Intermittent thoracic resuscitative endovascular balloon occlusion of the aorta improves renal function compared to 60 min continuous application after porcine class III hemorrhage

Resuscitative Endovascular Balloon Occlusion of the Aorta (REBOA) may be considered for stabilization of trauma patients with hemorrhage from below the diaphragm who are not responding to standard resuscitation [1]. The balloon device is positioned in the thoracic, descending aorta (zone 1) or below the renal arteries in the abdominal aorta (zone 3) depending on the anatomic location of vascular injury. The minimally invasive procedure may replace open thoracic aortic cross-clamping after non-compressible torso hemorrhage, a situation associated with high mortality [2‐4]. Research during the past decade of global physiological effects of REBOA has shown that acute kidney injury (AKI) constitutes an independent risk factor for mortality and prolonged intensive care after trauma and may occur as a result of REBOA [5‐10]. Furthermore, myocardial injury occurs during prolonged complete thoracic REBOA [11]. Therefore, current recommendations state a maximum of 30 min, zone 1 REBOA intervention [1]. Organ damage and ischemic injury cause multiple organ dysfunction, a major cause of mortality > 24 h after trauma [12]. It is possible that permissive distal organ perfusion during REBOA reduces ischemic injury. REBOA with smaller devices and partial or intermittent distal aortic flow led to fewer complications in animal models and humans [13]. Today, there are two principal methods of achieving distal aortic flow: partial and intermittent REBOA. Partial REBOA is associated with improved mesenteric perfusion, decreased acidosis, decreased myocardial load [14, 15] and decreased ischemic organ injury compared to intermittent REBOA [16]. Intermittent REBOA is associated with prolonged survival and decreased acidosis compared to continuous REBOA [17]. However, both methods have limitations. Partial REBOA is technically challenging and may not be feasible in situations with limited capabilities such as in a prehospital environment or in combat operations. Intermittent REBOA may cause rebleeding and hemodynamic instability. The relation between intermittent REBOA and distal organ perfusion, and the amount of reperfusion required for improved renal function have not been sufficiently elucidated. The regulation of organ blood flow is complex and multifactorial in REBOA applied after trauma and hemorrhagic shock, and the reduction in renal blood flow may continue after REBOA release [18]. Therefore, we investigated circulatory effects, renal artery blood flow and markers for renal injury, Osteopontine, NGAL and Vimentin, for 60 min of intermittent (iREBOA) versus continuous, complete (cREBOA) zone 1 REBOA, in a model of potentially lethal and uncontrolled hemorrhage. We hypothesized that iREBOA would improve the blood flow in the renal arteries and decrease markers of renal injury.

This study was approved and conducted in accordance with the Swedish regional ethics approval board for animal research (ethical approval ID S3-15 and 1470). Castrated crossbred male swine, (n = 12, mean weight 60.3 kg), were utilized in a randomized (closed envelope), prospective experimental study design including five phases (Fig. 1). Animals were assigned to continuous, complete 60-min zone 1 aortic occlusion (cREBOA; n = 6) or a 3-min full deflation interval every 10 min for 60 min (iREBOA; n = 6) followed by 60-min reperfusion after REBOA intervention.

Baseline parameters did not differ between groups (data not shown). One REBOA balloon malfunctioned during first inflation and was quickly replaced. A zone 1 location of the REBOA balloons was confirmed in all animals during post-mortem laparotomy. Average blood loss was 1360 mL (cREBOA: 1382 mL; iREBOA: 1337 mL). No significant rebleeding from the femoral artery was observed during reperfusion intervals in the iREBOA animals. Total ischemic time was lower in the iREBOA group (48.6 min ± 1.3 min versus 60 min ± 0 min in cREBOA) (p ≤ 0.001) (Fig. 2a). Survival was 100% (iREBOA; n = 6 of 6) and 83% (cREBOA; n = 5 of 6) (Fig. 2b). The death occurred within 5 min of the reperfusion phase, after total balloon deflation. Ventricular fibrillation was observed in EKG before the criterion of death was reached and no additional resuscitation was performed. Available data were included, and the animal was excluded in further analyses for the remainder of the study. The ratios of successful three minutes reperfusion intervals are shown in Fig. 2c and d. The need for rescue occlusion due to severe hypotension (Figs. 2c, 3a) showed a linear increase during the 60 min intervention. A full 3-min reperfusion was successfully conducted in 63% of total, ranging from 83 to 33% between 10- and 50-min intervention time. Individual values of reperfusion times are presented in Table 1.

