EXPERIMENTAL AND CLINICAL TRANSPLANTATION

Objectives: The effect of a functioning arteriovenous fistula on cardiac function in kidney transplant recipients has not been thoroughly investigated.

Materials and Methods: We retrospectively evaluated cardiac function in 99 renal transplant recipients using transthoracic echocardiography, with available follow-up at baseline and 2 and 5 years posttransplant. Patients were divided into 2 groups: a control group (n = 47) with no functioning arteriovenous fistula immediately after transplant and an arteriovenous fistula group (n = 52) with a functioning arteriovenous fistula for at least 5 years after transplant. Left ventricular ejection fraction, diastolic thickness of the interventricular septum, and left ventricular end-diastolic diameter were assessed.

Results: In our study, patients (62.6% men, 7.1% with diabetes, mean age of 55.6 ± 11.5 years), we observed no significant differences with respect to baseline left ventricular ejection fraction and interventricular septum; however, in the arteriovenous fistula group, baseline left ventricular end-diastolic diameter was marginally higher than that shown in the control group (50.6 ± 5.4 vs 48.6 ± 4.4 mm; P = .054). In multivariate analysis, functioning fistula and peripheral arterial disease were negatively associated with left ventricular ejection fraction at 5 years posttransplant, whereas baseline left ventricular ejection fraction had a minimal positive effect: B (95% confidence interval) of -2.186 (-4.312 to -0.061) (P = .044), -5.304 (-9.686 to -0.922) (P = .018), and 0.247 (0.047 to 0.446) (P = .016), respectively. Functioning fistula also emerged as associated with larger left ventricular end-diastolic diameter at 2 and 5 years posttransplant: B (95% confidence interval) of 3.047 (1.470-4.625) (P < .001) and 2.122 (0.406-3.838) (P = .016), respectively.

Conclusions: Maintenance of a functioning fistula in kidney transplant recipients may be associated with adverse long-term effects on left ventricular ejection fraction and left ventricular end-diastolic diameter.

Key words : Cardiac function, Left ventricular ejection fraction, Left ventricular end-diastolic diameter, Renal transplant

The results of mixed-design ANOVA showed a significant effect of fistula patency status on SBP overall (F[1,97] = 7.4; P = .008), with patients in the AVF group having higher SBP than controls of 134.5 mm Hg (95% CI, 131.3-137.8 mm Hg) and 128.0 mm Hg (95% CI, 124.6-131.5 mm Hg), respectively. There was no significant effect of time on SBP (F[2,194] = 0.37; P = .7) with patients’ SBP being 131.5 mm Hg (95% CI, 128.1-134.9 mm Hg), 132.0 mm Hg (95% CI, 128.6-135.4 mm Hg), and 130.3 mm Hg (95% CI, 127.2-133.4 mm Hg) at baseline and at 2 and 5 years posttransplant, respectively. We observed no significant interaction between time and fistula patency status in terms of SBP (F[2,194] = 0.07; P = .9). Changes in SBP by time and fistula patency status are depicted in Figure 2A.

The results of the mixed-design ANOVA showed no significant effect of fistula patency status on DBP overall (F[1,97] = 0.2; P = .7), with patients in AVF group having similar DBP compared with controls of 81.9 mm Hg (95% CI, 79.9-83.8 mm Hg) and 82.5 mm Hg (95% CI, 80.5-84.5 mm Hg), respectively. There was no significant effect of time on DBP (F[2,194] = 0.33; P = .7) with patients’ DBP being 82.3 mm Hg (95% CI, 80.0-84.6 mm Hg), 82.7 mm Hg (95% CI, 80.6-84.9 mm Hg), and 81.6 mm Hg (95% CI, 79.5-83.7 mm Hg) at baseline and at 2 and 5 years posttransplant, respectively. We observed no signi-ficant interaction between time and fistula patency status in terms of DBP (F[2,194] = 0.02; P = .98). Changes in DBP by time and fistula patency status are depicted in Figure 2B.

Multivariate analyses
With regard to LVEF at 2 years posttransplant (Table 3), our linear regression model was statistically significant (F = 2.69; P = .005) and accounted for approximately 25% of the LVEF variance in the study population (R² =0.254, adjusted R² = 0.160). Only baseline LVEF was significantly associated with LVEF at 2 years.

With regard to LVEF at 5 years posttransplant (Table 3), our linear regression model was statistically significant (F = 2.899; P = .003) and accounted for approximately 27% of the LVEF variance in the study population (R² = 0.268, adjusted R² = 0.176). Fistula patency and PAD were negatively associated with LVEF at 5 years (P = .044 and .018, respectively). Baseline LVEF was also significantly associated with LVEF at 5 years.

Our multivariate model for LVEDD at 2 years posttransplant (F = 8.495; P < .001, R2 = 0.518, adjusted R² = 0.457) showed fistula patency to be independently associated with larger LVEDD, whereas use of ACE inhibitors/ARB therapy emerged as a protective factor (Table 4). At 5 years posttransplant (F = 5.475; P < .001, R² = 0.409, adjusted R² = 0.334), fistula patency remained a negative predictor of LVEDD.

With regard to IVS at 2 years posttransplant (Table 5), our linear regression model was statistically significant (F = 4.124; P < .001, R² = 0.343, adjusted R2 = 0.260). Except for IVS at baseline, no factor was found to be significantly correlated with 2-year IVS results. At 5 years (F = 2.125; P = .026, R2 = 0.212, adjusted R² = 0.112), except for male sex, no other factor was found to be significantly correlated with IVS.

