Exposure to hypoglycemia and risk of stroke

Abstract

In the treatment of both Type 1 and Type 2 diabetes mellitus, maintaining a euglycemic state represents one of the key challenges. Improper dosing and administration of glucose-lowering drugs is associated with an increased risk of recurrent hypoglycemia episodes. In addition, the risk of adverse cardiovascular events in diabetic patients, particularly myocardial infarctions and strokes, is well-established. Current research indicates a potential link between the baseline risk of cardio/cerebrovascular events in diabetic patients and exposure to hypoglycemia. In this review of the literature we aim to determine if a relationship exists between recurrent hypoglycemia and adverse neurovascular events.

Introduction

Chronic hyperglycemia in diabetic patients can lead to a variety of long-term complications^1–5. Of particular importance is the increased risk of cardiovascular diseases, which is reported as the leading cause of death in diabetic patients^{6, 7}. Tight and aggressive glucose management treatment with pharmacological agents has been shown to reduce the risk for some macrovascular and microvascular pathologies, but such treatments also put patients at increased risk for subclinical and severe hypoglycemia⁸. During hypoglycemia, the autonomic nervous system acts to increase release of catecholamines in an attempt to restore normal concentrations by increasing hepatic production of glucose⁹. These increased systemic catecholamine levels can lead to unintended consequences, such as increased platelet aggregation, fatal arrhythmias and chronic cardiac inflammation¹⁰. Repeated episodes of hypoglycemia can weaken the body’s neuroglycopenic response to low blood glucose¹¹. Preclinical studies indicate that hypoglycemia may activate procoagulant pathways. However, not all clinical studies support the notion that exposure to hypoglycemia increases the risk of cardiovascular complications. In this review, we aim to review the existing literature to determine if hypoglycemia has any effect on cerebrovascular events in diabetic patients.

Incidence of hypoglycemia in diabetes management

The American Diabetes Associated Workgroup on Hypoglycemia defines hypoglycemia as “all episodes of abnormally low plasma glucose concentration that expose the individual to potential harm”¹². Several categories of hypoglycemia exist, including severe hypoglycemia (requiring another individual to assist with treatment), symptomatic and asymptomatic hypoglycemia (both of which have blood glucose ≤70 mg/dL but with variable symptoms), probable hypoglycemia (symptoms reported, but not confirmed by glucose measurement) and relative hypoglycemia (symptoms present, but glucose is still >70 mg/dL)¹².

Although hypoglycemia presents with similar symptoms in type 1 diabetes (T1D) and type 2 diabetes (T2D), the differing pathologies of the diseases cause subtle differences in the course of these hypoglycemic episodes. Repeated episodes of hypoglycemia in a short amount of time increase the risk of development of hypoglycemia-associated autonomic failure (HAAF), in which the body’s neurogenic responses to low blood sugar are blunted, leading to defective glucose counter-regulation and hypoglycemia unawareness¹³. These repeated hypoglycemic episodes shift the systemic thresholds needed to induce these autonomic responses to lower blood glucose levels¹¹. Owing to the onset of hypoglycemia unawareness in HAAF as well as the inability to properly measure and control blood glucose during sleep, it has been hypothesized that both T1D and T2D patients experience mild and moderate hypoglycemic episodes much more frequently than previously thought. Macleod et al.¹⁴ performed a retrospective clinical study on a randomly selected sample of diabetic patients (both T1D and T2D) and showed evidence of severe hypoglycemia at an event rate of 170 per 100 patient-years and 73 per 100 patient-years in insulin-treated T1D and T2D patients, respectively¹⁴. Several other studies also made similar observations^{15, 16}. Studies involving continuous blood glucose monitoring have shown that as many as 68% of T1D (3 days of monitoring)¹⁷ and 49% of T2D (5 days of monitoring)¹⁸ recorded systemic glucose levels below 60 mg/dl.

Impaired autonomic responses in old age may increase older diabetic patients’ susceptibility to experiencing recurrent hypoglycemia¹⁹. An earlier study to evaluate incidences of hypoglycemia in sulfonylurea (SU)- or insulin-treated older diabetic patients observed that elderly persons using multiple medications and those who are frequently hospitalized, are at greater risk for hypoglycemia^19,20. Although difficult to estimate, repeated exposure to hypoglycemia in older individuals may lead to deterioration of general heath, disability and poor outcomes^19,21.

