Advertisement

Appropriateness of Statins in Patients Aged ≥80 Years and Comparison to Other Age Groups

      In patients aged ≥80 years without previous coronary artery disease, peripheral vascular disease, or cerebrovascular disease, no evidence has shown a benefit from statin therapy. We examined the prevalence of statin use in patients aged ≥80 years for the indication of primary prevention. We reviewed the comprehensive electronic health records at the Geisinger Health System in Pennsylvania for all patients aged >55 years with ≥1 primary care encounter from January 24, 2004 and December 31, 2009. The records were scrutinized for the use of a statin, active medical diagnoses, and laboratory values. Patients without a previous diagnosis of coronary artery disease, peripheral vascular disease, or cerebrovascular disease were considered to have a primary prevention indication for statin therapy. The prevalence of statin use was examined, and a multivariate analysis was conducted to determine the predictors of use. A total of 89,086 patients were included in the analysis, with 22,646 patients aged ≥80 years. Of all the patients, 26% were prescribed a statin, of whom, 71% (n = 16,687) received it for primary prevention. Of the 14,604 patients aged ≥80 years with a primary prevention indication, 3,145 (22%) received a statin. A plot of 5-year age cohorts from 55 to >90 years demonstrated an n-shaped relation between age and statin use for primary prevention (18%, 23%, 27%, 29%, 28%, 26%, 21%, and 12%, p <0.001). Compared to patients aged <65 years, the ratio of statin prescription for secondary to primary prevention was 31% lower in patients aged ≥80 years (1.3 vs 1.9). Those aged ≥80 years with a primary prevention indication had, with treatment, a mean low-density lipoprotein level of 84 ± 26 mg/dl. In conclusion, many patients aged ≥80 years receive statin therapy for primary prevention and are treated to aggressive low-density lipoprotein levels. Because the efficacy is uncertain and the potential adverse effects are many, we urgently need to define the cost, benefit, and risk of statin use in the very elderly.
      By 2030, approximately 20 million people living in the United States will be ≥80 years old.
      • He W.
      • Sengupta M.
      • Velko V.A.
      • DeBarros K.A.
      Current population reports, P23-209, 65+ in the United States: 2005.
      In those aged 75 to 85 years, it is estimated that ≥33% of men and ≥20% of women will be taking a statin.
      • Qato D.M.
      • Alexander G.C.
      • Conti R.M.
      • Johnson M.
      • Schumm P.
      • Lindau S.T.
      Use of prescription and over-the-counter medications and dietary supplements among older adults in the United States.
      Although there is good evidence for the benefits of secondary prevention in patients aged ≥80 years, limited data are available for statins as primary prevention therapy. The Third Report of the Expert Panel on the Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults has stated that for men aged >65 years and women aged >75 years, “clinical judgment assumes increased importance in [the] choice of LDL [low-density lipoprotein]-lowering therapies.”
      National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III)
      Third report of the national cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (adult treatment Panel III) final report.
      This statement simply reflects the paucity of evidence showing benefit in this age group. Given the lack of data, the first step is to establish the prevalence of statin therapy in the general population.

      Methods

      The present study was conducted at Geisinger Health System, an integrated delivery system that serves residents in 43 counties in central and northeastern Pennsylvania. All clinics have used the EpicCare electronic health record since 2001, and the electronic health record database contains information for >3 million patients.
      The Geisinger Health System electronic health record was reviewed for eligible patients. All patients aged >55 years with ≥1 primary care encounter from January 24, 2004 to December 31, 2009 were included in the present study. Patients with a statin on their active medication list at their most recent clinical encounter were considered to be using a statin. The list of used statins included pravastatin, simvastatin, atorvastatin, lovastatin, rosuvastatin, and fluvastatin.
      Patient problem lists, clinical encounter data, and medication-associated diagnoses were electronically reviewed to determine the prevalence of clinical diagnoses in this population. Specifically, patient encounters were reviewed for “International Classification of Diseases, 9th Revision, Clinical Modification” codes, including 410 to 414 (ischemic heart disease), 430 to 438 (cerebrovascular disease), 440 to 448 (diseases of the arteries, arterioles, or capillaries), 250.0 to 250.93 (diabetes mellitus), 401 to 405 (hypertension), and 272 (hyperlipidemia). The record was reviewed for previous or active smoking.