Systolic blood pressure increased from a mean 56 mm Hg to 164 mm Hg (198%) in both groups with balloon inflation (cREBOA: mean 57 ± 19 mm Hg to 163 ± 34 mm Hg; iREBOA: mean 55 ± 16 mm Hg to 165 ± 29 mm Hg). SBP decreased in the iREBOA group during balloon deflations (Fig. 3a) with a total of 11 hypotensive events (Fig. 2d). Five minutes after final balloon deflation, the cREBOA animals had lower blood pressure (Fig. 3a). Cardiac performance showed no significant differences, however, there was a clear trend toward lower heart rate in the iREBOA group during balloon deflations with no difference in cardiac output (Fig. 3b, c, e). Norepinephrine infusion time increased after reperfusion and continuous occlusion (cREBOA: mean 12.5 min; iREBOA: mean 5 min; p < 0.05) (Fig. 3d). Troponin I increased in cREBOA animals (mean 666 ng/L) after 60 min intervention compared to iREBOA (mean 187 ng/L). Troponin increased during REBOA intervention and showed a similar trend toward normalization after 60 min reperfusion (cREBOA: 352 ng/L; iREBOA: 127 ng/L) in both groups (Fig. 3f).

Renal blood flow increased during reperfusion intervals in iREBOA animals and after 60 min reperfusion (Fig. 4a). The return of blood flow in the renal artery was 27.5% of baseline values during the first 3 min reperfusion where after it decreased gradually to 13.5%. Peak return of RBF after reperfusion was 61% (iREBOA) and 39% (cREBOA), respectively (Fig. 4a). The increase in renal artery blood flow during reperfusion corresponded to a decrease in proximal systolic blood pressure and increased blood flow distal to the balloon (Fig. 4b). Both groups showed a restoration of urine output after 60 min reperfusion, however, the cumulative urine output increased after iREBOA (mean difference 118 mL (± 17 mL; SE 30.38 mL); p < 0.005) (Fig. 4c–e). Mean plasma levels of creatinine did not differ between groups at 60 min intervention or after reperfusion. In both groups, creatinine levels increased after 60 min REBOA compared to baseline (Fig. 4f). Creatinine increased 90% (cREBOA; p < 0.05) and 51% (iREBOA; p ≤ 0.001) between baseline and 60-min c/iREBOA, respectively (Fig. 4f). During reperfusion, plasma creatinine clearance was 27% in both groups (cREBOA:43 µmol/L; iREBOA 40 µmol/L).

There was a trend toward increased metabolic acidosis in the cREBOA group during reperfusion phase, however, the mean levels of lactate, base excess or pH did not reach statistical significance (Fig. 5a–c). During reperfusion, cREBOA animals were administered higher norepinephrine doses (cREBOA: mean 75 µg; iREBOA: mean 30 µg; p < 0.05) to remain normotensive (Fig. 3d). Crystalloid and autologous whole blood resuscitation did not differ between groups according to the study protocol. PaO₂ and renal vein PO₂ showed no differences between groups during the experiments. There was a trend toward increased PaCO₂ during reperfusion phase in the cREBOA group, and renal vein PCO₂ increased after cREBOA during first 30 min after balloon deflation (Fig. 5d, h).

The integrated intensity (arbitrary units) of Osteopontin, in NKCC2-positive cells (the thick ascending loop of Henle), was significantly higher (p = 0.0256) in cREBOA (10184932 ± 8018146) vs iREBOA (1275219 ± 1450814). No significant difference (p = 0.9875) in levels of vimentin between cREBOA (1488056 ± 1277875) and iREBOA (1464134 ± 2730189) was detected. Neither was there any difference (p = 0.9709) in levels of NGAL in cREBOA (1226757 ± 2168597) compared with iREBOA (1270533 ± 1535600) (Fig. 6a–i).