Despite recent advances in new regimens for blood glucose control, the fear of hypoglycemia still looms large²². There have been a small number of clinical trials in recent years showing that the incidence of hypoglycemia with the use of insulin analogs and oral hypoglycemic agents is an undeniable reality^{23, 24}. The landmark clinical trials, such as Action to Control Cardiovascular Risk in Diabetes (ACCORD)²⁵ and Diabetes Control and Complications Trial (DCCT)²⁶, mostly concentrated on the hypoglycemia incidence in developed nations. However, it is as much of a reality in developing nations with huge population of diabetic patients, including India and China. The treatment of diabetes and the availability of appropriate clinical expertise to manage the condition varies slightly from country to country. The under-reporting of self-treated hypoglycemia events only adds to the seriousness of the issue as the risks and incidence of hypoglycemia can be underestimated. The Hypoglycemia Assessment Tool (HAT) study conducted recently in developing countries on 27,585 insulin-treated diabetic patients showed an incidence rate of hypoglycemia higher than previously reported²⁷. There have been a few European studies that have also demonstrated the presence of self-treated hypoglycemia^{28, 29}. As pointed out in a recent review on hypoglycemia in diabetic patients, the protocols of close glucose monitoring followed rigorously in clinical trials may not be an accurate representation of the real incidence of hypoglycemia in actual clinical practice³⁰.

The effect of hypoglycemia on procoagulant mechanisms

In this section we discuss potential mechanisms by which exposure to hypoglycemia (single or repeated) accelerate procoagulant mechanisms. Mechanisms described below are summarized in Figure 1.

Figure 1.

Schematic diagram summarizing procoagulant mechanisms activated by hypoglycemia.

The relation of hypoglycemia with thrombotic and hemostatic mechanisms

Disturbances in platelet activity can lead to bleeding disorders and thrombosis. Hypoglycemia appears to be associated with platelet activation^{31, 32}. Hypoglycemia accelerates vascular complications in diabetes by increasing platelet aggregation as well as fibrinogen formation. Studies have shown that acute hypoglycemia impairs fibrinolytic balance, increases pro-inflammatory responses, platelet activation and coagulation biomarkers as well as reduces NO-mediated endothelial function³³. Further studies have also shown that hypoglycemia induces an increase in circulating levels of vascular adhesion molecules (markers of endothelial cell damage), interleukin (IL)-6, and P-selectin (markers of platelet activation)^34,35. Injury to endothelial cells and activation of platelets can result in a procoagulant state. Previous studies have confirmed that hypoglycemia has an inhibitory effect on fibrinolytic mechanisms³⁵. Hypoglycemia also has a detrimental effect on calcium homeostasis and platelet mitochondrial integrity³⁴. Altered mitochondrial calcium homeostasis and mitochondrial dysfunction can result in platelet activation³⁶. Hypoglycemia decreases the activated partial thromboplastin time (aPTT) and increases levels of fibrinogen and factor VIII³⁷. With recurrent transient hypoglycemic episodes, repeated occurrences may have additive effects that accentuate inflammation-based processes that include atherogenesis and other thrombotic complications. Other pro-inflammatory changes include an increase in the plasma concentration of IL-6, IL-8,IL-1β and other pro-inflammatory mediators that includes leukocytosis, reactive oxygen species (ROS) generation, lipid peroxidation, and increased levels of tumor necrosis factor-α (TNF-α)³⁸. It is well established that proinflammatory mediators can increase systemic thrombogenicity by affecting multiple pathways^{39, 40}.

Atherosclerosis is one of the determining factors associated with ischemic stroke. Platelets have been reported as one of the key factors in developing atherothrombosis^{41, 42}. Platelets play a significant role in the pathology of cerebral ischemia through their participation in the generation of thromboemboli that may initiate stroke symptoms. More often, activated platelets particularly are a major contributor to the large-vessel subtype of ischemic stroke that occurs following atherosclerotic plaque rupture⁴³. These detrimental effects would add to the previously demonstrated relationship between hypoglycemia on ischemic stroke. Collectively, therefore, hypoglycemia can trigger a sequence of events that may induce stroke.

