Abstract
BACKGROUND
Central arteriovenous fistula (cAVF) has been investigated as a therapeutic measure for treatment-resistant hypertension in patients without advanced chronic kidney disease (CKD). There is considerable experience with the use of AVF for hemodialysis in patients with end-stage renal disease (ESRD). However, there is sparse data on the blood pressure (BP) effects of an AVF among patients with ESRD. We hypothesized that AVF creation would significantly reduce BP compared with patients who did not have an AVF among patients with ESRD before starting hemodialysis.
METHODS
BPs were compared during the 12 months before hemodialysis initiation in 399 patients with an AVF or AV graft created and 4,696 patients without either.
RESULTS
After propensity score matching 1:2 ratio (AVF to no AVF), repeated measures analysis of variance revealed significant reductions of –1.7 mm Hg systolic and –3.9 mm Hg diastolic BP 12 months in patients after AVF creation; P = 0.025 and P < 0.001, respectively, compared with those with no AVF.
CONCLUSIONS
These findings suggest that AVF creation results in modest BP reduction in patients with pre-dialysis ESRD who require AVF for eventual hemodialysis therapy. Preferential diastolic BP reduction suggests that greater work is needed to characterize the ideal patient subset in which to use cAVF for treatment-resistant hypertension in those without advanced CKD.
Keywords: arteriovenous fistula, blood pressure, central arteriovenous fistula, chronic kidney disease, hypertension, therapeutics, treatment-resistant hypertension
Treatment-resistant hypertension (TRH) is currently defined as blood pressure (BP) that remains above goal on at least 3 classes of antihypertensive drugs, with one of those being a diuretic.1–3 TRH is common among US adults, approximately 20% of the population based on the most recent hypertension (HTN) guidelines.1 When high-risk conditions are considered, such as chronic kidney disease (CKD), the presence of TRH is nearly 30%.4 In addition, nearly one-quarter of patients on ≥4 medications remain above BP goal. The limitations of medical control in TRH have prompted the search for novel and nonpharmacological means to control TRH.
The hemodynamic effect of an arteriovenous fistula (AVF)—decreasing systemic vascular resistance (SVR)—has recently been explored as a treatment for TRH in individuals without end-stage renal disease (ESRD). The physiology of an AVF in relation to BP control has been elegantly reviewed by Bertog et al.5 Briefly, after the creation of an AVF there is an immediate drop in SVR, which leads to a compensatory increase in heart rate (HR) to maintain cardiac output and mean arterial pressure.5 However, over time, if the flow through the AVF is high enough and SVR remains sufficiently suppressed, the net effect is a reduction in BP. The therapeutic device is called the Rox Coupler (a metal shunt that is inserted between the femoral vein and artery),6 resulting in significant decreases in office systolic (S) BP (mean decrease –23.2 mm Hg) and ambulatory SBP (–13 mm Hg) and office diastolic (D) BP (mean decrease –17.7 mm Hg and ambulatory DBP (–13.4 mm Hg) in patients with TRH.
Despite long-term experience with the use of AVF in patients with ESRD, who usually have suboptimal HTN control,3 data on the BP-lowering effects of AVF are scarce.7–9 This report addresses the question of whether the creation of an AVF or arteriovenous graft (AVG) in the year before hemodialysis initiation in patients with ESRD results in meaningful reductions in SBP and DBP compared with those with ESRD who start hemodialysis with no AVF/AVG in place. It thereby evaluates physiological plausibility for the use of an AVF for better BP control in ESRD patients with TRH, and potentially in TRH without advanced CKD.
METHODS
ESRD cohort was assembled using the Veterans Affairs Informatics and Computing Infrastructure, which allowed access to de-identified data on all veterans within the US Veterans Health Administration administrative database, the Corporate Data Warehouse. In addition, information related to ESRD was obtained from the Veteran’s Affairs subset of the United States Renal Data Service (USRDS), which is managed through the Veteran’s Information and Resource Center. The Columbia (SC) Veteran Affairs Healthcare Center Institutional Review Board and Research and Development committee approved this project under expedited review.
Cohort creation
All veterans who underwent AVF/AVG creation (current procedural terminology [CPT] codes 36818, 36819, 36820, 36821, 36825, 36830) between 2010 and 2015 were eligible for inclusion. A total of 8,170 unique individuals were identified (Figure 1). Figure 2 demonstrates the timing of the 2 subsets in relation to dialysis initiation. For patients with an AVF/AVG (cohort label), only those who had a date of surgery 12 months (±2 months) before hemodialysis initiation and documentation of a functioning AVF/AVG at dialysis initiation, or maturing AVF/AVG despite catheter use at dialysis initiation, were included. Second, a comparison cohort was created of individuals starting hemodialysis with a catheter and no functional AVF/AVG. Individuals with dialysis initiation before 2010 were excluded as there were no BP values available. In addition, given that the primary outcomes of interest were the change in SBP and DBP over 12 months, patients with missing SBP or DBP values over this period were excluded, leaving 399 patients with an AVFAVG and 4,696 patients with a catheter (Figure 1) for analysis. All BP values were obtained from outpatient clinics (emergency department and inpatient values were excluded). All BP values were averaged within the time frames evaluated. The BP measurements were not standardized; BP was assessed per local VA clinic routine.
Figure 1.
Flow diagram for patient selection. Abbreviations: AVF, arteriovenous fistula; AVG, arteriovenous graft; BP, blood pressure; CKD, chronic kidney disease; USRDS, United States Renal Data Service.
