Severe Hypoglycemia–Induced Fatal Cardiac Arrhythmias Are Mediated by the Parasympathetic Nervous System in Rats

A limitation in the treatment of diabetes is hypoglycemia. Although mortality due to severe iatrogenic hypoglycemia is rare, the rate of mortality due to hypoglycemia in people with insulin-treated diabetes is up to 10% (1–3). Our laboratory has demonstrated in a rat model that sudden death induced by severe hypoglycemia is mediated by cardiac arrhythmias (4). Consistent with these animal findings, clinical studies have demonstrated that hypoglycemia leads to alterations on electrocardiograms (ECGs) that can increase the risk of cardiac arrhythmias in people with diabetes (5). The classical response to hypoglycemia is activation of the sympathetic nervous system (SNS). In addition to increasing blood glucose levels, epinephrine secreted from the adrenal medulla also acts to increase heart rate. This sympathoadrenal response is thought to be harmful in that it can cause cardiac arrhythmias and QTc prolongation; thus, blocking SNS activity with β-blockers has been shown to be protective (4,6,7). The extent to which the autonomic nervous system directly leads to fatal arrhythmias during severe hypoglycemia is unknown. Because β-blockers reduce the incidence of lethal cardiac arrhythmias, it is speculated that adrenomedullary release of epinephrine, or perhaps excess release of the sympathetic neurotransmitter norepinephrine to the heart, are principal mediators of hypoglycemia-induced arrhythmias.

Additionally, in response to hypoglycemia, bradycardia has been observed both clinically and in animal models, suggesting that the parasympathetic nervous system (PSNS) is also activated during hypoglycemia (8–10). Activation of the PSNS leads to the release of acetylcholine (ACh) that acts on postganglionic muscarinic and nicotinic receptors in the heart. The involvement of the PSNS in regulating hypoglycemia-induced cardiac rhythms has recently been shown in sleeping individuals with type 1 diabetes (9). Moreover, excessive vagal tone has been shown to trigger cardiac arrhythmias (11). Although diminished parasympathetic tone has traditionally been thought to be protective against cardiovascular events in such cases as myocardial infarction and heart failure, the role of the PSNS in mediating cardiac arrhythmias during insulin-induced severe hypoglycemia is not known.

To understand how severe hypoglycemia can lead to fatal cardiac arrhythmias, multiple protocols were investigated in rats with the overarching hypothesis being that in the setting of severe hypoglycemia, excess activation of the autonomic nervous system induces fatal cardiac arrhythmias. To test the SNS involvement, rats underwent chemical sympathectomy or adrenal demedullation, and to test the PSNS involvement, rats had surgical vagotomy or pharmacological blockade of the muscarinic or nicotinic receptors. Additionally, ex vivo heart perfusions were performed so that determinants of arrhythmias could be ascertained in the absence of cardiac innervation.

Normal, adult, male Sprague-Dawley rats (Charles River, Wilmington, MA) at 7–9 weeks of age (250–300 g) were housed individually in 12:12 h light:dark cycles and received standard rat chow fed ad libitum. All studies were done in accordance with and approved by the Animal Studies Committee at University of Utah School of Medicine and Washington University School of Medicine.

Rats underwent vessel cannulation surgery and ECG placement surgeries ∼1 week prior to the hypoglycemic clamp described below. The carotid artery and jugular veins were isolated and cannulated, and ECG leads were placed as previously described (4).