      Patients with a diagnosis of coronary artery disease, peripheral vascular disease, or cerebrovascular disease were categorized as having a secondary prevention indication for statin use and those without were considered to have a nonsecondary or primary prevention indication.
      From this database, information was also available to assess and calculate the Charlson co-morbidity index for all patients in the present study. The Charlson index predicts the 10-year mortality of a patient according to the diagnosis of a wide range of co-morbid conditions.
      • Charlson M.E.
      • Pompei P.
      • Ales K.L.
      • MacKenzie C.R.
      A new method of classifying prognostic comorbidity in longitudinal studies: development and validation.
      • Hall W.H.
      • Ramachandran R.
      • Narayan S.
      • Jani A.B.
      • Vijayakumar S.
      An electronic application for rapidly calculating Charlson comorbidity score.
      Each condition is assigned a score of 1, 2, 3, or 6, depending on the risk of dying associated with that condition. The scores are then summed and given a total score that predicts mortality. A total of 17 co-morbidities were assessed in this index, including myocardial infarction, heart failure, peripheral vascular disease, cerebrovascular disease, dementia, chronic pulmonary disease, connective tissue disease, peptic ulcer disease, mild liver disease, diabetes without complications, diabetes with complications, paraplegia, renal disease, cancer, moderate or severe liver disease, metastatic cancer, and human immunodeficiency virus. These co-morbidities were used in a multivariate analysis to determine the predictors of statin use.
      The lipid levels were reviewed when available (the data were available for >85% of the patients.) The patient demographics were also examined.
      All analyses were performed using a standard statistical package (SPSS, version 16.0, SPSS, Chicago, Illinois). Continuous variables are expressed as the mean ± SD. Categorical variables are expressed as numbers and percentages. We performed univariate analyses of the continuous variables using a 2-tailed Student t test and chi-square test for categorical variables. One-way analysis of variance with a post hoc Bonferroni test was used to compare the mean values of the continuous variables among multiple groups. We calculated the predictors of statin use according to logistic regression analysis. Univariate and multivariate logistic regression models were developed to analyze the use of statins for primary and secondary prevention. Variables were inserted into the model as forward Wald conditional (i.e., only those variables that had clinical and statistical significance were included). P Values were considered significant at <0.05.

      Results

      A total of 89,086 patients were included in the analysis, with 22,646 patients aged ≥80 years. Of these, 23,602 (26%) were prescribed a statin, of whom, 16,687 (71%) received the statin for primary prevention and 6,915 (29%) for secondary prevention.
      The prevalence of statin use in those aged ≥80 years relative to those aged 55 to 79 years was 24% versus 27% (5,501 of 22,646 vs 18,101 of 66,440 patients, p <0.001) with 22% versus 25% (3,145 of 14,604 vs 13,542 of 55,043 patients, p <0.001) taking a statin for primary prevention and 29% versus 40% (2,356 of 8,042 vs 4,559 of 11,387 patients, p <0.001) for secondary prevention. The prevalence of vascular disease and co-morbidities, as assessed by the Charlson index, was greater (1.7 vs 1.0, p <0.001) in the age ≥80-year cohort (Table 1). Diabetes mellitus and hypertension were more prevalent in the older group, and hyperlipidemia and smoking history were less prevalent.
