Maintaining a functioning vascular access (VA) is of vital importance for hemodialysis (HD) patients. The native arteriovenous fistula (native AVF) is the preferred option in most cases because it presents fewer complications and has greater durability than prosthetic arteriovenous fistulae (prosthetic AVF) and central venous catheters (CVC).1 The radiocephalic arteriovenous fistula (AVF) at the wrist is preferred due to its numerous advantages, including fewer complications, greater long-term patency, and better preservation of the proximal venous territory for another fistula in the event that new VA creation is required.2–7

The most common complication of AVF is stenosis, which can result in poorer HD quality and thrombosis, which is one of the most feared complications. It is important to identify stenoses at high risk of thrombosis in order to perform elective treatment to restore their function.8–12

Over time, various methods of AVF surveillance have been developed. AVF monitoring and/or surveillance screening methods are classified into 2 major groups: first- and second-generation.13–17 First-generation methods include physical examination, difficulty with puncture and/or cannulation of the AVF, aspiration of clots during puncture, increased negative pre-pump arterial pressure (AP), inability to achieve the prescribed blood pump flow (QB), increased return or venous pressure (VP), prolonged hemostasis time in the absence of excessive anticoagulation, native AVF overload test by blood pump flow (QB), dynamic venous pressure, equivalent or normalized static intra-access pressure, recirculation percentage, and unexplained decrease in hemodialysis adequacy (Kt, Kt/V, urea recirculation percentage).8,13–17 Second-generation methods are based on measuring AVF blood flow (QA) by indirect methods using dilution methodology8,18–26 or directly by Doppler ultrasound (DU).8,27,28 Among the dilution methods are those performed with reversal of HD blood lines: ultrasound dilution method18 hematocrit/hemoglobin dilution (delta H) associated with ultrafiltration,19,20 dialysate conductivity difference (ionic dialysance),21,22 thermodilution (thermal gradient),23,24 and those performed without reversal of blood lines: transcutaneous optical method25 and glucose perfusion method.26 Several studies have demonstrated agreement among dilution screening methods when performed simultaneously, and there is consensus in the literature that no technique is superior to another.21–26,29–33 Good correlation has also been demonstrated when QA is measured using dilution techniques and Doppler ultrasound.34–40

The dilution screening method using the temperature gradient technique (TGT) was described in 2007 by Wjinen et al.24 and was validated against the ultrasound dilution method. The TGT method calculates QA by analyzing the thermal recirculation produced when the arterial and venous lines of the HD circuit are reversed. For this purpose, the blood temperature monitor (BTM) sensor incorporated into the dialysis machine is used.24,30 In addition, a device can be used that shortens the measurement time by allowing reversal of the HD blood lines without the need to disconnect the needles or stop the blood pump of the HD machine.30 Studies have demonstrated that for QA calculation, agreement is similar when comparing the TGT method with other dilution methods.24,30

The objective of this study was to assess the degree of agreement between Doppler ultrasound and TGT for QA measurements, a comparison not previously performed. As a secondary objective, the relative importance of the different monitoring and/or surveillance methods for detecting significant stenosis was evaluated using a predictive model.

Our study demonstrates that flow calculated by TGT and by Doppler ultrasound shows a high correlation (96.7%), which allows both techniques to be used for AVF QA follow-up in our area. Our work is consistent with other studies in which good correlation has been demonstrated when QA measurement is performed using dilution techniques and Doppler ultrasound.34–40

In the study by Fontseré et al.,39 which assessed the agreement between Doppler ultrasound and thermodilution, a high correlation between both methods was shown. However, for QA values > 1000 ml/min, there was greater dispersion of results. In contrast, our study did not confirm this dispersion, showing agreement greater than 95% between Doppler ultrasound and TGT in this value subset. This demonstrates that for any AVF QA value, both methods show good correlation.