Hyperinsulinemic-severe hypoglycemic clamps were performed 1 week after recovery from the respective surgeries. After an overnight fast, awake and freely mobile rats were subjected to a hyperinsulinemic (0.2 units/kg/min) severe hypoglycemic (10–15 mg/dL) clamp (4). Insulin (Humulin R; Eli Lilly, Indianapolis, IN) and glucose (50% dextrose; Hospira, Lake Forest, IL) were coinfused intravenously to lower blood glucose to 10–15 mg/dL for 3 h, unless otherwise noted. An ECG (PowerLab 26T; LabChart; ADInstruments, Colorado Springs, CO) was recorded throughout the clamp (4). Respirations and seizure-like activity were quantified. Blood glucose was measured throughout the clamp with a glucometer (Ascensia Contour BG monitors; Bayer Healthcare, Mishawaka, IN). ELISAs were performed for epinephrine and norepinephrine (Abnova, Taipei City, Taiwan) and insulin (Crystal Chem, Downers Grove, IL). Glucagon was measured by radioimmunoassay (Millipore, Billerica, MA). After the clamps, hearts were excised, frozen immediately, and later ground up for catecholamine analysis.

Multiple protocols were carried out to test SNS versus PSNS regulation of severe hypoglycemia–induced cardiac arrhythmias. See Fig. 1 for experimental designs.

Surgeries were modified from Borchard and Vogt (12) and Borkowski and Kelly (13). During vessel cannulation and ECG placement surgery, rats underwent a sham surgery (n = 10) or an adrenal demedullation surgery (n = 11). Incisions were made bilaterally just below the rib cage. The abdominal cavity was exposed and the adrenal glands were located. A small incision was made in the adrenal glands and the medulla squeezed out with forceps. The remaining adrenal cortex was left in place. The abdominal cavity was sutured followed by the outer skin layer. Sham rats did not have an incision made in the adrenal glands. All animals were given 0.9% saline water.

Rats received injections of either 6-hydroxydopamine hydrobromide (6OHDA) (150 mg/kg, n = 13) (Sigma-Aldrich, St. Louis, MO) dissolved in 0.9% NaCl containing 0.1% ascorbic acid (Sigma-Aldrich) or vehicle (control [CON]; n = 10) twice intraperitoneally (IP) over a period of 5–7 days, with the second injection administered 2–4 days after surgery (14,15) (see Fig. 3A). Between 2 and 4 days following the first IP injection, vessel cannulation and ECG placement surgery were performed as described. Between 2 and 4 days following the second and final IP injection, rats were subjected to a clamp as described above, but they were held at 10–15 mg/dL for 3.5 h.

During the vessel cannulation surgery (described above), a left vagotomy (n = 15) was performed. The left vagus nerve was isolated at the level of the left carotid artery in the neck and was completely severed. Sham rats (n = 13) underwent a similar surgery but without transection of the vagus nerve. ECG recordings during the surgery showed no changes in heart rhythm (data not shown). At 1 week after surgery, all rats underwent a severe hypoglycemic clamp as described above.

At 1 week after surgery, rats were divided into three groups just prior to the hypoglycemic clamp: 1) control (n = 25), 2) nicotinic receptor blocker (N blocker, mecamylamine [Sigma-Aldrich], 7.5 mg/kg [16,17]; n = 17), or 3) muscarinic receptor blocker (M2 blocker; AQ-RA 741 [Tocris Biosciences, Minneapolis, MN], 2.8 mg/kg [18,19]; n = 13).

Control rats were injected IP with ketamine/xylazine/heparin, and the heart and aorta were dissected and isolated in 4°C Tyrode’s solution. The aorta was isolated and set up for Langendorff perfusion as previously described (20,21). Hearts were perfused with warm (37°C; pH 7.4) normal glucose for 30 min to equilibrate. Hearts were then perfused for 3 h with either 1) normal glucose (5 mmol), 2) low glucose (0.1 mmol), 3) normal glucose + ACh (700 nmol/L), 4) low glucose + ACh, or 5) low glucose + ACh + mecamylamine (N blocker; 2.5 mg/L). The dose of ACh was chosen from the literature (22) and after a dose response was performed.

Student t test, one-way ANOVA, and repeated measures ANOVA were used for differences between and among the groups (Prism, version 6.07, Graphpad, San Diego, CA). Post hoc analysis was performed with Tukey test. Fisher exact test with Freeman-Halton extension was used for analysis of mortality and incidence of arrhythmias. Statistical significance was at P < 0.05.

