Hemorrhagic shock and gut ischemia are associated with high mortality. In rat models of shock, 2 h of hypovolemia or splanchnic arterial occlusion (SAO) followed by return of shed blood or release of the occlusion usually causes a decline in blood pressure and death within a few hours after the onset of reperfusion (1, 2). However, the mechanisms of blood pressure reduction that lead to death are currently unknown.
We have obtained evidence in models of shock, including 2 h SAO followed by reperfusion, hemorrhagic shock, and endotoxic shock, that central blood pressure reduction is associated with translocation of digestive enzymes from the lumen into the wall of the intestine and with formation of inflammatory mediators (1, 3-5). The enzymes and inflammatory mediators are transported via the portal venous circulation, the lymphatics, and the peritoneum into the central circulation and cause remote organ injury beyond the intestine (4, 6, 7). Inhibition of the pancreatic digestive enzymes by different digestive enzyme inhibitors results in significantly improved blood pressure and reduced inflammation (1, 3-5).
However, in addition to death from a gradual reduction of blood pressure after gut ischemia and reperfusion due to hypovolemia or SAO, we observed that death may also occur during ischemia following a precipitous drop in blood pressure. We designate such pressure drops as fast fatal pressure drops (FFPDs). Although the gradual pressure drop and death that occur after reperfusion have been linked to pancreatic digestive enzymes, there may be a mechanism involved in the FFPDs that is not directly linked to digestive enzymes.
The autonomic nervous system controls blood pressure by adjusting vascular tone. The sympathetic subsystem has the ability to facilitate moment-to-moment blood pressure adjustments, such as quickly raising pressure during rapid transition from supine to standing position. The parasympathetic system, in contrast, can lower systemic pressure and also buffers rapid fluctuations in blood pressure triggered by changes in the sympathetic nervous system (8). Gut ischemia is known to activate the sympathetic nervous system via spinal nerve afferents resulting in increased blood pressure (9).
Furthermore, the autonomic nervous system detects glucose absorption in intestinal epithelium and the presence of Glucose in the portal vein, resulting in increased parasympathetic nervous system activity (10, 11). In pilot studies of SAO, Glucose added into the lumen of the intestine increased incidence of mortality by FFPD before reperfusion. Consequently, we hypothesize that the mechanism for FFPDs in SAO may involve the autonomic nervous system and is similar to that of "neurogenic shock," a form of hypotension that results from loss of sympathetic control of vascular tone with a relative bradycardia and uncompensated parasympathetic activity, as in spinal injury (12).
Given that reperfusion is not always immediately possible, e.g., on the battlefield, in remote locations, in surgical conditions when occlusion is necessary (e.g., during aortic reconstructions), or when reperfusion would lead to further bleeding, we wished to determine the cause of rapid death during ischemia. The aims of this study were therefore to determine in a SAO model of intestinal ischemia the incidence of death by an FFPD and to determine whether the autonomic nervous system or pancreatic digestive enzymes are involved in the blood pressure drops. To accomplish this, we analyzed in detail blood pressure histories during gut ischemia in conjunction with a variety of treatments including Glucose in the intestinal lumen and/or intramuscularly administered xylazine to stimulate the parasympathetic nervous system; total subdiaphragmatic vagotomy (TSV) or addition of glycopyrrolate, an anticholinergic, to interfere with the parasympathetic nervous system; and 6-amidino-2-naphthyl p-guanidinobenzoate dimethanesulfonate (ANGD, nafamostat mesilate; Torii Pharmaceutical, Chiba, Japan) to block pancreatic digestive enzymes in the lumen of the intestine.
MATERIALS AND METHODS
To identify the treatments and conditions tested, animal group designations were selected to be modular in nature. For example, animals in the Xyl Gly Glucose group received xylazine, glycopyrrolate, and Glucose before undergoing SAO.
Animals and surgical procedures
All animal protocols were reviewed and approved by the University of California San Diego Animal Subjects committee, and the experiments were performed in adherence to the National Institutes of Health Guidelines on the Use of Laboratory Animals. Male Wistar rats, after a minimum of 5 days' nonstressful existence in our facility, were given general anesthesia (sodium pentobarbital, 50 mg/kg, i.m.) and kept warm on a heated surgical table. Rats in the Xyl groups were also tranquilized with xylazine (4 mg/kg, i.m.) 20 min before general anesthesia. One group of Xyl animals also received glycopyrrolate (0.5 mg/kg, i.m., Xyl Gly Glucose group) immediately after xylazine injection. Because reported cardiovascular effects of xylazine include initial hypertension followed by sustained hypotension and increased vagal tone-related bradycardia (13), we also included groups that did not receive xylazine (Saline and Glucose groups) to determine the effects of ischemia without parasympathetic nervous system stimulation by xylazine.
All animals were cannulated via the left femoral vein for supplemental anesthetics as needed by reflex test and via the left femoral artery for fluid collection and pressure measurement. Arterial blood was collected in 10 U/mL (final) of sodium heparin and centrifuged (1,000 g, 10 min) to obtain plasma, which was frozen (−20°C) for later measurement of plasma Glucose levels and protease activity.
