Epilepsy & Behavior 79 (2018) 82–86
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Impact of the modiﬁed Atkins diet on cardiovascular health in adults with epilepsy Tanya J.W. McDonald a, Elizabeth V. Ratchford b, Bobbie J. Henry-Barron c, Eric H. Kossoff a,d, Mackenzie C. Cervenka a,⁎ a
Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, MD, United States Johns Hopkins Center for Vascular Medicine, Johns Hopkins University School of Medicine, Baltimore, MD, United States Institute for Clinical and Translational Research, Johns Hopkins University, Baltimore, MD, United States d Department of Pediatrics, Johns Hopkins University School of Medicine, Baltimore, MD, United States b c
a r t i c l e
i n f o
Article history: Received 16 August 2017 Revised 24 October 2017 Accepted 27 October 2017 Available online xxxx Keywords: Epilepsy Carotid intima-media thickness Modiﬁed Atkins diet Ketogenic diet Low-density-lipoprotein
a b s t r a c t Aim: The current study investigated biochemical and vascular markers of cardiovascular health in adult patients with epilepsy treated with long-term (greater than 1 year) ketogenic diet therapy compared with controls. Method: Anthropometric measures, serum fasting lipid panel, apolipoproteins A-1 and B, lipoprotein subfractions as well as common carotid intima-media thickness (cIMT), and plaque presence were assessed in 20 adult patients with epilepsy on a modiﬁed Atkins diet (MAD) for N 1 year started as an adult compared with 21 adult patients with epilepsy naïve to diet therapy. Results: Patients treated with MAD had signiﬁcantly lower weight, body mass index, waist and hip circumference, percent body fat, and serum triglyceride levels when compared with control patients. In contrast, they had signiﬁcantly higher serum levels of small low-density-lipoprotein (LDL) particles and were signiﬁcantly more likely to have LDL pattern B in which small LDL particles predominate when compared with controls. However, there was no signiﬁcant difference in cIMT or plaque presence between groups. Conclusion: Our results provide clinical evidence demonstrating the cardiovascular safety of a high-fat, lowcarbohydrate diet used in adults with epilepsy for at least 12 months. It also highlights potential markers of cardiovascular risk – small dense LDL particles – that should be closely monitored in adults treated with diet therapy long-term. © 2017 Elsevier Inc. All rights reserved.
1. Introduction Ketogenic diet (KD) therapy, a mainstay in the treatment of epilepsy for nearly one century and in obesity treatment for over 50 years, has demonstrated increasing therapeutic potential for a wide variety of pathological conditions and neurologic diseases over the last decade . Ketogenic diets stimulate fat metabolism, measured by ketone body production, through high-fat and low-carbohydrate intake. Similar to the classic KD, the modiﬁed Atkins diet (MAD) has been increasingly used in adults with drug-resistant epilepsy since 2003 [2,3]. However, one barrier to prescribing a high-fat diet among adult neurologists, general practitioners, and patients is the negative perception in the health care community regarding the potential risk of dyslipidemia, obesity, and cardiovascular disease. Contrary to this belief, studies of carbohydrate-restricted diets have shown improvements in markers
⁎ Corresponding author at: Department of Neurology, Johns Hopkins Hospital, 600 North Wolfe Street, Meyer 2-147, Baltimore, MD 21287, United States. E-mail addresses: [email protected]
(T.J.W. McDonald), [email protected]
(E.V. Ratchford), [email protected]
(B.J. Henry-Barron), [email protected]
(E.H. Kossoff), [email protected]
https://doi.org/10.1016/j.yebeh.2017.10.035 1525-5050/© 2017 Elsevier Inc. All rights reserved.
of cardiovascular health in adults including fasting lipid proﬁles, lowdensity-lipoprotein (LDL) particle size, and serum Apolipoprotein B (ApoB)/Apolipoprotein A-1 (ApoA-1) ratio [1,4]. Moreover, studies in children and young adults treated with the classic KD have either shown no negative effect on measures of subclinical disease despite dyslipidemia  or an adaptation of cardiovascular biomarkers and reversibility in vascular dysfunction with long-term diet treatment . This study investigated an extended cardiovascular risk proﬁle, both biochemical and vascular, of adult patients with epilepsy treated with greater than 1 year of MAD compared with controls. 2. Material and methods 2.1. Patients Consecutive adults (age ≥ 18 years) with epilepsy seen in the Johns Hopkins Adult Epilepsy Diet Center (AEDC) from March 2016 through November 2016 were screened for eligibility. Patients following a 20 g per day net carbohydrate MAD  for ≥ 12 months were included in the diet group. Consecutive patients naïve to KD therapy who presented to the AEDC to initiate MAD for refractory epilepsy were included in the
T.J.W. McDonald et al. / Epilepsy & Behavior 79 (2018) 82–86
control group. The study was approved by the Johns Hopkins Institutional Review Board. Written informed consent was obtained from all patients or a legally authorized representative. 2.2. Demographics and clinical measures Collected demographic measures included age at study and diet onset, gender, self-identiﬁed ethnicity, number of prior and current antiseizure drugs (ASDs) used, ASDs at study enrollment, smoking history, personal history of diabetes, hypertension, cardiovascular disease, myocardial infarction, stroke, and dyslipidemia, physical exercise levels, and family history of myocardial infarction and stroke. Patient anthropometric parameters according to guidelines set forth by the International Society for the Advancement of Kinanthropometry included height and weight for body mass index (BMI) calculation, waist and hip circumference, and skin fold thickness at four sites (triceps, biceps, subscapular, and iliac crest) for percent body fat calculation . Blood pressure and serum levels of total cholesterol (TC), triglycerides (TG), high-density-lipoprotein (HDL), calculated LDL, ApoB and ApoA-1 were measured. