Pressure natriuresis refers to the phenomenon that increased pressure in the renal arteries, which leads to an increase in total sodium excretion.1 This mechanism is the major way by which the kidney can compensate for systemic arterial hypertension, regardless of the etiology of the hypertension. The inverse holds true as well, which decreased pressure to the renal artery leads to sodium avidity in the kidney and decreased sodium excretion.2

Although pressure natriuresis typically implies a mechanism by which the kidney is able to slowly reduce systemic blood pressure in response to hypertension over a period of time, it has been described acutely in experimental settings as well.3 We report a case of polyuria and natriuresis resulting from high pressures delivered locally to the renal artery by blood flow from the arterial catheter of an extracorporeal membrane oxygenation (ECMO) circuit.

Case Report

A 40-year-old woman presented with weight loss and dysphagia. Her hospital course was significant for vomiting and hypokalemia and an EKG showed a prolonged QTc interval of 513 ms. She underwent upper endoscopy evaluation but subsequently experienced cardiac arrest due to polymorphic ventricular tachycardia. She developed repeat VT arrests which were felt to be due to prolonged QTc from hypokalemia and a possible underlying genetic long QT syndrome leading to “R on T phenomenon.” Coronary angiography demonstrated normal coronary arteries and echocardiography revealed left ventricular (LV) ejection fraction of 15%. Due to progressive cardiogenic shock, she was cannulated for VA-ECMO with a 15 Fr arterial cannula placed in the right femoral artery and a 23 Fr cannula placed in the right femoral vein. The ECMO flow was initiated at 3.6 L/min and, given LV dysfunction, LV venting was provided via an intracardiac Impella heart pump at 1.5 L/min. At the initiation of ECMO, right atrial (RA) pressure was 0 mm Hg, pulmonary arterial pressures was 14/4 mm Hg, and pulmonary capillary wedge pressure was 7 mm Hg. The lactic acid level was 2.8 mmol/L.

After initiation of ECMO, she developed marked polyuria. Notably, before ECMO cannulation, urine output (UOP) was roughly 1 L/d. In the first 24 hours after ECMO initiation, she urinated 25.9 L. The ECMO flow, UOP, and sodium excretion are depicted in Figure 1. Her initial urine sodium was 144 meq/L, urine chloride was 135 meq/L, and urine osmolality 297 mOsm/kg. Due to concerns for iatrogenic volume expansion causing a solute diuresis, she had 2/3 of her urine volume repleted with Lactated Ringer’s. However, this resulted in hypotension, increased pressor requirement, low-flow alarms, and ECMO circuit “chattering.” During this time, her RA pressure was 5–7 mm Hg. The administration of increased IV fluids promptly increased her blood pressure, reduced pressor requirements, and resolved ECMO alarms.

Figure 1.:

Relationship between VA-ECMO flows, sodium excretion, and urine output over time. VA-ECMO flow rates (LPM; red line), sodium excretion (mEq/d; green line), urine output (L; blue line), and net fluid balance (L; gray bars) are plotted in 24 hour intervals during which the patient was supported by VA-ECMO. Day 0 represents when ECMO was initiated. Both urine output and total sodium excretion decrease to a corresponding decrease in ECMO flow rates while total net fluid balance remains constant. LPM, liters per minute; UOP, urine output; VA-ECMO, venoarterial extracorporeal membrane oxygenation.;

On the subsequent day, the patient urinated 15.6 L and urine sodium was 149 meq/L. Plasma renin activity was 9.38 ng/ml/h. RA pressure was 5 mm Hg. The trend of central venous pressure, renin activity, urine sodium, and serum lactate are shown in Figure 2. Her serum creatinine remained at her prearrest level of 0.8–0.9 mg/dl. She did not receive diuretics. Her blood urea nitrogen was less than 10 mg/dl, and serum glucose was consistently under 150 mg/dl.

Figure 2.:

Relationship between hemodynamic parameters during mechanical support by VA-ECMO. CVP (mm Hg; blue line), urinary sodium (mEq/L; red line), plasma renin activity (ng/mL/h; orange circles), and lactate concentrations (mmol/L; green line) are plotted during the period of support by VA-ECMO. Day 0 represents when ECMO was initiated. Soon after cannulation, urine sodium concentration was high despite low CVP, high plasma renin activity, and high serum lactate. CVP, central venous pressure; VA-ECMO, venoarterial extracorporeal membrane oxygenation.

The clinical assessment was that she was hypovolemic, yet her renal response of polyuria with high urine sodium was incongruent with the clinical assessment. Given the temporal relation of polyuria to the initiation of ECMO, it was considered if ECMO itself could be the cause of the polyuria. To assess the relation between the ECMO flows (and thus the pressure at the femoral arterial cannula) and urine volume, the ECMO flows were reduced and the effect on UOP was assessed. When the ECMO flows were decreased from 3.2 to 2.5 L/min her UOP decreased by 50% from 8.4 to 4.2 L/d (Figure 1). Notably, her urine osmolality remained the same, demonstrating that this was a decrease in total natriuresis, not simply an increase in urine concentration. As evidenced in Figure 1, the ECMO flow was correlated with UOP in this patient. The only time the urine flow increased while ECMO flow decreased was after a single dose of furosemide that was administered on day 6. Subsequently, her ECMO flow was decreased further to below 2 L/min and her UOP dropped below 2 L/d. Extracorporeal membrane oxygenation was then discontinued and the patient was decannulated. No further polyuria was noted during her hospitalization. A definitive etiology of her weight loss and dysphagia was not determined and evaluation for genetic causes of prolonged QTc is planned.