Hypoglycemia stimulates the autonomic nervous system and brings about sympathoadrenal activation to facilitate the release of catecholamines⁴⁴. This promotes secondary effects on the circulatory system, including a decrease in plasma volume⁴⁵; an increase in leukocyte and erythrocyte counts^{46, 47}, and platelet aggregation⁴⁸. In addition to these changes in the formed elements of blood, hypoglycemia has substantial hemorheological effects. Hypoglycemia is shown to increase blood coagulation factor VIII activity via a β₂ adrenoceptor-mediated process⁴⁹. This was further confirmed and supported by other relevant studies^50,51. Hypoglycemia leads to a rise in Von Willebrand factor⁵² and accelerates the rate of thrombin generation⁵⁰.

Endothelin-1 (ET-1) is one of the major endothelins that acts as a potent vasoconstrictor. In an earlier clinical trial, Wright et al. investigated the effect of acute symptomatic hypoglycemia in adult T1D patients on plasma ET-1 response⁵³. The goal of the study was to determine the mechanism by which acute insulin-induced hypoglycemia leads to vasoconstriction and triggers microvascular and macrovascular events. Plasma ET-1 levels were observed to increase substantially during hypoglycemia. Though these results adequately help to implicate the role of ET-1 in altered vascular hemodynamics during hypoglycemia, whether the observed effect is directly due to hypoglycemia, or secondary due to hypoglycemia-induced hormonal response, or insulin administration remains to be determined. In another study, Trovati et al. observed that insulin-induced hypoglycemia results in an increased platelet sensitivity to aggregating agents like ADP and thrombin in vitro potentially due to hypoglycemia-related alpha-adrenoceptor activation⁵⁴.

Another clinical study determined the difference that euglycemia and hypoglycemia had on fibrinolytic balance and pro-inflammatory mechanisms³⁵. This study included thirty five healthy and twenty four T1D volunteers who previously never experienced any episode of hypoglycemia and had intact autonomic function. Pro-inflammatory markers, such as vascular cel adhesion molecular (VCAM), intercellular adhesion molecular (ICAM), E-Selectin and vascular endothelial growth factor (VEGF), were increased during the hyperinsulinemic hypoglycemia episodes. These results suggest the possibility of hypoglycemia leading to atheroma formation, as these proteins are normally present on cell surfaces and are expressed during inflammation to promote leukocyte adhesion which can lead to plaque and atheroma formation. Another study observed increased CD40 expression and platelet–monocyte aggregation during acute hypoglycemia in both non-diabetic and T1D subjects⁵⁵. A recent study on 45 healthy subjects compared inflammatory and pro-atherosclerotic biomarkers released during euinsulinemic hyperglycemia observed an increase in plasminogen activator inhibitor-1 (PAI-1) and TNF-α levels during hypoglycemia⁵⁶. PAI-1 is a protein that inhibits activators of plasminogen and fibrinolysis like tissue plasminogen activator (tPA) while TNF-α has been shown to facilitate activation of the coagulation cascade⁵⁷. These studies demonstrate activation of hemostatic mechanisms by exposure to hypoglycemia.

Hemostatic activation is a major contributor to the progression of ischemic stroke. Elevated levels of several inflammatory and rheological variables are reported to be associated with ischemic stroke. Past studies have provided strong evidence of increased thrombin generation and fibrin turnover with associated altered fibrinolytic activity and derailed endothelial function in acute stroke^58–63. Further studies have revealed an elevated thrombin–antithrombin complex (TAT) and fibrin D-dimer levels in patients with progressing stroke⁶³. This literature supports the notion that exposure to hypoglycemia may increase the risk of ischemic stroke via activation of hemostatic mechanisms.