The data sets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request. No applicable resources were generated or analyzed during the current study.

To test if circulating epinephrine causes arrhythmias during severe hypoglycemia, rats underwent adrenal demedullation to remove the source of circulating epinephrine (Adx) while sham surgery rats served as controls. After surgery, rats recovered normally, and body weight was similar in both groups. During the hypoglycemic clamp, glucose levels were evenly matched between sham (mean ± SD 11 ± 0.4 mg/dL) and Adx (11 ± 0.3 mg/dL) groups (Fig. 2A). The mean glucose infusion rate required to maintain this glycemic level during severe hypoglycemia was 2.5-fold higher in Adx rats (3.14 ± 0.3 mg/kg/min; P < 0.001) compared with sham rats (1.25 ± 0.1 mg/kg/min), consistent with impaired counterregulation to hypoglycemia (Fig. 2B). During severe hypoglycemia, Adx rats did not have a significant increase in epinephrine (580 ± 301 pg/mL) compared with basal levels (249 ± 42 pg/mL), indicating successful removal of the adrenal medulla, while sham rats had an increase in epinephrine from 505 ± 79 pg/mL in the basal period to a peak of 4,508 ± 291 pg/mL during severe hypoglycemia (Fig. 2C). Plasma norepinephrine was similar between the two groups; however, at 3 h of severe hypoglycemia, norepinephrine was significantly lower in Adx rats (597 ± 132 pg/mL; P < 0.03) compared with sham rats (3,779 ± 1,158 pg/mL) (Fig. 2D). An impairment of plasma norepinephrine levels to rise in Adx rats during severe hypoglycemia suggests that circulating norepinephrine may come from the adrenal medulla and not from spillover from adrenergic nerve terminals.

Surprisingly, adrenal demedullation did not reduce mortality during severe hypoglycemia (sham 20% vs. Adx 18%) (Fig. 2E). Consistent with a lack of an effect on mortality, the incidence of all types of cardiac arrhythmias measured during severe hypoglycemia (premature ventricular contractions and first-, second-, and third-degree heart block) was not different between the two groups (Fig. 2F–H).

Basal heart rate was similar in control (380 ± 6 beats per minute [bpm]) and Adx (396 ± 4 bpm) rats. During severe hypoglycemia, heart rates decreased to 268 ± 7 and 271 ± 10 bpm in control and Adx rats, respectively, with no differences between the groups. QTc was also similar between the two groups at baseline (sham: 135 ± 8 ms; Adx: 148 ± 6 ms) and during severe hypoglycemia (sham: 175 ± 10 ms; Adx: 174 ± 6 ms). Seizure occurrence and respirations were similar between the two groups (not shown).

Since reductions in circulating epinephrine had no effect on severe hypoglycemia–induced mortality or arrhythmias, we sought to test whether sympathetic innervation of the heart contributes to fatal cardiac arrhythmias during severe hypoglycemia. Chemical sympathectomy was achieved with 6OHDA (14) and was considered successful as shown by a 69% reduction in heart epinephrine and a 74% reduction in heart norepinephrine in 6OHDA-treated rats compared with controls (P < 0.05) (Fig. 3A–C). During the severe hypoglycemic clamp, glucose levels were evenly matched (CON: 11 ± 0.2 mg/dL; 6OHDA: 11 ± 0.1 mg/dL) (Fig. 3D). The 6OHDA-treated rats required more glucose infusion (3.95 ± 0.5 mg/kg/min; P < 0.02) during the clamp compared with CON rats (2.24 ± 0.2 mg/kg/min) (Fig. 3E). Plasma epinephrine and norepinephrine were increased to a similar extent during the severe hypoglycemia period compared with the basal period in both CON and 6OHDA groups (Fig. 3F and G). Peak glucagon during severe hypoglycemia was also similar (Fig. 3H).

