Metabolic Rehabilitation: Science Gathers to Support a New Intervention to Prevent Stroke

Type 2 diabetes is a well-recognized risk factor for cerebrovascular disease. In this issue of Stroke, Thacker and colleagues demonstrate that earlier abnormalities in carbohydrate metabolism, insulin resistance and impaired glucose tolerance, are also associated with increased risk for cerebral ischemic events¹. These findings are reliable and timely, and they have immediate implications for the design of the next generation of stroke prevention trials.

Abnormal carbohydrate metabolism comprises a spectrum of disorders characterized by impaired energy utilization. Early manifestations include obesity, decreased insulin sensitivity, impaired fasting glucose (IFG) and impaired glucose tolerance (IGT). The latter two conditions are often referred to as ‘pre-diabetes’ – a stark description of their typically progressive nature. Its most advanced manifestation, type 2 diabetes, results from the inexorable loss of insulin secretory capacity in the face of well-established insulin resistance. These metabolic states have become endemic in most countries as westernized lifestyles are adopted. In the US alone, there are now 26 million individuals with diabetes and 79 million with pre-diabetes. It is estimated that by 2030, there will be 366 million people with diabetes throughout the world.

Clinicians who care for stroke patients know that metabolic disease is also prevalent in this population. Among patients with ischemic cerebrovascular disease 18%-30% are obese, 25%-30% are insulin resistant, 23%-28% have impaired glucose tolerance, and 13%-36% have diabetes^2-5.

Figure 1 illustrates a model of impaired glucose regulation and its potential causal relatinship to stroke. Each of its cardinal manifestations (i.e., obesity, insulin resistance, hyperglycemia, diabetes) have been associated with increased risk for stroke⁶. However, the evidence is strongest for obesity and diabetes, possibly because these are easiest to measure and classify. Obesity (BMI ≥ 30kg/m²) doubles the risk for ischemic stroke, although the effect appears to be substantially mediated by coexistant cardiovascular risk factors^{7, 8}. Importantly, not all obese patients exhibit these complications; indeed about 25% of those with obesity lack additional evidence for the metabolic syndrome; that is, they do not have hypertension, dyslipidemia, hyperglycemia, or evidence of vascular inflammation, insulin resistance, or endothelial dysfunction that typically characterizes this condition⁹. Adipose tissue topography may explain some of these differences, since patients with so-called “uncomplicated obesity” may have fat distributed preferentially to more metabolically quiescent sites such as arms, legs, and buttocks⁹, as opposed to the more metabolically active centripetal sites, including liver and omentum. It is not surprising, therefore, that waist-hip-ratio (a reflection of central obesity) has a closer association with stroke risk than BMI (odd ratio adjusted for risk factors including diabetes and BMI=3.0, 95% CI 2.1-4.2)¹⁰.

figure 1.

A model of impaired carbohydrate regulation and its potential causal relation to ischemic stroke

Obesity is a major contributing risk factor to the development of insulin resistance leading to type 2 diabetes. Patients with diabetes have 2-3 times the risk of ischemic stroke compared with non-diabetic patients after adjusting for blood pressure and other risk factors^{6, 11}. In addition, on average patients with diabetes tend to have strokes about two years earlier than non-diabetic patients¹².

The distinct importance of the study by Thacker and colleagues is in confirming that pre-diabetic conditions, including insulin resistance and post-load hyperglycemia, are associated with increased risk for stroke. The investigators conducted an observational cohort study of 3442 community-based participants in the Cardiovascular Health Study who were free of stroke and diabetes at baseline. To classify glucose and insulin resistance status, participants underwent an oral glucose tolerance test. Insulin and glucose measures at baseline and 2 hours were used to calculate the Gutt insulin sensitivity index. This measure of insulin sensitivity is more accurate than the fasting insulin alone or the popular HOMA index, which is calculated from fasting insulin and glucose concentrations and reflects predominantly hepatic insulin resistance. The Gutt index factors in post-load glucose and insulin levels which reflect, to a large degree, peripheral (i.e., skeletal muscle) insulin resistance. This is important because peripheral insulin resistance is more closely aligned with vascular risk. The main outcome, ischemic stroke, was confirmed by medical record review. Participants were followed from enrollment in 1989/90 to 2007 (17 years).

Risk for ischemic stroke (adjusted for age, sex, race, renal function, coronary disease, atrial fibrillation and peripheral arterial disease) was significantly increased for persons in the fourth quartile of insulin sensitivity (i.e., more insulin resistant) compared with the first (RR= 1.64, 95% CI 1.24-2.16). Further adjustment for blood pressure and cholesterol attenuated the risk but it remained significant (RR 1.39, 95% CI 1.05-1.86). Two-hour serum glucose was also associated with increased risk for ischemic stroke in the fully adjusted model (RR = 1.57, 95% CI 1.18-2.09).