Animals in the TSV (Xyl TSV Glucose) group also received after laparotomy a left subcostal incision (after ligation of the superior epigastric vessels to prevent bleeding). In these groups, we ligated the esophagus in two spots approximately 3 mm apart below the diaphragm and performed an esophagotomy between the ligatures, thus cutting both the anterior and posterior trunks of the vagus nerve positioned alongside the esophagus.
In all groups, we exteriorized the small intestine and cecum and sandwiched them between two pieces of gauze soaked in warm Saline (37°C), placed alongside of the animal, and covered in plastic wrap to maintain body temperature and moisture. The superior mesenteric and celiac arteries, which serve to perfuse the small intestine and in part the liver, pancreas, spleen, and stomach, were isolated with umbilical tape. Six to 10 mL (volumes were adjusted to provide fluid along the entire length of the small intestine with minimal stretching of the intestinal wall) of Saline (37°C) (Saline groups), 100 mg/mL Glucose in Saline (Glucose groups), or the protease inhibitor ANGD (2 mg/mL in 100 mg/mL Glucose in Saline; Glucose ANGD group) was injected into the intestinal lumen at two or three sites along the small intestine. Luminal injections were completed in less than 5 min. Animals in the Xyl Gly Glucose group received additional glycopyrrolate (0.5 mg/kg, i.v.) at this time.
Within seconds of completing the luminal injections, SAO was initiated (with the exception of the Xyl Gly Glucose group, which was allowed to adjust to the i.v. glycopyrrolate for 15 min before occlusion). Occlusion was achieved by tying the lengths of umbilical tape isolating the superior mesenteric and celiac arteries around short pieces of polyethylene tubing placed adjacent to the arteries to occlude with minimal damage to the blood vessels. Because of its close proximity to the superior mesenteric artery, the main lymphatic vessel exiting the intestine was also isolated and occluded together with the superior mesenteric artery. Pilot studies (n = 3) showed that the weight of the small intestine plus its luminal content does not change from fluid loss or gain during a 2-h ischemia, even if followed by reperfusion for 2 h, excluding the possibility that the central blood pressure reduction was due to fluid loss into the lumen of the intestine. No artery/arteriolar or venous/venular bleeders were detected in any of the animals in this study. An additional group (n = 3 animals) received xylazine and Glucose treatments but no occlusion of the arteries (Xyl Glucose sham group).
The wet gauze and plastic wrap were adjusted to cover the intestine and the abdominal opening, leaving the small intestine outside the body for the remainder of the experiment to allow direct observation of hemorrhage within the intestine and prevent possible leakage of mediators from the wall of the intestine into the peritoneal space. Each animal was then left undisturbed throughout the experiment. Death was declared when mean arterial blood pressure (MABP) reached less than 20 mmHg with pulse pressure less than 1 mmHg.
After 4-h ischemia or at death, whichever came first, an aliquot of blood was collected into sodium heparin (10 U/mL, final) via the arterial catheter if blood pressure was greater than 40 mmHg (enough pressure to allow blood collection through the catheter) or via atrial cardiac puncture if the pressure was less than 40 mmHg or if the animal was deceased. At 4 h, surviving animals were killed (120 mg/kg sodium pentobarbital, i.v.).
Whole-blood samples were centrifuged (1,000g, 10 min), and the plasma stored (−20°C). High levels of hemolysis were observed in the blood obtained from cardiac puncture from two animals, and these were excluded from protease analysis because of interference with the zymographic techniques. Organs were inspected for macroscopic signs of edema, tissue hemorrhage, and other lesions.
Plasma Glucose levels were obtained with a commercial Glucose meter (Contour Meter; Bayer Healthcare, Tarrytown, NY). Plasma protease activity levels were determined with a fluorescently labeled casein substrate (E6639 Enzchek Protease Assay Kit; Invitrogen, Carlsbad, Calif) after 90-min reaction at 37°C in a 96-well plate fluorescence reader (Spectramax Gemini XS; Molecular Devices, Sunnyvale, Calif).
Continuous arterial blood pressure histories were digitally recorded every 0.001 s (Maclab; ADInstruments, Colorado Springs, Colo), from which heart rate and systolic, diastolic, and mean blood pressures for each pulse were computed and stored for statistical analysis (Microsoft Excel).
Arterial blood pressure analysis
To examine early differences between groups, we noted hemodynamic parameters immediately after femoral artery catheterization, immediately before arterial occlusion, at peak initial MABP (usually within the first 5 min of occlusion), and at 20, 40, and 60 min during occlusion.
To analyze the full pressure histories, we fit the MABP trace with a series of linear segments. This approach provides a close approximation to the measurements because most MABP traces are piecewise linear and had clearly identifiable transitions between segments with different slopes (Fig. 1). In some animals, the blood pressure rapidly fluctuated around an average trend for varying lengths of time (e.g., Fig. 1, bottom). In those cases, we placed the line segment along that average trend.