In this study, we demonstrate that the iREBOA protocol was survivable, did not cause rebleeding, decreased the total ischemic time and increased the renal blood flow and urine output compared to cREBOA.

The rationale for the iREBOA protocol was that an intermittent reperfusion strategy may restore oxygen levels and decrease the anaerobic metabolism before severe ischemic injury could occur, and the rationale for the 60 min application time was to investigate the physiological effects of the reperfusion intervals in relation to a possible extended safe time of supra-celiac REBOA by intermittent balloon deflations. During the development of the protocol, one third of the animals died on prompt deflation of the REBOA balloon after 60 min continuous occlusion, which suggested that deflation of cREBOA may lead to severe circulatory consequences. While this study was not powered to determine survival primarily, our results suggest that while both methods may cause circulatory adverse effects on removal, iREBOA had a lower requirement for norepinephrine and may therefore lead to lower risks on removal. Nevertheless, zone 1 REBOA balloon deflation, even after a short period of total occlusion, should be performed with hemodynamic monitoring and guided by a blood pressure threshold. Until a robust strategy to mitigate the combined effects of hypotension and ischemic reperfusion with associated acidosis and electrolyte derangements has been established [6], the time limit of 30 min continuous (total) zone 1 REBOA should not be exceeded.

Conflicting data have been reported regarding the benefits and risks of intermittent balloon deflation during REBOA [16, 17, 19, 20]. A strict time-based protocol for intermittent reperfusion during REBOA may be easy to manage but risks severe hypotension and should be utilized together with hemodynamic monitoring and guided by blood pressure limits. In this study, the balloon was promptly deflated and re-inflated at each reperfusion interval. It is possible that this approach led to more severe hemodynamic changes than a more stepwise and gradual return of aortic flow. Such techniques have been described and utilized in partial REBOA. However, the return of aortic blood flow during graded deflation of the balloon may be unpredictable and not linearly corresponding to the balloon volume. Moreover, the volume is not transferrable between individuals [21]. Therefore, our technique of prompt balloon deflation allowed for a uniform effect in the animals. The ratio of rescue occlusions was 37%, which suggested that patients should be monitored with invasive arterial blood pressure and that deflation of the balloon should be performed in combination with resuscitation capabilities including volume infusion and vasopressors, even after short intervention times. The ratio of rescue occlusions also suggested that a pressure-based intermittent reperfusion protocol may be superior to a strictly time-based protocol to prevent severe hypotension [16, 17]. We detected a trend toward shorter tolerance of reperfusion times during the 60-min intervention. This may be ameliorated by adjunctive fluid resuscitation. The optimal volume of resuscitation during iREBOA may be guided by hemodynamic parameters and blood gases and should be the focus of future investigations.

Management of REBOA without standard intra-hospital monitoring of the patient’s physiology requires knowledge of the associated organ damage [22]. The kidneys play a vital role in the response to hemorrhage shock. First, the reduced blood pressure secondary to hypovolemia causes decreased glomerular filtration and decreased urine output. Second, the combined effects of decreased renal perfusion and neuro-hormonal activation cause the kidneys to release renin, which stimulates angiotensin II production and causes vasoconstriction to increase blood pressure. In addition, it increases renal reabsorption but this effect may not be sufficient to increase systemic blood pressure in the acute setting, but may increase renal oxygen consumption and cause further hypoxia [23, 24]. The return of renal blood flow during reperfusion intervals was lower than expected. It is likely that the decrease in proximal blood pressure after balloon deflation was caused by a sharp return of blood flow in the abdominal aorta and a significant reduction in cardiac afterload. The renal flow only restored to 31% of baseline. Hence, the renal blood flow could not be predicted by the systemic blood pressure. The mechanism behind the decreased return of renal artery blood flow is likely multifactorial and may be related to increased circulating volume area secondary to redistribution of blood flow in vascular beds in the abdomen and lower extremities. Also, hyperemia after ischemia may contribute to the decreased RBF [25]. The restoration of renal blood flow after the 60-min reperfusion phase increased in iREBOA but only reached 61% of baseline. The mechanism may be related to renal vascular constriction due to endothelial hypoxia and possibly also to central regulation due to baroreceptor signaling from the sharp decrease in blood pressure following balloon deflation, which should be the focus of future investigations. The regulation of renal blood flow is complex and previous studies have reported a full return of renal blood flow at 45 min after zone 1 cREBOA [18, 26]. Nevertheless, reducing the progressing ischemic injury to distal organs during REBOA is likely important for organ function after definite surgical care. The increased urine output after iREBOA may reflect the degree of active nephrons and thus global renal function. It is possible that the decreased ischemic time improved the renal function and increased urine output already in the immediate intensive care phase. We detected decreased renal vein pCO₂ after reperfusion in iREBOA which suggests that the renal ischemia was lower.

