Frequency and Determinants of Lipid Testing in Ischemic Stroke and Transient Ischemic Attack
Findings From Get With The Guidelines–Stroke
Background and Purpose—National guidelines recommend lipid testing for all patients with ischemic stroke and transient ischemic attack. This study examined the frequency and predictors for in-hospital low-density lipoprotein testing using data from a nationwide stroke registry.
Methods—Between 2003 and 2008, Get With The Guideline–Stroke (GWTG–Stroke) hospitals (n=981) contributed 479 284 consecutive ischemic stroke or transient ischemic attack admissions. Logistic regression models were used to determine patient and hospital characteristics associated with lipid testing.
Results—Frequency of LDL measurement increased from 54.3% in 2003 to 81.9% in 2008 (P<0.001), the adjusted OR for LDL measurement was 1.23 per additional calendar year (95% CI, 1.18 to 1.29; P<0.001). The frequency of LDL measurement also increased with longer hospital program participation; the adjusted OR was 1.17 per additional year of GWTG–Stroke participation (95% CI, 1.12 to 1.23; P<0.001). LDL measurement was lower in women, nonsmokers, those with atrial fibrillation, those with a history of stroke or transient ischemic attack, and in those with transient ischemic attack (versus ischemic stroke; all P<0.001). LDL ≥100 mg/dL was seen in 52.1% of those tested, including in 35.5% of patients already prescribed lipid-lowering therapy before admission.
Conclusions—Rates of LDL measurement in hospitalized patients with ischemic stroke and transient ischemic attack have improved dramatically in this large quality improvement program, although disparities in testing still exist. Testing frequently revealed an LDL level that could prompt a change in clinical management.
Lipid testing and treatment is well established for secondary prevention of ischemic cardiovascular and cerebrovascular disease. Studies show that HMG-CoA reductase inhibitors (statins) prevent ischemic stroke in persons with dyslipidemia, coronary heart disease, or stroke.1,2 A recent meta-analysis of randomized trials of statins, including >165 000 individuals, showed that each 39-mg/dL decrease in low-density lipoprotein (LDL) decreased the relative risk for stroke by 21.1%.1 Most of these patients did not have a history of stroke or transient ischemic attack (TIA) at the time of enrollment, however. The recent Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) study showed that patients with stroke randomized to 80 mg atorvastatin per day had fewer recurrent strokes as well as a substantially fewer coronary events.3 The American Heart Association/American Stroke Association (AHA/ASA) advises in-hospital lipid testing with initiation of statin therapy for most patients with ischemic stroke and TIA with elevated cholesterol, comorbid coronary artery disease, or stroke or TIA of atherosclerotic origin.4–6
Previous stroke registries show that many hospitalized patients with ischemic stroke do not receive adequate investigation for modifiable risk factors.7 In particular, prior studies show that lipid levels are not tested in many patients despite the accumulating evidence that lipid management is important for stroke prevention8,9 and guidelines that recommend lipid testing in hospitalized patients with stroke.6,10 There are few data on patient and hospital characteristics predictive of lipid testing in patients hospitalized with ischemic stroke and TIA. Furthermore, little is known on whether lipid testing is increasing in patients hospitalized with stroke, especially given the recent changes in clinical guidelines.
The Get With The Guidelines–Stroke (GWTG–Stroke) program is a national stroke registry and quality improvement program designed to foster better adherence to guideline-based stroke care in patients hospitalized with stroke and TIA.7,11 We sought to determine the frequency of lipid testing in hospitals participating in the GWTG–Stroke program and whether testing is increasing over time and with longer program participation. Additionally, we sought to determine patient and hospital characteristics associated with LDL measurement and with the presence of elevated LDL levels that could prompt consideration of lipid-lowering therapy (LLT).
