National Trends in the Achievement of Recommended Strategies for Stroke Prevention in U.S. Adults With Type 2 Diabetes, 2001–2018

In the U.S., stroke ranks fifth among all causes of death and is the leading neurological cause of long-term disability, which caused a compelling economic burden of ∼$49.8 billion in 2016–2017 (1,2). As one of the leading causes of morbidity and mortality in individuals with diabetes, stroke is among the largest contributors to health care costs of diabetes (3–7). The median annual health care cost per patient with type 2 diabetes (T2D) for stroke was 322% higher compared with that for a T2D patient without stroke (8). On average, treating a patient with both T2D and stroke resulted in an annual health care cost increase from $3,755 to $13,460 in comparison with treating a patient with T2D alone (8). Common conditions coexisting with T2D such as hypertension and dyslipidemia are risk factors for stroke, but diabetes itself confers an independent, major risk for both incident and recurrent stroke (3,9). Compared with those without diabetes, individuals with diabetes have a twofold excess risk for incident stroke and an ∼45% increased risk for recurrent stroke (3,10,11). Individuals with diabetes have a 25% reduction in favorable outcomes following a stroke; people with diabetes and stroke are less able to function independently in activities of daily living and are more likely to develop vascular cognitive impairment and die of stroke (3).

The American Diabetes Association, American Heart Association, and American Stroke Association have recommended effective intervention strategies for primary and secondary prevention of stroke and early treatment of acute ischemic stroke in people with diabetes (12–18). Most of these interventions, when implemented either in an individual or in an integrated way, are cost-effective (19–25). However, the age-standardized incidence of stroke among people with diabetes remained substantial—53 cases per 10,000 U.S. adults with diagnosed diabetes per year in 2010. This is higher than the incidence of acute myocardial infarction, amputation, end-stage renal disease, or death from hyperglycemic crisis (26). Moreover, the prevalence of stroke among U.S. adults with newly diagnosed T2D remained unchanged from 1988 to 2018 (6.8% to 6.4%) (27).

Recognizing the extent to which U.S. adults with T2D achieve current guideline-recommended strategies for primary and secondary prevention of stroke is critical for clinical management and health care policy planning, but it remains unclear. A better understanding of the gaps in comprehensive diabetes care may help identify opportunities to improve clinical outcomes for patients with diabetes. The objective of this study was to assess the national prevalence of and trends in the achievement of current guideline-recommended strategies for primary and secondary prevention of stroke among U.S. adults with T2D. To accomplish this, we conducted serial cross-sectional analyses of data from the 2001–2018 National Health and Nutrition Examination Survey (NHANES).

NHANES is a nationally representative, cross-sectional survey designed to monitor the health of the U.S. population. During each survey cycle, a sample of participants is selected from the civilian, noninstitutionalized U.S. population with use of a complex, stratified, multistage, and probability cluster sampling design. Data are collected from participants through in-home interviews, physical examinations, and laboratory tests administered by the National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention (28). Study protocols were approved by the NCHS Research Ethics Review Board, and written informed consent was obtained from all NHANES participants. We analyzed data from nine continuous survey periods of NHANES (2001–2002, 2003–2004, 2005–2006, 2007–2008, 2009–2010, 2011–2012, 2013–2014, 2015–2016, and 2017–2018). The response rate for the NHANES cycles between 2001 and 2018 ranged from 52% to 84% (29).

We restricted our analytic sample to participants ≥45 years of age with T2D. To select participants with T2D, we first identified participants with self-reported diabetes or use of glucose-lowering medications and then excluded those with diabetes diagnosed before 30 years of age or who initiated insulin within 1 year of diabetes diagnosis. To evaluate primary and secondary prevention for stroke, we further stratified our sample by stroke history, defined as self-reported stroke.

Based on recommendations by the American Diabetes Association, American Heart Association, and American Stroke Association (12–17,30–32) and data availability from NHANES, we defined eight outcomes to create a cascade of current guideline-recommended strategies (Supplementary Table 1) for stroke prevention among patients with T2D in each NHANES period. For optimal glycemic control, an individualized HbA_1c level target was defined between 7.0% and 8.0%, depending on a participant’s age and presence of stroke history. For participants aged 45–64 years, target HbA_1c level was <7.0% in the absence of stroke history or <8.0% in the presence of stroke history. For participants aged ≥65 years, target HbA_1c level was <7.5% in those without stroke history or <8.0% in those with stroke history. Optimal blood pressure (BP) control was defined as systolic/diastolic BP level <140/90 mmHg. All individuals with T2D aged ≥45 years were recommended for statin treatment. The use of aspirin or clopidogrel for prophylaxis of atherosclerotic cardiovascular disease (ASCVD) was considered for people whose risk was high (10-year ASCVD risk >10%). Smoking cessation advised to all individuals. Optimal weight control was defined as a BMI target <25 kg/m². The use of an ACE inhibitor or angiotensin receptor blocker (ARB) was recommended for diabetes patients with hypertension and/or albuminuria.

