Visual Abstract
Keywords: ESKD, vascular access
Abstract
Background
Twenty to 60% of newly created hemodialysis arteriovenous fistulas do not mature adequately for use. One barrier to developing interventions to improve fistula outcomes is a lack of standardized criteria for maturation.
Methods
Using data from the multicenter, prospective Hemodialysis Fistula Maturation (HFM) Study, we determined sensitivities, specificities, and positive and negative predictive values of multiple candidate maturation criteria using the HFM Study maturation criteria as the reference. We also compared, across the maturation criteria, relationships between maturation and fistula survival using Cox proportional hazards models.
Results
We included 535 of the 602 HFM Study participants. The median (interquartile range) age was 57 (47–65) years, 70% were men, and 45% were Black participants. Depending on the criterion and time frame for ascertainment (3, 4, 5, 6, or 9 months), sensitivities ranged from 57% to 100%, specificities ranged from 85% to 100%, positive predictive values ranged from 88% to 100%, and negative predictive values ranged from 65% to 100%. For all criteria, areas under the curve for the 6-month (0.90–0.97 for unassisted maturation and 0.89–0.95 for overall maturation) and 9-month time frames were similar. Attainment of unassisted maturation was associated with lower risks of fistula abandonment, with hazard ratios ranging from 0.10 to 0.40 depending on the criterion and time frame. Eliminating dialysis adequacy indicators, or simplifying the criteria in other ways, had little effect on performance characteristics.
Conclusions
High performance characteristics are maintained with maturation criteria that are simpler and less burdensome to ascertain than the HFM Study outcome measure.
Introduction
Patients undergoing maintenance hemodialysis require a well-functioning vascular access to receive their dialysis treatments. The arteriovenous fistula is the preferred form of vascular access for hemodialysis because of lower rates of thrombosis or infection compared with synthetic grafts or central venous catheters.1 However, 20%–60% of newly created fistulas do not mature,2–4 meaning that the fistula does not undergo sufficient vasodilation and wall remodeling to support the hemodialysis blood circuit. While numerous studies have been conducted to elucidate underlying mechanisms, identify risk factors, or test interventions for fistula maturation failure, an important challenge in interpreting and applying the findings of these studies is the lack of consistency in defining fistula maturation. A systematic review found that 43 different criteria for fistula maturation and 15 different time frames for meeting the criteria, ranging from 4 to 100 weeks, have been used in randomized clinical trials.5 Developing and validating standardized criteria for maturation is one of the research priorities included in the 2019 Kidney Disease Outcomes Quality Initiative Clinical Practice Guideline for Vascular Access.1
Professional societies and expert committees have attempted to standardize outcome measures for functional maturation, a term referring to the ability to use the fistula successfully for hemodialysis treatments; however, the criteria proposed by these groups are based on expert opinion rather than data, and there has not been widespread adoption of any particular criterion from among those proposed.6,7 In addition, the predictive utilities of various maturation criteria for long-term fistula function have not been established. The National Institute of Diabetes and Digestive and Kidney Diseases Hemodialysis Fistula Maturation (HFM) observational cohort study used a maturation criterion that required use of the fistula for at least 75% of sessions within a 4-week period with a dialysis machine blood pump speed (BPS) of at least 300 ml/min or a measured single pool Kt/V of at least 1.4.8 While this definition of maturation is stringent and clinically compelling, the complex nature and incorporation of multiple data elements make its routine ascertainment highly burdensome for either research or clinical care.7,9
As an initial step toward developing a fistula maturation outcome measure for clinical trials, we used data from the HFM Study to compare the performance characteristics of multiple candidate maturation criteria and the associations of those criteria with long-term fistula usability. Establishing an outcome measure that is both clinically meaningful and feasible to ascertain should facilitate the comparison of studies, development of clinical practice guidelines, and clinical decision making.
Methods
Overall Approach
We generated seven candidate maturation criteria from existing criteria and by creating a priori criteria. Using data from the HFM Study, and applying the HFM maturation criterion as the reference, we compared performance characteristics of the candidate maturation criteria using sensitivity, specificity, and positive and negative predictive values (NPVs). In addition, we evaluated relationships between fistula maturation defined by each criterion and long-term usability.
Data Source
The HFM Study was a prospective cohort study designed to elucidate predictors and mechanisms of fistula maturation. The study enrolled 602 patients with dialysis dependence or advanced CKD who underwent fistula creation with a single-stage surgical procedure. Participants were enrolled from seven centers in the United States between 2010 and 2013 and were followed for up to 4 years. During the pre-, intra-, and postoperative periods, granular data about clinical, anatomical, and biological attributes were collected as previously described.8 Demographic variables, including race and ethnicity, were ascertained in the HFM Study through participant self-report and are provided in this report to facilitate assessment of the generalizability of the findings. For the current analyses, all HFM participants were included unless they (1) had not yet initiated dialysis by the end of follow-up or (2) could not have fistula maturation ascertained because of death, transplantation, change to peritoneal dialysis, or loss to follow-up. The Institutional Review Board of the University of Pennsylvania approved the study (protocol number 831994).
Reference for Evaluating the Fistula Maturation Criteria
We used the original HFM fistula maturation criterion as the reference to evaluate the performance characteristics of the candidate criteria. In the HFM Study, maturation was defined as the use of the fistula with two needles for ≥75% of dialysis sessions during a 4-week period with either (1) four consecutive dialysis sessions with mean machine BPS ≥300 ml/min or (2) single-pool Kt/V ≥1.4 as a measure of solute clearance. The first of the four qualifying consecutive dialysis sessions, or the qualifying Kt/V measurement, had to occur within 9 months of fistula creation surgery unless the participant had not started dialysis by 9 months after fistula creation, in which case the first session of the 4-week ascertainment period had to take place within 4 weeks of dialysis initiation. Fistula maturation in the HFM Study was assessed using rolling 4-week periods until the maturation outcome was met, the fistula was abandoned, or the study ended. An outcome of fistula abandonment in the HFM Study required a determination by the treating nephrologist or surgeon that the fistula would not be used in the future.