Subject Population and Study Measurements
The GWTG–Stroke program has been previously described.7,11 Participating hospitals use an Internet-based Patient Management Tool (Outcome Sciences Inc, Cambridge Mass) to enter data, receive decision support, and obtain feedback through on-demand reports of performance on quality measures. One of the 7 key performance measures of the GWTG–Stroke program is discharge on LLT based on existing AHA/ASA and National Cholesterol Education Panel Adult Treatment Panel III (NCEP-ATPIII) guidelines.12 Furthermore, measurement of LDL is a quality measure that is also monitored in the program.7 Measurement of the LDL cholesterol fraction is a critical component of the GWTG–Stroke quality measure for lipid testing because LDL is the primary target for goal-directed therapy according to NCEP-ATPIII guidelines.12 An LDL target of <100 mg/dL was recommended in 1999 AHA/ASA guidelines for secondary prevention of stroke,13 and adherence to NCEP-ATPIII guidelines was recommended in the 2006 update of the American Heart Association guidelines.4
Hospitals were instructed to record data from consecutive stroke and TIA admissions. Case ascertainment was through clinical identification during hospitalization, retrospective identification by International Classification of Diseases, 9th Revision codes, or both. The eligibility of each case was confirmed at chart review before abstraction. Trained hospital personnel abstracted data, including LDL and other lipid levels if measured, using the Internet-based Patient Management Tool with standardized data definitions and detailed coding instructions. The Internet-based system performs checks to ensure that the reported data are complete and internally consistent. In addition, data quality is monitored for completeness and accuracy.
Only patients with ischemic stroke and TIA were included in this analysis. Between April 1, 2003, and September 30, 2008, there were 1215 hospitals that contributed data on 727 564 stroke and TIA discharges. Patients with hemorrhagic stroke or unknown stroke type were excluded (n=126 353). Patients who died, were discharged to hospice care, were transferred to another acute care hospital, left against medical advice, or in whom discharge status was missing were excluded because these patients would not be eligible for LLT at discharge and therefore lipid testing may not have been warranted (n=68 784). Therefore, there were 532 427 eligible patients. Among the eligible patients, 61 458 were excluded because of missing data on demographics or medical history (11.5%) leaving 470 969 patients from 981 hospitals for analysis.
Each participating hospital received either human research approval to enroll cases without individual patient consent under the common rule or a waiver of authorization and exemption from subsequent review by their Institutional Review Board. Outcome Sciences, Inc serves as the data collection and coordination center for GWTG. The Duke Clinical Research Institute serves as the data analysis center and has Institutional Review Board approval to analyze the aggregate deidentified data for research purposes.
Using χ2 tests for categorical variables and Wilcoxon rank sum test for continuous variables, patient characteristics and hospital characteristics were compared among those with or without measurement of LDL and among those with and without elevated LDL defined as ≥100 mg/dL. LDL was counted as “measured” in patients in which very high levels of triglycerides (>400 mg/dL) prevented indirect calculation of LDL using the Friedewald formula14; these patients were excluded from analyses of LDL levels, however. Univariate changes in the use of lipid testing over time were tested using the Cochran-Mantel-Haenszel row-mean score test for categorical variables.
Multivariable logistic regression models were performed to determine the independent predictors of LDL measurement and LDL ≥100 mg/dL. The generalized estimating equations approach was used to account for within-hospital clustering.15 Covariates in all the models included the baseline patient characteristics age (as a continuous variable), sex, race (categorized as white, black, Hispanic, or other), type of index event (ischemic stroke versus TIA), comorbid medical conditions (atrial fibrillation, prosthetic heart valve, previous stroke/TIA, coronary artery disease or prior myocardial infarction, carotid stenosis, diabetes, peripheral vascular disease, hypertension, dyslipidemia, and smoking), use of LLT before admission, and hospital characteristics (bed size, annual number of stroke discharges, academic teaching status, and geographical region). Hospital bed size was entered as a continuous variable, whereas annual number of stroke discharges were categorized as 0 to 100, 101 to 300, or >300. Hospital teaching status and hospital region (defined as northeast, midwest, south, or west) were determined using statistics published by the American Hospital Association.16 Two time variables were included in the models: calendar time, reflecting secular trends in lipid testing, and time in the GWTG–Stroke program, reflecting the impact of longer duration of program participation.11 Time in GWTG–Stroke was categorized as the number of quarters (ie, 90 days) of program participation after the hospital had submitted the first 30 admissions, which were used as a baseline; the linear trend by quarter of participation was then tested. Because hospitals entered the GWTG–Stroke program in a staggered fashion, duration of program participation is not highly correlated with calendar time, allowing estimates of the effects of calendar time and time in GWTG–Stroke in the same model.11
All probability values are 2-sided. Analyses were performed using SAS Version 9.13 (SAS Institute, Cary, NC).