Taken together, chemical sympathectomy does not alter the counterregulatory response to hypoglycemia. Despite the reduction in heart catecholamine levels, overall mortality was similar between CON (60%) and 6OHDA (46%; P < 0.15) rats (Fig. 3I). Similarly, cardiac arrhythmias during severe hypoglycemia were not different between the two groups (Fig. 3J and K). Heart rate was increased at baseline in 6OHDA (378 ± 5 bpm; P < 0.04) compared with CON (365 ± 3 bpm) rats, and mean heart rate during severe hypoglycemia was significantly higher in 6OHDA (334 ± 7 bpm; P < 0.001) compared with CON (310 ± 7 bpm). Baseline QTc was also increased in 6OHDA (179 ± 5 ms; P < 0.001) compared with CON (129 ± 3 ms) and remained higher throughout most of the clamp. Hypoglycemia increased QTc in both 6OHDA and CON groups compared with their respective baseline levels. However, mean QTc during severe hypoglycemia was significantly increased in 6OHDA (207 ± 4 ms; P < 0.02) compared with CON (194 ± 4 ms). Respirations and seizures were similar between the two groups (data not shown).

To assess the role of the vagus nerve in mediating hypoglycemia-induced arrhythmias, left vagotomized rats were compared with sham surgery controls. Rats that underwent vagotomy were healthy and had body weights similar to those of controls (sham: 299 ± 8; vagotomy: 305 ± 8 g). Glucose levels were similar between sham (12 ± 0.2 mg/dL) and vagotomy (11 ± 0.3 mg/dL) rats (Fig. 4A). The glucose infusion rates were similar between the two groups until 105 min into severe hypoglycemia, when rats with vagotomy required more glucose for the remainder of the clamp (Fig. 4B). However, plasma epinephrine, norepinephrine, and glucagon as well as respiration were all similar between the two groups throughout the clamp (Fig. 4C–F).

Rats with vagotomy had a 6.9-fold decrease in mortality during severe hypoglycemia compared with sham (sham: 46%; vagotomy: 6.7%; P < 0.05) (Fig. 5A). Decreased mortality was associated with a 4.6- and 4-fold decrease in first- and second-degree heart block, respectively, in vagotomy rats (Fig. 5B and C). Third-degree heart block occurred in 42% of sham rats, whereas no rats with vagotomy experienced third-degree heart block (P < 0.05) (Fig. 5D).

Basal heart rate on the day of the clamp was significantly increased in vagotomy rats (423 ± 7 bpm; P < 0.001) compared with sham rats (391 ± 5 bpm) (Fig. 5E). During severe hypoglycemia, the mean heart rate was similar for sham (264 ± 8 bpm) and vagotomy (280 ± 5 bpm; P < 0.07) rats, possibly due to the influence of the right vagus nerve, which was left intact. The QTc interval was similar during the basal period between the two groups and increased during severe hypoglycemia, with vagotomy rats having increased QTc at several time points compared with sham rats (Fig. 5F).

On the basis of the above experiments indicating a role for the PSNS in mediating hypoglycemia-induced arrhythmias, we wished to delineate whether the PSNS was acting through muscarinic (M2) or nicotinic (N) receptors. Therefore, pharmacologic blockade of the M2 or N receptors was performed. Glucose levels during severe hypoglycemia were evenly matched in control (11.3 ± 0.3 mg/dL), M2 blocker (11.2 ± 0.2 mg/dL), and N blocker (11.3 ± 0.2 mg/dL) (Fig. 6A). The glucose infusion rates required to maintain severe hypoglycemia were 2.8-fold higher in N blocker (4.8 ± 0.3 mg/kg/min; P < 0.01) versus control (1.7 ± 0.4 mg/kg/min) rats (Fig. 6B). M2 blocker rats (2.3 ± 0.3 mg/kg/min) had glucose infusion rates similar to those of controls. Consistent with increased glucose infusion rates in N blocker rats was their decreased epinephrine (2,172 ± 308 pg/mL; P < 0.01) compared with controls (5,951 ± 431 pg/mL) (Fig. 6C) and decreased norepinephrine (830 ± 106 pg/mL; P < 0.01) compared with controls (2,434 ± 372 pg/mL) (Fig. 6D). M2 blocker rats had epinephrine and norepinephrine levels similar to those of controls. Glucagon and insulin increased during severe hypoglycemia to a similar extent in all groups (Fig. 6E and F).