Candidate Criteria for Fistula Maturation
We included three previously proposed maturation criteria and created, a priori, four new criteria (Table 1). The previously proposed criteria included:
Table 1.
Candidate fistula maturation criteria
| Criterion | Definition | Time Frame |
|---|---|---|
| HFM | Two-needle cannulation for ≥75% of dialysis sessions during a rolling 4-wk period with either: 1) Four consecutive dialysis sessions with mean BPS ≥300 ml/min or 2) Single-pool Kt/V ≥1.4 Time frame: within 9 mo of fistula creation |
Must meet components within X mo after fistula creation. If patient initiates dialysis after X mo, the maturation criteria must be satisfied within 1 mo of dialysis therapy initiation. X=3, 4, 5, 6, or 9 mo |
| NAVAC | Two-needle cannulation for two thirds of dialysis sessions within a month and adequate delivery of the prescribed dialysis Adequate delivery of dialysis is defined as: • Average BPS of 300 ml/min if conventional intermittent hemodialysis, 3.5–4.5 h, 3 times per week • Average BPS of 250 ml/min if short-daily or nocturnal dialysis |
|
| KHI | Two-needle cannulation of fistula for 75% of dialysis sessions within a 4-wk period | |
| No use of vascular access other than fistula for ≥1 mo | The use of fistula for dialysis without an alternative vascular access for a minimum of 1 mo | |
| Four consecutive uses of fistula with dialysis BPS ≥300 ml/min | Two-needle cannulation of fistula for four consecutive dialysis sessions with dialysis pump speed of ≥300 ml/min | |
| Four consecutive uses of fistula | Two-needle cannulation of fistula for four consecutive dialysis sessions | |
| ≥1 use of fistula | Two-needle cannulation of fistula for a minimum of one dialysis session |
HFM, Hemodialysis Fistula Maturation; BPS, blood pump speed; NAVAC, North America Vascular Access Consortium; KHI, Kidney Health Initiative.
HFM Criterion
We modified the HFM criterion by incorporating time frames for achieving maturation that were shorter than the 9-month time frame used for the HFM Study, on the basis of the rationale that earlier determination of the maturation outcome has advantages for both clinical care (e.g., by promoting earlier revision or replacement of a fistula with maturation failure) and research (e.g., by reducing the duration of follow-up).5 The modified time frames of 3, 6, and 9 months for meeting maturation were established a priori. After finding that performance characteristics were similar when using the 6- and 9-month time frames, but different when using the 3-month time frame, we also evaluated time frames of 4 and 5 months.
North American Vascular Access Consortium Criterion
The North American Vascular Access Consortium (NAVAC) defines a “fistula used successfully for hemodialysis” as a fistula that “can be used with two-needle cannulation for two-thirds or more of all dialysis runs for 1 month and deliver the prescribed dialysis within the prescribed time frame.”6 For conventional hemodialysis, NAVAC defined adequate dialysis as dialysis with an average BPS of 300 ml/min in a 3.5-hour dialysis session, or 250 ml/min for daily or nocturnal dialysis. To meet the NAVAC criterion, a patient must have their dialysis catheter removed and receive catheter-free dialysis for at least 1 month. Because some patients might have an alternative arteriovenous access rather than a catheter, we modified the catheter removal component of the NAVAC criterion to dialysis with no use of an alternative vascular access for at least 1 month. As we did for the HFM maturation outcome, we evaluated multiple time frames (3, 4, 5, 6, and 9 months) for meeting the NAVAC criterion.
Kidney Health Initiative Criterion
The Kidney Health Initiative (KHI) Clinical Trial Endpoints for Hemodialysis Vascular Access Project defined a clinically functional fistula as a fistula that can be cannulated with two dialysis needles for at least 75% of dialysis sessions within a 4-week period to achieve the prescribed dialysis.7 As we did for the HFM maturation outcome, we evaluated multiple time frames (3, 4, 5, 6, and 9 months) for meeting the KHI criteria.
New Criteria
The new criteria that we developed a priori included one or multiple components reflecting (1) use for dialysis, (2) small solute clearance, and (3) dialysis BPS. We combined these components to generate four candidate maturation criteria as presented in Table 1. Again, we evaluated the time frames of 3, 4, 5, 6, and 9 months for meeting these maturation criteria.
Outcomes
The performance characteristics of the candidate criteria were evaluated for the following HFM Study outcomes: (1) unassisted maturation, defined as fistula maturation without a maturation-promoting procedure, such as angioplasty or surgical revision; (2) overall maturation, defined as fistula maturation with or without a prior maturation-promoting intervention; and (3) overall fistula survival, defined as the time from fistula creation until fistula abandonment.6,10 A fistula was classified as abandoned when it had not been used for a period of at least 2 weeks and the primary nephrologist or surgeon determined that the fistula would not be able to be used for future dialysis, meaning an alternative access was required.8
Statistical Analyses
Descriptive statistics included means and SDs for normally distributed continuous measures, medians and interquartile ranges (IQRs) for skewed continuous measures, and proportions for categorical measures. We estimated the sensitivities, specificities, positive predictive values (PPVs), and NPVs, with 95% confidence intervals (CIs), for the candidate maturation criteria using the HFM criterion as the reference. To evaluate the ability of candidate criteria to discriminate between fistulas that met maturation and those that did not meet maturation, we generated receiver-operating characteristic curves with corresponding areas under the curve (AUCs). To determine whether the maturation criteria provide prognostic information about long-term fistula survival, we fit Cox proportional hazards models for fistula abandonment, in which fistula maturation status was treated as a time-varying covariate such that the value of maturation status changed from “not mature” to “mature” once the fistula satisfied the candidate maturation criterion.11 The follow-up time started at the date of fistula surgery with kidney transplantation, death, and the end of the study treated as censoring events. In a post hoc analysis, we assessed the level of agreement among candidate criteria using the Cohen kappa statistic. All statistical analyses were performed using STATA version 14.