To compare pressure histories between experimental groups, we grouped line segments into different time/pressure periods based on common characteristics among the pressure histories (Fig. 1). For example, immediately after occlusion, blood pressure rises sharply (for less than a minute), followed usually by a slower rise (for several minutes). We designated this, collectively, the initial rise period. After a few minutes of relatively stable MABP, blood pressure begins to drop in the initial drop period. In animals treated with xylazine, the pressure on average drops more sharply at first then levels off as it approaches the pressure before occlusion (Fig. 1, upper and middle left). The end point of the initial drop was determined as the first time following its onset at which the trend of the MABP maintained a stable or positive slope (i.e., roughly greater than −0.5 mmHg/min for ≥4 min).
Following the initial drop period, the pressure histories vary from animal to animal as MABP goes through a series of stable pressures, gradual pressure drops, gradual pressure recoveries, and occasional sharp drops followed usually by a rapid pressure recovery of similar magnitude. Despite frequent pressure drops and recoveries in many animals, the overall pressure change in this period is usually small.
The fatal pressure drop represents the final blood pressure reduction bringing the MABP to less than 40 mmHg, a level at which death follows soon afterward. All but two of the animals exhibited a fatal pressure drop. The majority of fatal pressure drops fit into either of two classes: a fast drop (Fig. 1, upper left and middle right) or a slow drop (Fig. 1, upper right and middle left) as defined by the length of time between the start of the fatal pressure drop and death (Table 1). The transition into the start of an FFPD was usually very clear and could be pinpointed to within a minute (and sometimes within a second). The transition into a slow fatal pressure drop (SFPD) was not always fully evident, as sometimes there was more than one transition from line segment to line segment that could be viewed as the onset. In those cases, we chose the transition involving the larger change in slope. These potential onset times were usually within 10 min of each other.
Although the last few line segments following the fatal pressure drop vary from animal to animal, with some accelerating toward no detectable blood pressure and others leveling off for a while or even briefly recovering, a MABP of less than 40 mmHg was rapidly followed by death regardless of the type of fatal pressure drop.
When more than one line segment in a pressure history was part of a given period, we calculated the period's total duration and pressure change as the sum of the segment durations and pressure changes. For example, in Figure 1 top left, while the initial drop period is shown as a single line that is not superimposed over the blood pressure trace, it represents the sum of three sequential line segments that did superimpose over the trace. The period's overall slope was calculated as the total pressure change divided by the total duration.
Results are presented as mean (SD). The average blood pressure histories were plotted using the starting times, starting pressures, durations, and pressure changes of the initial rise, initial drop, and fatal pressure drop periods as well as the time of death. However, all statistical analyses were performed comparing the exact start and end times, start and end blood pressures, durations, slopes, and pressure change of each period.
For statistical comparisons between two groups (e.g., FFPD animals vs. SFPD animals), a Student t test was used, with P < 0.05 considered significant. To compare mean values between multiple groups, we first performed ANOVA over all seven SAO groups, then used a Bonferroni corrected Student t test for pairwise comparisons. As each group was compared with either one or two others (e.g., Xyl Saline ↔ Xyl Glucose ↔ NoXyl Saline ↔ NoXyl Glucose ↔ NoXyl Glucose ANGD), P < 0.025 was considered significant.
Within the 4-h observation period, 13 of the total of 77 animals subjected to SAO clearly started a fatal pressure drop but did not complete it (i.e., they were still alive at 4 h) (Figs. 2, 3, and 9). For these animals, we estimated the time of death by extrapolating the pressure drop at the final recorded rate to a value less than 20 mmHg MABP, because we observed animals always died within a few minutes of reaching 20 mmHg.
Two animals had no fatal pressure drops within the 4-h observation period and were therefore excluded from fatal-drop and time-of-death comparisons (Figs. 3 and 9). These animals did not differ within their groups by weight, time spent in vivarium, or details in the surgical technique.
Significant changes in FFPD incidence were determined by chi-square analysis (P < 0.025 considered significant as this was also a comparison between multiple groups).
Effects of xylazine and Glucose on hemodynamics and blood Glucose levels
Both xylazine and Glucose revealed effects suggesting parasympathetic nervous system activation. Animals receiving xylazine had a significantly lower MABP (79.1 [SD, 6.2] vs. 94.3 [SD, 11.3] mmHg), heart rate (246.2 [SD, 37.3] vs. 333.1 [SD, 35.9] beats per minute [bpm]), and pulse pressure (40.5 [SD, 3.4] vs. 43.9 [SD, 6.1] mmHg) before laparotomy compared with animals that did not receive xylazine. With the exception of a transient pulse pressure increase immediately after ischemia in the xylazine-treated animals, the pulse pressures and heart rates continued unchanged for at least the first hour of ischemia (this point is illustrated in Fig. 4, although it splits the animals into FFPD and SFPD animals).
In contrast to xylazine, Glucose in the intestinal lumen had little effect on preischemic MABP values (time = 0 in Figs. 2 and 3), possibly because it had less time to act. However, it did decrease preischemic pulse pressure (48.1 [SD, 8.3] mmHg for Saline vs. 40.3 [SD, 6.5] mmHg for Glucose, P < 0.015; 39.0 [SD, 7.2] mmHg for Xyl Saline vs. 34.0 [SD, 3.9] mmHg for Xyl Glucose, not statistically significant, P = 0.066). It also tended to decrease heart rate (in the absence of the stronger effect of xylazine; 359 [SD, 48] bpm for Saline vs. 321 [SD, 52] bpm for Glucose, not statistically significant, P = 0.075).
