Mechanistic Basis for Intradialytic Hypertension with Hemodialysis

In this issue of CJASN, Elsayed et al. describe the “Association of Bioimpedance Parameters with Increases in Blood Pressure during Hemodialysis.”¹ Hand on heart, what have you told your 50-year-old patient lately who came in with systolic BP of 130 mm Hg before hemodialysis, completed treatment with 180 mm Hg, asked why this was happening, what should be done, and whether he could go home, despite headache? We recently replied that he might be fluid overloaded, so we should start lowering dry weight. The gentleman, however, did not feel fluid overloaded and showed us a list with stable weights and residual urine volumes. We ensured that dialysate sodium was below serum sodium, but were reluctant about administering more than two antihypertensive drugs. After another episode of headache, the patient presented to a community hospital, had a normal computed tomography and magnetic resonance imaging result, but arrived back at our center with five different antihypertensives. Intradialytic hypertension did not fully disappear. Recently, he has been asking whether he could reduce the number of drugs: the quest continues.

Mechanisms for Intradialytic Hypertension

In an earlier review,² Chazot and Jean postulated that there might be eight potential mechanisms for intradialytic hypertension: renin–angiotensin system activation, sympathetic overactivity, intradialytic K⁺/Ca²⁺ variations (electrolyte imbalance), erythropoietin-stimulating agents, fluid overload, and increased cardiac output. In a small study, thoroughly discussed in the review by Chazot, several of these mechanisms could not be confirmed.³ The study by Elsayed et al.¹ focused on bioimpedance-derived measures (for fluid overload), reporting the relationship between shorter bioimpedance vector length and pre- to post-dialysis rise in BP in Frequent Hemodialysis Network participants. We, therefore, suggest reviewing the fundamentals of bioimpedance.

Technical Basics of (Bio)impedance

Imagine you are using two electrodes at the wrist and ankle to apply a weak alternating current of a given frequency to the human body (Figure 1A, red). The opposition to the alternating electrical current is denoted by the complex quantity termed impedance, Z. Its cartesian representation consists of real resistance R and imaginary reactance X (Z=R+jX, where j is the imaginary unit) and is analogous to the polar representation Z=${\left|Z\right|e}^{j\theta }$. Here, $\left|Z\right|$ is the impedance magnitude and $\theta$ the phase angle (Figure 1C). Impedance can be measured by placing two additional sensing electrodes within the circuit, which only measure the corresponding voltage drop (Figure 1A, blue). Human tissue is typically modeled as an equivalent electrical circuit of resistors and capacitors. In circuits solely comprising resistors, reactance X and phase angle $\theta$ are zero, and impedance Z is then only characterized by resistance R and $\left|Z\right|$. At the frequency used in the study by Elsayed et al.,¹ the resistance is directly proportional to conductor length, indirectly proportional to ion concentration and the conductor's cross-sectional area, and therefore an indirect, inverse measure of fluid volume. The cell membrane between extracellular and intracellular water compartments may be considered as a capacitor, insulating at zero and conducting current with a phase angle $\theta$ at finite frequencies, thereby introducing an imaginary reactance X (Figure 1B). Thus, in the frequency range used in bioimpedance analysis, impedance depends on body size, electrode configuration, electrolyte concentration, fluid volume, and cell mass.

Figure 1.

Setup, signal, R–X graph, and confidence ellipses of bioimpedance measurements. (A) A typical setup for bioimpedance measurements with two current carrying electrodes (red) and two sensing electrodes (blue). (B) Schematically shows the sinewaves of current and voltage that are phase-shifted by $\theta$ (phase angle). (C) Illustrates the representation of the complex impedance value in the resistance–reactance graph and in vector form. Vector length and phase angle must not be interpreted separately, but should always be regarded as one value (which they are per definition). (D) Represents the idea behind the 50%, 75%, and 95% confidence ellipses for impedance vectors standardized by body height according to Piccoli et al.⁴ The upper and lower pole of the confidence ellipse include fluid-depleted and fluid-overloaded patients, respectively, while the left and right poles are more dependent on cell mass (left: obese and athletic; right: cachectic and anorexic). Piccoli⁵ suggested that stable patients receiving hemodialysis with relatively short vectors may be brought to euvolemia by adapting ultrafiltration prescription if their phase angle is otherwise within the target range. In patients with phase angles on the edge of the confidence ellipse, however, focus on ultrafiltration alone is not expected to solve the issue, but rather will require nutritional treatment as well. Such additional information might allow for more accurate description and treatment of hemodialysis subgroups in the future. j, imaginary unit; $\theta$, phase angle; R, resistance; V, voltage; X, reactance; Z, impedance; $\left|Z\right|$, impedance magnitude.

