Hypertension and Lipid Management in Prediabetic States


Alan J Garber, MD, PhD, Department of Medicine, Division of Endocrinology, Diabetes and Metabolism, Baylor College of Medicine, One Baylor Plaza Houston, TX 77030
E-mail: agarber@bcm.edu


Patients with prediabetes have impaired fasting glucose, impaired glucose tolerance, and the metabolic syndrome. They have similar coronary heart disease risk rates as patients with type 2 diabetes. Goals and agents for blood pressure and lipid management should be the same for patients with prediabetes as those for patients with overt diabetes. Multiple antihypertensive agents are likely necessary for normotension. Aggressive statin usage is the first-line lipid therapy and concomitant fibrate and nicotinic acid usage should be reserved for failures in primary prevention. J Clin Hypertens (Greenwich). 2011;13:270–274. © 2011 Wiley Periodicals, Inc.

Although there are 28 million patients with diagnosed or undiagnosed diabetes in the United States, there are another 77 million patients with defined prediabetic states at great risk for future diabetes. These prediabetic states include impaired glucose tolerance (IGT), impaired fasting glucose (IFG), and the metabolic syndrome (Third Report of the Adult Treatment Panel National Cholesterol Education Program [NCEP ATP-III]).1 These states individually increase future diabetes risk approximately 5-fold. In combination, any 2 of the 3 prediabetic states together increase future diabetes risk by nearly 20-fold.1 Each alone increases coronary heart disease (CHD) risk by about 2- to 3-fold. This is the same degree of increased CHD risk as is noted with diagnosed diabetes. This being the case, it seems reasonable to mandate the same vigorous CHD risk reduction measures for prediabetes as are already implemented for overtly diabetic patients. Recently, levels of hemoglobin A1c (HbA1c) have become accepted as secondary criteria for the identification of diabetic as well as prediabetic states. While levels of HbA1c >6.5% are generally consistent with overt diabetes, levels between 5.7% and 6.4% are thought to indicate a prediabetic state.3 Nonetheless, there are few prospective data by which to sustain these levels of HbA1c as diagnostic for prediabetes, and the same cautions apply to prediabetes and HbA1c relationships, as have already been noted for diabetes and HbA1c for diagnosis. A significant number of patients are either rapid or slow “glycosylators” and therefore have an undefined but idiosyncratic relationship between ambient glucose levels and HbA1c levels, which precludes conclusions regarding actual glycemia in such individuals. Since diabetes is a disease of glucose levels and not glycosylation rates, HbA1c levels are a secondary diagnostic criteria for diabetes diagnosis and require glucose confirmation for accurate diagnosis. Nevertheless, it should be noted that HbA1c levels <5.5% are clearly noted to have substantially elevated risks for CHD and total mortality, which are further increased little by the progressions into overtly diabetic levels of HbA1c.4

Targets for Blood Pressure in Prediabetic Patients

Almost no randomized and controlled multicenter clinical trials of blood pressure (BP) reduction have been conducted in populations of patients with the metabolic syndrome or other forms of prediabetes. Indeed, there is no recognition of differences in BP targets in prediabetes as compared with overt diabetes. Targets may be thought of as falling into two different classes: namely, in patients with and without concomitant chronic kidney disease (CKD). In the absence of CKD, the Hypertension Optimal Treatment (HOT) trial best defines diastolic BP targets for diabetic patients. In this study, 3 different diastolic targets of 90 mm Hg, 85 mm Hg, and 80 mm Hg were chosen.5 Approximately 19,100 patients were enrolled, of whom 8% had diabetes. In the diabetic cohort, patients randomized to the 80-mm Hg target had a two-thirds reduction in diabetic mortality, whereas those randomized to the 85 mm Hg target had an intermediate benefit. These striking differences in outcomes were clearly statistically and biologically significant and were more impressive than the smaller benefits noted for the same targets in the nondiabetic population of the study.