We used the mean of all available BP measurements to estimate each participant’s systolic and diastolic BP levels. We defined hypertension based on self-reported hypertension or elevated mean BP levels ≥140/90 mmHg. Participants were categorized as currently nonsmoking if they had serum cotinine levels ≤10 ng/mL or self-reported no cigarette or tobacco use. Albuminuria was defined as a urine albumin-to-creatinine ratio ≥30 mg/g. Pharmacological treatments such as statins, aspirin, clopidogrel, ACE inhibitors, and ARBs were identified based on NHANES prescription medication files, which contain detailed personal interview information on use of prescription medications.

We performed analyses with SAS software (version 9.4) to account for the complex survey design of NHANES and used appropriate sample weights to make our results representative of the civilian, noninstitutionalized U.S. adult population ≥45 years of age with T2D. We report the estimated percentage and 95% CI for categorical variables and the mean or geometric mean and 95% CI for continuous variables. We estimated the population number and proportion of U.S. adults ≥45 years of age with T2D who achieved recommended strategies for stroke prevention, with stratification by stroke history, using the combined data of nine continuous survey cycles of NHANES and the data of each 2-year survey cycle. We calculated the population number estimates using the annual estimates of the resident population for selected age-groups in 2018 provided by the U.S. Census Bureau Population Division. We used logistic regression models to assess changes in the achievement of recommended strategies for stroke prevention across the nine survey cycles (2001–2018), assuming that the logit of the prevalence of adherence to these recommendations was a linear function of time. A two-sided P value of <0.05 was considered statistically significant.

There were 4,834 NHANES participants aged ≥45 years with T2D in 2001–2018. Of these, 4,300 had no stroke history and 534 had a stroke history. Compared with participants without a stroke history, those with a stroke history were older, more likely to be non-Hispanic White or Black, and less likely to be uninsured. They were also more likely to have a longer duration of diabetes, use insulin, and have hypertension, albuminuria, and higher ASCVD risk. Summary statistics of demographic and clinical characteristics of the study population by stroke history in NHANES 2001–2018 are presented in Supplementary Tables 2–4.

In 2018, an estimated 14.4 million U.S. adults ≥45 years of age with T2D had no stroke history and 1.7 million had a stroke history (Table 1). During the 2001–2018 period, among those without a stroke history, the proportion who achieved current guideline-recommended strategies for primary prevention of stroke ranged from 8.4% (95% CI 6.8–10.1) for treatment with aspirin/clopidogrel in those with high ASCVD risk to 80.5% (95% CI 78.8–82.2) for nonsmoking. Among those with a stroke history, the proportion who achieved current guideline-recommended strategies for secondary prevention of stroke ranged from 11.8% (95% CI 8.7–14.8) for achievement of the BMI target to 80.0% (95% CI 74.9–84.9) for achievement of the HbA_1c target. Compared with U.S. adults ≥45 years of age with T2D and without stroke history, those with stroke history were more likely to achieve the HbA_1c target (80.0% vs. 60.2%), take a statin (61.4% vs. 54.3%), and use aspirin/clopidogrel if at high ASCVD risk (24.8% vs. 8.4%) and were less likely to achieve the BP target (63.4% vs. 70.6%) (Table 1).

Among U.S. adults ≥45 years of age with T2D, the proportions who achieved the HbA_1c and BP targets were unchanged over time in the population without (Table 2) and with (Table 3) stroke history during the 2001–2018 period. Across the 2001–2018 period, the proportion who achieved the HbA_1c target was consistently higher in the population with stroke history than those without stroke history, while the proportion of achieving the BP target was generally lower in the population with stroke history than those without stroke history, except in 2009–2014 (Fig. 1).

The proportion taking a statin increased significantly over time among both the population without and population with stroke history. Among those without stroke history, the proportion increased from 32.9% (95% CI 26.9–39.0) in 2001–2002 to 58.1% (95% CI 51.4–64.8) in 2017–2018 (P_trend < 0.0001), and among those with stroke history it increased from 39.4% (95% CI 23.9–55.0) in 2001–2002 to 62.1% (95% CI 44.8–79.4) in 2017–2018 (P_trend = 0.01) (Tables 2 and 3 and Fig. 1).

Approximately 72.8% (95% CI 70.7–75.1) of U.S. adults ≥45 years of age with T2D and without stroke history and 88.6% (95% CI 85.1–92.2) with stroke history had a 10-year ASCVD risk >10% and thus received recommendation for treatment with aspirin/clopidogrel (Supplementary Table 2). Among these patients, the proportion who took aspirin/clopidogrel increased over time but the change was not statistically significant: from 4.0% (95% CI 0.7–7.4) in 2001–2002 to 12.7% (95% CI 4.9–20.4) in 2017–2018 for those without stroke history (P_trend = 0.06) and from 3.1% (95% CI 2.0–4.3) in 2001–2002 to 27.4% (95% CI 10.6–44.1) in 2017–2018 for those with stroke history (P_trend = 0.08) (Tables 2 and 3 and Fig. 1).

The proportion of nonsmokers increased significantly, from 78.2% (95% CI 71.6–84.9) in 2001–2002 to 82.9% (95% CI 76.4–89.4), in 2017–2018 for the population without stroke history (P_trend = 0.01) but did not change significantly for those with stroke history (P_trend = 0.35) (Tables 2 and 3 and Fig. 1).

The proportion who achieved the BMI target declined significantly from 20.9% (95% CI 14.6–27.2) in 2001–2002 to 10.1% (95% CI 5.8–14.3) in 2017–2018 in the population without stroke history (P_trend < 0.001) but did not change significantly in that with stroke history (P_trend = 0.56) (Tables 2 and 3 and Fig. 1).

Approximately 76.2% (95% CI 74.1–78.9) of U.S. adults ≥45 years of age with T2D and without stroke history and 88.1% (95% CI 84.5–91.7) of those with stroke history had hypertension and thus were recommended to receive an ACE inhibitor or ARB (Supplementary Table 2). Among these patients, the proportion who used an ACE inhibitor or ARB increased significantly from 48.2% (95% CI 41.2–55.1) in 2001–2002 to 63.2% (95% CI 56.5–69.8) in 2017–2018 among those without stroke history (P_trend = 0.04) but did not change significantly among those with stroke history (P_trend = 0.49) (Tables 2 and 3 and Fig. 1).

Approximately 23.7% (95% CI 21.8–25.5) of U.S. adults ≥45 years of age with T2D and without stroke history and 38.1% (95% CI 32.5–43.7) with stroke history had hypertension and albuminuria and were thus recommended to receive an ACE inhibitor or ARB (Supplementary Table 2). Among these patients, the proportion who used an ACE inhibitor or ARB was not found to change significantly from 2001–2002 to 2017–2018 among either those without stroke history (P_trend = 0.31) or those with stroke history (P_trend = 0.72) (Tables 2 and 3 and Fig. 1).

Results of analyses of the nationally representative, population-based, multiyear samples of U.S. adults ≥45 years of age with T2D and without a stroke history showed that use of a statin, refraining from smoking, and use of an ACE inhibitor or ARB for the primary prevention of stroke increased significantly, but there was a significant decline in the achievement of the BMI target. Among those with a stroke history, only the use of a statin for the secondary prevention of stroke increased significantly. Overall, the achievement of the eight strategies for stroke prevention was generally higher in the population with a stroke history (secondary prevention of stroke) than in the population without stroke history (primary prevention of stroke). However, the level of achieving current guideline-recommended strategies for primary and secondary prevention of stroke was suboptimal. In 2001–2018, only two recommended strategies for stroke prevention reached an achievement rate of ≥80% among U.S. adults ≥45 years of age with T2D: 80.5% with not smoking for primary stroke prevention and 80.0% with well-controlled HbA_1c for secondary stroke prevention.