Definitions for Intradialytic Hypertension

Back to the clinic. There is no standardized definition for intradialytic hypertension.⁶ However, during the past decades, the most frequent definition used for studies dealing with intradialytic hypertension was a 10 mm Hg increase in post- compared with pre-dialysis systolic BPs, legitimized by significantly increased mortality.⁷ Recently, the Sarafidis group added a >150 mm Hg threshold for post-dialysis systolic BP⁸ as a useful and pragmatic criterion. Elsayed et al.¹ defined intradialytic hypertension as any intradialytic increase in systolic BPs. Nineteen percent sessions fulfilled this criterion, and shorter vector length (per 50 Ω/m) was associated with an adjusted odds ratio of 1.66 (95% confidence interval, 1.07 to 2.55) for such sessions. An increase of >10 mm Hg systolic BP pre- to post-dialysis was observed in 7% sessions, and shorter vector length (per 50 Ω/m) was associated with an adjusted odds ratio of 2.17 (95% confidence interval, 0.88 to 5.36) for such sessions. In various analyses, shorter vector length, and hence increased fluid volume, was associated with higher systolic BP after dialysis, also.

Bioimpedance Vector Analysis, BP, and Fluid Overload

Conventional bioimpedance analysis and bioimpedance spectroscopy methods (single- and multifrequency, respectively) require tissue models for interpretation.⁹ These models allow estimating fluid compartments, body composition, and, in some cases, even fluid overload.¹⁰ Because these equations are empirically derived through linear regression models, their validity is at the mercy of reference populations, adding uncertainty. Bioimpedance vector analysis, however, does not rely on empirically derived equations: Its interpretation takes place in the resistance–reactance graph, exclusively. On the basis of the fundamental work by Piccoli, resistance R and reactance X are typically normalized for body height.⁵ A patient's impedance vector is depicted in relation to a sex- and age-matched reference individual (Figure 1D). In Piccoli's earlier analysis,⁵ patients with frequent intradialytic hypotension were characterized by longer, less steep impedance vectors measured pre-dialysis (implying volume depletion), compared with stable patients. In the analysis by Elsayed et al., patients with hypertension were found to have shorter impedance vectors. In addition, patients in the lowest tertile of vector length had larger phase angles, suggesting that hypertension is not solely associated with fluid, but also with cell mass.

Fluid Overload, Endothelium, and Sodium Gradient

The association between intradialytic hypertension and fluid overload has previously been reported, as acknowledged by Elsayed et al.,¹ whose work on vector length (and phase angle) is additionally insightful. Fluid overload is, therefore, among the most plausible mechanisms for intradialytic hypertension. Two research groups have, moreover, shown higher ambulatory BP levels among patients with intradialytic hypertension compared with those without, indicating that intradialytic hypertension is an exacerbation of hypertension in general.¹¹ Mechanistically, there is another important relationship, linking fluid overload and intradialytic hypertension with the endothelium, through vascular refilling. In the study mentioned above,³ patients in the intradialytic hypertension group had a significant pre- to post-dialysis increase of peripheral resistance, plasma endothelin, and lower post-dialysis nitric oxide over endothelin-1 ratio, compared with the control group. Intradialytic hematocrit changes were lower in the intradialytic hypertension group,³ suggesting that increased refilling from the interstitial space might trigger the impairment of nitric oxide/endothelin balance and vasoconstriction. The dialysate-to-serum sodium gradient (adjusted for in the study by Elsayed et al.¹) may play an additional role. Its association with intradialytic hypertension¹² may share the same pathophysiology. A randomized crossover trial with dialysate sodium 5 mEq/L lower versus 5 mEq/L higher than serum sodium and at least three subsequent studies with individualized or low versus standard dialysate sodium concentrations showed an effect of these interventions on intradialytic hypertension and/or systolic and diastolic BP before/after hemodialysis.¹¹

Treatment Conclusion

How should our patient(s) be treated? Strong evidence for anything is lacking, but checking the dialysate-to-serum sodium gradient, potentially decreasing it if positive, is likely worthwhile. Fluid removal with ultrafiltration has, decades ago, been shown efficient to correct intradialytic hypertension.² To convince reluctant patients, it is helpful to collect data in favor of fluid overload: bioimpedance; intradialytic blood volume monitoring, a flat or positive curve indicating active intradialytic refilling; and ambulatory BP monitoring, high interdialytic BP reflecting fluid excess. Antihypertensives with proven benefit¹¹ may not be avoidable, but fluid removal is worth a try. The interesting study by Elsayed et al.¹ reminds us that it is time to act.²

Disclosures

M. Hecking reports employment with Kuratorium for Dialysis and Transplantation Germany; research funding from Astellas Pharma, Boehringer Ingelheim, Eli Lilly, and Siemens Healthcare; and honoraria from Bayer AG, Fresenius Medical Care, Janssen-Cilag, and Vifor. M. Hecking has served as a speaker and/or consultant for Astellas Pharma, AstraZeneca, Eli Lilly, Fresenius Medical Care, Janssen-Cilag, Siemens Healthcare, and Vifor and has previously received academic study support from Astellas Pharma, Boehringer Ingelheim, Eli Lilly, Nikkiso, and Siemens Healthcare (not related to the present work). All remaining authors have nothing to disclose.

Funding

This work was supported by the Vienna Science and Technology (WWTF) Precision Medicine Grant LS20-079.

Author Contributions

Conceptualization: Manfred Hecking.

Project administration: Manfred Hecking.

Resources: Manfred Hecking.

Visualization: Manfred Hecking, Sebastian Mussnig.

Writing – original draft: Charles Chazot, Manfred Hecking, Sebastian Mussnig.

Writing – review & editing: Charles Chazot, Manfred Hecking, Sebastian Mussnig.