The study population consisted of 331 080 patients with ischemic stroke and 139 889 patients with TIA (Table 1). Mean age was 70.3±14.5 years and 53.6% were women. Admissions were submitted by 981 hospitals (median hospital bed size 293, interquartile range 185 to 421); 485 hospitals (49.4%) were identified as teaching hospitals based on data from the American Hospital Association.
Overall, LDL was measured in a high proportion of patients (74.9%). The rates of LDL testing by hospital varied substantially (10th to 90th percentile 44.0% to 89.1%). The proportion of cases with LDL measurement was lower in 2003 (54.3%) but increased significantly over time to a high of 89.1% in 2008 (Figure 1). Characteristics associated with LDL measurement in univariate analysis are shown in Table 1. Patients without LDL measurement were more likely to be older, female, present with TIA rather than ischemic stroke, have more medical comorbidities, including prior cardiovascular disease, be discharged to a skilled nursing facility, and to be admitted to smaller nonacademic hospitals with fewer stroke discharges per year. Patients with known dyslipidemia were more likely to have LDL measured, but patients on LLT were less likely to have LDL measured. Multivariable-adjusted OR for lipid testing are shown in Table 2. Most of the univariate predictors were significant in the multivariable-adjusted analysis, and the directions of effect were the same. Both calendar time and the hospital duration of participation in GWTG–Stroke were associated with an increased likelihood of LDL measurement in the multivariable analysis (Table 2). Exploratory analyses to see whether the characteristics associated with LDL measurement were changing over time failed to show clinically meaningful differences.
Among patients who had LDL measured, the mean LDL was 106.5±39.9 mg/dL; the distribution of LDL levels is shown in Figure 2. There were 16.8% with LDL <70 mg/dL, 30.8% with LDL 70 to 99 mg/dL, 27.5% with LDL 100 to 129 mg/dL, 15.6% with LDL 130 to 159 mg/dL, and 9.3% with LDL >160 mg/dL. Thus, of those tested, 52.4% had elevated LDL (ie, >100 mg/dL). The mean LDL was 107.5±40.3 mg/dL in patients with ischemic stroke and 104.1±38.6 mg/dL in patients with TIA (P<0.001). Among patients not previously taking LLT, the mean LDL was 114.7±39.2 mg/dL compared with 93.4±37.4 mg/dL for patients taking LLT before hospitalization (P<0.001).
Univariate analyses of characteristics associated with LDL ≥100 mg/dL are shown in Table 3, and multivariable-adjusted analysis results are shown in Table 4. Patients with cardiovascular comorbidities and risk factors, including previous stroke, myocardial infarction, and diabetes mellitus, were less likely to have LDL ≥100 mg/dL (Table 4). A history of dyslipidemia was associated with increased adjusted odds of LDL ≥100 mg/dL, whereas preadmission LLT was associated with substantially decreased adjusted odds of LDL ≥100 mg/dL (Table 4). Even so, 35.5% of the patients taking LLT before admission had LDL ≥100 mg/dL. Testing identified LDL ≥100 mg/dL in 62.6% of patients not previously taking LLT, of whom 84 284 of 134 090 (62.9%) were newly started on LLT by discharge.
We found that the frequency of LDL measurement in patients with ischemic stroke and TIA was initially low but rapidly increased in the context of hospital participation in a quality improvement program. By contrast, in a Michigan stroke registry in 2002, the frequency of lipid testing was only 38.8% among hospitalized patients with ischemic stroke and TIA not previously taking LLT.8 A combined analysis of data from Paul Coverdell National Acute Stroke Registries, collected in 2001 to 2002, also showed low levels of in-hospital LDL testing.9 The multi-center Vienna Stroke Registry found that although 90% of hospitalized ischemic stroke and TIA survivors enrolled in 1998 to 2001 had measurement of total cholesterol, only 48% had measurement of LDL.17 The somewhat higher initial frequency of LDL measurement in the GWTG–Stroke program in 2003 may reflect the voluntary enrollment of hospitals interested in participating in a quality improvement program.