Results
Baseline Characteristics
After excluding the HFM Study participants who could not have ascertainment of fistula maturation, 535 of the 602 HFM Study participants were included in the cohort created for these analyses (Figure 1). The median age was 57 years with an IQR of 47–65 years, 70% were men, 45% were Black participants, 58% had diabetes, and 66% were being treated with hemodialysis at the time of fistula creation (Table 2). The median (IQR) follow-up duration was 26 (17–35) months.
Figure 1.
Composition of the analytic cohort. HFM, Hemodialysis Fistula Maturation.
Table 2.
Baseline characteristics
| Characteristic | n (%) or Median (IQR) |
|---|---|
| Age at fistula creation surgery, yr | 57 (47–65) |
| Sex, n (%) | |
| Male | 372 (70) |
| Female | 163 (30) |
| Race, n (%) | |
| American Indian | 11 (2) |
| Asian | 11 (2) |
| Black | 243 (45) |
| Native Hawaiian or Pacific Islander | 11 (2) |
| White | 244 (46) |
| Multiple races | 7 (1) |
| Unknown | 8 (2) |
| Ethnicity, n (%) | |
| Hispanic | 73 (14) |
| Non-Hispanic | 458 (86) |
| Unknown | 4 (1) |
| Dialysis status at fistula creation surgery, n (%) | |
| No treatment with dialysis | 176 (33) |
| Treatment with hemodialysis | 355 (66) |
| Treatment with peritoneal dialysis | 4 (1) |
| Serum creatinine level among those not receiving dialysis, mg/dl | 4.6 (3.8–5.4) |
| Comorbid conditions, n (%) | |
| Diabetes | 311 (58) |
| Hypertension | 517 (97) |
| Congestive heart failure | 146 (27) |
| History of myocardial infarction | 66 (12) |
| History of angina | 75 (14) |
| History of coronary artery bypass surgery | 56 (10) |
| Prior percutaneous coronary intervention | 50 (9) |
| Prior carotid endarterectomy | 3 (1) |
| Prior carotid artery angioplasty | 3 (1) |
| History of stroke or transient ischemic attack | 70 (13) |
| History of claudication | 34 (6) |
| History of lower extremity angioplasty or bypass surgery | 14 (3) |
| Smoking status, n (%) | |
| Non-smoker | 250 (47) |
| Former smoker | 190 (36) |
| Current smoker | 90 (17) |
| Unknown | 5 (1) |
IQR, interquartile range.
Unassisted Maturation and Overall Maturation
Depending on the maturation criterion and time frame used, the proportion of fistulas that achieved unassisted maturation ranged from 32%–62%, and the proportion that achieved overall maturation ranged from 36%–81% (Supplemental Table 1). Eighty-eight of the 535 fistulas (16%) were never cannulated during follow-up and, accordingly, were classified as not achieving either unassisted maturation or overall maturation.
Performance Characteristics of Candidate Criteria for Unassisted Maturation
The sensitivities and specificities of the candidate maturation criteria, with the original HFM criterion as the reference, are presented in Figure 2 and Table 3. Criteria with fewer components, such as single use of the fistula (an a priori criterion) or use of the fistula for 75% of sessions within a 4-week period (Kidney Health Initiative criterion), had modestly higher sensitivities but similar specificities compared with those that also incorporate dialysis BPS or Kt/V. Extending the time frame for ascertaining maturation from 3 to 6 months increased the range of sensitivities from 57%–78% to 86%–98%, and extending the maturation time frame from 6 to 9 months increased the range of sensitivities from 86%–98% to 90%–100% (Figure 2 and Table 3). The specificities of the candidate criteria ranged from 88%–100%, 86%–100%, and 85%–98%, using the 3-, 6-, and 9-month time frames, respectively. The ranges for the AUC for the candidate criteria using the 3-, 6-, and 9-month time frames were 0.78–0.85, 0.90–0.97, and 0.92–0.96, respectively (Figure 3, Table 3). The 95% CIs for the AUCs overlapped, suggesting no statistical difference between the AUCs for the candidate criteria within the 3-, 6-, and 9-month time frames. Using the 3-, 6-, and 9-month time frames, the PPVs ranged from 88%–100%, 90%–100%, and 90%–98%, respectively, and the NPVs ranged from 65%–77%, 84%–97%, and 89%–100%, respectively (Table 4).
Figure 2.
Sensitivities and specificities for the candidate maturation criteria for unassisted fistula maturation. The time frames for meeting the maturation criteria are within 3 months (A), within 6 months (B), and within 9 months (C). BPS, blood pump speed; KHI, Kidney Health Initiative; NAVAC, North American Vascular Access Consortium.
Table 3.