The NHANES data have been used to assess levels of achievement of risk factor control goals and receipt of preventive practices for ASCVD among U.S. adults with newly diagnosed T2D, diagnosed diabetes, or undiagnosed diabetes (27,30,31,33–40). Approximately 25%–89% of U.S. adults with diagnosed diabetes achieved HbA_1c targets, 39%–77% achieved BP targets, 4%–20% achieved BMI targets, 28%–75% used a statin, 12%–64% used an ACE inhibitor or ARB if there was albuminuria, and 76%–86% were not current smokers. Nonetheless, none of these studies included stratification of analyses by stroke history of patients to examine the status of implementing primary and secondary prevention strategies for stroke. Direct comparisons of our study results with those of previous studies could not be made because of discrepancies in the analytic nature and definition of the study populations, outcomes, and survey cycles/periods.

Our study has a number of strengths. We used the contemporary NHANES data to provide unbiased, nationally representative estimates for the entire U.S. noninstitutionalized adult population ≥45 years of age with T2D. We described demographic and clinical characteristics of patients with T2D without and with stroke history in detail including diabetes duration, HbA_1c, BP, ASCVD risk, medication use, and comorbidity status. The assessment of these data were based on objective, rigorous, and systematic measurement. To our knowledge, our study is the first to provide empirical data about national trends in prevalence of achieving current guideline-recommended risk factor control goals, behavioral modifications, and pharmacotherapies for primary and secondary prevention of stroke among U.S. adults with T2D.

Several limitations of this study deserve mention. First, we used cross-sectional survey examination and laboratory data from noninstitutionalized civilian adults to approximate stroke preventive care for the U.S. resident population. The analysis may thus underrepresent some segments of the U.S. population. The use of cross-sectional data provides a snapshot of preventive practices for stroke during the survey periods. Second, diabetes was defined as self-reported diabetes or use of glucose-lowering medications. Self-reported diabetes is known to be highly specific (41,42), although some misclassification may occur. Self-reported diabetes status and other variables might be subject to diagnosis, recall, and social desirability bias. Third, as with previous research, we used more recent guidelines to define risk factor control goals and prevention practices and applied those definitions retrospectively, albeit consistently, across all survey periods. We framed our analyses as how many people are meeting current guideline-recommended strategies to get an overall picture of comprehensive stroke prevention and care over time. Fourth, there was a small sample size for adults with T2D with stroke history in the study cohort, which may provide limited power to make inferences about this population. Moreover, although there seemed to be obvious changes in the prevalence of achievement of some stroke prevention strategies among those with stroke history between the NHANES 2001–2002 wave and the NHANES 2003–2004 wave, the relative stability in prevalence over time after the NHANES 2003–2004 wave might in part contribute to the insignificant changing trend over time from data being assessed longitudinally across nine survey cycles. Lastly, clinical health outcomes were not assessed in this study. Although we found improvements in trends of achieving some current guideline-recommended strategies for stroke prevention, the causal relationship between such improvements and better health outcomes could not be established due to the nature of the cross-sectional study design.

In summary, there were improvements in the achievement of current guideline-recommended strategies for prevention of stroke among U.S. adults with T2D between 2001 and 2018. However, the prevalence of not achieving implementation of current guideline-recommended behavior, treatment goals, or pharmacotherapies for stroke prevention remained high—ranging from 20% to 92%. The increased level of achieving current guideline-recommended strategies for stoke prevention may reduce the risk of stroke among people with diabetes. There is substantial room to improve implementation of strategies for primary and secondary prevention of stroke among adults with T2D, especially weight control, aspirin/clopidogrel treatment for those at high ASCVD risk, statin treatment, and glycemic and BP control. Considerable work remains to be done by health systems, health care providers, and patients to increase recognition of the importance of preventive strategies for stroke and other ASCVD among individuals with T2D.

Funding. The project described was supported by the Diabetes Interdisciplinary Studies (DISP) program by the National Institute of Diabetes and Digestive and Kidney Diseases grant P30DK092926 (Michigan Center for Diabetes Translational Research). We also thank Dr. Huang-Tz Ou from the National Cheng Kung University for her assistance and the Ministry of Science and Technology in Taiwan for the financial support (grant no. 109-2320-B-006-047-MY3).

The sponsors had no role in the design or conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.

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

Author Contributions. S.K., W.H.H., L.D.L., and W.Y. contributed to study concept and design. All authors contributed to acquisition, analysis, or interpretation of data. S.K., C.-T.Y., and W.Y. contributed to drafting of the manuscript. W.H.H. and L.D.L. contributed to critical revision of the manuscript for important intellectual content. S.K., C.-T.Y., and W.Y. contributed to statistical analysis. W.Y. contributed administrative, technical, and material support. L.D.L. and W.Y. contributed to study supervision. S.K., C.-T.Y., and W.Y. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.