Hypoglycemia and free radicals

Hypoglycemia plays a major role in enhancing diabetes-related vascular complications¹⁰. There is evidence that indicates a deleterious impact of hypoglycemia on the blood-brain barrier (BBB)^64–66. Compromised BBB function during hypoglycemia has been reported to trigger the pathogenesis of secondary brain injuries. Importantly, lower glucose levels inhibit the protective physiological effects of flow-induced shear stress on vasculature and endothelial responses to flow by abundant accumulation of ROS⁶⁷. Reduced glutathione functions as a major anti-oxidant in the brain. In hypoglycemia, impaired synthesis of glutathione subjects the brain to considerable oxidative stress⁵³. Glucose reintroduction after hypoglycemic coma has been reported to generate superoxide and nitrotyrosine immunoreactivity and is suggested to be the main cause of stimulation of oxidative stress through the activity of NADPH oxidase^68,69. Selective increase in lipo-peroxidation due to hypoglycemia leads to oxidative stress⁷⁰ that has been confirmed in animal studies⁷¹. Animal studies have also confirmed increased mitochondrial ROS production during both hypoglycemia⁷² and in cultured neurons during glucose deprivation⁷³. ROS are also known to activate several intermediates of thrombosis, including platelets⁷⁴. These studies suggest that the hypoglycemic condition promotes the production of ROS and reactive nitrogen species (ROS/RNS) that might create a pro-thrombotic state.

Incidence and morbidity of stroke in diabetes

The importance of diabetes as a risk factor for the occurrence of stroke is well-known. There have been several studies proving an association between hyperglycemia and an increased risk of stroke^75,76. The Framingham study was the first to report an increased occurrence of ischemic stroke in both men and women with diabetes⁷⁷. The Copenhagen Stroke Study and the GCNKSS study provided epidemiological link between diabetes and stroke^78,79. There are other studies that also support this link^80,81. The U.S. Nationwide Inpatient Sample has shown an increase of about 27% in the absolute number of hospitalizations for acute ischemic stroke in patients with co-morbid diabetes (T2D) when compared to an overall 17% decrease in acute ischemic stroke hospitalizations without diabetes from 1997 to 2006⁸². The rate of stroke in diabetic patients has declined by about 53% between 1990 and 2010⁸³.

Incidence of hypoglycemia and the risk of cerebrovascular events

Preclinical studies provide strong evidence that hypoglycemia may accelerate procoagulant mechanisms. In view of this, below we provide a summary of studies that support and oppose this premise.

Effect of prior hypoglycemia and cerebrovascular events

A study conducted by Rathmann et al. comparing a group of T2D patients receiving a dipeptidyl peptidase-4 inhibitor (DPP4-I) and another group receiving a SU only for two years observed a 5-fold decrease in frequency of recorded hypoglycemia in DPP4-I users compared to those on SU⁸⁴. The hazard ratio for stroke/transient ischemic attack (TIA) was 0.57 (p < 0.001) in DPP4-I users compared to SU users. Similarly, hazard ratios for macrovascular disease, coronary heart disease, myocardial infarction, and peripheral vascular disease in DPP4-I treated patients were 0.74 (p < 0.001), 0.74 (p < 0.001), 0.81 (p < 0.05), and 0.73 (p < 0.001) compared to SU-treated patients, respectively. This study suggests that increased incidences of hypoglycemia may be linked to increased risk of stroke/TIA and other macrovascular diseases in SU-treated diabetic patients. Another prospective study was conducted by Gitt et al. to confirm earlier studies that treatment with DPP4-I resulted in decreased incidences of hypoglycemia. Patients received dual treatment with either metformin and sulfonylureas (SUs) or metformin and DPP4 inhibitors⁸⁵. These two groups were followed for 12 months to study the frequency of hypoglycemia in these patients. It was observed that the patients on metformin + SU had a greater frequency of episodes of hypoglycemia than the patients on metformin + DPP4-I. The group on metformin + SU showed a substantial increase in stroke/TIA incidences when compared to patients on DPP4-I + metformin (2% vs 0.2%; p < 0.05). It is plausible that increased incidences of stroke/TIA in the SU + metformin group may be due to higher incidences of hypoglycemia observed in this group. It should be noted that these two studies were observational ones that did not adequately explore if hypoglycemia was an independent predictor of stroke/TIA. Although basic experimental data suggest the association of hypoglycemia and stroke, future clinical studies aimed to determine this association are warranted. The Veterans Affairs Diabetes Trial (VADT) reported significantly more episodes of hypoglycemia in the intensive-therapy group than in the standard-therapy group⁶. Patients with BMI of ≥27 and <27 were treated with metformin plus rosiglitazone or glimepiride plus rosiglitazone, respectively. However, this study did not find any significant effect of intensive therapy on the rates of major cardiovascular events when compared to standard therapy.