      Table 1Summary of baseline characteristics
      VariableAll Patients (n = 89,086)Age (years)p Value
      55–79 (n = 66,440)≥80 (n = 22,646)
      Mean age (years)71 ± 1067 ± 785 ± 4<0.001
      Women49,312 (55%)35,668 (54%)13,644 (60%)<0.001
      Race<0.001
       White87,485 (98%)65,016 (98%)22,469 (99%)
       Black633 (0.7%)574 (0.9%)59 (0.3%)
       Hispanic447 (0.5%)389 (0.6%)58 (0.3%)
       Asian American294 (0.3%)258 (0.4%)36 (0.2%)
      Coronary artery disease8,836 (10%)5,307 (8%)3,529 (16%)<0.001
      Peripheral vascular disease6,373 (7%)3,717 (6%)2,656 (12%)<0.001
      Cerebrovascular disease8,830 (10%)4,753 (7%)4,077 (18%)<0.001
      Coronary artery disease, peripheral vascular disease, or cerebrovascular disease19,429 (22%)11,387 (17%)8,042 (36%)<0.001
      Diabetes13,893 (16%)9,821 (15%)4,072 (18%)<0.001
      Hypertension30,647 (34%)21,113 (32%)9,534 (42%)<0.001
      Ever smoker44,146 (50%)34,608 (52%)9,538 (42%)<0.001
      Hyperlipidemia25,058 (28%)19,007 (29%)6,051 (27%)<0.001
      Mean Charlson index1.2 ± 1.41.0 ± 1.31.7 ± 1.5<0.001
      Prescribed statin23,602 (26%)18,101 (27%)5,501 (24%)<0.001
       Primary prevention16,687 (24%)13,542 (25%)3,145 (22%)<0.001
       Secondary prevention6,915 (36%)4,559 (40%)2,356 (29%)<0.001
      For primary prevention, a plot of the 5-year age cohorts (Figure 1) from 55 to >90 years demonstrated an n-shaped relation between age and statin use (18%, 23%, 27%, 29%, 28%, 26%, 21%, 12%, p <0.001). Statin use for secondary prevention was significantly lower in patients aged ≥80 years, with a declining trend seen with increasing age. Compared to those aged <65 years, the ratio of statin prescription for secondary to primary prevention was 31% lower in the patients aged ≥80 years (1.3 vs 1.9).
      Figure thumbnail gr1
      Figure 1Prevalence of statin use by indication, primary versus secondary, in all age groups. Percentages represent all patients in that age group with particular indication (denominator) and number of patients receiving a statin (numerator). The table lists the absolute number of patients in each age group receiving a statin and the ratios of patients receiving a statin for secondary prevention versus primary prevention. p <0.001 for trends across ages in all 3 subgroups.
      Multivariate analysis showed patients with a greater Charlson co-morbidity index were less likely to be taking a statin for either primary prevention (relative risk [RR] 0.94, 95% confidence interval [CI] 0.91 to 0.97) or secondary prevention (RR 0.94, 95% CI 0.93 to 0.96). Specifically, those aged ≥80 years with a primary prevention indication were significantly less likely to be taking a statin if they had a diagnosis of dementia (RR 0.77, 95% CI 0.64 to 0.94), cancer (RR 0.83, 95% CI 0.74 to 0.92), chronic pulmonary disease (RR 0.57, 95% CI 0.50 to 0.66), mild liver disease (RR 0.13, 95% CI 0.04 to 0.47), connective tissue disease (RR 0.53, 95% CI 0.42 to 0.67) or heart failure (RR 0.63, 95% CI 0.54 to 0.73). All multivariate predictors of statin use in the primary prevention group are listed in Table 2.
      Table 2Predictors of statin use in patients aged ≥80 years for primary prevention
      PredictorUnivariate RR95% CIp ValueMultivariate RR95% CIp Value
      Age (years)0.910.89–0.92<0.0010.930.91–0.97<0.001
      Women1.231.13–1.34<0.0011.641.48–1.81<0.001
      White11
      Black1.630.84–3.140.1432.591.18–5.650.02
      Hispanic0.520.20–1.330.1720.550.21–1.480.28
      Asian American0.690.24–2.020.501.080.32–3.720.90
      Hypertension0.740.68–0.81<0.0010.690.63–0.75<0.001
      Diabetes1.721.55–1.90<0.0011.121.00–1.260.046
      Hyperlipidemia1.651.52–1.80<0.0011.461.33–1.61<0.001
      Ever smoking1.010.93–1.100.841.050.95–1.160.31
      Low-density lipoprotein0.980.98–0.98<0.0010.980.98–0.98<0.001
      Low-density lipoprotein group (mg/dl)
       <7011
       70–1000.780.70–0.87<0.0010.730.65–0.82<0.001
       101–1300.330.29–0.37<0.0010.290.25–0.32<0.001
       131–1600.160.13–0.19<0.0010.130.11–0.16<0.001
       161–2000.210.15–0.28<0.0010.160.12–0.22<0.001
       >2000.200.10–0.39<0.0010.150.07–0.30<0.001
      Charlson index1.031.00–1.060.030.940.91–0.97<0.001
      In those aged ≥80 years with a secondary prevention indication, the patients were significantly less likely to be prescribed a statin if they were diagnosed with heart failure (RR 0.86, 95% CI 0.76 to 0.97), dementia (RR 0.63, 95% CI 0.52 to 0.77), connective tissue disease (RR 0.73, 95% CI 0.59 to 0.92), mild liver disease (RR 0.13, 95% CI 0.03 to 0.59), or cancer (RR 0.79, 95% CI 0.70 to 0.89) and more likely if they had cerebrovascular disease (RR 1.27, 95% CI 1.14 to 1.42) or peptic ulcer disease (RR 1.75, 95% CI 1.14 to 1.42).