Sensitivities and specificities for candidate criteria for unassisted fistula maturation
| Time Frame | 3 mo | 4 mo | 5 mo | 6 mo | 9 mo | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Performance Characteristic | Sens. | Spec. | AUC (95% CI) | Sens. | Spec. | AUC (95% CI) | Sens. | Spec. | AUC (95% CI) | Sens. | Spec. | AUC (95% CI) | Sens. | Spec. | AUC (95% CI) |
| % (95% CI) TP/(TP+FN) |
% (95% CI) TN/(TN+FP) |
% (95% CI) TP/(TP+FN) |
% (95% CI) TN/(TN+FP) |
% (95% CI) TP/(TP+FN) |
% (95% CI) TN/(TN+FP) |
% (95% CI) TP/(TP+FN) |
% (95% CI) TN/(TN+FP) |
% (95% CI) TP/(TP+FN) |
% (95% CI) TN/(TN+FP) |
||||||
| HFM | 61 (55 to 66) 179/296 |
100 (99 to 100) 239/239 |
0.80 (0.77 to 0.83) | 79 (74 to 84) 234/296 |
100 (99 to 100) 239/239 |
0.90 (0.87 to 0.92) | 90 (86 to 93) 234/296 |
100 (98 to 100) 239/239 |
0.95 (0.93 to 0.97) | 94 (90 to 96) 234/296 |
100 (99 to 100) 239/239 |
0.97 (0.95 to 0.98) | Ref | Ref | Ref |
| NAVAC | 63 (57 to 68) 186/296 |
96 (93 to 98) 230/239 |
0.80 (0.77 to 0.83) | 76 (70 to 81) 224/296 |
95 (92 to 98) 228/239 |
0.86 (0.83 to 0.88) | 84 (80 to 88) 249/296 |
95 (91 to 97) 227/239 |
0.90 (0.87 to 0.92) | 86 (81 to 90) 254/296 |
95 (91 to 97) 227/239 |
0.90 (0.88 to 0.93) | 90 (86 to 93) 267/296 |
94 (90 to 97) 225/239 |
0.92 (0.90 to 0.94) |
| KHI | 75 (69 to 80) 221/296 |
88 (83 to 92) 210/239 |
0.81 (0.78 to 0.8) | 86 (81 to 88) 253/296 |
86 (81 to 90) 206/239 |
0.86 (0.83 to 0.89) | 94 (91 to 97) 279/296 |
86 (81 to 90) 206/239 |
0.90 (0.88 to 0.93) | 97 (94 to 99) 287/296 |
86 (81 to 90) 206/239 |
0.92 (0.89 to 0.94) | 100 (99 to 100) 296/296 |
85 (80 to 89) 203/239 |
0.92 (0.90 to 0.95) |
| No use of alternative vascular access for ≥1 mo | 62 (56 to 68) 184/296 |
96 (92 to 98) 229/239 |
0.79 (0.76 to 0.82) | 78 (73 to 83) 232/296 |
95 (92 to 98) 228/239 |
0.87 (0.84 to 0.90) | 89 (85 to 92) 263/296 |
95 (92 to 98) 228/239 |
0.92 (0.90 to 0.94) | 93 (89 to 95) 274/296 |
95 (92 to 98) 228/239 |
0.94 (0.92 to 0.96) | 97 (94 to 99) 287/296 |
95 (91 to 97) 226/239 |
0.96 (0.94 to 0.98) |
| Four consecutive uses of fistula with BPS ≥300 ml/min | 57 (51 to 63) 168/296 |
98 (96 to 100) 235/239 |
0.78 (0.75 to 0.80) | 74 (69 to 79) 220/296 |
98 (95 to 99) 234/239 |
0.86 (0.83 to 0.89) | 85 (81 to 89) 252/296 |
98 (95 to 99) 234/239 |
0.92 (0.89 to 0.94) | 88 (84 to 92) 261/296 |
98 (95 to 99) 234/239 |
0.93 (0.91 to 0.95) | 93 (90 to 96) 277/296 |
98 (95 to 99) 233/239 |
0.96 (0.94 to 0.97) |
| Four consecutive uses of fistula | 73 (67 to 78) 215/296 |
95 (92 to 98) 228/239 |
0.84 (0.81 to 0.87) | 86 (82 to 90) 255/296 |
94 (90 to 96) 224/239 |
0.90 (0.87 to 0.92) | 93 (89 to 96) 275/296 |
94 (90 to 96) 224/239 |
0.93 (0.91 to 0.95) | 97 (94 to 99) 287/296 |
94 (90 to 96) 224/239 |
0.95 (0.94 to 0.97) | 100 (98 to 100) 295/296 |
92 (87 to 95) 219/239 |
0.96 (0.94 to 0.97) |
| ≥1 use of fistula | 78 (73 to 83) 231/296 |
93 (88 to 96) 221/239 |
0.85 (0.82 to 0.88) | 89 (85 to 92) 263/296 |
90 (85 to 94) 215/239 |
0.89 (0.87 to 0.92) | 95 (91 to 97) 280/296 |
89 (84 to 92) 212/239 |
0.92 (0.89 to 0.94) | 98 (95 to 99) 289/296 |
88 (84 to 92) 211/239 |
0.93 (0.91 to 0.95) | 100 (99 to 10) 296/296 |
86 (81 to 90) 205/239 |
0.93 (0.91 to 0.95) |
Sens., sensitivity; Spec., specificity; CI, confidence interval; TP, true positive; FN, false negative; TN, true negative; FN, false positive; AUC, area under the curve; HFM, Hemodialysis Fistula Maturation; Ref, reference; NAVAC, North America Vascular Access Consortium; KHI, Kidney Health Initiative; BPS, blood pump speed.
Figure 3.
AUCs for candidate criteria for unassisted maturation. The time frames for meeting maturation criteria are within 3 months (A), within 6 months (B), and within 9 months (C). AUC, area under the curve; CI, confidence interval. Figure 3 can be viewed in color online at www.cjasn.org.
Table 4.