Figure 2.
Timing of blood pressure measurements in the AVFAVG and catheter cohorts. Abbreviations: AVF, arteriovenous fistula; AVG, arteriovenous graft.
Variables
Variables evaluated demographic characteristics, dialysis start date, pre-dialysis care, vital signs, laboratory data, medication use history, and comorbidity information. Medications were assessed at 12 ± 3 months before dialysis initiation (baseline) and again 6 ± 3 months before dialysis initiation.
Statistical analysis
All statistical analyses were performed using R statistical software (version 3.5.1 The R Foundation for Statistical Computing). General descriptive statistics were performed to compare baseline characteristics of the study population. For the purposes of analysis, baseline is defined as the time of AVF/AVG creation, which was 12 ± 2 months before dialysis initiation, in the AVFAVG cohort, and at this same timeframe before dialysis initiation in the catheter cohort. Thus, BP and HR assessments were made at 12 months (baseline), 6 months, and within 1 month before dialysis initiation (immediately before dialysis initiation).
Propensity score matching was performed to balance baseline characteristics. Propensity matching in a 1 AVFAVG:2 Catheter manner was performed using the nonrandom package in R. The process of care outcome for the propensity score was having an AVF or AVG created. The predictor variables chosen for the propensity score matching were the following: age, institutionalized status, current illegal drug use, non-VA insurance, currently employed, currently receiving nephrology and/or dietician care, duration of nephrology care, and race. The comparison between the study cohorts within the matched and unmatched analyses is displayed as standardized mean difference (SMD); an SMD < 0.1 was considered nonsignificant.10 The primary outcomes were the changes in SBP and DBP from baseline to the just before dialysis initiation for all participants. BP parameters were compared at baseline (12 ± 2 months before dialysis initiation), 6 months later, and just before dialysis initiation (12 months after baseline). For univariate comparisons of baseline demographic and clinical parameters, parametric tests were used to compare continuous, normally distributed variables; nonparametric tests were used to compare continuous, non-normally distributed variables. Categorical variables were compared using χ2 test, or Fisher’s exact test for variables with fewer than 5 events per category. Repeated measures analysis of variance (ANOVA), using the nlme and car packages in R, was used to compare changes in SBP, DBP, and HR over time between the 2 comparator cohorts (catheter and AVFAVG). Sensitivity analysis was performed to see if immature fistulas were not able to provide enough change in SVR to cause a change in BP. All analyses were repeated after excluding individuals who started dialysis with a catheter but a maturing fistula. Furthermore, to assess for differential effects on subjects with and without TRH, separate analyses were performed in a cohort with TRH and without TRH at baseline. TRH was defined according to the 2008 American Heart Association statement on resistant HTN: ≥140 SBP or ≥90 DBP and on 3 antihypertensive agents of different classes.11 Two-tailed P values < 0.05 were set as the limit for statistical significance in univariate and multivariate analysis.
RESULTS
Sociodemographics/comorbidities
After propensity score matching, 399 AVFAVG patients remained and 798 matches for these patients were identified from the catheter population. Table 1 displays the baseline characteristics of the propensity score-matched population. Baseline characteristics were well balanced between the 2 groups, after matching. Table 2 displays the renal and dialysis-related parameters in the 2 groups. In the AVFAVG group, most patients (~75%) had AVF at the start of hemodialysis.
Table 1.
Baseline sociodemographic and clinical variables
| Unmatched | Matched | |||||
|---|---|---|---|---|---|---|
| Catheter | AVFAVG | SMD | Catheter | AVFAVG | SMD | |
| N | 4,696 | 399 | 798 | 399 | ||
| Sociodemographic characteristics | ||||||
| Gender (Male) | 4575 (97.4) | 390 (97.7) | 0.021 | 782 (98.0) | 390 (97.7) | 0.017 |
| Age (years) | 69.78 (13.24) | 66.90 (9.67) | 0.249 | 67.21 (13.35) | 66.90 (9.67) | 0.027 |
| Ethnicity (Hispanic) | 283 (6.0) | 27 (6.8) | 0.037 | 59 (7.4) | 27 (6.8) | 0.024 |
| Race | 0.16 | 0.082 | ||||
| Asian | 71 (1.5) | 9 (2.3) | 24 (3.0) | 9 (2.3) | ||
| Black | 1,442 (30.7) | 148 (37.1) | 299 (37.5) | 148 (37.1) | ||
| Native American | 27 (0.6) | 2 (0.5) | 1 (0.1) | 2 (0.5) | ||
| White | 3,149 (67.1) | 240 (60.2) | 474 (59.4) | 240 (60.2) | ||
| Comorbidity | ||||||
| Hypertension | 4,145 (88.3) | 396 (99.2) | 0.466 | 794 (99.5) | 396 (99.2) | 0.032 |
| Type 2 diabetes mellitus | 2,967 (63.2) | 291 (72.9) | 0.21 | 586 (73.4) | 291 (72.9) | 0.011 |
| Congestive heart failure | 1,868 (39.8) | 145 (36.3) | 0.071 | 277 (34.7) | 145 (36.3) | 0.034 |
| Coronary artery disease | 1,179 (25.1) | 158 (39.6) | 0.314 | 305 (38.2) | 158 (39.6) | 0.028 |
| Cerebrovascular accident | 526 (11.2) | 56 (14.0) | 0.085 | 99 (12.4) | 56 (14.0) | 0.048 |
| Cancer | 508 (10.8) | 45 (11.3) | 0.015 | 90 (11.3) | 45 (11.3) | <0.001 |
| Peripheral vascular disease | 740 (15.8) | 45 (11.3) | 0.131 | 89 (11.2) | 45 (11.3) | 0.004 |
| Chronic obstructive pulmonary disease | 771 (16.4) | 32 (8.0) | 0.259 | 68 (8.5) | 32 (8.0) | 0.018 |
| Illicit drug use | 92 (2.0) | 9 (2.3) | 0.021 | 17 (2.1) | 9 (2.3) | 0.009 |
| Cigarette use | 400 (8.5) | 40 (10.0) | 0.052 | 84 (10.5) | 40 (10.0) | 0.017 |
| Weight at baseline (kg) | 96.09 (23.57) | 98.67 (24.62) | 0.107 | 97.17 (22.81) | 98.67 (24.62) | 0.063 |
| Weight at 12 months (kg) | 89.21 (26.09) | 91.92 (23.89) | 0.108 | 90.76 (22.41) | 91.92 (23.89) | 0.05 |
Abbreviations: AVF, arteriovenous fistula; AVG, arteriovenous graft; SMD, standardized mean difference (< 0.1 considered non-significant difference); VA, Veterans Affairs.
