High doses of erythropoietin stimulating agents may be a risk factor for AV-fistula stenosis

1 Introduction

Patients treated with hemodialysis (HD) are dependent on a patent vascular access. According to guidelines the recommended first choice is a native arterio-venous fistula (AVF), while AV graft and central dialysis catheter are less preferred. Although the patency of the AVF is superior to a graft, complications arise such as thrombosis, stenosis and infection [1, 2]. Stenosis, a major hurdle, is associated with vessel anatomy, surgical technique, gender, older age and diabetes mellitus [3, 4]. Besides surgical technique most of the variables are not able to influence by the clinician [5–7]. In addition, there appears to be other unexplained reasons for differences, such as why some patients are more prone to recurrent stenosis or thrombosis than others [8, 9]. At present there are few clinical recommendations that prevent patients from recurrent AVF stenosis.

This case-control study aimed to clarify if any medical and/or laboratory factors, that can be altered, could be associated to AVF stenosis.

2 Material and methods

This retrospective observational case-control single center study included 66 patients with a native lower arm AVF undergoing chronic HD treatment (age range 18–90 years). Controls consisted of 33 patients with a patent AVF without need of intervention during a two-year period. Controls were selected matched by diagnosis and age compared with 33 patients (Cases) that had at least one radiological examination/intervention due to suspected AVF malfunction (case-control mode 1:2).

Statistical analyses were processed with SPSS 22 program, Analysis of categorical data was made with Fisher's test, and group statistics were performed with Student t-test and Mann-Whitney test. Paired statistics were performed with the Wilcoxon test. Data were expressed as mean (±SD) and median values and a two-tailed p value of less than 0.05 considered significant.

All patients had an end-to-side AVF and most a left arm placement (Controls: 88% , Cases: 85%). In Cases invasive radiological investigations were performed most (except one) by fistulography or angiography. A significant stenosis (>50%) was detected in 27 of the cases (82%), all in the venous limb of the access. In six other cases with suspicion of AVF flow problems, no significant stenosis was detected by the use of fistulography (n = 1), ultrasound (n = 1), or angiography (n = 4); in one patient an aneurysm was observed using angiography.

Three vascular surgeons had performed operation to the same extent on both groups (Controls/Cases: 6/5, 11/11, 10/13). The hemoglobin levels were similar before intervention, but were lower after invasive intervention in Cases versus Controls at the follow up measurement (mean value 110 g/L versus 117 g/L, respectively, p = 0.007). Cases, had a higher weekly dose of ESA than Controls, both before intervention (8312±7119 U/w versus 4348±3790 U/w, p = 0.005) and also after intervention (7656±6795 U/w versus 4477±3895 U/w, p = 0.018). The results were confirmed by paired statistics (before: p = 0.004, after: p = 0.017). The s-phosphate was higher in Cases versus Controls before (1.8±0.7 versus 1.5±0.4, p = 0.010), but not after intervention.

There was a similar distribution of blend of ESA between the two groups; Epoetin beta (Neorecormon ^®) was the most frequent used ESA (27 patients among cases and 29 among controls).

There was no difference in variables such as dosing of low molecular weight heparin (LMWH) during HD, duration of HD (hours/week), weekly dialysis dosing or efficacy (Kt/V), parathyroid hormone level (PTH), albumin, C-reactive protein (CRP), ferritin, calcium or BMI. There was no difference in gender, smoking habits, or other medication except that the administered doses of intravenous iron were lower among cases than controls after intervention (p = 0.043). The access blood flow was improved after intervention among cases from a mean of 496 ml/min to 992 ml/min (p = 0.047).

4 Discussion

Similar to our previous report of a smaller case-control material [10] this extended case-control study exhibited almost double doses of ESA, with similar hemoglobin values, in Cases versus Controls. No other factors of significance were found before intervention except a higher serum phosphate that was corrected after the stenoses were dilated (restoring blood flow). This indicated that the rise in phosphate was due to somewhat impaired dialysis efficacy due to fistula problems. However, the higher phosphate level was within the recommended target values [11] The lower b-hemoglobin after intervention compared to Controls at year two could be explained by the blood loss that appeared during the intervention. There were similar levels of b-hemoglobin and CRP before intervention. Therefore the high doses of ESA in Cases indicate less response of the patient to the ESA to increase b-hemoglobin. This lack of response seemed not related to inflammation (using CRP as a rough indicator). This can indicate that the Cases held a partial resistance to ESA. If so, a lowered response to ESA of the patient (genetic or other reasons) may in addition result in AVF stenosis. This is not evidently shown in our material, since the diagnoses responsible for the kidney disease were similar in both groups. Another reason could be that ESA in a higher dose may support vascular disturbances such as increasing intimal hyperplasia induced by the local punctures, flow turbulence, increased blood flow in stenotic areas, and local pressure.

Previous studies also found high ESA doses in patients with vascular failure [12–14]. Those studies included patients with AVF problems while patients without problems were not matched for variables such as age, diabetes mellitus, and other diagnoses related to kidney injury, that differed significantly between the groups. Ikegaya et al. noted significant elevation in expression of erythropoietin receptors and TGF-beta1 in patent AVF compared to the cutaneous veins. In addition enhancement of the receptor expression was detected in the stenotic fistulae. They hypothesized that increased recombinant erythropoietin binding thereby may appear and contribute to the development of AVF stenosis caused by intimal hyperplasia and extracellular matrix accumulation in response to increased TGF-beta1 expression [15]. Those data indicate that prevention of turbulence and stenosis of even smaller magnitude may be beneficial. We noted that Far infrared therapy may help to improve vascular diameter [16] and prolong patency of AVF, [17] an option that might be of value in clinical use. Another option would be to maintain a lower b-hemoglobin instead of pushing ESA doses towards high levels in patients at risk.

A limitation of the present study was the retrospective design. However, it was not possible to make a prospective design since the risk factors to investigate were not clarified in advance. By using a case-control design an attempt was made to neutralize external factors as much as possible. The study population allowed significant power.

In conclusion almost double doses of ESA were administered in patients with AVF stenosis. Since ESA is currently the drug of choice in treating anemia among HD patients, the question arises if ESA may cause local hypertrophic effects on the vascular endothelium, that would motivate clinician to prescribe lower doses of ESA in patients at risk.

Conflict of interest

The authors have declared that no conflict of interest exists.

The study was approved by the ethics committee in Gothenburg, Sweden, EPN, DNR, 402-14, 2014-08-22, T-037-15, 2015-01-09, and was performed according to Declaration of Helsinki.

Obtained funding

The Department of Research and Development at Skaraborg Hospital supported the study.