Calendar time and duration of participation in GWTG–Stroke were both independently associated with increased odds of LDL measurement (Table 2). This suggests that both national secular trends and participation in the GWTG–Stroke program contributed to the rising frequency of lipid testing. We speculate that participation in GWTG–Stroke may have facilitated lipid testing by the on-demand web-based performance measure reporting system, web-based access to evidence-based guidelines, or other aspects of the program, although we do not have data on the frequency of hospital use of various aspects of these various web-based tools.
There was substantial hospital variation in rates of LDL testing. Patients who underwent testing differed from patients who did not (Table 2). The magnitudes of many differences were small despite the highly significant probability values, reflecting the very large sample size. However, more substantial differences were seen for several variables, including less testing in older persons, women, nonsmokers, patients with atrial fibrillation or a history of stroke or TIA, patients with TIA (versus ischemic stroke), and in smaller nonacademic hospitals with fewer stroke discharges per year. Many of these observations are consistent with other studies. For example, a previous population-based study of ischemic stroke survivors also found decreased lipid testing in older persons, women, and in smaller nonteaching hospitals.8 Testing was less frequent in patients with TIA, although TIAs are caused by similar pathophysiological mechanisms as ischemic stroke, and patients with TIA are at high risk for early stroke recurrence. The reasons for less frequent lipid testing in patients with TIA cannot be determined from this study but could include uncertainty about the clinical diagnosis of TIA. Less frequent testing in patients with medical comorbidities, controlling for use of LLT, could partly reflect an increased likelihood that recent lipid test results were available to the admitting physician.
A limitation of this study is that we do not have detailed information on why lipid testing was not ordered nor do we have information on physician characteristics that might influence testing patterns. We also cannot rule out that lipid testing was done in follow-up soon after discharge.
The mean LDL among patients hospitalized with ischemic stroke and TIA was 106.5 mg/dL. Although this estimate may be affected by selection bias, because not all patients had lipid testing, it is remarkably similar to that reported in a prior study of >136 000 patients hospitalized with coronary artery disease (104.9 mg/dL).18 More than half had LDL ≥100 mg/dL, which is higher than the NCEP-ATPIII-defined LDL goal for patients with coronary risk equivalents or a sufficiently elevated Framingham risk score and is higher than the minimum LDL level criterion used in the SPARCL trial.3 Based on the SPARCL trial, AHA/ASA guidelines were revised just before the end of the study period to recommend intensive LLT with a statin in patients with ischemic stroke or TIA of atherosclerotic origin.5 Because presumed mechanism of stroke is not a mandatory GWTG–Stroke data element, we do not have data on the frequency of ischemic stroke or TIA of atherosclerotic origin; however, it is likely that many patients with LDL ≥100 mg/dL would have qualified for intervention.8,19,20
Patients with a history of preadmission cardiovascular disease were less, rather than more, likely to have LDL ≥100 mg/dL, although elevated LDL is a risk factor for atherosclerosis. This is probably explained by an association between history of atherosclerotic diseases and use of LLT (which were strongly associated with decreased odds of having LDL ≥100 mg/dL). After adjustment for the use of LLT and other variables, the associations between known atherosclerotic disease and decreased odds of LDL ≥100 mg/dL were relatively small but still present (Table 4). Although we controlled for the prior use of LLT, we did not have data on the dose and duration of therapy with preadmission LLT, which might explain some of the residual association between preadmission atherosclerotic diseases and decreased odds of having LDL ≥100 mg/dL. The decreased odds of finding LDL ≥100 mg/dL in patients with atrial fibrillation or a prosthetic heart valve probably indicates a greater likelihood of presentation with nonatherosclerotic cardioembolic stroke in those groups. Despite these differences between groups, an important finding of the study was that a substantial proportion of patients in all groups had LDL >100 mg/dL (Table 3), supporting the wide use of lipid testing as recommended by guidelines.6,10
There are other limitations to this study. Our results may not be valid for non-GWTG–Stroke hospitals, because hospital participation in GWTG–Stroke is voluntary and therefore limited to hospitals with an interest in stroke quality improvement. Study data were collected based on the medical record and may not be completely accurate if clinical documentation was incomplete or data abstraction was unreliable. We were unable to explore the relationship between stroke severity and odds of LDL measurement because the National Institutes of Health Stroke Scale is not routinely collected in clinical practice and therefore incompletely available in the registry. Our data suggests that GWTG–Stroke participation improves rates of LDL measurement by showing that duration of program participation is associated with increased rates of measurement after adjusting for hospital characteristics, case mix, and secular time trends; however, the study does not include a comparison control group.