Table 2.
Dialysis-related characteristics before and after propensity matching
| Unmatched | Matched | |||||
|---|---|---|---|---|---|---|
| Catheter | AVFAVG | SMD | Catheter | AVFAVG | SMD | |
| N | 4,696 | 399 | 798 | 399 | ||
| Pre-dialysis characteristics | ||||||
| Nephrology care before dialysis initiation | 0.755 | 0.068 | ||||
| Yes | 2,755 (58.7) | 354 (88.7) | 708 (88.7) | 354 (88.7) | ||
| No | 1,195 (25.4) | 16 (4.0) | 41 (5.1) | 16 (4.0) | ||
| Unknown | 746 (15.9) | 29 (7.3) | 49 (6.1) | 29 (7.3) | ||
| Nephrology care duration before dialysis initiation | 0.83 | 0.041 | ||||
| None | 1,941 (41.4) | 45 (11.3) | 90 (11.3) | 45 (11.3) | ||
| < 6 months | 642 (13.7) | 38 (9.5) | 77 (9.6) | 38 (9.5) | ||
| > 12 months | 1379 (29.4) | 245 (61.4) | 477 (59.8) | 245 (61.4) | ||
| 6–12 months | 732 (15.6) | 71 (17.8) | 154 (19.3) | 71 (17.8) | ||
| Dialysis characteristics | ||||||
| Dialysis type at initiation (hemodialysis) | 4682 (99.7) | 398 (99.7) | 0.021 | 794 (99.5) | 398 (99.7) | 0.041 |
| Primary cause of ESRD | 0.333 | 0.092 | ||||
| Cystic Kidney | 41 (0.9) | 17 (4.3) | 23 (2.9) | 17 (4.3) | ||
| Diabetes | 2,359 (50.2) | 206 (51.6) | 410 (51.4) | 206 (51.6) | ||
| Glomerulonephritis | 211 (4.5) | 28 (7.0) | 64 (8.0) | 28 (7.0) | ||
| Hypertension | 1381 (29.4) | 115 (28.8) | 235 (29.4) | 115 (28.8) | ||
| Unknown Cause | 159 (3.4) | 7 (1.8) | 17 (2.1) | 7 (1.8) | ||
| Type of access at dialysis start | 2.892 | 2.892 | ||||
| AVF | 0 (0.0) | 299 (74.9) | 0 (0.0) | 299 (74.9) | ||
| Catheter | 4696 (100.0) | 77 (19.3) | 798 (100.0) | 77 (19.3) | ||
| Graft | 0 (0.0) | 23 (5.8) | 0 (0.0) | 23 (5.8) | ||
| Arteriovenous fistula maturing if catheter* | 5.148 | 5.148 | ||||
| Yes | 0 (0.0) | 73 (18.3) | 0 (0.0) | 73 (18.3) | ||
| Arteriovenous graft maturing if catheter* | 3.093 | 3.093 | ||||
| Yes | 0 (0.0) | 8 (2.0) | 0 (0.0) | 8 (2.0) | ||
| Days before dialysis initiation AVFAVG created | NA | 427.47 (60.58) | NA | NA | 427.47 (60.58) | NA |
| eGFR at baseline (ml/min/1.73 m2) | 26.1 (11.46) | 15.6 (4.26) | 1.218 | 23.4 (10.64) | 15.56 (4.26) | 0.974 |
| eGFR at 12 months (ml/min/1.73 m2) | 11.3 (5.08) | 9.6 (3.96) | 0.363 | 10.1 (4.52) | 9.6 (3.96) | 0.109 |
| Change in eGFR over 12 months (ml/min/1.73 m2) | –15.75 (15.61) | –6.93 (4.98) | 1.015 | –14.24 (10.37) | –6.93(4.98) | 0.898 |
*Subsets of patients may have been inadvertently indicated as having more than one type of access on the dialysis initiation form. Abbreviations: AVF: arteriovenous fistula; AVG: arteriovenous graft; eGFR: estimated glomerular filtration rate.