A prospective study of participants with diagnosed diabetes from the Atherosclerosis Risk in Communities (ARIC) aimed to determine the link between severe hypoglycemia and cardiovascular disease observed that prior sever hypoglycemia exposure increases the risk of coronary heart disease, all-cause mortality and cardiovascular mortality⁸⁶. However, no association between prior exposure to severe hypoglycemia and stroke was observed⁸⁶. Recently published secondary analysis of the double-blind Trial Comparing Cardiovascular Safety of Insulin Degludec vs Insulin Glargine in Patients with Type 2 Diabetes at High Risk of Cardiovascular Events (DEVOTE) trial also evaluated the association between exposure to severe hypoglycemia and a major adverse cardiovascular event (MACE), defined as either cardiovascular death, non-fatal myocardial infarction or non-fatal stroke⁸⁷. This study observed an association between exposure to severe hypoglycemia and all-cause mortality. However a non-significant difference was observed in the risk of MACE in individuals who experienced severe hypoglycemia (hazard ratio 1.38, p = .080) vs. those who did not. Similarly a non-significant increase in the risk for non-fatal stroke was observed in individuals those who had (hazard ratio 1.81, p = .085) vs. those who had not experienced severe hypoglycemia. However, a recently published study using data from the Examination of Cardiovascular Outcomes with Alogliptin versus Standard of Care (EXAMINE) study observed that the adjusted hazard ratio for MACE risk in patients who experienced serious hypoglycemia events (hazard ratio 2.42, p = .007) was significantly higher than those without serious hypoglycemia exposure⁸⁸. However, this study did not report effects on non-fatal stroke separately. Another recent study evaluated the association between hypoglycemia and cardiovascular disease in SUs-treated T2D observed that an increasing frequency of hypoglycemia exposure was associated with increased stroke risk⁸⁹. The strongest association between exposure to hypoglycemia and stroke risk was observed when patients were exposed to 3 or more hypoglycemic events (relative risk 1.57)⁸⁹. It should be noted that the effect of mild or moderate hypoglycemia was not evaluated in the ARIC, DEVOTE, or EXAMINE populations. It is plausible that the risk of cardio- and cerebrovascular events may be already higher in studied patients owing to their high risk for exposure to mild or moderate hypoglycemia, and if so, this may also mask (i.e. lower the hazard ratio) the effect of severe hypoglycemia on studied cardio- and cerebrovascular events.

The effects of intensive glucose control, involving the use of gliclazide (modified release) as compared to standard glucose control in T2D patients was undertaken with 11,140 patients from over 200 medical centers spanning over 20 countries by the ADVANCE Collaborative group. After a follow-up period of 5 years, there was a significant decrease in the overall HbA1C values in the intensive group (6.5%) than in the standard control group (7.3%). Hypoglycemia was more common in the intensive control group (2.7%) than the standard control group (1.5%) [Hazard ratio 1.86; 95% CI 1.4 to 2.4 and p < .001]. However, the study showed no statistically significant difference in the outcome of nonfatal stroke in the two groups⁹⁰. In the ACCORD Trial, participants belonging to the intensive glucose control group experienced more episodes of hypoglycemia (self-monitored blood glucose < 70 mg/dl) than the standard control group (1.06 and 0.29 episodes per week, respectively)^{91, 92}. In both original and nine year follow-up studies, the odds ratio for nonfatal stroke was lower in the intensive glucose control group^91–93. However, this difference was not statistically significant. These effects may be explained as gliclazide has anti-oxidant properties and thus may also lower the impact of procoagulant pathways activated by hypoglycemia exposure⁹⁴. Similarly, studies also reported that other SUs also improve platelet function^95,96. For example, glibenclamide and glimepiride was able to suppress thrombin-stimulated increase in intracellular calcium and thrombin-induced arachidonic acid metabolism in platelets⁹⁵. However, gliclazide showed no effect on above-mentioned parameters⁹⁵. Another study reported that both gliclazide and glyburide treatment in noninsulin-dependent diabetic subjects normalized ADP-induced platelet aggregation when compared to a diet alone group of noninsulin-dependent diabetic subjects⁹⁶. It should be noted that this study did not observe any correlation between plasma glucose levels and platelet function suggesting that these effects may be from the drug and not plasma glucose level⁹⁶. It is possible that beneficial effects of glucose-lowering medications used to treat T2D patients may mask some of the deleterious effects of hypoglycemia on cerebrovascular outcomes.