In some animals, we measured blood Glucose levels before ischemia and at time of death. Xylazine had a powerful effect on blood Glucose levels (preischemic values: 401 [SD, 53] mg/dL for Xyl animals] vs. 122 [SD, 15] mg/dL for animals without xylazine; n = 17 and 13 respectively, P <1.4 × 10−14) in agreement with reports in the literature that xylazine, as a parasympathetic α2-adrenergic receptor agonist, inhibits pancreatic insulin release causing hyperglycemia (14). Interestingly, the effect of xylazine appears to be greater than having Glucose or Saline in the lumen of the intestine as inferred from time-of-death blood Glucose values (55 [SD, 33] mg/dL for Saline and 115 [SD, 60] mg/dL for Glucose animals and 135 [SD, 113] mg/dL for Xyl Saline and 277 [SD, 163] mg/dL for Xyl Glucose animals). It is one reason we included animals without xylazine in this study.
Hemodynamic effects of ischemia
As expected, given the ability of gut ischemia to activate the sympathetic nervous system (9), all groups responded to ischemia with an immediate and pronounced increase in blood pressure (Figs. 2, 3, 9, 10). The four nontreatment groups increased their blood pressure to approximately the same pressure upon arterial occlusion (Figs. 2 and 3).
Parasympathetic "stimulation" lowers blood pressures and may increase incidence of FFPDs
Animals in the Xyl Glucose group had a significantly larger drop in MABP during their initial drop period than animals in the Saline group, and their MABP stayed significantly lower at many points throughout their pressure histories before their fatal pressure drops (Xyl Glucose shown in Fig. 2 vs. Saline in Fig. 3). Also, we saw that Glucose animals with SFPD had a lower MABP than Saline animals with SFPD at the time their fatal pressure drop started (Fig. 3, bottom). MABP was also lower in the Glucose vs. Saline group at the end of the initial rise and the beginning and end of the initial drop, but not significantly (P = 0.08 for all three points). This depression by glucose was not seen in the presence of xylazine, which presumably masks the effect with its more powerful depressing effect on MABP. Although neither xylazine nor luminal Glucose is a simple "stimulator" of the parasympathetic system, the depression of MABP seen here supports the reports that Glucose and xylazine can activate the parasympathetic nervous system (10, 11, 13, 15).
In preliminary studies with xylazine and 2 h SAO, we observed rapid pressure drops and subsequent death within minutes before the end of the 2-h observation period in animals with Glucose added to the intestinal lumen, but not with Saline alone. This observation led to the hypothesis that an autonomic mechanism may be active in SAO and the decision to compare animals with Glucose or Saline in the intestinal lumen. Similarly, of the 10 animals in the XylSaline group, only two animals had an FFPD as compared with 7 of 14 FFPDs in the Xyl Glucose group (not significant, P = 0.065). This difference is significant (P < 0.02) if we include only the first 2 h of ischemia (as in the pilot study), because only one of the two Xyl Saline FFPDs occurred within the 2-h time point as compared with all seven of the Xyl Glucose FFPDs. Note that the fatal pressure drops of Xyl Glucose animals, in general, occurred sooner than that of Xyl Saline animals (90 [SD, 40] min for Xyl Glucose vs. 147 [SD, 43] min for Xyl Saline, P < 0.004, when pooling FFPD and SFPD animals).
In the absence of xylazine, Glucose in the lumen of the intestine had no apparent effect on FFPD incidence with 30% of the animals in the Glucose group and 33.3% of animals in the Saline group experiencing an FFPD within 4 h of SAO (Table 2; with the current number of animals, the Saline group was also not significantly different from the Xyl Glucose group).
Given the powerful hyperglycemia caused by xylazine and the fact that FFPDs occurred both in its presence and absence, we saw no correlation between blood Glucose levels before ischemia and the incidence of an FFPD (correlation coefficient of −0.097).
To determine if intestinal distension is necessary for FFPDs to occur, we also subjected an animal to SAO without xylazine and without any injection directly into the intestinal lumen. The animal had an FFPD, indicating that FFPDs are possible even without fluid in the intestinal lumen.
Gross morphological organ damage
Animals subjected to SAO, but not shams, had damaged small intestines with extensive hemorrhagic lesions (Fig. 5). The extent of hemorrhage increased with prolonged survival time, i.e., animals that died rapidly had minimal damage to their intestines, whereas animals that survived for the duration or died near the end of the 4 h showed extensive damage. This damage, visible to the unaided eye, was reduced with Glucose in the lumen. In most SAO, but not sham animals, the pancreas developed hemorrhagic lesions, and the liver had large pale areas reflective of a lack of red blood cell flow out of the intestine through the portal vein.