Antihypertensive medications
After propensity score matching, the medication classes that demonstrated persistent differences between groups at baseline included angiotensin-converting enzyme inhibitor (ACEI) and aldosterone blocker use being higher among patients with catheter as compared to AVFAVG (Table 3). At 6 months before dialysis initiation, only ACEI remained higher in the catheter group, but more patients with AVFAVG were on loop diuretics and alpha blockers. The total number of antihypertensive medications was similar at baseline and 6 months before dialysis between the groups.
Table 3.
Antihypertensive medications 12 months and 6 months before dialysis
| Unmatched | Matched | |||||
|---|---|---|---|---|---|---|
| Catheter | AVFAVG | SMD | Catheter | AVFAVG | SMD | |
| N | 4,696 | 399 | 798 | 399 | ||
| Antihypertensive meds 12 months before dialysis | ||||||
| ACEI | 1,804 (38.6) | 130 (32.7) | 0.125 | 315 (39.5) | 130 (32.7) | 0.143 |
| ARB | 999 (21.4) | 85 (21.4) | 0.001 | 193 (24.2) | 85 (21.4) | 0.068 |
| BB | 3,216 (68.9) | 283 (71.1) | 0.049 | 549 (68.9) | 283 (71.1) | 0.049 |
| Non-dihydropyridine CCB | 323 (6.9) | 28 (7.0) | 0.005 | 64 (8.0) | 28 (7.0) | 0.038 |
| Dihydropyridine CCB | 2,748 (58.8) | 256 (64.3) | 0.113 | 488 (61.2) | 256 (64.3) | 0.064 |
| Alpha blocker | 1,252 (26.8) | 118 (29.6) | 0.063 | 234 (29.4) | 118 (29.6) | 0.006 |
| Loop diuretic | 2,983 (63.9) | 275 (69.1) | 0.111 | 530 (66.5) | 275 (69.1) | 0.056 |
| Thiazide type diuretic | 827 (17.7) | 53 (13.3) | 0.121 | 113 (14.2) | 53 (13.3) | 0.025 |
| Aldosterone blocker | 380 (8.1) | 19 (4.8) | 0.137 | 60 (7.5) | 19 (4.8) | 0.115 |
| Total number of medications | 3.09 (1.4) | 3.13 (1.4) | 0.022 | 3.19 (1.4) | 3.13 (1.4) | 0.047 |
| Antihypertensive meds 6 months before dialysis | ||||||
| ACEI | 1,306 (27.8) | 82 (20.6) | 0.169 | 224 (28.1) | 82 (20.6) | 0.175 |
| ARB | 809 (17.2) | 62 (15.6) | 0.045 | 148 (18.5) | 62 (15.6) | 0.079 |
| BB | 3,338 (71.1) | 276 (69.3) | 0.038 | 571 (71.6) | 276 (69.3) | 0.048 |
| Non-dihydropyridine CCB | 331 (7.0) | 25 (6.3) | 0.031 | 66 (8.3) | 25 (6.3) | 0.077 |
| Dihydropyridine CCB | 2,844 (60.6) | 265 (66.6) | 0.125 | 523 (65.5) | 265 (66.6) | 0.022 |
| Alpha blocker | 1,201 (25.6) | 125 (31.4) | 0.129 | 200 (25.1) | 125 (31.4) | 0.141 |
| Loop diuretic | 3,364 (71.6) | 321 (80.7) | 0.213 | 604 (75.7) | 321 (80.7) | 0.12 |
| Thiazide type diuretic | 838 (17.8) | 67 (16.8) | 0.027 | 135 (16.9) | 67 (16.8) | 0.002 |
| Aldosterone blocker | 318 (6.8) | 17 (4.3) | 0.11 | 39 (4.9) | 17 (4.3) | 0.029 |
| Total number of medications | 3.06 (1.4) | 3.11 (1.3) | 0.039 | 3.15 (1.3) | 3.11 (1.3) | 0.029 |
Abbreviations: ACEI: angiotensin-converting enzyme inhibitor; ARB: angiotensin receptor blockers; BB: beta blockers; CCB: calcium channel blockers.
BP, HR, and weight
SBP.
At baseline, the SBP was similar between groups (Table 4). At 6 months and just before dialysis, SBP was lower in the AVFAVG group compared with catheter as evidenced by SMD > 0.1 at each of those time points- 6 months: 142.6 vs. 144.6 mm Hg just before dialysis: 141.3 vs. 144.6 mm Hg. Repeated measures ANOVA revealed a significant main effect for the access type (P = 0.025), but not for time (P = 0.54). However, there was a significant interaction between access type and time (P = 0.049) suggesting that the 2 groups had different trajectories of SBP values over the 12-month observation period as demonstrated in Figure 3a.
Table 4.
Changes in blood pressure and heart rate over the observation period
| Catheter | AVFAVG | SMD | |
|---|---|---|---|
| N | 798 | 399 | |
| Systolic blood pressure (mm Hg) | |||
| Baseline | 143.5 (17.8) | 143.0 (14.0) | 0.032 |
| 6 months | 144.6 (17.2) | 142.6 (16.5) | 0.118 |
| 12 months | 144.6 (19.3) | 141.3 (16.5) | 0.184 |
| Change over 6 months | 1.1 (17.5) | –0.4 (17.2) | 0.084 |
| Change over 12 months | 1.1 (21.2) | –1.7 (17.0) | 0.146 |
| Diastolic blood pressure (mm Hg) | |||
| Baseline | 76.4 (11.7) | 76.3 (10.4) | 0.008 |
| 6 months | 75.9 (11.2) | 72.3 (11.1) | 0.327 |
| 12 months | 75.8 (11.7) | 72.4 (11.2) | 0.297 |
| Change over 6 months | –0.4 (9.5) | –3.7 (8.7) | 0.365 |
| Change over 12 months | –0.6 (11.1) | –3.9 (9.0) | 0.328 |
| Heart rate (beats per minute) | |||
| Baseline | 73.8 (12.0) | 71.0 (9.9) | 0.248 |
| 6 months | 72.2 (11.9) | 70.1 (10.6) | 0.187 |
| 12 months | 75.3 (11.4) | 72.1 (10.8) | 0.29 |
| Change over 6 months | –1.0 (9.7) | 0.1 (10.4) | 0.114 |
| Change over 12 months | 1.5 (11.8) | 0.9 (9.9) | 0.052 |
Abbreviations: AVF: arteriovenous fistula; AVG: arteriovenous graft.