Systolic pressure treatment targets for the patient with diabetes are somewhat more controversial. The recent Action to Control Cardiovascular Risk in Diabetes (ACCORD) BP trial targeted an intensive treatment arm with a target systolic pressure of <120 mm Hg, where conventional therapy allowed systolic pressures to range as high as 135 mm Hg.6 The primary major adverse cardiac event (MACE) end point of fatal and nonfatal CHD and stroke was not benefited by more intensive systolic BP reduction. Interestingly, there was a significant 41% stroke reduction in the intensive arm (P<.001), which was a prespecified end point. Although adverse therapy–related events were more prevalent among the intensively treated patients, the reduction in stroke may be more than sufficient to provide benefit, which offsets those risks attributable to the increased adverse drug–related events. Prior hypertension trials such as the Systolic Hypertension in Europe (Sys-Eur) and others did not target systolic pressures below 130 mm Hg. Most prior studies, including the UK Prospective Diabetes (UKPDS) BP trial, concluded that systolic pressure reduction to at least that level was clearly beneficial with regard to CHD end points. With stroke, there may be added benefit in lower systolic pressures. In patients with underlying CKD, systolic, diastolic, and mean arterial targets may be lower still. Some have suggested systolic pressures <120 mm Hg to minimize or even arrest the progressive loss of glomerular filtration rate seen in diabetic patients with CKD.7 Others have focused on mean arterial pressures of <93 mm Hg.

Nature of the Antihypertensive Agent(s)

Identification of the best, first-line antihypertensive agent in various treatment populations has been the principal focus of most randomized clinical trials in hypertension. Patients with diabetes were an important population of study for a number of these studies. Many trials used thiazide diuretics as the active comparator. Most studies found various members of the angiotensin-converting enzyme inhibitor or angiotensin receptor blocker classes of antihypertensive agents to be superior to diuretics in secondary parameters such as altered glucose control or lipid levels. On the other hand, the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) found diuretics to be superior agents, compared with calcium channel antagonists or renin-angiotensin system (RAS) inhibitors.8 However, more than one third of patients in the trial were African American and might therefore have had less benefit from RAS agents than from diuretics.

For a time, calcium channel antagonists were thought to be deleterious in patients with diabetes. Numerous subsequent studies have since found them to be just as effective as other, comparative agents. Cardioselective β-adrenergic antagonists seemed to have some deleterious effects on dyslipidemia and insulin resistance, which were not found with the vasodilating β-adrenergic antagonists.9 Nevertheless, all β-adrenergic antagonists do have antiarrhythmic effects, which might explain their apparent protection in the immediate post–myocardial infarction (MI) period and the consequent reduction in post-MI mortality with those agents despite their negative metabolic characteristics. These metabolic characteristics become more important in prediabetic patients because they influence the rate of conversion of prediabetes to overt diabetes. In prediabetic patients, thiazides and β-adrenergic antagonists may thus be less desirable than in patients with completely normal carbohydrate metabolism. RAS inhibitors, calcium channel antagonists, and loop diuretics may be preferred in this highly specialized population.

Much of this discussion is obviated in diabetic as well as prediabetic patient populations, owing to the need for large numbers of antihypertensive agents in both groups. Thus, numerous studies have found that patients with diabetes require 3 to 5 antihypertensive agents to achieve the goals of BP management. For the most part, clinical trials have studied the best monotherapies and, on occasion, the best 2 drug combinations for special populations of patients such as those with diabetes or the metabolic syndrome. Studies with 3 to 5 drugs have as yet to be performed. Since there are a limited number of antihypertensive therapeutic categories, agents from virtually all of those categories are used routinely in patients with diabetes. Discussions of the best single agent or even 2-drug combinations become moot when viewed in the context of the likely patient who will consume agents from nearly all of the major therapeutic categories. Thus, the prediabetic patient with significant hypertension is likely to receive a RAS inhibitor, a calcium channel antagonist, a β-adrenergic antagonist, and a diuretic or loop diuretic. This leaves few remaining categories from which to choose the remaining therapeutic choices if BP remains uncontrolled. Usually, centrally acting agents are used next, although direct-acting vasodilators such as minoxidil may be necessary if moderate or severe hypertension remains.