Sixteen of the 25 animals with an FFPD, but only 1 of the 41 animals with an SFPD, exhibited a postmortem gastric reflux (as assessed by the presence of a pool of yellow liquid forming by the head as fluid exited the mouth and nostrils; typically ∼0.5 mL) within a few minutes of death (P < 2.0 × 10−8 by chi-square test). Gastric reflux was the only event we saw by unaided inspection tied almost exclusively to FFPDs. The single SFPD animal with a postmortem gastric reflux was in the Glucose ANGD group and was one of the two animals that survived the 4 h with no sign of a fatal pressure drop. In that case, reflux occurred after euthanasia.
Animals with FFPDs have earlier fatal pressure drops and death than animals with SFPDs
Of 77 rats subjected to SAO, all but two animals had a fatal pressure drop start within 4 h of the onset of ischemia. By pooling their onset times, we saw that animals undergoing an FFPD initiated a fatal pressure drop significantly sooner than those with an SFPD (86 [SD, 47] min, n = 25, for FFPD vs. 140 [SD, 40] min, n = 49, for SFPD; P < 1.5 × 10−5). Animals with an FFPD also had significantly earlier death than animals with an SFPD (91 [SD, 48] min vs. 241 [SD, 45] min; P < 3.7 × 10−17). Excluding the Xyl Saline group, which did not have a sufficient number of FFPDs (n = 2) to test the issue, this earlier death with FFPD remained statistically supported within the groups as well.
The onset of the FFPDs often occurred within a single heartbeat (Fig. 6). Mean arterial blood pressures were already significantly lower within 10 s after the onset of the FFPD and decreased further over the next minute (Fig. 7, top left). Pulse pressure, in contrast, did not drop until the next minute (P < 0.05) when the animal was near death; even then, the pulse pressure drop was proportionately less than the MABP drop (Fig. 7, middle left). Heart rate was reduced in the first 10 s (P < 0.05) but over the next minute did not drop further (Fig. 7, bottom left), and in fact, the heart often continued to beat after no pulse was detectable in the arterial catheter (not the case with SFPD animals). The relatively small loss of heart rate and pulse pressure compared with MABP suggested the MABP reduction was due more to reduction of venous return than to loss of cardiac function.
We examined whether the pressure history may serve to identify animals at risk for an FFPD. The FFPDs were not predicted using MABP, pulse pressure, or heart rate before laparotomy, at time zero (just before ischemia), at the time of the peak MABP, or at 20, 40, and 60 min after ischemia (pooling Xyl and NoXyl animals separately) (Fig. 4). Neither did pressures in the initial rise and initial drop periods differ in FFPD animals versus SFPD animals (Figs. 2 and 3, top vs. bottom rows).
We occasionally saw sharp drops and recoveries before the fatal drop, in some cases of a magnitude similar to that of the fatal pressure drops (Fig. 1). We wished to determine if their incidence, magnitude, or other characteristics of the time between the end of the initial drop and the start of the fatal drop could predict an FFPD. Therefore, every line segment in this time period was classified, based on its slope, as a stable segment, fast drop, fast recovery, slow drop, or slow recovery (see Supplemental Table 1, Supplemental Digital Content 1, at https://links.lww.com/SHK/A78, for cut-off values). For each animal group demonstrating both FFPDs and SFPDs (see Supplemental Table 1, Supplemental Digital Content 1, at https://links.lww.com/SHK/A78), or pooled FFPD and SFPD animals (see Supplemental Table 2, Supplemental Digital Content 1, at https://links.lww.com/SHK/A78), we determined within this period: (i) the number of line segments falling into each category, (ii) the percentage of time spent in each category, and (iii) the segments' average durations, and (iv) pressure changes, to determine, among other things, whether the fast drops of animals that went on to have an FFPD were more frequent or of greater magnitude than those with an SFPD.
Comparing animal groups (see Supplemental Table 1, Supplemental Digital Content 1, at https://links.lww.com/SHK/A78), we saw that animals without xylazine had shorter stable line segments than Xyl animals (ANGD reversed this). In addition, all pressure drops and recoveries were more frequent, and the slopes of the fast drops and fast recoveries were greater in magnitude, in the Saline group than in the Xyl Glucose group (reversed somewhat in the Glucose group and more so in the ANGD group). These were indicative of numerous high-frequency pressure fluctuations in the Saline group and is likely a sign of sympathetic stimulation not buffered by parasympathetic activity (8).
When animals were grouped according to the presence of FFPD or SFPD (see Supplemental Table 2, Supplemental Digital Content 1, at https://links.lww.com/SHK/A78), we saw that (i) animals that went on to an FFPD had fewer fast pressure drops and recoveries than SFPD animals before the fatal pressure drop. Moreover, the slopes of fast drops and recoveries were lesser in magnitude in FFPD animals than SFPD animals, although the FFPD animals had greater pressure changes. (ii) In the absence of xylazine, even slowly dropping/recovering segments were more frequent in the SFPD animals than in the FFPD animals. (iii) The SFPD animals have more stable segments than FFPD animals.
In summary, pressure fluctuated more in animals that went on to have an SFPD. Based on this particular analysis in the rat, the pressure history is not suitable for predicting individuals at risk for an FFPD in SAO.
We could only determine the starting point of SFPDs to within about a minute (as opposed to a second in the FFPD groups). When comparing 5 min before vs. 5 min after the approximate beginning of the SFPD, we saw a significant drop in MABP and a significant increase in heart rate (Fig. 6, right). Pulse pressure did not change in this time.