Figures 3. (a–c).
Systolic blood pressure, diastolic blood pressure, and heart rate changes over the 12-month observation period.
DBP.
DBP was also similar between groups at baseline. DBP was lower in the AVFAVG group than the catheter group at both 6 months (72.3 vs. 75.9 mm Hg, respectively) and just before dialysis initiation (72.4 vs. 75.9 mm Hg, respectively; SMD > 0.1 at both time points). The change in DBP from baseline was greater at 6 months (–3.7 vs. –0.4 mm Hg, respectively) and just before dialysis initiation (–3.9 vs. –0.6 mm Hg, respectively) in the AVFAVG group compared with the catheter group (SMD > 0.2 at each comparison). Repeated measures ANOVA revealed a significant main effect for both access type (P < 0.001) and time (P < 0.001), and a significant interaction between access type and time (P < 0.001). Pairwise t-test with Bonferroni correction for multiple comparisons revealed significant differences from baseline at both 6 months (P = 0.002) and just before dialysis initiation (P < 0.001).
HR/weight.
HR was different between groups at each time point with the AVFAVG group having a slightly lower value at each time point: baseline: 71 vs. 73.8 beats per minute (bpm), respectively; 6 months: 70.1 vs. 72.2 bpm, respectively; just before dialysis initiation: 72.1 vs. 75.3 bpm, respectively (P = 0.023 for access type main effect in repeated measures ANOVA; and P = 0.003 for main effect of time). However, there was no interaction between access type and time as the change from baseline to 6 months and then to 12 months was in a similar direction for both groups (Figure 3c). Weight was similar between cohorts at baseline and just before dialysis initiation; both cohorts demonstrated a mean reduction of 7 kg from baseline at 12 months.
Sensitivity analysis after removing patients with a maturing fistula but started HD with a catheter
A total of 77 patients (19%) were removed for starting HD with a catheter and a maturing fistula or graft. Patients who started with a catheter and maturing fistula/graft, than those who started with a mature AVF/AVG, were significantly older (69 vs. 65 years, P < 0.05), more likely to have CHF as a comorbidity and more likely to have HTN vs. diabetes as the cause of ESRD than those starting HD with a AVF/AVG. There was no difference between the 2 groups in the time of AVF/AVG creation before HD initiation (415 vs. 430 d, P = 0.058). Patients with maturing AVF at dialysis initiation had a significantly greater decline in SBP 12 months following the creation of AVF than those with a mature AVF/AVG at dialysis initiation (–5.2 vs. –0.7 mm Hg, respectively, P = 0.02). Change in DBP was similar between the 2 groups.
Effect of AVF in patients without and with TRH at baseline
Of patients without TRH at baseline, 188 were in the AVFAVG cohort and 376 in the catheter cohort. Of patients with TRH at baseline, 211 were in the AVFAVG cohort and 422 in the catheter cohort. At baseline, patients with TRH were more likely to be Black (42 vs. 33%, P = 0.008), more likely to have a history of diabetes (79 vs. 68%, P < 0.001), heavier (100 vs. 95 kg, P < 0.001), have a lower baseline eGFR (19.7 vs. 21.4 ml/min/1.73 m2, P = 0.005), and were on more antihypertensives (3.7 vs. 2.7 meds, P < 0.001). Supplementary Tables 1 and 2 display the SBP, DBP, and HR values in the 2 groups.
Supplementary Figures 1 and 2 display the DBP and SBP, respectively, changes over time in those without TRH. ANOVA revealed significant effects of access type, time and interaction between access type and time for both DBP and SBP in those without TRH at baseline.
Supplementary Figures 3 and 4 display DBP and SBP, respectively, changes over time in those with TRH at baseline. In contrast to those without TRH at baseline, the effect of an AVF/AVG was only significant for DBP changes over time, not for SBP. For SBP in patients with TRH at baseline, SBP decrease over time was similar in the AVFAVG and catheter cohorts. SBP was lower at both time points after baseline in AVFAVG compared with catheter, but the difference was not significant (P = 0.05).
DISCUSSION
We have reported on the longest analysis (12 months) of BP changes following AVF/AVG (heretofore only AVF mentioned as the effect is the same) creation in advanced CKD patients transitioning to ESRD, compared with a cohort who transitioned to ESRD without an AVF. The current analysis suggests that AVF creation leads to modest but significant reductions in SBP and DBP over 12 months.