The findings by Thatcher et al are driven by the two-hour glucose and insulin values. Fasting glucose and fasting insulin concentrations were not associated with increased risk for ischemic stroke. The greater relative importance of IGT is expected^{6, 13, 14}, but the absence of any observed effect of IFG on risk for ischemic stroke is surprising. An important body of research suggests that fasting glucose in the non-diabetic range remains positively associated with vascular disease risk¹⁵ , including ischemic stroke specifically⁶. The same is true for fasting insulin¹⁶. Indeed fasting and post-prandial glucose and insulin levels are linked, and the variable findings from studies exploring this question may be the result of methodology rather than true biological differences. Hemoglobin A1c, which incorporates fasting and post-prandial hyperglycemia, also appears to be a more robust marker of vascular risk than fasting glucose¹⁵.

While the research methods used by Thatcher and colleagues are generally sound, their results may have actually underestimated the association between insulin resistance and risk for stroke due to over-adjustment in their second multivariable model. Insulin resistance can increase blood pressure through mechanisms that include up-regulation of endothelin 1, microvascular dysfunction, and two properties of insulin which appear to be unimpaired in insulin resistant states, namely renal sodium retention and activation of the sympathetic nervous system. Hyperinsulinemia, especially in a hyperglycemic milieu, also increases liver triglyceride synthesis and thereby indirectly lowers HDL-cholesterol concentrations. Hypertension and dyslipidemia, therefore, may represent intermediate mechanisms by which insulin resistance increases stroke risk; adjustment for them may, therefore, attenuate the association statistically.

The paper by Thatcher and colleagues is one of several studies^16-18 to examine the link between insulin resistance and risk for ischemic stroke, but only the second¹⁸ to measure peripheral insulin resistance using more robust techniques based on glucose challenge, either with oral loading or glucose infusion. All studies have suggested an association and most report relative risks or hazard ratios at about 2.0 or higher. In the context of these other studies, Thatcher and colleagues report a weaker association (RR=1.64, 95% CI 1.24-2.16 in the first adjusted model), although it still reaches clinical significance. IGT has also been associated with increased risk for stroke, although the data are more scarce^{6, 19}.

This confirmatory paper by Thacker and colleagues should lay to rest any doubt that insulin resistance and IGT are important risk factors for ischemic stroke. More importantly, this work should provide new impetus to developing effective approaches to ‘metabolic rehabilitation’ both for prevention of first stroke and secondary prevention of vascular disease of all types in patients with a recent ischemic stroke or TIA. Abnormal carbohydrate metabolism harms blood vessels over many years. To have the greatest impact in preventing macrovascular events, therefore, metabolic rehabilitation may need to target those with the earliest manifestations (e.g. obesity, insulin resistance, pre-diabetes).

Models for effective metabolic rehabilitation have been developed, but none are adequate for prevention of first or recurrent stroke. Cardiac rehabilitation programs typically last 12 weeks and emphasis physical activity. They have been associated with reduced all-cause mortality, but sustained effects on weight, long-term vascular outcomes, and diabetes prevention have not been demonstrated²⁰. Similarly, diet programs directed at weight loss generally produce results in the short term, but weight is typically regained within a few years and the clinical significance of the loss has not been determined²¹. Unsupported office-based advice to increase physical activity or lose weight is generally ineffective²². For sustained weight loss, intensive programs involving personalized coaching for both fitness and nutrition, meal provision (at least initially), and continuous participation over years seem to be required^{23, 24}. Such intensive efforts, as exemplified by the Diabetes Prevention Program²³, are expensive and require a high level of commitment from participants. Even these intensive programs produce only modest weight loss over the long term. Nevertheless, sustained multi-disciplinary program-based models may be the best we have to offer in the near term.

Because changing human behavior is difficult, pharmacotherapy and surgery are popular among patients, clinicians, and researchers. Pharmacotherapy has a second-tier role in diabetes prevention (next to lifestyle modification) and a marginal role as adjunctive therapy for weight loss. The NINDS is currently funding the Insulin Resistance Intervention after Stroke (IRIS) trial to determine the effectiveness of the insulin-sensitizer, pioglitazone, in prevention of recurrent stroke and myocardial infarction among non-diabetic patients with a recent ischemic stroke or TIA and insulin resistance (NCT 00091949). If this study proves positive, a new avenue of clinical investigation may open. Other pharmacological agents with insulin sensitizing effects include other thiazolidinediones, non-thiazolidinedione PPAR agonists, 11β hydroxysteroid dehydrogenase-1 inhibitors, protein tyrosine phosphatase-1b inhibitors, and acetyl-coA Carboxylase-1 and -2 inhibitors. For selected patients, bariatric surgery is successful in preventing progression to diabetes²⁵, but effects on vascular events and mortality remain uncertain²⁶.

Thacker and colleagues have added to the scientific basis for a new generation of research on metabolic disease and stroke. Until these investigations bear fruit, we recommend that patients with ischemic stroke be screened for diabetes and pre-diabetes and that those with abnormal findings are encouraged to enter and remain in structured programs for fitness and good nutrition.