier S, Yewetu E, Cross J, Davenport A.
Arteriovenous fistula survival with buttonhole (constant site)
cannulation for hemodialysis access. ASAIO J. 2014;60(1):95-98.
52. Pauly RP, Boivin JF, Chan CT. Methodological considerations for research in nocturnal home hemodialysis. J Nephrol.
2007;20(4):381-387.
Am J Kidney Dis. 2014;64(6):918-936