Mean heart rate was elevated during severe hypoglycemia in N blocker (276 ± 3 bpm; P < 0.03) and M2 blocker (364 ± 4 bpm; P < 0.01) rats compared with control rats (260 ± 7 bpm) (Fig. 6G). QTc was also increased in M2 blocker (209 ± 2 ms; P < 0.01) compared with control (173 ± 4 ms) rats during severe hypoglycemia, but there was no difference in N blocker rats (170 ± 3 ms) (Fig. 6H). However, hypoglycemia increased QTc in all groups compared with their respective basal values.

Severe hypoglycemia–induced mortality was completely prevented in N blocker rats compared with 25% mortality in control rats (P < 0.05) with no difference in mortality with M2 blocker rats (15.4%) (Fig. 7A). Overall, the frequency of cardiac arrhythmias was reduced in N blocker rats compared with control rats, while M2 blocker rats had arrhythmias similar to those of control rats. Premature ventricular contractions; first-, second-, and third-degree heart block; and nonsustained ventricular tachycardia were nearly completely prevented in the N blocker group (Fig. 7B–F). The M2 blocker rats had arrhythmias similar to those of control rats. Seizures and respiration were similar among the groups (data not shown).

To further determine if hypoglycemia-induced adverse cardiac arrhythmias are mediated by the parasympathetic neurotransmitter ACh, hearts were excised from rats and perfused with normal glucose or low glucose with or without ACh. Heart rate did not change in the hearts either in normal or low glucose alone, with average heart rates of 251 ± 2 and 244 ± 4 bpm, respectively, during the 3 h perfusion (Fig. 8A). When ACh was added, heart rate decreased as expected in normal (202 ± 6 bpm) and low (181 ± 6 bpm) glucose settings (P < 0.001). Hearts treated with low glucose + ACh + N blocker also had decreased heart rates (198 ± 10 bpm; P < 0.001). Hearts perfused in normal glucose, normal glucose + ACh, and low glucose + ACh + N blocker all survived the 3 h. However, hearts in low glucose had an 87.5% survival rate, and hearts in low glucose + ACh had an 85.7% survival rate (Fig. 8B). These denervated hearts had no arrhythmias when placed in normal glucose or low glucose alone. In addition, ACh alone did not cause arrhythmias. Interestingly, it was the combination of low glucose and ACh that induced second-degree heart block, which was prevented by blocking the nicotinic receptors with mecamylamine (Fig. 8C). While third-degree heart block only occurred in hearts with low glucose + ACh, the difference between groups was not statistically significant (Fig. 8D).

Fortunately, mortality due to severe hypoglycemia is rare, but hypoglycemia has been reported as the cause of death in ∼10% of fatalities in people with insulin-treated diabetes (1–3). To reduce the incidence of severe hypoglycemia–induced mortality, a better understanding of the pathophysiological mechanisms that cause severe hypoglycemia–induced fatal cardiac arrhythmias is needed. Thus, several experimental protocols were undertaken to determine the extent to which the SNS or PSNS mediates cardiac arrhythmias in the setting of severe hypoglycemia. Surprisingly, neither reducing circulating epinephrine nor chemical sympathectomy decreased arrhythmias or mortality as hypothesized. Interestingly, left vagotomy and pharmacological blockade of the nicotinic receptors of the PSNS reduced arrhythmias and mortality during severe hypoglycemia. Furthermore, denervated hearts had no arrhythmias under low glucose conditions, but the combination of low glucose and ACh was necessary to induce arrhythmias. These arrhythmias were prevented by blocking the nicotinic receptors. Taken together, the studies suggest that the PSNS is the main mediator of severe hypoglycemia–induced fatal cardiac arrhythmias.