Positive and negative predictive values for candidate criteria for unassisted fistula maturation
| Time Frame | 3 mo | 4 mo | 5 mo | 6 mo | 9 mo | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| Performance Characteristic | PPV | NPV | PPV | NPV | PPV | NPV | PPV | NPV | PPV | NPV |
| % (95% CI) TP/(TP+FP) |
% (95% CI) TN/(TN+FN) |
% (95% CI) TP/(TP+FP) |
% (95% CI) TN/(TN+FN) |
% (95% CI) TP/(TP+FP) |
% (95% CI) TN/(TN+FN) |
% (95% CI) TP/(TP+FP) |
% (95% CI) TN/(TN+FN) |
% (95% CI) TP/(TP+FP) |
% (95% CI) TN/(TN+FN) |
|
| HFM | 100 (98 to 100) 179/179 |
67 (62 to 72) 239/356 |
100 (98 to 100) 234/234 |
79 (74 to 84) 239/301 |
100 (99 to 100) 265/265 |
89 (84 to 92) 239/270 |
100 (99 to 100) 277/277 |
93 (89 to 96) 239/258 |
Ref | Ref |
| NAVAC | 95 (91 to 98) 186/195 |
68 (62 to 73) 230/340 |
95 (92 to 98) 224/235 |
76 (71 to 81) 228/300 |
95 (92 to 98) 249/261 |
83 (78 to 87) 227/274 |
96 (92 to 98) 254/266 |
84 (80 to 89) 227/269 |
95 (92 to 97) 267/281 |
89 (84 to 92) 225/254 |
| KHI | 88 (84 to 92) 221/250 |
74 (68 to 79) 210/285 |
88 (84 to 92) 253/286 |
83 (78 to 87) 206/249 |
89 (86 to 93) 279/312 |
92 (88 to 96) 206/223 |
90 (86 to 93) 287/320 |
96 (92 to 98) 206/215 |
89 (85 to 92) 296/332 |
100 (98 to 100) 203/203 |
| No use of alternative vascular access for ≥1 mo | 95 (91 to 98) 184/194 |
67 (62 to 72) 229/341 |
96 (92 to 98) 232/243 |
78 (73 to 83) 228/292 |
96 (93 to 98) 263/274 |
87 (83 to 91) 228/261 |
96 (93 to 98) 274/285 |
91 (87 to 94) 228/250 |
96 (93 to 98) 287/300 |
96 (93 to 98) 226/235 |
| Four consecutive uses of fistula with BPS ≥300 ml/min | 98 (94 to 99) 168/172 |
65 (60 to 70) 235/363 |
98 (95 to 99) 220/225 |
76 (70 to 80) 234/310 |
98 (96 to 99) 252/257 |
84 (79 to 88) 234/278 |
98 (96 to 99) 261/266 |
87 (82 to 91) 234/269 |
98 (95 to 99) 277/283 |
93 (89 to 95) 233/252 |
| Four consecutive uses of fistula | 95 (92 to 98) 215/226 |
74 (69 to 79) 228/309 |
94 (91 to 97) 255/270 |
85 (80 to 89) 224/265 |
95 (92 to 97) 275/290 |
91 (87 to 95) 224/245 |
95 (92 to 97) 287/302 |
96 (93 to 98) 224/233 |
94 (90 to 96) 295/315 |
100 (98 to 100) 219/220 |
| ≥1 use of fistula | 93 (89 to 96) 231/249 |
77 (72 to 82) 221/286 |
92 (88 to 95) 263/287 |
87 (82 to 91) 215/248 |
91 (88 to 94) 280/307 |
93 (89 to 96) 212/228 |
91 (88 to 94) 289/317 |
97 (94 to 99) 211/218 |
90 (86 to 93) 296/330 |
100 (98 to 100) 205/205 |
PPV, positive predictive value; NPV, negative predictive value; CI, confidence interval; TP, true positive; FP, false positive; TN, true negative; FN, false negative; HFM, Hemodialysis Fistula Maturation; Ref, reference; NAVAC, North America Vascular Access Consortium; KHI, Kidney Health Initiative; BPS, blood pump speed.
Performance Characteristics of Candidate Criteria for Overall Maturation
For overall maturation (unassisted or assisted), criteria with fewer components had modestly higher sensitivities and similar specificities compared with criteria with many components. Extending the time frame for ascertaining maturation from 3 to 6 months increased the range of sensitivities from 49%–69% to 87%–95%, and extending the time frame from 6 to 9 months increased the range of sensitivities from 87%–95% to 94%–100%. The specificities of the candidate criteria ranged from 90%–100%, 83%–100%, and 77%–100%, using the 3-, 6-, and 9-month time frames, respectively (Supplemental Table 2). The ranges for the AUCs for the candidate criteria using the 3-, 6-, and 9-month time frames were 0.74–0.80, 0.89–0.95, and 0.88–0.98, respectively (Supplemental Table 2). The 95% CIs for the AUCs overlapped, suggesting no statistical difference between the AUCs for the candidate criteria within the 3-, 6-, and 9-month time frames. The PPV of the candidate criteria ranged from 96%–100%, 94%–100%, and 93%–100%, for the 3-, 6-, and 9-month time frames, respectively, and the NPV ranged from 40%–50%, 94%–100%, and 93–100%, using the 3-, 6-, and 9-month time frames, respectively (Supplemental Table 3).
Subgroup Analysis
In a post hoc analysis, we evaluated subgroups on the basis of age (65 years and older versus younger than 65 years). As shown in Supplemental Tables 4 and 5, the performance characteristics for the maturation criteria were similar for the subgroups.
Agreement among Criteria That Do Not Require BPS or Kt/V
Because the performance characteristics did not look markedly better for the criteria that incorporate BPS and Kt/V compared with the four criteria that do not include those components, we assessed, in a post hoc analysis, the degree of agreement for unassisted maturation among the four simpler criteria. As shown in Supplemental Table 6, the κ coefficients for all pairwise comparisons of the four criteria range from 0.81 to 0.94 for unassisted maturation and 0.80 to 0.96 for overall maturation, indicating a high level of agreement among these four criteria (Supplemental Table 6).