In summary, our study shows dramatic increases in LDL measurement in this large hospital-based quality improvement program. Because the GWTG–Stroke program has enrolled a significant fraction of all estimated stroke cases in the United States in later years (in 2008, the enrolled cases represent approximately 23.5% of all ischemic strokes as estimated by American Heart Association statistics21), these results suggest a major national change in ischemic stroke management. In parallel with increasing rates of lipid testing, a prior study of the GWTG–Stroke program shows increasing use of LLT at discharge in hospitalized patients with ischemic stroke.11 The frequency of lipid testing in the GWTG–Stroke program now compares favorably with the frequency of lipid testing in patients hospitalized for coronary artery disease.18 However, some disparities in LDL measurement in GWTG–Stroke persist, including lower odds of measurement in women, older patients, and patients with TIA. Our finding that LDL measurement was increasing with each calendar year, as well as with each additional year of GWTG–Stroke participation, suggests that both national secular trends and program-specific effects contributed to the observed increase in LDL measurement. These findings suggest that participation in programs that monitor performance and provide a process for quality improvement are associated with improved evaluation of LDL, yet opportunities for further improvement still exist.
The GWTG is funded by the American Heart Association and the American Stroke Association. The program is also supported in part by unrestricted educational grants to the American Heart Association by Pfizer, Inc, New York, NY, and the Merck-Schering Plough Partnership (North Wales, Pa), who did not participate in the design, analysis, manuscript preparation, or approval.
E.E.S. serves as a member of the Get With The Guidelines (GWTG) Science Subcommittee and receives research support from the National Institutes of Health (National Institute of Neurological Diseases and Stroke R01 NS062028) and the Canadian Stroke Network and salary support from the Heart and Stroke Foundation of Canada and the Canadian Institutes for Health Research. W.P., E.D.P., and D.O. are members of the Duke Clinical Research Institute, which serves as the American Heart Association (AHA) GWTG data coordinating center. M.J.R. has received salary support from the Michigan Stroke Paul Coverdell Registry and serves as a member of the AHA’s GWTG Quality Improvement Subcommittee. E.D.P. reports research support from BMS-Sanofi. G.C.F. chairs the AHA GWTG Steering Committee; serves as a consultant to Pfizer, Merck, Schering Plough, Bristol Myers Squibb, and Sanofi-Aventis; receives speaker honoraria from Pfizer, Merck, Schering Plough, Bristol Myers Squibb, and Sanofi-Aventis; and receives research support from Pfizer and the National Institutes of Health. L.H.S. serves as vice chair of the AHA GWTG Steering Committee; serves as a consultant to the Research Triangle Institute, CryoCath, and to the Massachusetts Department of Public Health; and has provided expert medical opinions in malpractice lawsuits regarding stroke treatment and prevention.
- Received September 11, 2009.
- Accepted September 17, 2009.
Sacco RL, Adams R, Albers G, Alberts MJ, Benavente O, Furie K, Goldstein LB, Gorelick P, Halperin J, Harbaugh R, Johnston SC, Katzan I, Kelly-Hayes M, Kenton EJ, Marks M, Schwamm LH, Tomsick T. Guidelines for prevention of stroke in patients with ischemic stroke or transient ischemic attack. Stroke. 2006; 37: 577–617.
Adams RJ, Albers G, Alberts MJ, Benavente O, Furie K, Goldstein LB, Gorelick P, Halperin J, Harbaugh R, Johnston SC, Katzan I, Kelly-Hayes M, Kenton EJ, Marks M, Sacco RL, Schwamm LH, American Heart Association, American Stroke Association. Update to the AHA/ASA recommendations for the prevention of stroke in patients with stroke and transient ischemic attack. Stroke. 2008; 39: 1647–1652.