We found that a patent AVF in patients post-transplant had a long-term impact on LVEDD and LVEF, whereas administration of ACE inhibitors/ARBs seemed to have a protective effect on left ventricle size.

Closure of an AVF in patients with end-stage renal disease and conversion to a tunnelled central venous catheter for hemodialysis access have been reported to result in beneficial changes in left ventricle size and function.¹¹ However, there is conflicting evidence regarding the effects of both kidney transplant per se on cardiomyopathy and left ventricular hypertrophy^14,15 and the effects of AVF closure on cardiac size and function, with a few relatively small and without long-term follow-up studies showing an improvement in the above-mentioned parameters.^10,16-21 On the other hand, there is a reasonable debate regarding the potentially beneficial effects of routine ligation of a patent AVF in kidney transplant recipients, as this could deprive patients from valuable access to hemodialysis should a graft failure ensue.²²

In our study, all baseline clinical and laboratory variables examined (apart from C-reactive protein) did not differ between the 2 groups. Baseline C-reactive protein was found to be significantly higher in transplanted patients with thrombosed fistulas than in those with patent fistulas. At 2 years after transplant, LVEDD was significantly higher in the AVF group than in the control group, and this difference was maintained at 5 years posttransplant. We presume that this difference in left ventricle size reflected chronic left ventricle volume loading mediated by the functional AVF, as the latter emerged as a significant predictor of LVEDD in multivariate analysis independent of sex, age (per 10-y increase), presence of diabetes, prior CAD, hypertension, diuretic use, and posttransplant estimated glomerular filtration rate (per 10 mL/min/1.73m² increase). Small echocardiographic studies, without long-term follow-up, have also shown reduced left ventricle size following AVF closure in kidney transplant patients compared with controls,^17-20 a finding that has been recently verified by a longitudinal cardiac magnetic resonance imaging study.¹⁶ Moreover, the slightly, albeit significantly higher SBP in the AVF group at 2 years could be attributed to the higher stroke volume as implied by the larger LVEDD with a similar LVEF in the 2 groups. In line with this, we did not observe any differences in DBP in contrast to previous findings from other groups describing increased DBP after AVF closure.²³

At 5 years posttransplant, a borderline statistically significant reduction in left ventricle systolic function as estimated by LVEF appeared in the AVF group. Independent predictors of LVEF were the presence of a patent AVF and history of PAD. We speculate that long-term volume loading of the left ventricle by the patent AVF could be associated with this mild reduction of left ventricle systolic function in this group of patients and could be indicative of progressive cardiomyopathy and a failing heart with decreasing cardiac output.²⁴ Additionally, the detrimental effects of PAD on left ventricle systolic function could reflect a higher left ventricle afterload, which could also negatively affect the former. Most importantly, administration of ACE inhibitors/ARBs seemed to have a protective role regarding left ventricle dilatation in this study. This observation is in line with both the presumed pathophysiology described above and evidence regarding the beneficial effects of this class of drugs in left ventricle remodeling.

Finally, IVS diastolic thickness did not differ between the 2 groups at 2 and 5 years posttransplant. We presume that this could be because left ventricular volume loading from a patent AVF may be enough to lead to a mild, albeit statistically significant, dilatation of the former, although not to a degree to cause a detectable difference in IVS thickness on echocardiography. The latter is an insensitive index of left ventricle mass, and others have also shown no change in its values after AVF closure, whereas left ventricle mass and mass index determined by echocardiography improved com-pared with that shown in controls.17 In accordance with that, a study using cardiac magnetic resonance imaging, which is much more accurate in deter-mining small changes in left ventricle mass compared with echocardiography, showed reduction of left ventricle mass after closure of AVF in transplanted patients compared with controls.¹⁶

There are several limitations to our study. Concerning design, this was a retrospective study as far as echocardiographic data collection and analysis; by definition, it could not be randomized regarding the presence of a patent AVF posttransplant. Echocardiographic data were restricted to available information from echocardiograms already con-ducted in our patients; therefore, newer and more sensitive left ventricle function indexes of left ventricle systolic function, like global longitudinal strain, were not available. No quantitative assessment of LVEF was performed; however, visual estimation of LVEF by experienced operators has been shown to be highly correlated with quantitative LVEF deter-mination.²⁵ We observed no difference in IVS thickness over the long term between the 2 groups despite left ventricle dilatation in the patent AVF group. The IVS is an insensitive index of left ventricle mass; therefore, estimation of the latter might be preferable. However, to detect small differences in left ventricle mass with echocardiography would require a large number of patients, and in this regard cardiac magnetic resonance imaging would be preferable. Additionally, the number of patients was relatively small, although our follow-up was quite long. Lack of AVF flow measurements also constituted a limitation of our study.²⁶ Nevertheless, the results of this study could be tested against a well-designed randomized study of AVF closure versus maintenance in patients with patent AVF at short-term posttransplant.

A patent AVF in kidney transplant patients may lead to left ventricle volume loading and even left ventricle systolic dysfunction over the long term, contributing to the observed high risk of cardiovascular disease, which implies that closure of AVF in selected patients could be beneficial. On the other hand, routine administration of ACE inhibitors/ARBs in this popu-lation could be considered, as it seems to confer a protective role.

Volume: 17 Issue: 5 October 2019

FULL TEXT