      The mean low-density lipoprotein (LDL) level was significantly lower in patients taking a statin across all age cohorts (Figure 2). The LDL levels declined significantly with increasing age, regardless of whether the patients were taking a statin. The mean LDL level for those aged ≥80 years treated with a statin was 84 ± 26 mg/dl vs 100 ± 34 mg/dl for those not taking a statin (p <0.001).
      Figure thumbnail gr2
      Figure 2Mean LDL levels in those with primary prevention indication for statin use. p <0.001 for trends in statin and no-statin groups; p <0.001 for difference in mean LDL with and without statin for all age groups.

      Discussion

      The principle findings of our study were that despite a lack of solid evidence, a significant number of patients aged ≥80 years are prescribed a statin for primary prevention. The ratio of those treated for primary prevention to those treated for secondary prevention was significantly greater in those aged ≥80 years than in those aged 55 to 65 years. The relatively greater number of patients treated for primary prevention with increasing age could suggest that clinicians are as likely to treat older patients, regardless of the underlying indication. A similar preponderance of primary over secondary prevention was reported in an Italian population aged >75 years. Two thirds of these patients had an indication for primary prevention.
      • Trifirò G.
      • Alacqua M.
      • Corrao S.
      • Tari M.
      • Arcoraci V.
      Statins for the primary prevention of cardiovascular events in elderly patients: a picture from clinical practice without strong evidence from clinical trials.
      However, the small decline in statin usage in patients aged ≥80 years suggests clinicians are taking additional factors other than age into consideration in their decision-making. Conceivably, global risk assessment with traditional coronary risk factors plays a role because patients with a diagnosis of diabetes mellitus or hyperlipidemia were more likely to be prescribed a statin. However, clinicians appear to be weighing these factors against existing co-morbidities as implied by the decreased likelihood of statin use in those with an increased Charlson index score.
      Those aged ≥80 years with a diagnosis of hypertension were significantly less likely to be prescribed a statin. The cause of this association is unclear. One possibility is that patients with hypertension might be prescribed a greater number of medications and, therefore, were less likely to be prescribed a statin owing to the significant pill burden.
      We also found statins to be used less frequently for secondary prevention in patients aged ≥80 years, which has also been shown in previous studies.
      • Ko D.T.
      • Mamdani M.
      • Alter D.A.
      Lipid-lowering therapy with statins in high-risk elderly patients: the treatment-risk paradox.
      Patients with dementia, cancer, liver disease, or connective tissue disease were significantly less likely to be taking a statin.
      The treatment goals, with respect to lipid levels, remain unclear in this group.
      • Petersen L.K.
      • Christensen K.
      • Kragstrup J.
      Lipid-lowering treatment to the end? A review of observational studies and RCTs on cholesterol and mortality in 80+-year olds.
      Although lowering cholesterol has shown some benefit in secondary prevention in the very elderly, low cholesterol has also been correlated with greater noncardiovascular mortality and overall mortality.
      • Schatz I.J.
      • Masaki K.
      • Yano K.
      • Chen R.
      • Rodriguez B.L.
      • Curb J.D.
      Cholesterol and all-cause mortality in elderly people from the Honolulu Heart Program: a cohort study.
      • Newson R.S.
      • Felix J.F.
      • Heeringa J.
      • Hofman A.
      • Witteman J.C.
      • Tiemeier H.
      Association between serum cholesterol and noncardiovascular mortality in older age.
      Our data have shown that, on average, patients aged ≥80 years are treated to aggressive LDL levels.