Given these findings, it is unclear why our previous studies indicated a role for β-blockers in reducing hypoglycemia-induced arrhythmias (4,7). The fact that adrenal demedullation had no effect on arrhythmias indicated that circulating epinephrine levels were not the mediators of cardiac arrhythmias in the setting of severe hypoglycemia. Thus, it was hypothesized that perhaps direct sympathetic innervation of the heart was responsible for the observed mortality and arrhythmias. However, when chemical sympathectomy was achieved (and confirmed by a significant reduction in cardiac epinephrine and norepinephrine levels), there were no differences in mortality or arrhythmias during severe hypoglycemia. These surprising results suggest that other aspects of the autonomic nervous system, such as the PSNS or low glucose per se, might mediate severe hypoglycemia–induced arrhythmias.

The next hypothesis tested was that the PSNS causes arrhythmias during severe hypoglycemia. Our previous animal studies consistently noted a decrease in heart rate during severe hypoglycemia, which was consistent with the notion that activation of the PSNS was the overriding determinant of cardiac heart rate (4,7,23,24). Similarly, in clinical studies, activation of the PSNS is suggested by consistent reports of bradycardia noted in response to hypoglycemia (10,25). Thus, there is evidence that activation of the PSNS during severe hypoglycemia acts to lower heart rate and possibly increase the risk of fatal atrioventricular heart block (4,7,23). Consistent with a detrimental PSNS effect on the heart, it has been noted previously that excess vagal tone can lead to arrhythmias (11). Taken together, it was hypothesized that the vagus nerve leads to arrhythmias during severe hypoglycemia. The right and left vagus nerves innervate both sinoatrial and atrioventricular nodes of the heart, but the left vagus nerve primarily innervates the AV node (26). Since, in our rat model, severe hypoglycemia leads to fatal atrioventricular heart blocks, the left vagus nerve was hypothesized to be a primary mediator of severe hypoglycemia–induced mortality. Notably, the left vagus nerve transection did not result in elevated heart rate during severe hypoglycemia, possibly because the right vagus nerve was kept intact or activation of the SNS masked the effects of PSNS blockade during severe hypoglycemia. Our results indicate that surgical vagotomy significantly reduced arrhythmias and mortality, thus implicating an important role for the PSNS in mediating hypoglycemia-induced arrhythmias.

To further delineate the role of parasympathetic neurotransmitters in mediating hypoglycemia-induced arrhythmias, pharmacologic blockade was investigated with the use of muscarinic and nicotinic receptor blockers separately. The M2 muscarinic receptors are located throughout the heart (27). We chose to block M2 receptors because they have been found to be involved in tachycardia and heart failure (28). Additionally, M2 receptor knockout mice are protected against agonist-induced bradycardia (29), and the M2 receptor is a known mediator of parasympathetic activity in the heart (30). The M2 receptor blocker used in this study is relatively cardioselective with minimal ability to cross the blood-brain barrier (18). It was surprising that blocking the M2 receptors had no effect on mortality or arrhythmias. However, there are other muscarinic receptor subtypes, such as M1, found within the heart that may be involved in severe hypoglycemia-induced arrhythmias. Future studies are needed to clarify this.