Management of Dyslipidemia and Hyperlipidemia in Prediabetic States

Prediabetes is a common subtype of patients with obesity. These patients have distinctive forms of dyslipidemia characterized by low high-density lipoprotein (HDL) cholesterol levels and modestly elevated levels of triglyceride-rich lipoproteins. There may or may not be elevations of low-density lipoprotein (LDL) cholesterol levels, although commonly there is a dysmetabolism of triglyceride-rich lipoproteins leading to an overabundance of smaller, denser LDL particles. Thus, there is almost always an increased level of apolipoprotein B containing lipoproteins and, as a result, findings of increased non-HDL cholesterol levels in prediabetic patients, including those with overt metabolic syndrome (NCEP ATP-III). This has led to some considerable controversy in recommendations to reduce the excess risk of CHD seen in prediabetes. There has been no systematic trial of lipid-based interventions to reduce CHD risk in this patient population, although there has been an identification of this subgroup in retrospective post hoc analyses that are, at least, hypothesis-generating in nature.

The lynchpin of CHD risk reduction in the metabolic syndrome and in other prediabetic states is LDL cholesterol reduction. Although not studied with resins or absorption inhibitors, LDL reduction has been studied with statins in numerous trials that have included subpopulations of patients with characteristics of the metabolic syndrome or prediabetes. One of the earliest of these studies was the Air Force/Texas Coronary Atherosclerosis Prevention Study (AFCAPS/TEXCAPS) trial conducted in patients with normal LDL and low HDL cholesterol.10 The intervention was lifestyle plus lovastatin compared with lifestyle alone. The statin lovastatin reduced an expanded MACE end point by 37% as the apparent result of a 25% reduction of LDL cholesterol and a 6% increase in HDL cholesterol. A variety of other statin-based trials such as the Cholesterol and Recurrent Events (CARE) trial, the Scandinavian Simvastatin Survival Study (4S), the Heart Protection Study (HPS), and the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT) all included large subpopulations of patients with diabetes. In each, there was a clear and convincing demonstration of efficacy of statin administration on lipid levels and CHD risk reduction. Another trial, the Collaborative Atorvastatin Diabetes Study (CARDS), was organized but stopped early in 2838 patients with diabetes.11 Treatment with 10-mg atorvastatin reduced MACE events by 37%.

Patients with the metabolic syndrome have not been a prespecified subgroup in many large-scale clinical trials. In the Justification for the Use of Statins in Prevention: An Intervention Trial Evaluating Rosuvastatin (JUPITER), 17,800 patients with normal LDL and slightly elevated C-reactive protein were randomized to rosuvastatin 20 mg or placebo.12 Patients with the metabolic syndrome constituted 41% of the randomized population in the trial and were the largest subgroup. Patients with diabetes were excluded per protocol. This trial was also stopped early with a 47% reduction in MACE events in treated patients having an on-treatment LDL cholesterol level of 54 mg/dL as compared with the placebo control patients having a normal LDL cholesterol of 108 mg/dL. Although this finding is also post hoc in nature and thus hypothesis-generating, it does provide insight into treatment approaches and goals for prediabetic and metabolic syndrome patients.

The NCEP update of 2004 was intended to focus on high-risk patient subpopulations and treatment goals appropriate for those subpopulations.13 Thus, diabetic patients with heart disease were found to be at highest CHD risk and therefore merited greatest LDL cholesterol reduction. For them, a goal of <70 mg/dL was proposed. The same goal was also selected for high-risk patients with multiple risk factors for CHD. For uncomplicated patients with diabetes alone, a higher LDL target of <100 mg/dL was thought appropriate. Since prediabetic patients had the same cardiovascular risk as patients with overt diabetes, and since metabolic syndrome patients have multiple risk factors, those same goals for patients with diabetes that were formulated in the 2004 update are most appropriate. Support for this concept is provided by Pravastatin or Atorvastatin Evaluation and Infection Therapy—Thrombolysis in Myocardial Infarction 22 (PROVE-IT/TIMI 22). In this study, performed in patients with acute coronary syndromes (ACS), who are largely insulin-resistant diabetic or prediabetic patients, patients were randomized to pravastatin 40 mg/dL (on-treatment LDL cholesterol, 95 mg/dL) or atorvastatin 80 mg (on-treatment LDL cholesterol, 62 mg/dL). Atorvastatin resulted in a 16% further reduction in recurring coronary events, supporting the target of <70 mg/dL as the goal for multiple risk factor patients.