The time of death from SFPD did not vary significantly between groups regardless of the presence of xylazine or Glucose.
ANGD prevented changes in plasma protease activity and reduced intestinal damage but did not prevent FFPDs
When we examined plasma protease activity in our four nontreatment SAO groups, we saw that the change in protease activity in animals with an SFPD was significantly greater at the time of collection (death or 4-h ischemia) than in FFPD animals, which exhibited no significant change in plasma protease activity (−0.09 [SD, 1.67] relative fluorescent units [RFUs] for FFPD animals [n = 11] vs. 4.83 [SD, 2.85] RFUs for SFPD animals [n = 16]; P < 7.7 × 10−6). This increase was likely due to the increased duration of ischemia, as time of collection correlated well with change in protease activity (Fig. 8; correlation coefficient = 0.76; FFPD time of collection = 86 [SD, 47] min; SFPD time of collection = 226 [SD, 22] min).
ANGD treatment almost completely prevented visible signs of damage to the intestine, even after 4 h, suggesting successful inhibition of luminal digestive enzymes. However, ANGD had no significant effect on the change in plasma protease activity other than a small increase in FFPD animals (P < 0.027) that may also be due to a later time of collection (1.71 [SD, 1.40] RFUs and 105 [SD, 65] min [n = 7] for ANGD FFPD animals, and 2.60 [SD, 2.28] RFUs and 240 [SD, 0] min [n = 4] for ANGD SFPD animals).
ANGD treatment, despite the absence of xylazine, increased the rate of death by FFPD to 67% (not significant, P = 0.04, Table 2), although it had no effect on the hemodynamics of the FFPD animals (Fig. 9, top). However, ANGD significantly delayed the onset of SFPDs (Fig. 9, bottom).
Parasympathetic blockade prevents FFPDs
To determine whether interference with the parasympathetic nervous system may affect the incidence of FFPDs, we performed a TSV, or alternatively added glycopyrrolate, to SAO animals treated with xylazine and Glucose to increase their incidence in FFPDs. Both vagotomy (Xyl TSV Glucose) and treatment with glycopyrrolate (Xyl Gly Glucose) prevented the FFPDs (P < 0.008; Table 2).
Total subdiaphragmatic vagotomy had no effect on pressure history other than to prevent all FFPDs, as seen by the almost complete overlap of the Xyl TSV Glucose pressure history with that of the Xyl Glucose SFPD subset pressure history (Fig. 10). Glycopyrrolate, in contrast, increased MABP even before ischemia (80 [SD, 8] vs. 67 [SD, 15] mmHg in the pooled Xyl Glucose FFPD and SFPD animals, P < 0.019), yielded a higher pressure than any other group at the end of the initial rise, and delayed the onset of the large initial drop caused by xylazine, suggesting that it inhibited the parasympathetic nervous system (Fig. 11). The difference in early hemodynamic effects between the two treatments may be because glycopyrrolate may have more effects on the peripheral vasculature than the subdiaphragmatic vagotomy. However, like the vagotomy group, the later events in the Xyl Gly Glucose group mirrored that of the Xyl Glucose SFPD animals, suggesting that the most prominent effect of the glycopyrrolate at later time points was to prevent FFPDs. Glycopyrrolate had no apparent effect on gross intestinal morphology.
The current results in an acute model of severe gut ischemia by SAO show that systemic blood pressure first increases in response to ischemia then decreases past its initial value, ending in death. The fall in blood pressure may either be very rapid (occurring over seconds) or slow (over hours). Neither blood pressure/heart rate histories nor the prior existence of sharp drops and recoveries predicted rapid fatal pressure drops. Intramuscularly administered xylazine and Glucose in the lumen of the intestine, activators of the parasympathetic nervous system, increase the incidence of FFPDs, whereas TSV or addition of glycopyrrolate, a parasympathetic inhibitor, prevents them. Digestive enzyme inhibition prevents neither FFPDs nor SFPDs, although it delays the slow ones.
Evidence contradicting a hypovolemic mechanism
We were concerned that some of the hypotension we observed in FFPDs or SFPDs could be attributed to hypovolemia from loss of fluid through the intestinal wall. However, pilot studies showed no change of weight in the intestine for hours after ligature of the superior mesenteric and celiac arteries. Whereas Glucose in the intestinal lumen would increase its osmolarity and thus its potential for pulling fluid from the vasculature, Glucose did not increase the incidence of FFPDs in the absence of xylazine; nor did Glucose change survival time for SFPD animals. These results suggest the osmotic effects of Glucose did not play a major role for FFPDs or SFPDs, and thus it is less likely that hypovolemia contributed to the hypotension.