Because blockade of muscarinic receptors did not improve outcomes, it was hypothesized that nicotinic receptors might mediate severe hypoglycemia–induced arrhythmias. Nicotinic ACh receptors (nAChR) in the heart receive synaptic input that leads to a reduction in heart rate (31). Since the α3β4 nAChR is the most ubiquitous receptor type within the intracardiac parasympathetic ganglion neurons (31–33), this receptor was targeted for blockade with mecamylamine. Mecamylamine is noncompetitive and has high binding potency to α3β4 receptors (34). As hypothesized, blocking the nicotinic receptors completely prevented mortality and arrhythmias during severe hypoglycemia. A limitation to this systemic pharmacological approach is that the α3β4 receptor is also found in adrenal glands (35) (see Fig. 1). Consistent with an effect of adrenal gland nAChR activation in mediating the adrenomedullary response to hypoglycemia, blocking of the nicotinic receptors with mecamylamine led to a blunted rise in epinephrine and norepinephrine during severe hypoglycemia, which might have affected the outcomes. However, since complete abrogation of the adrenomeduallary response in the demedullation experiments did not alter mortality or arrhythmias (see Fig. 2), we propose that the beneficial net effect of blocking the nicotinic receptors in preventing fatal cardiac arrhythmias during severe hypoglycemia was mediated through direct action on the heart, rather than mediated via suppression of the adrenomedullary response.

The ex vivo studies again point to multifactorial causes of hypoglycemia-induced arrhythmias. Under the conditions studied, the following occurred: 1) a denervated heart had no arrhythmias, 2) ACh alone did not cause arrhythmias, 3) low glucose alone did not cause arrhythmias, 4) it was only the combination of ACh and low glucose that increased arrhythmias, and 5) these induced arrhythmias were abrogated by blocking the nicotinic receptors. Thus, in the setting of glucose deprivation, increased parasympathetic neurotransmitter signaling, via the nicotinic receptors, results in bradyarrhythmias. Taken together, these data suggest that in the setting of an energy-depleted heart, the autonomic nervous system’s release of ACh leads to hypoglycemia-induced arrhythmias.

The level of severe hypoglycemia (10–15 mg/dL) achieved in these rodent studies was found to be necessary to consistently and reproducibly induce cardiac arrhythmias. Fortunately, such low levels of hypoglycemia are rarely seen in people with insulin-treated diabetes. However, 1) iatrogenic hypoglycemia can be fatal (1–3); 2) glucose levels of 10 mg/dL have been reported in a patient found dead in bed (36); and 3) conditions of less severe hypoglycemia can induce ECG abnormalities (6).

It can be concluded that the PSNS is the primary mediator of severe hypoglycemia–induced cardiac arrhythmias in this nondiabetic rodent model because 1) neither abrogating the rise in circulating epinephrine nor cardiac chemical sympathectomy alters mortality or arrhythmias associated with severe hypoglycemia, 2) surgical and pharmacological blockade of the PSNS protects against arrhythmias, and 3) under low glucose conditions, ACh induces arrhythmias in denervated hearts ex vivo. Clinically, drugs that target the PSNS may be promising to protect against severe hypoglycemia–induced fatal cardiac arrhythmias.

Acknowledgments. The authors thank Dr. Phil Cryer and Krishan Jethi from Washington University in St. Louis for assistance with the cardiac catecholamine assays.

Funding. The authors acknowledge funding from the National Institutes of Health (grant 5T32-DK-091317 to C.M.R. and grant HL-128752 to D.J.D.), JDRF (grant 3-APF-2017-407-A-N to C.M.R.), and the University of Utah Diabetes and Metabolism Research Center (grant RO1 NS070235 to S.J.F.).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. C.M.R. designed and conducted the experiments and wrote the manuscript. J.B., Y.H., and M.O. conducted the experiments and wrote the manuscript. A.M.H. conducted the experiments. D.J.D. designed the experiments and edited the manuscript. S.J.F. designed the experiments and edited the manuscript. S.F.F. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this manuscript were presented in abstract form at the 78th Scientific Sessions of the American Diabetes Association, Orlando, FL, 22–26 June 2018, and the 79th Scientific Sessions of the American Diabetes Association, San Francisco, CA, 7–11 June 2019.