The potentially beneficial effects of AVF creation in patients with ESRD have been assessed in only a few studies. The benefits of AVF creation on BP derive primarily from a reduction in SVR.5 This would result in a reduction in mean arterial pressure, thus affecting both SBP and DBP. Korsheed et al.7 analyzed the hemodynamic effects of patients before and after AVF creation and before initiation of hemodialysis. They found a significant decrease in both SBP (~13 mm Hg) and DBP (~9 mm Hg) 3 months following AVF creation. This study only included only a small number of patients, especially at the 3-month time point (21 compared with the 30 patients at the AVF creation); in addition, there was no comparator cohort at 3 months. More recently, Reddy et al.9 examined a cohort of 137 patients from the Mayo clinic who underwent AVF creation and had echocardiograms performed shortly before and then > 6 months after the surgery. The focus of this analysis was on changes in right ventricle (RV) size and function. However, they did demonstrate a significant 16 and 10 mm Hg drop in SBP and DBP, respectively, at a median of 3 years following AVF creation. This study did not include a comparator cohort and the follow-up BP readings were performed after dialysis initiation. A recent meta-analysis by Scholz et al.12 has nicely summarized the available literature on the BP effects of AVF/AVG creation or ligation. They identified 14 studies with a total of 412 patients. The main differences identified from our cohort are the short duration of follow-up (maximum 6 months—and in this study (Dundon et al.13) there was no change in SBP after AVF/AVG creation—146 ± 19 mm Hg before AVF/AVG creation and 146 ± 17 mm Hg after AVF/AVG creation13), and 4 of these studies were after hemodialysis initiation.
The findings from the current analysis are concordant with these prior analyses although smaller in magnitude. The availability of longer follow-up time before dialysis initiation allows separation of the effects of dialysis (minimization of the effects of total body sodium and water content to blood pressure), from the effect of the AVF itself on BP control; as well as an assessment of the durability of the BP-lowering effect of AVF. In the current cohort, SBP was reduced by ~2 mm Hg and DBP ~4 mm Hg over 12 months, with an increase of 1 mm Hg and decrease of 0.6 mm Hg, respectively, in the catheter cohort over 12 months.
An important difference between these prior studies and the current analysis, that may account for the difference in magnitude of BP reduction, is the method of BP assessment.7–9 Korsheed et al. and Ori et al. all conducted formal, research-oriented BP measurements in subjects. The current analysis uses clinic values obtained during clinical visits. Higher values for SBP and DBPs have been routinely demonstrated for in-clinic values as opposed to ambulatory or research-oriented BPs, in the non-dialysis CKD and non-CKD cohorts.14–16
In the current analysis, there was a greater effect of AVF/AVG on DBP compared with SBP. A difference in response in DBP compared with SBP to treatment has been described in well-established cohorts of BP treatment. Analyses of both the Framingham and Antihypertensive and Lipid-Lowering Treatment to prevent Heart Attacks Trial (ALLHAT) cohorts revealed that >90% of subjects in both cohorts were able to achieve DBP targets whereas 49 and 67%, respectively, achieved SBP targets.11 The strongest predictors of lack of SBP control included older age (> 75 years; in both cohorts) and the presence of CKD (identified in the ALLHAT cohort). On the contrary, younger patients (<60 years) were more likely to achieve control for both SBP and DBP.17 If subsets of patients, especially the elderly, manifest greater decrements in DBP compared with SBP, then the potentially detrimental effects of widened pulse pressure and excessively low DBP need to be considered.18 This is particularly important in that the effect of the central AVF (cAVF) would be persistent and not easily modulated (i.e., the cAVF cannot be “stopped” as easily as a medication). The sensitivity analyses removing those with a maturing AVF/AVG and separating those with TRH from those without TRH, further support the notion that the long-term effect of AVF/AVG in an older cohort with high comorbidity seems to be on persistent DBP lowering. The effect on SBP seems to be much more variable and less significant.
This analysis did find a slower rate of renal function decline in the AVF/AVG group compared with the catheter group. This is consistent with prior analyses that have demonstrated a slowing of renal function decline following creation of an AVF/AVG.19,20 The current analysis was not designed to test the association between BP changes and renal function change.
Additional lessons can be gleaned from the long history with AVF use for ESRD. The increase in returning blood volume to the RV is acknowledged as an important change in systemic hemodynamics by Bertog et al.,5 and the long-term adverse effects on RV in ESRD have been demonstrated by Reddy et al.9 In addition, the changes in the RV found by Reddy et al. did not improve following ligation of the AVF, suggesting long-term changes may persist once established. Elevations in pulmonary artery pressures, steal syndrome, and acute decompensated heart failure have also been documented following AVF creation in hemodialysis patients.21–23 Long-term follow-up data will be important in the original cohort from the open-labeled randomized trial (original follow-up 6 months) of the Rox-Coupler,6 the first of which demonstrated 33% venous stenosis over the initial 12 months following device insertion.24
Limitations
Differences in medication classes were noted at baseline and 6 months, despite similar number of medications between cohorts after matching. No measure of cardiac hemodynamic parameters, such as stroke volume and cardiac output, was performed or was available for the current analysis. Access blood flow rate was not available for analysis. A recent meta-analysis has demonstrated that access blood flow rate is associated with BP reductions in patients with AVF/AVG.12 It is very likely that the average blood flow rate in the AVF/AVG group was at least 600 ml/min, which is the minimal blood flow rate considered optimal for use during dialysis.25 In addition, in-office, non-standardized BP measures were used to assess changes. Despite these limitations, the current analysis represents a realistic estimation of longer-term changes of SBP and DBP. Another limitation is that patients started dialysis within 1 year of AVF creation. These patients may have been sicker and less able to demonstrate potential beneficial effects of AVF on BP than those who remained dialysis-free for longer periods. Finally, this analysis was performed in a US veteran population, which is primarily male and Caucasian. This will limit the generalizability of the findings.