In the present study, the application of first-generation methods (negative pre-pump arterial pressure, recirculation, Kt, Kt/V) did not alert to the presence of significant stenosis, as no statistically significant differences were obtained. Only venous pressure in the first sampling showed a statistically significant decrease in patients with stenosis. This finding alone lacks value, as venous pressure depends on needle location in AVFs with stenosis8. In contrast, the second-generation methods used in our study (Doppler ultrasound and TGT) showed a statistically significant QA reduction in patients with stenosis, both in the first and third samplings, enabling the scheduling of elective therapeutic intervention in these patients. This reaffirms the importance of applying second-generation techniques for early detection of significant stenosis.43 According to the literature, some authors have found that first-generation methods, when showing altered values, may indicate severe AVF stenosis, but they are not the best methods for early detection of significant stenosis.8,44,45

Furthermore, in our study, a predictive model was developed using first-generation and second-generation methods. If group membership classification capacity is understood as sensitivity through the calculated functions, when stenosis prediction is performed using only first-generation methods, this prediction is not statistically significant. However, using QA measurement including both second-generation methods provides very high discriminative power, with a sensitivity of 100% and a specificity of approximately 74%, this prediction being statistically significant. The existence of approximately 26% false positives is acceptable, since imaging tests (ultrasound or fistulography) will rule out the lesion.46 Most importantly, QAcalculation using second-generation methods detects stenosis in 100% of cases, allowing scheduled treatment and preventing fistula loss. The results of our study are consistent with data from the clinical trial by Aragoncillo et al.,43 which demonstrates that flow measurement performed in combination with Doppler ultrasound and ultrasound dilution screening methods reduces the thrombosis rate and increases assisted primary patency after one year of follow-up compared with classical monitoring. Furthermore, Scaffaro et al.47 and Tonelli et al.48 confirm that performing second-generation techniques for native AVF surveillance can decrease the incidence of thrombosis, resulting in decreased morbidity and mortality.

Our findings corroborate the results obtained by other published studies confirming that first-generation methods are not sensitive for early stenosis, because they do not generate hemodynamic problems that affect dialysis quality or physical examination.49,50

Another interesting finding of our study is the fact that second-generation techniques alone show different specificities. While QA measured solely by Doppler ultrasound shows a slightly lower overall specificity than that obtained by combining the two second-generation methods, QA measured solely by TGT shows an evident decrease in specificity. Although it remains a statistically significant model, and in either case it will be able to detect all patients with stenosis, the TGT method will yield nearly half of the results indicating stenosis when it does not actually exist.

Our results regarding QAmeasured by Doppler ultrasound correlate with the study by Moreno Sánchez et al., where Doppler ultrasound shows a sensitivity of 98% and a specificity of 74%.28 However, for QA measured by TGT, there are no studies that specifically evaluate diagnostic performance parameters; therefore, our study adds information about its diagnostic utility.

Our study has some limitations that should be considered. First, it is an observational study with a small sample size, which may limit the statistical power to detect small differences and affect the generalizability of results. Therefore, additional studies would be needed, preferably randomized clinical trials with a larger sample size, to obtain more robust conclusions. Despite these limitations, the study provides preliminary evidence on the importance of establishing an adequate AVF follow-up and monitoring program that enables early detection of stenosis and elective treatments to prevent AVF thrombosis. In this regard, the results highlight the relevance of incorporating second-generation methods, demonstrating greater sensitivity for early stenosis detection, although these findings should be confirmed in larger prospective studies. Furthermore, the included population is well-characterized, and the study was conducted following a systematic and rigorous methodology based on direct comparison between first- and second-generation methods, which strengthens the internal validity of the results.

In conclusion, in our study, the agreement between QA measurement by the temperature gradient technique and Doppler ultrasound shows a high correlation at all flow levels. Furthermore, the predictive model demonstrates that second-generation methods are superior to first-generation methods for the early detection of clinically significant stenosis.

No funding was received for this study.