We detected hemodynamic effects in both iREBOA and cREBOA similarly to previously reported. The proximal blood pressure increased from 75 to 124 mm Hg after initiation of supra-celiac REBOA in humans [27]. We also detected an increase in Troponin I in cREBOA after 60 min. Cardiac myocyte ischemia may occur after 60 min cREBOA [11]. It is possible that iREBOA reduced injury in the myocardium. An immediate decrease in heart rates occurred after balloon deflation, which was possibly a consequence of normalized intra-thoracic cardiovascular anatomy. Correspondingly, beta-blockers reduce myocardial injury in REBOA [28]. The increased need for vasopressor support on release of cREBOA may also be connected to a cardiac ischemic reperfusion injury and a decrease in central blood volume. We detected a similar effect when comparing zone 3 REBOA and an Abdominal Aortic and Junctional Tourniquet [29].

To further assess renal physiology, we analyzed renal venous blood gases in comparison with systemic arterial blood. A metabolic acidosis occurred in both cREBOA and iREBOA, showing a decrease in pH and base excess, and an increase in lactate, both systemically and in renal venous blood. There was a trend of increased metabolic acidosis in the cREBOA, although not statistically significant. Intracellular adaptation to hypoxia and ischemic preconditioning may result in decreased reperfusion injury after iREBOA which should be investigated in future studies [30]. We then analyzed immunohistochemical markers of kidney injury [31]. Osteopontin increased in the kidney cortex, and vimentin and NGAL were not affected. Osteopontin and NGAL are elevated in ischemic injury, and vimentin is elevated in AKI [32]. We specifically analyzed levels of injury markers leaked into the nephrons, by co-staining with NKCC2. Sodium–potassium-chloride cotransporters, NKCC2-pumps are located the thick ascending limb of loop of Henle [33]. It is therefore possible that the decreased urine output in cREBOA was a consequence of both decreased renal blood flow and local ischemia in the nephrons.

There are some limitations to be discussed. First, the hemorrhage protocol allowed for a standardized induction of hemorrhagic shock. However, no definitive surgical control was investigated, and no restoration of hind leg blood flow was assessed. Second, hemostasis from the transected femoral artery during reperfusion would likely not occur in humans. Extravasation through collateral pathways below the level of aortic occlusion has been described in CT scans [34], and swine without vascular pathology likely have a better vasoconstriction, in combination with a hypercoagulability, compared to humans [35]. Third, a longer follow-up during reperfusion may further characterize the return of renal blood flow and allow for a more comprehensive grading of AKI. Fourth, transferring results from animal models to humans should always be done with caution [36].

REBOA may improve mortality rates in comparison with resuscitative thoracotomy, in trauma and severe hemorrhage [37, 38]. Further improvement of resuscitation strategies and methods of REBOA to preserve renal function may improve long-term outcomes, and iREBOA shows potential as one of these methods.

iREBOA was survivable, did not cause rebleeding, decreased the total ischemic time and increased the renal blood flow, urine output and decreased renal ischemic injury compared to cREBOA. Intermittent reperfusions during REBOA may be preferred to continuous, complete occlusion in prolonged application to improve renal function.