Discussion

In normal physiology, hypervolemia increases the mean arterial pressure, suppressing sodium-retaining hormones (renin, angiotensin, aldosterone, and catecholamines), and increasing the renal blood flow.1 As a result, urine volume, urine sodium concentration, and total natriuresis increase. As hypervolemia resolves, there is a consequent reduction in mean arterial pressure causing an increase in sodium-retaining hormones, a reduction in renal blood flow and reduction in natriuresis. In a dog model of unilateral renal artery stenosis, increased systemic blood pressure resulted in increased natriuresis from the nonstenotic contralateral kidney due to exposure to increased pressure.3

This patient developed marked polyuria after the initiation of VA-ECMO. Considering her elevated urine osmolality, her polyuria was caused by a solute diuresis and not a water diuresis. It was not driven by glucose or urea clearance but by high levels of urine sodium and chloride. Her degree of natriuresis was extreme. In the first 24 hours after VA ECMO initiation, she excreted 25.9 L of urine with urine sodium concentration of 144 meq/L, a total of 3730 meq of sodium, which is equal to more than 85 g of sodium. The initial hypothesis was that due to ongoing shock, she was administered large amounts of salt-rich intravenous fluids leading to hypervolemia and an appropriate diuresis. This hypothesis was inconsistent with her consistently low cardiac filling pressures and her elevated renin level. Additionally, attempts to replacing only a fraction of her UOP resulted in low-flow alarms on her VA-ECMO circuit, ECMO chatter, and rapid increase in pressor requirement. Thus, her clinical status suggested that she was not hypervolemic but hypovolemic, and her natriuresis was therefore inappropriate.

As can be seen in Figure 1, the onset of the polyuria coincided with the initiation of ECMO, improved when ECMO was weaned, and terminated when ECMO was stopped. We believe the ECMO was directly responsible for the polyuria. We postulate that the high-pressure jet from the VA-ECMO catheter delivered high pressures to the renal arteries despite the low MAP measured at the radial artery. This local increase in pressure sensed at the renal arteries resulted in pressure natriuresis, polyuria, and clinical volume depletion, which can be termed ECMO-associated salt wasting. Although the exact pressure was at the ECMO cannula tip in the femoral artery is unknown, the postmembrane pressure in the ECMO circuit was as high as 262 mm Hg. We believe that even with an expected pressure drop across the catheter, the intraarterial pressure in proximity to the catheter tip would be significantly elevated. This further explains the elevated renin levels, which were incongruous with a hypothesis of clinical hypervolemia. Although decreased renal blood flow is one stimulus for renin secretion, beta-1 activation from hypovolemia and decreased stretch in the carotid artery can also stimulate renin secretion.4

To our knowledge, this is the first report which describe ECMO-associated salt wasting. It is not clear what was unique about this patient that led to the pressure natriuresis, whereas many other patients on similar ECMO settings do not develop it. It is noteworthy that nearly 80% of patients on ECMO develop acute kidney injury (AKI),5 with up to 60% requiring renal replacement therapy.6 Significant AKI would preclude the development of polyuria in most patients on VA-ECMO. Our patient had normal renal function despite her profound shock, an important factor permitting development of pressure natriuresis. Although x-ray imaging demonstrated the arterial ECMO cannula overlaying the common femoral artery, it is possible that the catheter position or angle was more proximate to the renal arteries than other patients. Finally, it is possible that there could have been a summation of pressures near the renal arteries from the retrograde pumping ECMO cannula and the anterograde pumping Impella. In fact, aortic pumps are in development to utilize this mechanism to improve renal perfusion and induce pressure natriuresis as a treatment for heart failure.7

A limitation of our case is the inability to measure the pressure at the femoral artery cannula tip. Nonetheless, we believe our thorough investigation of the patient’s polyuria excluded all accepted causes of polyuria and leaves the diagnosis of pressure natriuresis from the high-pressure output of the ECMO cannula as the best explanation for this patient’s polyuria in the face of clinical hypovolemia.

Conclusion

In patients who develop polyuria while on VA-ECMO, the possibility of pressure natriuresis be considered. If confirmed, therapeutic interventions include immediate volume expansion to correct clinical hypovolemia, retraction of the arterial cannula if placed in inappropriate proximity to the renal arteries and, as tolerated by patient condition, weaning of ECMO flow.

Acknowledgment

The authors thank Dr. Qais Al-Awqati, who reviewed the manuscript and guided the authors through the physiology of this case.

References 1. Guyton AC: Blood pressure control–special role of the kidneys and body fluids. Science. 252: 1813–1816, 1991. 2. Selkurt EE, Hall PW, Spencer MP: Influence of graded arterial pressure decrement on renal clearance of creatinine, p-aminohippurate and sodium. Am J Physiol. 159: 369–378, 1949. 3. Mizelle HL, Montani JP, Hester RL, Didlake RH, Hall JE: Role of pressure natriuresis in long-term control of renal electrolyte excretion. Hypertension. 22: 102–110, 1993. 4. Kopp U, DiBona GF: Interaction of renal beta 1-adrenoceptors and prostaglandins in reflex renin release. Am J Physiol. 244: F418–F424, 1983. 5. Chen YC, Tsai FC, Chang CH, et al.: Prognosis of patients on extracorporeal membrane oxygenation: The impact of acute kidney injury on mortality. Ann Thorac Surg. 91: 137–142, 2011. 6. Kielstein JT, Heiden AM, Beutel G, et al.: Renal function and survival in 200 patients undergoing ECMO therapy. Nephrol Dial Transplant. 28: 86–90, 2013. 7. Shabari FR, George J, Cuchiara MP, et al.: Improved hemodynamics with a novel miniaturized intra-aortic axial flow pump in a porcine model of acute left ventricular dysfunction. ASAIO J. 59: 240–245, 2013.