Conclusions
This report demonstrates modest but significant reductions in BP in the 12 months following AVF creation in a large cohort of pre-dialysis ESRD patients. This BP reduction was most notable for DBP.
PERSPECTIVES
The findings from this analysis will be informative for investigations looking into the best clinical application of cAVF for TRH. The current analysis supports the use of cAVF to lower BP in non-advanced CKD subjects with TRH. However, certain patient groups may manifest disproportionate reductions in DBP, as evident in the current cohort of older individuals with advanced CKD. Further analysis should be taken to identify ideal populations in which to use cAVF to lower BP, to ensure proportionate lowering of SBP and DBP.
Supplementary Material
ACKNOWLEDGMENTS
The authors note no conflicts of interest related to this analysis. The authors acknowledge the Veterans Affairs for the data used for this analysis. The Veterans Integrated Networking and Computing Infrastructure and the VA Information Resource Center provided the data used. All analyses and interpretations presented are the responsibility of the authors and do not necessarily represent the views of the Department of Veterans Affairs, National Heart, Lung, and Blood Institute, National Institutes of Health, or the U.S. Department of Health and Human Services.
DISCLOSURES
None.
REFERENCES
- 1. Carey RM, Sakhuja S, Calhoun DA, Whelton PK, Muntner P. Prevalence of apparent treatment-resistant hypertension in the United States. Hypertension 2019; 73:424–431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Carey RM, Calhoun DA, Bakris GL, Brook RD, Daugherty SL, Dennison-Himmelfarb CR, Egan BM, Flack JM, Gidding SS, Judd E, Lackland DT, Laffer CL, Newton-Cheh C, Smith SM, Taler SJ, Textor SC, Turan TN, White WB; American Heart Association Professional/Public Education and Publications Committee of the Council on Hypertension; Council on Cardiovascular and Stroke Nursing; Council on Clinical Cardiology; Council on Genomic and Precision Medicine; Council on Peripheral Vascular Disease; Council on Quality of Care and Outcomes Research; and Stroke Council Resistant hypertension: detection, evaluation, and management: a scientific statement from the American Heart Association. Hypertension 2018; 72:e53–e90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Braam B, Taler SJ, Rahman M, Fillaus JA, Greco BA, Forman JP, Reisin E, Cohen DL, Saklayen MG, Hedayati SS. Recognition and management of resistant hypertension. Clin J Am Soc Nephrol 2017; 12:524–535. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Thomas G, Xie D, Chen HY, Anderson AH, Appel LJ, Bodana S, Brecklin CS, Drawz P, Flack JM, Miller ER III, Steigerwalt SP, Townsend RR, Weir MR, Wright JT Jr, Rahman M; CRIC Study Investigators Prevalence and prognostic significance of apparent treatment resistant hypertension in chronic kidney disease: report from the chronic renal insufficiency cohort study. Hypertension 2016; 67:387–396. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Bertog SC, Sobotka NA, Sobotka PA, Lobo MD, Sievert K, Vaskelyte L, Sievert H, Schmieder RE. Percutaneous creation of a Central Iliac arteriovenous anastomosis for the treatment of arterial hypertension. Curr Hypertens Rep 2018; 20:18. [DOI] [PubMed] [Google Scholar]
- 6. Lobo MD, Sobotka PA, Stanton A, Cockcroft JR, Sulke N, Dolan E, van der Giet M, Hoyer J, Furniss SS, Foran JP, Witkowski A, Januszewicz A, Schoors D, Tsioufis K, Rensing BJ, Scott B, Ng GA, Ott C, Schmieder RE; ROX CONTROL HTN Investigators Central arteriovenous anastomosis for the treatment of patients with uncontrolled hypertension (the ROX CONTROL HTN study): a randomised controlled trial. Lancet 2015; 385:1634–1641. [DOI] [PubMed] [Google Scholar]
- 7. Korsheed S, Eldehni MT, John SG, Fluck RJ, McIntyre CW. Effects of arteriovenous fistula formation on arterial stiffness and cardiovascular performance and function. Nephrol Dial Transplant 2011; 26:3296–3302. [DOI] [PubMed] [Google Scholar]
- 8. Ori Y, Korzets A, Katz M, Perek Y, Zahavi I, Gafter U. Haemodialysis arteriovenous access–a prospective haemodynamic evaluation. Nephrol Dial Transplant 1996; 11:94–97. [PubMed] [Google Scholar]
- 9. Reddy YNV, Obokata M, Dean PG, Melenovsky V, Nath KA, Borlaug BA. Long-term cardiovascular changes following creation of arteriovenous fistula in patients with end stage renal disease. Eur Heart J 2017; 38:1913–1923. [DOI] [PubMed] [Google Scholar]
- 10. Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med 2009; 28:3083–3107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Calhoun DA, Jones D, Textor S, Goff DC, Murphy TP, Toto RD, White A, Cushman WC, White W, Sica D, Ferdinand K, Giles TD, Falkner B, Carey RM; American Heart Association Professional Education Committee Resistant hypertension: diagnosis, evaluation, and treatment: a scientific statement from the American Heart Association Professional Education Committee of the Council for High Blood Pressure Research. Circulation 2008; 117:e510–e526. [DOI] [PubMed] [Google Scholar]