Relationships between Candidate Criteria for Unassisted Maturation and Fistula Survival
Within this analytic cohort, the incidence rate for fistula abandonment was 0.13 per patient-year. Figure 4 shows the hazard ratios (HRs) for fistula abandonment for fistulas that met unassisted maturation relative to those that did not meet maturation on the basis of each candidate criterion and applying the 6-month time frame (the time frame with the highest AUC for several of the criteria). Attainment of unassisted maturation was associated with a lower risk of fistula abandonment, with HRs ranging from 0.14 to 0.25, depending on the maturation criterion. Among all candidate criteria, when applying the 6-month time frame, fistulas that met the criterion of “no use of alternative vascular access for ≥1 month” had the lowest risk of fistula abandonment (HR, 0.14; 95% CI, 0.10 to 0.21). When using candidate criteria with the 3- and 9-month time frames, the attainment of unassisted maturation was associated with lower risk of fistula abandonment with HRs ranging from 0.24 to 0.39 (Supplemental Figure 1) and 0.10 to 0.23 (Supplemental Figure 2), respectively, and, for both time frames, the criterion with the lowest HR was “no use of alternative vascular access for ≥1 month.”
Figure 4.
Hazard ratios for fistula abandonment for fistulas with versus without unassisted maturation on the basis of each maturation criterion applying the 6-month time frame for attaining maturation.
Relationships between Candidate Criteria for Overall Maturation and Fistula Survival
Attainment of overall maturation was associated with lower risk of fistula abandonment with HRs ranging from 0.24 to 0.34, 0.12 to 0.17, and 0.09 to 0.16, with the 3-, 6-, and 9-month time frames, respectively (Supplemental Figures 3–5).
Discussion
Using the HFM Study fistula maturation criterion as the reference, we evaluated the performance characteristics of alternative maturation criteria that incorporate several indicators of fistula usability and multiple time frames. Across the candidate criteria, we found a broad range of sensitivities (57%–100%) and NPVs (65%–100%) and a relatively narrow range of specificities (85%–100%) and PPVs (88%–100%). Overall, our analyses suggest that (1) there is not a clear advantage of extending the maturation ascertainment period beyond 6 months; (2) incorporating dialysis machine BPS or solute clearance measures does not substantially improve specificity; and (3) both the original HFM criterion and simpler criteria provide prognostic information about fistula long-term outcomes.
For all of the candidate criteria, the sensitivity for fistula maturation increased substantially when the ascertainment time frame was extended from 3 to 6 months, but only marginally, or not at all, when the time frame was extended from 6 to 9 months. This finding, together with specificities for the candidate criteria that were stable across the full range of time frames, suggests that 6 months is generally sufficient to determine the maturation outcome for a fistula and that extending the time for fistula maturation from 6 to 9 months does not provide additional benefit. Whether it is appropriate to shorten the time frame for determining the maturation outcome from 6 to 5 or 4 months depends on the relative importance of avoiding misclassifying some fistulas as having maturation failure versus (1) the clinical advantages of earlier implementation of maturation-enhancing interventions or fistula replacement or (2) the research advantages of increased efficiency and decreased cost resulting from earlier ascertainment of a study outcome.
Our finding that including dialysis machine pump speed and Kt/V in the maturation outcome criteria increased the specificity only modestly suggests that these indicators of dialysis adequacy are not critical components of maturation outcome measures. The minimal effect of these factors on test characteristics may reflect the multiple determinants other than fistula performance, such as dialysis session duration and body size, that influence prescribed BPS and achieved solute clearance.12 Given the variability in practice patterns for the dialysis prescription and solute clearance targets, an advantage of criteria that do not incorporate indicators or determinants of solute clearance is the greater relevance to the full population of patients receiving hemodialysis. Thus, on the basis of broader applicability, lower burden for ascertainment, and minimal effect on performance characteristics, it seems reasonable to use maturation criteria that do not include indicators or determinants of solute clearance. The results from our assessment of agreement among the four criteria that do not include these components suggest that the classification of fistula maturation with any of these criteria, at least when using the 6-month time frame, would be nearly identical.
The HFM Study maturation criteria were designed to classify as mature those fistulas that support repeated dialysis over a short period (4 weeks) because the goal was to elucidate predictors and underlying mechanisms for the process of fistula maturation rather than fistula durability. However, for maximal clinical utility, criteria for maturation should also identify fistulas with long-term usability. Our comparison, across the candidate criteria, of the hazard of fistula abandonment associated with meeting versus not meeting the maturation outcome indicate that criteria with high sensitivity and specificity for maturation are also associated with marked reductions in the risk of fistula abandonment. While we expected to find a relationship between maturation attainment and fistula durability, we performed this analysis to provide information about the relative utilities of the maturation criteria not just for the HFM Study fistula maturation outcome measure but for the clinically important outcome of fistula survival. Using the 6-month time frame for maturation ascertainment, we found that the criterion associated with the lowest hazard of fistula abandonment was “no use of an alternative vascular access for at least 1 month.” As with the sensitivities and specificities for maturation, for any of the time frames evaluated, candidate criteria that incorporate dialysis machine BPS or KT/V did not have greater associations with fistula durability than criteria that do not incorporate solute clearance determinants or measures (Figure 4 and Supplemental Figures 1 and 2).
This study has several limitations. We used the HFM outcome measure as the reference, but acknowledge that there is no gold standard for defining a mature fistula. The HFM Study did not include fistulas created using a two-stage approach or an endovascular approach; thus, the results may not be applicable to fistulas that were created using these methods.13–17 In addition, the findings may not be generalizable to other populations or to data obtained retrospectively. The study also has important strengths. The data used for the analyses were collected prospectively in a standardized manner from multiple centers. Because of the extensive and highly granular nature of the data collected for the HFM Study, we were able to evaluate a range of maturation criteria, including several that were previously proposed by expert working groups. The continued follow-up of HFM participants after maturation of the study fistula allowed us to evaluate relationships between maturation criteria and long-term fistula outcomes. Importantly, we evaluated maturation criteria that are based on fistula usability, a parameter that is clinically meaningful and aligns with the priorities of clinicians, patients, and other stakeholders.18
In conclusion, using the HFM Study maturation criterion as the reference, we found that there is not a substantial loss of sensitivity or specificity for several fistula maturation criteria that are simpler and easier to ascertain and that there is not a clear benefit of extending the ascertainment period beyond 6 months. The findings from these analyses can inform clinical decisions, selection of outcome measures and data collection requirements for future research studies, and efforts to standardize criteria for fistula maturation.