      The only prospective study to date to examine statin use in a truly elderly population is PROspective Study of Pravastatin in the Elderly at Risk (PROSPER).
      • Shepherd J.
      • Blauw G.J.
      • Murphy M.B.
      • Bollen E.L.
      • Buckley B.M.
      • Cobbe S.M.
      • Ford I.
      • Gaw A.
      • Hyland M.
      • Jukema J.W.
      • Kamper A.M.
      • Macfarlane P.W.
      • Meinders A.E.
      • Norrie J.
      • Packard C.J.
      • Perry I.J.
      • Stott D.J.
      • Sweeney B.J.
      • Twomey C.
      • Westendorp R.G.
      PROSPER Study Group (PROspective Study of Pravastatin in the Elderly at Risk)
      Pravastatin in elderly individuals at risk of vascular disease (PROSPER): a randomised controlled trial.
      In patients aged 70 to 82 years with risk factors for, or a history of, vascular disease, randomization to pravastatin versus placebo showed a 15% risk reduction in cardiovascular events during 3 years. However, the study showed no benefit in all-cause mortality (10.3% with pravastatin vs 10.5% without, p = 0.74). In a subsequent review of unpublished PROSPER data by Afilalo et al,
      • Afilalo J.
      • Duque G.
      • Steele R.
      • Jukema J.W.
      • de Craen A.J.
      • Eisenberg M.J.
      Statins for secondary prevention in elderly patients: a hierarchical bayesian meta-analysis.
      the outcomes were stratified further by the primary and secondary prevention cohorts. The patients with a history of coronary artery disease were shown to derive a significant benefit in all-cause mortality (RR 0.82, 95% CI 0.69 to 0.98). However, although the study was not powered to detect all-cause mortality, the investigators noted no benefit was found in the primary prevention group. Pooling of published data (from PROSPER and Afilalo et al
      • Afilalo J.
      • Duque G.
      • Steele R.
      • Jukema J.W.
      • de Craen A.J.
      • Eisenberg M.J.
      Statins for secondary prevention in elderly patients: a hierarchical bayesian meta-analysis.
      ) demonstrated a trend toward increased mortality in the primary prevention group (9.6% with pravastatin versus 8.8% without; 188 of 1,957 vs 178 of 2,014 patients; p = 0.40).
      In contrast, most of our current risk stratification models, such as the Framingham Risk, include age as the most potent predictor of future cardiac events. The prevalence of cardiovascular disease, including asymptomatic disease, is known to increase with age. Thus, patients aged ≥80 years arguably represent the highest risk group, regardless of traditional coronary artery disease risk factors and, therefore, would stand to benefit the most from statin use. Adding uncertainty to the issue is that these models might not be as accurate at predicting risk in older patients.
      • de Ruijter W.
      • Westendorp R.G.
      • Assendelft W.J.
      • den Elzen W.P.
      • de Craen A.J.
      • le Cessie S.
      • Gussekloo J.
      Use of Framingham risk score and new biomarkers to predict cardiovascular mortality in older people: population based observational cohort study.
      Although the evidence for statin use for secondary prevention in older patients is good, the data to support their use for primary prevention are limited. When treating for primary prevention, consideration should be given to the decreased life expectancy, increased co-morbidities, risk of polypharmacy, and increased risk of adverse reactions in the geriatric population. In addition, the cost implications play an increasingly important role as this demographic continues to increase.
      Without adequate data, clinicians are left to sort out the conundrum of whether to continue, initiate, or even discontinue statin therapy for their older patients with little guidance.
      The intent of the present study was not to advocate for or against the use of statins. Rather, we sought to describe the statin usage patterns in the elderly to understand the clinical factors that might be guiding the decision to prescribe statins.
      Our study had several limitations, chief of which was the use of an electronic health record database for analysis. The accuracy of medical diagnoses is largely dependent on the diligence and accuracy of the clinical documentation. However, the accuracy of the Geisinger Health database has been validated against manual chart review in a previous study.
      • Hsiao P.
      • Wood G.C.
      • Hartman T.
      • Still C.
      • Jensen G.
      PS1-19: using electronic data extraction to identify subjects with metabolic syndrome: a validation using manual chart review.