- 12. Scholz SS, Vukadinović D, Lauder L, Ewen S, Ukena C, Townsend RR, Wagenpfeil S, Böhm M, Mahfoud F. Effects of arteriovenous fistula on blood pressure in patients with end-stage renal disease: a systematic meta-analysis. J Am Heart Assoc 2019; 8:e011183. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Dundon BK, Torpey K, Nelson AJ, Wong DT, Duncan RF, Meredith IT, Faull RJ, Worthley SG, Worthley MI. The deleterious effects of arteriovenous fistula-creation on the cardiovascular system: a longitudinal magnetic resonance imaging study. Int J Nephrol Renovasc Dis 2014; 7:337–345. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Sarafidis PA, Mallamaci F, Loutradis C, Ekart R, Torino C, Karpetas A, Raptis V, Bikos A, Papagianni A, Balafa O, Siamopoulos K, Pisani G, Morosetti M, Del Giudice A, Aucella F, Di Lullo L, Tripepi R, Tripepi G, Jager K, Dekker F, London G, Zoccali C. Prevalence and control of hypertension by 48-h ambulatory blood pressure monitoring in haemodialysis patients: a study by the European Cardiovascular and Renal Medicine (EURECA-m) working group of the ERA-EDTA. Nephrol Dial Transplant 2018. doi:10.1093/ndt/gfy147. [DOI] [PubMed] [Google Scholar]
- 15. Banegas JR, Ruilope LM, de la Sierra A, Vinyoles E, Gorostidi M, de la Cruz JJ, Ruiz-Hurtado G, Segura J, Rodríguez-Artalejo F, Williams B. Relationship between clinic and ambulatory blood-pressure measurements and mortality. N Engl J Med 2018; 378:1509–1520. [DOI] [PubMed] [Google Scholar]
- 16. Minutolo R, Gabbai FB, Agarwal R, Chiodini P, Borrelli S, Bellizzi V, Nappi F, Stanzione G, Conte G, De Nicola L. Assessment of achieved clinic and ambulatory blood pressure recordings and outcomes during treatment in hypertensive patients with CKD: a multicenter prospective cohort study. Am J Kidney Dis 2014; 64:744–752. [DOI] [PubMed] [Google Scholar]
- 17. Cushman WC, Ford CE, Cutler JA, Margolis KL, Davis BR, Grimm RH, Black HR, Hamilton BP, Holland J, Nwachuku C, Papademetriou V, Probstfield J, Wright JT Jr, Alderman MH, Weiss RJ, Piller L, Bettencourt J, Walsh SM; ALLHAT Collaborative Research Group Success and predictors of blood pressure control in diverse North American settings: the antihypertensive and lipid-lowering treatment to prevent heart attack trial (ALLHAT). J Clin Hypertens (Greenwich) 2002; 4:393–404. [DOI] [PubMed] [Google Scholar]
- 18. Kovesdy CP, Bleyer AJ, Molnar MZ, Ma JZ, Sim JJ, Cushman WC, Quarles LD, Kalantar-Zadeh K. Blood pressure and mortality in U.S. veterans with chronic kidney disease: a cohort study. Ann Intern Med 2013; 159:233–242. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Golper TA, Hartle PM, Bian A. Arteriovenous fistula creation may slow estimated glomerular filtration rate trajectory. Nephrol Dial Transplant 2015; 30:2014–2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Sumida K, Molnar MZ, Potukuchi PK, Thomas F, Lu JL, Ravel VA, Soohoo M, Rhee CM, Streja E, Yamagata K, Kalantar-Zadeh K, Kovesdy CP. Association between vascular access creation and deceleration of estimated glomerular filtration rate decline in late-stage chronic kidney disease patients transitioning to end-stage renal disease. Nephrol Dial Transplant 2017; 32:1330–1337. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Agarwal AK. Systemic effects of hemodialysis access. Adv Chronic Kidney Dis 2015; 22:459–465. [DOI] [PubMed] [Google Scholar]
- 22. Korn A, Alipour H, Zane J, Gray K, Ryan T, Kaji A, De Virgilio C, Bowens N. Predictors of steal in hemodialysis access. Am Surg 2017; 83:1099–1102. [PubMed] [Google Scholar]
- 23. Raza F, Alkhouli M, Rogers F, Vaidya A, Forfia P. Case series of 5 patients with end-stage renal disease with reversible dyspnea, heart failure, and pulmonary hypertension related to arteriovenous dialysis access. Pulm Circ 2015; 5:398–406. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Lobo MD, Ott C, Sobotka PA, Saxena M, Stanton A, Cockcroft JR, Sulke N, Dolan E, van der Giet M, Hoyer J, Furniss SS, Foran JP, Witkowski A, Januszewicz A, Schoors D, Tsioufis K, Rensing BJ, Scott B, Ng GA, Schmieder RE. Central Iliac arteriovenous anastomosis for uncontrolled hypertension: one-year results from the ROX CONTROL HTN trial. Hypertension 2017; 70:1099–1105. [DOI] [PubMed] [Google Scholar]
- 25. Mudoni A, Caccetta F, Caroppo M, Musio F, Accogli A, Zacheo MD, Burzo MD, Gallieni M, Nuzzo V. Echo color Doppler ultrasound: a valuable diagnostic tool in the assessment of arteriovenous fistula in hemodialysis patients. J Vasc Access 2016; 17:446–452. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.