Evidence supporting a neurogenic mechanism
There were a number of characteristics associated with FFPDs to suggest a neurogenic shock mechanism for pressure loss as opposed to a mechanism involving plasma mediators. First, we reduced mediator communication from the intestine by total obstruction of the two major arteries feeding the organ, occlusion of the major lymphatic vessel exiting the intestine, and retention of the intestine outside the body to prevent mediator release into the peritoneal space. Second, in hemorrhagic shock, which is affected by mediators from the intestine (16), hypotension is typically accompanied by tachycardia (17). Instead, we observed either a transient bradycardia or normocardia accompanying the FFPDs, a trait of neurogenic shock (18). Third, pulse pressure was preserved or dropped only to a small degree after an FFPD until the animals were near death. The heart muscle visibly continued to contract after FFPD but not SFPD, even at zero femoral arterial blood pressure and pulse pressure, which, combined with normocardia, suggests that pressure loss was due less to cardiac failure and more to a failure of peripheral resistance and venous return. Fourth, we observed, often within a single heartbeat, an immediate transition in the blood pressure slope at the start of fast fatal drops. This short time interval suggests a neurogenic event as opposed to an effect from an evolving circulating inflammatory mediator. Fifth, FFPDs were not prevented by protease inhibition in the intestinal lumen, which has been shown to improve survival and reduce the presence of shock mediators in the plasma of SAO with reperfusion animals (1, 3, 4). Sixth, gut ischemia is known to activate the sympathetic nervous system (9), and we observed a powerful and near-immediate increase in MABP upon SAO, suggesting that this is also present in our model. Seventh, activators of the parasympathetic nervous system tended to increase the incidence of FFPDs, whereas the parasympathetic nervous system disruption prevented them. Finally, we observed after most of the FFPDs, and almost never after SFPDs, a postmortem gastric reflux. Because loss of sympathetic stimulation results in increased gastric tone (19) and the lower esophageal sphincter is relaxed by parasympathetic stimulation (20), gastric reflux is also in line with an autonomic nervous system failure.
Although operating by a different mechanism than FFPDs, SFPDs may also have a neurogenic component. Slow fatal pressure drop animals (i) had restricted mediator release from the intestine, (ii) had sympathetic nervous system activated by SAO, (iii) had normocardia after SFPD onset, (iv) had preserved pulse pressure despite pressure loss, and (v) were not prevented by protease inhibition. However, death from SFPD occurred at approximately the same time regardless of xylazine, Glucose, TSV, or glycopyrrolate, suggesting independence from the parasympathetic nervous system. Thus, the SFPDs resemble traditional neurogenic shock, which only requires dysfunction of the sympathetic nervous system.
Classic neurogenic shock is caused by a failure of the sympathetic nervous system to maintain vascular tone. Vasodilation decreases venous return as blood pools in the periphery leading to decreased cardiac output. Neurogenic shock has mostly been reported to be associated with brain and spinal cord injury (12). Acute spinal cord injury, for example, can cause an explosive sympathetic discharge, followed by failure. However, neurogenic shock has also been associated with hypovolemia (21), emotional shock (22), burns (23), and gastric rupture (24). Recent results in rats show that administration of endotoxins produced by Bacillus anthracis, which produces shock-like cardiovascular alterations, first increases and then leads to a decrease in sympathetic nerve discharge and death at about the same time as our SFPD animals (25). The current results suggest for the first time neurogenic shock originating from severe gut ischemia.
The combined evidence suggests that prolonged, increased sympathetic discharge (whether from ischemia, spinal cord injury, or infection by B. anthracis) is not sustainable and can result in death, as the sympathetic nervous system ceases to function over the following hours when either the brain fails to send out sympathetic efferent signals or the blood vessels fail to respond to such signals. This hypothesis is in line with reports that one cause of the progressive hypotension in hemorrhagic shock is vascular hyporeactivity to norepinephrine, etc., that develops because of hyperpolarization of arterial smooth muscle cells via changes in ion-channel function (26). To determine whether the SFPDs occur by the same mechanism as traditional neurogenic shock, direct measurements of sympathetic and/or parasympathetic nerve activity and other additional experiments will be required.
Current models for neurogenic shock in the literature are related only to spinal cord or brain injuries or direct nerve and/or brain stimulation (PubMed search with keywords "neurogenic shock"). If direct nerve activity measurements confirm a neurogenic shock, gut ischemia may also serve as a useful model for other, more difficult to model, forms of neurogenic shock resulting from sustained sympathetic stimulation (e.g., emotion-triggered neurogenic shock).
Relevance to hemorrhagic shock
SAO followed by reperfusion is a model for hemorrhagic/hypovolemic shock with release of shock mediators from an ischemic gut region into the central circulation, as the gut may be ultimately responsible for death (16). However, if neurogenic shock is a cause of death in gut ischemia, a neurogenic mechanism may also be at least partially responsible for blood pressure loss in clinical cases of hemorrhagic shock or trauma. Supporting this hypothesis are studies showing that hemorrhagic shock activates the sympathetic nervous system (27), and septic shock is accompanied by loss of vasomotor tone and cardiovascular depression (28).
In addition, we observed animals in a 2-h hemorrhagic model of shock spontaneously die before reperfusion after a rapid pressure loss that did not respond to fluid resuscitation (2). This evidence suggests that FFPDs may also occur in hemorrhagic shock, although future studies will be required to determine whether the same mechanism is involved. If it is involved, then prophylactic parasympathetic inhibition, such as with glycopyrrolate, may be indicated in conditions in which resuscitation options are limited (battlefield remote locations, surgical requirements, etc.), leading to prolonged ischemia, although bleeding with increased pressure could remain a problem.