Smith SC Jr, Allen J, Blair SN, Bonow RO, Brass LM, Fonarow GC, Grundy SM, Hiratzka L, Jones D, Krumholz HM, Mosca L, Pasternak RC, Pearson T, Pfeffer MA, Taubert KA. AHA/ACC guidelines for secondary prevention for patients with coronary and other atherosclerotic vascular disease: 2006 update: endorsed by the National Heart, Lung, and Blood Institute. Circulation. 2006; 113: 2363–2372.
Mullard AJ, Reeves MJ, Jacobs BS, Kothari RU, Birbeck GL, Maddox K, Stoeckle-Roberts S, Wehner S. Lipid testing and lipid-lowering therapy in hospitalized ischemic stroke and transient ischemic attack patients: results from a statewide stroke registry. Stroke. 2006; 37: 44–49.
Reeves MJ, Broderick JP, Frankel M, LaBresh KA, Schwamm L, Moomaw CJ, Weiss P, Katzan I, Arora S, Heinrich JP, Hickenbottom S, Karp H, Malarcher A, Mensah G, Reeves MJ. The Paul Coverdell National Acute Stroke Registry: initial results from four prototypes. Am J Prev Med. 2006; 31: S202–S209.
Adams HP Jr, del Zoppo G, Alberts MJ, Bhatt DL, Brass L, Furlan A, Grubb RL, Higashida RT, Jauch EC, Kidwell C, Lyden PD, Morgenstern LB, Qureshi AI, Rosenwasser RH, Scott PA, Wijdicks EF. Guidelines for the early management of adults with ischemic stroke: a guideline from the American Heart Association/American Stroke Association Stroke Council, Clinical Cardiology Council, Cardiovascular Radiology and Intervention Council, and the Atherosclerotic Peripheral Vascular Disease and Quality of Care Outcomes in Research Interdisciplinary Working Groups. Stroke. 2007; 38: 1655–1711.
Schwamm LH, Fonarow GC, Reeves MJ, Pan W, Frankel MR, Smith EE, Ellrodt G, Cannon CP, Liang L, Peterson E, Labresh KA. Get With The Guidelines–Stroke is associated with sustained improvement in care for patients hospitalized with acute stroke or transient ischemic attack. Circulation. 2009; 119: 107–115.
Wolf PA, Clagett GP, Easton JD, Goldstein LB, Gorelick PB, Kelly-Hayes M, Sacco RL, Whisnant JP. Preventing ischemic stroke in patients with prior stroke and transient ischemic attack: a statement for healthcare professionals from the Stroke Council of the American Heart Association. Stroke. 1999; 30: 1991–1994.
American Hospital Association Guide Book. Chicago: American Hospital Association; 2008.
Lalouschek W, Lang W, Greisenegger S, Mullner M. Determination of lipid profiles and use of statins in patients with ischemic stroke or transient ischemic attack. Stroke. 2003; 34: 105–110.
Sachdeva A, Cannon CP, Deedwania PC, Labresh KA, Smith SC Jr, Dai D, Hernandez A, Fonarow GC. Lipid levels in patients hospitalized with coronary artery disease: an analysis of 136 905 hospitalizations in get with the guidelines. Am Heart J. 2009; 157: 111–117. e112.
Smith EE, Abdullah AR, Amirfarzan H, Schwamm LH. Serum lipid profile on admission for ischemic stroke: failure to meet National Cholesterol Education Program Adult Treatment Panel (NCEP-ATPIII) guidelines. Neurology. 2007; 68: 660–665.
Lloyd-Jones D, Adams R, Carnethon M, De Simone G, Ferguson TB, Flegal K, Ford E, Furie K, Go A, Greenlund K, Haase N, Hailpern S, Ho M, Howard V, Kissela B, Kittner S, Lackland D, Lisabeth L, Marelli A, McDermott M, Meigs J, Mozaffarian D, Nichol G, O'Donnell C, Roger V, Rosamond W, Sacco R, Sorlie P, Stafford R, Steinberger J, Thom T, Wasserthiel-Smoller S, Wong N, Wylie-Rosett J, Hong Y. Heart disease and stroke statistics—2009 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation. 2009; 119: 480–486.