Moderately elevated levels of triglyceride-rich lipoproteins and low levels of HDL cholesterol are the signature lesions of obesity and prediabetes. Intervention with fibric acid derivatives to lower triglyceride-rich lipid fractions and raise HDL cholesterol has long been advocated. One trial with gemfibrozil, the Veterans Administration HDL Intervention Trial (VA HIT) study,14 was notable for its performance in largely metabolic syndrome patients. Here, gemfibrozil reduced CHD events by 22%. Unfortunately, this particular fibric acid derivative cannot be used in combination with statins since it blocks glucuronidation of most statins and thereby raises statin blood levels by 50% to 200%, thereby increasing the risk for rhabdomyolysis. This agent has also been effective in other studies of dyslipidemia such as the Helsinki Heart Study. Because of the statin interaction, other fibrates have been advocated for use alone or with statins for combination therapy. Chief among these other fibrates has been fenofibrate. This has been studied in two major trials. The first of these, the Fenofibrate Intervention and Event Lowering in Diabetes (FIELD) study, was a great disappointment to those who anticipated similar benefit as seen with gemfibrozyl.15 Instead, with fenofibrate alone in a trial of 9800 diabetic patients, there was an insignificant 9% treatment benefit to fenofibrate administration. This lack of benefit was attributed to an unfortunately high rate of drop in therapy (approximately 10% of patients) with concomitant statin therapy during the blinded phase of the trial.

To further complicate this issue, the concept of potential incremental efficacy of combination statin-fibrate therapy was tested directly in the ACCORD-Lipid arm, which investigated combination fenofibrate plus 40 mg simvastatin vs 40 mg simvastatin alone in a total of 5518 high-risk patients with type 2 diabetes.16 There was no difference in MACE end points between the two arms of the study, indicating no treatment benefit to added fibrates beyond that achieved by statin alone. Since the average triglyceride level in randomized patients of this trial was 160 mg/dL, it seems likely that most diabetic patients would not benefit from fibrate therapy in combination with statins. Whether the conclusion applies to patients with very high triglyceride levels is unclear, but there did seem to be an insignificant trend toward benefit in the subgroup on post hoc analysis.

Finally, there has been much epidemiologic evidence to implicate low HDL cholesterol in the pathogenesis of the excess CHD risk in patients with diabetes. Unfortunately, there are no modern randomized, multicenter large-scale interventional trials to test therapies such as niacin for isolated low HDL cholesterol in these patients. Since niacin raises HDL cholesterol more than statins, it seems reasonable to consider it for overt high-risk diabetic patients failing statin interventions alone. However, this may not as yet be the case for patients with prediabetes since niacin worsens insulin resistance and raises HbA1c and fasting glucose levels. It may therefore hasten the deterioration to overt diabetes in patients with prediabetes and thus be counterproductive in these patient populations. Niacin should be reserved for use in combination with statins for secondary prevention of CHD events in patients failing a course of adequate statin therapy.

It seems clear that cardiovascular risk reduction in patients with prediabetes should proceed along the same pathways as risk reduction for patients with overt diabetes since the risks of both populations are identical as are the pathogenic mechanisms underlying the accelerated atherosclerosis of both states of dysglycemia. Aggressive control of hypertension must be addressed with the use of all available agents, especially those not worsening the underlying metabolic defects of prediabetes. For the control of dyslipidemia, vigorous statin interventions seem warranted. Additional lipid interventions are justified only if substantial, residual dyslipidemia remains despite adequate statin therapy and especially if intercurrent CHD events have appeared while on station therapy alone. Dysglycemia is a legitimate treatment target for the mitigation of microvascular complications, which arise in the prediabetic state. Dysglycemia modification does not, however, appear to mitigate the excess CHD risk in overt diabetes and likely will not succeed in prediabetes either.