      The present study was a retrospective study with no outcomes data available. The documentation of a statin prescription does not equate to its actual use by the patients. However, a prescription is reflective of clinician decision-making, which was the primary intent of our study.

      References

        • He W.
        • Sengupta M.
        • Velko V.A.
        • DeBarros K.A.
        Current population reports, P23-209, 65+ in the United States: 2005.
        (In: Bureau USC, ed.) U.S. Government Printing Office, Washington DC2005
        • Qato D.M.
        • Alexander G.C.
        • Conti R.M.
        • Johnson M.
        • Schumm P.
        • Lindau S.T.
        Use of prescription and over-the-counter medications and dietary supplements among older adults in the United States.
        JAMA. 2008; 300: 2867-2878
        • National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III)
        Third report of the national cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (adult treatment Panel III) final report.
        Circulation. 2002; 106: 3143-3421
        • Charlson M.E.
        • Pompei P.
        • Ales K.L.
        • MacKenzie C.R.
        A new method of classifying prognostic comorbidity in longitudinal studies: development and validation.
        J Chronic Dis. 1987; 40: 373-383
        • Hall W.H.
        • Ramachandran R.
        • Narayan S.
        • Jani A.B.
        • Vijayakumar S.
        An electronic application for rapidly calculating Charlson comorbidity score.
        BMC Cancer. 2004; 4: 94
        • Trifirò G.
        • Alacqua M.
        • Corrao S.
        • Tari M.
        • Arcoraci V.
        Statins for the primary prevention of cardiovascular events in elderly patients: a picture from clinical practice without strong evidence from clinical trials.
        J Am Geriatr Soc. 2008; 56: 175-177
        • Ko D.T.
        • Mamdani M.
        • Alter D.A.
        Lipid-lowering therapy with statins in high-risk elderly patients: the treatment-risk paradox.
        JAMA. 2004; 291: 1864-1870
        • Petersen L.K.
        • Christensen K.
        • Kragstrup J.
        Lipid-lowering treatment to the end?.
        Aging. 2010; 39: 674-680
        • Schatz I.J.
        • Masaki K.
        • Yano K.
        • Chen R.
        • Rodriguez B.L.
        • Curb J.D.
        Cholesterol and all-cause mortality in elderly people from the Honolulu Heart Program: a cohort study.
        Lancet. 2001; 358: 351-355
        • Newson R.S.
        • Felix J.F.
        • Heeringa J.
        • Hofman A.
        • Witteman J.C.
        • Tiemeier H.
        Association between serum cholesterol and noncardiovascular mortality in older age.
        J Am Geriatr Soc. 2011; 59: 1779-1785
        • Shepherd J.
        • Blauw G.J.
        • Murphy M.B.
        • Bollen E.L.
        • Buckley B.M.
        • Cobbe S.M.
        • Ford I.
        • Gaw A.
        • Hyland M.
        • Jukema J.W.
        • Kamper A.M.
        • Macfarlane P.W.
        • Meinders A.E.
        • Norrie J.
        • Packard C.J.
        • Perry I.J.
        • Stott D.J.
        • Sweeney B.J.
        • Twomey C.
        • Westendorp R.G.
        • PROSPER Study Group (PROspective Study of Pravastatin in the Elderly at Risk)
        Pravastatin in elderly individuals at risk of vascular disease (PROSPER): a randomised controlled trial.
        Lancet. 2002; 360: 1623-1630
        • Afilalo J.
        • Duque G.
        • Steele R.
        • Jukema J.W.
        • de Craen A.J.
        • Eisenberg M.J.
        Statins for secondary prevention in elderly patients: a hierarchical bayesian meta-analysis.
        J Am Coll Cardiol. 2008; 51: 37-45
        • de Ruijter W.
        • Westendorp R.G.
        • Assendelft W.J.
        • den Elzen W.P.
        • de Craen A.J.
        • le Cessie S.
        • Gussekloo J.
        Use of Framingham risk score and new biomarkers to predict cardiovascular mortality in older people: population based observational cohort study.
        BMJ. 2009; 338: a3083
        • Hsiao P.
        • Wood G.C.
        • Hartman T.
        • Still C.
        • Jensen G.
        PS1-19: using electronic data extraction to identify subjects with metabolic syndrome: a validation using manual chart review.
        Clin Med Res. 2011; 9: 168