Supplementary Material
Acknowledgments
The data for these analyses were provided by the Data Coordinating Center for the Hemodialysis Fistula Maturation Study at the Cleveland Clinic, and the work was performed in accordance with a data use agreement between the Cleveland Clinic and the University of Pennsylvania.
Footnotes
The list of nonauthor contributors is extensive and has been provided in the Supplemental Material.
See related editorial, “Keep Seeking the Holy Grail: Predictive Modeling of Arteriovenous Fistula Maturation and Survival,” on pages 1257–1259.
Contributor Information
Collaborators: H. Feldman, L. Dember, A. Farber, J. Kaufman, L. Stern, P. LeSage, C. Kivork, D. Soares, M. Malikova, M. Allon, C. Young, M. Taylor, L. Woodard, K. Mangadi, P. Roy-Chaudhury, R. Munda, T. Lee, R. Alloway, M. El-Khatib, T. Canaan, A. Pflum, L. Thieken, B. Campos-Naciff, T. Huber, S. Berceli, M. Jansen, G. McCaslin, Y. Trahan, M. Vazquez, W. Vongpatanasin, I. Davidson, C. Hwang, T. Lightfoot, C. Livingston, A. Valencia, B. Dolmatch, A. Fenves, N. Hawkins, A. Cheung, L. Kraiss, D. Kinikini, G. Treiman, D. Ihnat, M. Sarfati, I. Lavasani, M. Maloney, L. Schlotfeldt, J. Himmelfarb, C. Buchanan, C. Clark, C. Crawford, J. Hamlett, J. Kundzins, L. Manahan, J. Wise, G. Beck, J. Gassman, T. Greene, P. Imrey, L. Li, J. Alster, M. Li, J. MacKrell, M. Radeva, B. Weiss, K. Wiggins, C. Alpers, K. Hudkins, T. Wietecha, M. Robbin, H. Umphrey, L. Alexander, C. Abts, L. Belt, J. Vita, N. Hamburg, M. Duess, A. Levit, H. Higgins, S. Ke, O. Mandaci, C. Snell, J. Gravley, S. Behnken, R. Mortensen, G. Chertow, A. Besarab, K. Brayman, M. Diener-West, D. Harrison, L. Inker, T. Louis, W. McClellan, J. Rubin, J. Kusek, and R. Star
Disclosures
During the past 3 years, L.M. Dember has received compensation from the National Kidney Foundation for her role as Deputy Editor of the American Journal of Kidney Diseases; compensation for serving on Data and Safety Monitoring Boards for the National Institute of Diabetes and Digestive and Kidney Diseases, CSL Behring, and Vertex Pharmaceuticals; consultancy for AstraZeneca, Cara Therapeutics, CSL Behring, Merck, and Vertex; and research funding from AstraZeneca, Boehringer Ingelheim, Genentech, GlaxoSmithKline, Kyowa Kirin, Novo Nordisk, and Regeneron. L.M. Dember's spouse reports consultancy for Actelion, Alexion, Bristol Myers Squibb, ChemCentryx, Genentech/Roche, Genzyme/Sanofi, and GlaxoSmithKline; stock options from Kyverna; research funding from Actelion, Bristol Myers Squibb, Celgene, ChemoCentryx, Genentech/Roche, and GlaxoSmithKline; and compensation as author and section editor for UpToDate. J.H. Ng reports employment with Northwell Health; consultancy for Vifor Pharmaceuticals; ownership interest in PublishedMD Consulting LLC; research funding from NIH/NIDDK (award number: 1K23DK132459-01); roles as Editorial Board Member of ASN Kidney News (unpaid) and Editorial Board Member of Advances in Chronic Kidney Disease (unpaid); and other interests or relationships as founder of PublishedMD Consulting LLC. W. Yang reports currently serving as a statistical editor for the American Journal of Kidney Diseases.
Funding
J.H. Ng: National Institute of Diabetes and Digestive and Kidney Diseases (K23 DK132459-02 and T32DK007006) and Breslin Family Foundation. L.M. Dember: National Institute of Diabetes and Digestive and Kidney Diseases (U01DK082232).
Author Contributions
Conceptualization: Laura M. Dember, Jia Hwei Ng, Wei Yang.
Formal analysis: Jia Hwei Ng.
Investigation: Laura M. Dember, Jia Hwei Ng.
Methodology: Laura M. Dember, Jia Hwei Ng.
Supervision: Laura M. Dember, Wei Yang.
Visualization: Jia Hwei Ng.
Writing – original draft: Laura M. Dember, Jia Hwei Ng, Wei Yang.
Writing – review & editing: Laura M. Dember, Jia Hwei Ng, Wei Yang.
Data Sharing Statement
Data from the Hemodialysis Fistula Maturation Study are available at the NIDDK Central Repository.
Supplemental Material
This article contains the following supplemental material online at http://links.lww.com/CJN/B798.
Members of the HFM Study Group.
Supplemental Table 1. Number (%) with attainment of unassisted maturation and overall maturation on the basis of criterion and time frame.
Supplemental Table 2. Sensitivities and specificities of candidate criteria for overall maturation.
Supplemental Table 3. Positive and negative predictive values for candidate criteria for overall maturation.
Supplemental Table 4. Area under the curve (AUC) for candidate criteria for unassisted maturation for age subgroups.
Supplemental Table 5. Area under the curve (AUC) for candidate criteria for overall maturation for age subgroups.
Supplemental Table 6. Degree of agreement for unassisted and overall maturation among the candidate maturation criteria that do not include blood pump speed or adequacy indicators (6-month time frame).
Supplemental Figure 1. Hazard ratios for fistula abandonment for fistulas with versus without maturation (unassisted) on the basis of each candidate maturation criterion applying the 3-month time frame for attaining maturation.
