Predictors of Stroke in Patients With Impaired Glucose Tolerance
Results From the Nateglinide and Valsartan in Impaired Glucose Tolerance Outcomes Research Trial
Background and Purpose—Risk factors for stroke are well-established in general populations but sparsely studied in individuals with impaired glucose tolerance.
Methods—We identified predictors of stroke among participants with impaired glucose tolerance in the Nateglinide and Valsartan in Impaired Glucose Tolerance Outcomes Research (NAVIGATOR) trial. Cox proportional-hazard regression models were constructed using baseline variables, including the 2 medications studied, valsartan and nateglinide.
Results—Among 9306 participants, 237 experienced a stroke over 6.4 years. Predictors of stroke included classical risk factors such as existing cerebrovascular and coronary heart disease, higher pulse pressure, higher low-density lipoprotein cholesterol, older age, and atrial fibrillation. Other factors, including previous venous thromboembolism, higher waist circumference, lower estimated glomerular filtration rate, lower heart rate, and lower body mass index, provided additional important predictive information, yielding a C-index of 0.72. Glycemic measures were not predictive of stroke. Variables associated with stroke were similar in participants with no prior history of cerebrovascular disease at baseline.
Conclusions—The most powerful predictors of stroke in patients with impaired glucose tolerance included a combination of established risk factors and novel variables, such as previous venous thromboembolism and elevated waist circumference, allowing moderately effective identification of high-risk individuals.
Individuals with impaired glucose tolerance (IGT) are at increased risk for developing type 2 diabetes mellitus (T2DM) and cardiovascular disease, including stroke.1 Individuals with IGT exhibit many cardiovascular disease risk factors found in T2DM.2 However, longitudinal studies of well-characterized IGT populations have not been available to examine predictors of stroke. The Nateglinide and Valsartan in Impaired Glucose Tolerance Outcomes Research trial (NAVIGATOR) provides the opportunity to establish the factors associated with increased risk of stroke in IGT.3
NAVIGATOR was a multi-continental randomized, double-blind, placebo-controlled trial to determine whether treatment with nateglinide and valsartan, respectively (2×2 factorial design), would reduce progression to T2DM and incident cardiovascular events, respectively, in participants with IGT but not T2DM.3 A total of 9306 participants were followed for a median of 6.4 years. Doses of nateglinide and valsartan could be adjusted based on side-effects but not glucose levels or blood pressure readings, respectively. Patients with IGT (2-hour postchallenge glucose value, 7.8–11.0 mmol/L and fasting glucose, 5.3–6.9 mmol/L) were eligible if they had ≥1 cardiovascular risk factor and were aged ≥55 years or established cardiovascular disease and aged ≥50 years. Relevant qualifying cardiovascular disease risk factors are provided in previous publications.3 The trial was approved by local ethics committees, and participants provided written informed consent. The end point of stroke was adjudicated by an independent committee and trial definitions are provided in Table I in the online-only Data Supplement.
Cox proportional-hazard regression models were developed for occurrence of stroke. Thirty-nine candidate variables were included in the forward model selection (P<0.05; Table 1). Randomized treatments (valsartan versus placebo and nateglinide versus placebo) and race were forced into the model. Discrimination ability was assessed with optimism-corrected C-statistics. P values <0.05 were considered statistically significant, and all P values were 2-sided. Analyses were conducted independently at the Duke Clinical Research Institute (Durham, NC) using SAS version 9.2 software (SAS Institute Inc, Cary, NC).
Stroke occurred in 237 of the 9306 NAVIGATOR participants during a median follow-up of 6.4 years (202 nonhemorrhagic strokes). Among subjects without history of cerebrovascular disease (n=8570), 193 (2.3%) experienced a stroke, compared with 44 (6.0%) out of 736 with a history of cerebrovascular disease.
Major independent predictors of stroke included previous venous thromboembolism (pulmonary embolism or deep venous thrombosis; the strongest predictor), previous cerebrovascular disease (stroke or transient ischemic attack), higher pulse pressure, and higher waist circumference (Table 2). Other independent predictors were previous coronary heart disease, higher low-density lipoprotein cholesterol, lower estimated glomerular filtration rate, lower heart rate (below a threshold of 70 bpm), lower body mass index, older age, the presence of atrial fibrillation, and black race versus all other races. No glycemic measure was associated with stroke. The C-statistic for the prognostic model was 0.72. Results were similar in the 8570 patients without a history of stroke at baseline (C-statistic 0.72; Table II in the online-only Data Supplement).
Predictors of nonhemorrhagic stroke (202 events) were similar to predictors of all strokes. The variables most strongly associated with nonhemorrhagic stroke were cerebrovascular disease, previous venous thromboembolism, coronary heart disease, higher pulse pressure, and lower estimated glomerular filtration rate (Table III in the online-only Data Supplement).
Cigarette smoking, hypertension, age, T2DM, cholesterol, low high-density lipoprotein cholesterol, atrial fibrillation, and carotid stenosis are the recognized stroke risk factors.4 Results from NAVIGATOR largely reflected this, although surprisingly current smoking status was not predictive of stroke. Prior venous thromboembolism (pulmonary embolism or deep venous thrombosis) emerged as the strongest stroke risk factor, but this remains unexplained. Sensitivity analyses limited to only nonhemorrhagic strokes and to the subset of participants with no history of cerebrovascular disease at baseline were largely similar to the main analysis. Measures of glycemia were not predictive of stroke, possibly reflecting the relatively narrow range of dysglycemia in NAVIGATOR.
In summary, independent predictors of stroke in patients with IGT included not only established risk factors, but also other important variables, such as previous venous thromboembolism and elevated waist circumference.
Sources of Funding
Nateglinide and Valsartan in Impaired Glucose Tolerance Outcomes Research (NAVIGATOR) was sponsored by Novartis Pharma.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.113.001177/-/DC1.
- Received February 20, 2013.
- Revision received June 21, 2013.
- Accepted June 24, 2013.
- © 2013 American Heart Association, Inc.
- Laakso M,
- Lehto S
- Califf RM,
- Boolell M,
- Haffner SM,
- Bethel M,
- McMurray J,
- Duggal A,
- et al
- Roger VL,
- Go AS,
- Lloyd-Jones DM,
- Benjamin EJ,
- Berry JD,
- Borden WB,
- et al