Role of the parasympathetic nervous system in SAO
Both FFPD and SFPD may occur as a result of a large-scale vasodilation. Slow fatal pressure drops may occur because of a gradual failure of the sympathetic nervous system that may be irreversible as long as the trigger or need for an elevated sustained sympathetic nerve discharge remains. Fast fatal pressure drops may occur because of short-term decreases in the sympathetic output, which may become less stable when the sympathetic nervous system is stimulated by ischemia (e.g., the high-frequency fluctuations we observe in the absence of strong parasympathetic buffering by xylazine/Glucose, Supplemental Table 1, Supplemental Digital Content 1, at https://links.lww.com/SHK/A78). Our results with glycopyrrolate and TSV suggest that what separates a fatal fast drop in blood pressure from one that recovers is the degree of parasympathetic activity. Strong parasympathetic nervous system stimulation uncompensated by sympathetic nervous system activity, however brief, can cause blood pressure to drop below recoverable levels. This possibility is supported by the trend toward increasing incidence of FFPDs with the parasympathetic nervous system activators, xylazine and Glucose.
Protease inhibition with ANGD
ANGD in the lumen of the intestine attenuates the enzymatic destruction of the mucosal barrier and entry of digestive enzymes into the wall of the intestine (4). We saw here that it preserved gross intestinal morphology. Interestingly, ANGD in the intestinal lumen did not significantly prevent the increase in plasma protease activity with prolonged intestinal ischemia, suggesting that the proteases in the blood may not all be derived from the intestinal lumen. This makes sense, given that one goal of our surgical procedure was to prevent entry of mediators generated in the intestine into the rest of the circulation.
ANGD delayed the onset of SFPDs and may delay death during SFPDs (although a longer observation period might be required to determine the latter). One possible explanation for this observation is that, by preserving the morphology of the intestine, ANGD may attenuate the enzymatic destruction of nerve fibers such as sympathetic afferents in the wall of the intestine, which could then continue to detect and report gut ischemia, stimulating a longer pressure increase before failure. An intact intestine may also provide continued nutrition and thus maintain sympathetic nervous system function.
ANGD tended to increase the incidence of FFPDs (not significant, P = 0.04) without apparently altering the FFPD animals' pressure histories. This might also be due to protection of nerve fibers or even a more direct action on the autonomic nervous system promoting instability of sympathetic nervous system output or increasing parasympathetic activity over time. This hypothesis is suggested by the fact that ANGD tended to mimic the effects of xylazine in the period between the initial and fatal drops (see Supplemental Table 1, Supplemental Digital Content 1, at https://links.lww.com/SHK/A78) and by pilot studies with SAO and intravenous ANGD, which had depressed MABP in comparison to animals with SAO and no intravenous ANGD (not shown). Interestingly, we recently observed a rat that had both intramuscularly administered xylazine and intraluminal Glucose/ANGD experienced a fast drop in blood pressure even without ischemia, suggesting that a parasympathetic stimulus may cause an FFPD even without increased sympathetic activity.
Implications of FFPDs
Given the relatively short time interval between the onset of ischemia and death, a fast fatal drop in humans may not always be preventable by conventional resuscitation. Moreover, death would not necessarily be attributed to neurogenic shock and/or gut ischemia because the fast form of death in our model leaves little evidence behind for the actual cause of death. Therefore, it may be difficult to conclude with certainty what proportion, if any, of human deaths are due to an FFPD in the absence of continuous blood pressure monitoring. In the case of hospital patients with severe gut ischemia (e.g., due to a blood clot) or cases of prolonged hypotension, or when resuscitation is not possible, the current results suggest that analysis of the incidence of FFPD in humans is needed, and treatment with a parasympathetic inhibitor such as glycopyrrolate or atropine may be tested to minimize the incidence of FFPDs, if they occur.
The current model may also be relevant for analysis of sudden infant death syndrome (SIDS) or sudden cardiac death (SCD). Sudden infant death syndrome remains unexplained after forensic autopsies and detailed death scene investigations (29). Currently, the only significant extrinsic risk factor for SIDS is infant sleep position, which affects vasomotor tone (30). Sudden infant death syndrome has been linked to autonomic dysregulation (31), death can occur in less than 20 min (32), there is intestinal damage (33), and gastric reflux in 30% to 40% of cases (34), all characteristics shared by our model.
In SCD, death is frequently assumed to be the result of an arrhythmia. However, some of the same triggers of arrhythmia, namely, ischemia and sympathetic activation (35), are triggers for the mechanism of sudden death we describe here. Given that most SCDs occur outside the hospital and are resistant to cardiopulmonary resuscitation (35), it is possible that in many cases the cause of death may not have been due as much to cardiac failure as to an autonomic failure leading to vessel dilation and loss of venous return.
Although the effects of Glucose, xylazine, and ANGD are not completely understood, their use in combination with SAO, or other forms of sympathetic stimulation, may prove useful for the study of SIDS, SCD, or other forms of sudden death, by creating a reproducible form of sudden death that occurs with high incidence.