Supplemental Figure 2. Hazard ratios for fistula abandonment for fistulas with versus without maturation (unassisted) on the basis of each candidate maturation criterion using the 9-month time frame for attaining maturation.
Supplemental Figure 3. Hazard ratios for fistula abandonment for fistulas with versus without maturation (overall) on the basis of each candidate maturation criterion using the 3-month time frame for attaining maturation.
Supplemental Figure 4. Hazard ratios for fistula abandonment for fistulas with versus without maturation (overall) on the basis of each candidate maturation criterion using the 6-month time frame for attaining maturation.
Supplemental Figure 5. Hazard ratios for fistula abandonment for fistulas with versus without maturation (overall) on the basis of each candidate maturation criterion using the 9-month time frame for attaining maturation.
References
- 1.Lok CE Huber TS Lee T, et al. KDOQI clinical practice guideline for vascular access: 2019 update. Am J Kidney Dis. 2020;75(4):S1–S164. doi: 10.1053/j.ajkd.2019.12.001 [DOI] [PubMed] [Google Scholar]
- 2.Al-Jaishi AA Oliver MJ Thomas SM, et al. Patency rates of the arteriovenous fistula for hemodialysis: a systematic review and meta-analysis. Am J Kidney Dis. 2014;63(3):464–478. doi: 10.1053/j.ajkd.2013.08.023 [DOI] [PubMed] [Google Scholar]
- 3.Dember LM Beck GJ Allon M, et al. Effect of clopidogrel on early failure of arteriovenous fistulas for hemodialysis: a randomized controlled trial. JAMA. 2008;299(18):2164–2171. doi: 10.1001/jama.299.18.2164 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.United States Renal Data System. 2022 USRDS Annual Data Report: Epidemiology of Kidney Disease in the United States. National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 2022. [Google Scholar]
- 5.Viecelli AK O'Lone E Sautenet B, et al. Vascular access outcomes reported in maintenance hemodialysis trials: a systematic review. Am J Kidney Dis. 2018;71(3):382–391. doi: 10.1053/j.ajkd.2017.09.018 [DOI] [PubMed] [Google Scholar]
- 6.Lee T, Mokrzycki M, Moist L, Maya I, Vazquez M, Lok CE. Standardized definitions for hemodialysis vascular access. Semin Dial. 2011;24(5):515–524. doi: 10.1111/j.1525-139x.2011.00969.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Beathard GA Lok CE Glickman MH, et al. Definitions and end points for interventional studies for arteriovenous dialysis access. Clin J Am Soc Nephrol. 2018;13(3):501–512. doi: 10.2215/CJN.11531116 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Dember LM Imrey PB Beck GJ, et al. Objectives and design of the hemodialysis fistula maturation study. Am J Kidney Dis. 2014;63(1):104–112. doi: 10.1053/j.ajkd.2013.06.024 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Viecelli AK Teixeira-Pinto A Valks A, et al. Study protocol for Vascular Access outcome measure for function: a vaLidation study in hemoDialysis (VALID): a multi-center, multinational validation study to assess the accuracy and feasibility of measuring vascular access function in clinical practice. BMC Nephrol. 2022;23(1):372. doi: 10.1186/s12882-022-02987-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.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. doi: 10.1067/mva.2002.122025 [DOI] [PubMed] [Google Scholar]
- 11.Kleinbaum DG, Klein M. Extension of the Cox proportional hazards model for time-dependent variables. In: Survival Analysis: A Self-Learning Text, 3rd ed. Springer New York; 2012:241–288. [Google Scholar]
- 12.Lambie SH, Taal MW, Fluck RJ, McIntyre CW. Analysis of factors associated with variability in haemodialysis adequacy. Nephrol Dial Transplant. 2004;19(2):406–412. doi: 10.1093/ndt/gfg570 [DOI] [PubMed] [Google Scholar]
- 13.Bashar K Healy DA Elsheikh S, et al. One-stage vs. two-stage brachio-basilic arteriovenous fistula for dialysis access: a systematic review and a meta-analysis. PLoS One. 2015;10(3):e0120154. doi: 10.1371/journal.pone.0120154 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Jones RG, Morgan RA. A review of the current status of percutaneous endovascular arteriovenous fistula creation for haemodialysis access. Cardiovasc Interv Radiol. 2019;42(1):1–9. doi: 10.1007/s00270-018-2037-6 [DOI] [PubMed] [Google Scholar]
- 15.Yan Wee IJ, Yap HY, Tang TY, Chong TT. A systematic review, meta-analysis, and meta-regression of the efficacy and safety of endovascular arteriovenous fistula creation. J Vasc Surg. 2020;71(1):309–317.e5. doi: 10.1016/j.jvs.2019.07.057 [DOI] [PubMed] [Google Scholar]
- 16.Lok CE Rajan DK Clement J, et al. Endovascular proximal forearm arteriovenous fistula for hemodialysis access: results of the prospective, multicenter novel endovascular access trial (NEAT). Am J Kidney Dis. 2017;70(4):486–497. doi: 10.1053/j.ajkd.2017.03.026 [DOI] [PubMed] [Google Scholar]
- 17.Radosa CG Radosa JC Weiss N, et al. Endovascular creation of an arteriovenous fistula (endoAVF) for hemodialysis access: first results. Cardiovasc Interv Radiol. 2017;40(10):1545–1551. doi: 10.1007/s00270-017-1750-x [DOI] [PubMed] [Google Scholar]
- 18.Viecelli AK Tong A O'Lone E, et al. Report of the standardized outcomes in nephrology-hemodialysis (SONG-HD) consensus workshop on establishing a core outcome measure for hemodialysis vascular access. Am J Kidney Dis. 2018;71(5):690–700. doi: 10.1053/j.ajkd.2017.12.003 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data from the Hemodialysis Fistula Maturation Study are available at the NIDDK Central Repository.