Effect of acute post-stroke hypoglycemia and cerebrovascular events

The UK Glucose Insulin in Stroke Trial (GIST-UK) that determined if intensive insulin therapy (variable-dose-insulin by glucose-potassium insulin: GKI infusion) immediately after acute ischemic stroke reduces death at 90 days observed that episodes of hypoglycemia (<72 mg/dl over a period >30 min) occurred in 73 patients belonging to the GKI infusion group^{97, 98}. Plasma glucose of lower than 72 mg/dl was also observed in an additional 187 patients belonging to the GKI group; however, these patients did not meet the temporal requirement. This study also noted that 7 patients receiving GKI had a recurrent stroke within 72 h after the initiation of treatment compared to only 3 patients in the control arm. However, this difference was not statistically significant.

The Treatment of Hyperglycemia in Ischemic Stroke (THIS) trial: a randomized, multicenter, blinded feasibility, and tolerability trial evaluating the effect of aggressive hyperglycemia correction using intravenous insulin vs standard care during acute cerebral infarction, noted that 35% of patients receiving aggressive treatment had at least one episode of hypoglycemia (<60 mg/dL). Only 3% of these patients had neurologic symptoms, and majority of them (64%) were asymptomatic⁹⁹. This trial did not observe any correlation between the hypoglycemic episodes and worsened clinical outcomes. The Glucose Regulation in Acute Stroke Patients (GRASP) trial: a prospective, randomized, multicenter, trial evaluated the feasibility and safety of two insulin infusion protocols (tight glucose control (70–110 mg/dL) and loose glucose control (200 mg/dL)), and compared with usual care (glucose levels 70–300 mg/dL)¹⁰⁰. This trial reported hypoglycemic (<55 mg/dL) episodes in 30% of patients in the tight glucose control group. This trial also did not report any significant differences in serious adverse events among study groups.

In summary, some clinical evidence suggests that exposure to hypoglycemia increases the risk of cerebrovascular events in diabetic patients. It is also possible that hypoglycemia may indirectly affect stroke risk, as increased incidences of hypoglycemia may also reflect underlying poor health per se that predisposes to adverse vascular outcomes. A systematic review aimed to summarize literature on the effect of hypoglycemia on cardiovascular risk in diabetic patients found that several studies identified hypoglycemia, via increasing thrombotic tendency and other mechanisms, as contributing to cardiovascular risk in diabetic patients¹⁰. Another similar review by Moheet and Seaquist concluded that although acute hypoglycemia may lead to a prothrombotic state, current knowledge based on large clinical trials does not establish a link between hypoglycemia and adverse cardiovascular outcomes¹⁰¹. A recent review indicated a potential link between hemorrhagic transformation observed in stroke patients and exposure to hypoglycemia post-stroke¹⁰². Similarly, future pre-clinical and clinical studies aimed to determine association between hypoglycemia and the risk of stroke are warranted.

Conclusions

Diabetic patients compared to non-diabetic subjects have both a higher risk of stroke and poor stroke outcomes. Frequent exposure to hypoglycemia is observed in both T1D and T2D patients. Pre-clinical studies suggest that hypoglycemia may increase the risk of stroke in diabetic patients. This premise is not well-supported by clinical studies. The effects of hypoglycemia in diabetic patients on normal and pathological hemostatic pathways warrant further investigation. It would be also important to identify any sub-group of diabetic patients that are more susceptible to a hypoglycemia-induced pro-hypercoagulable state.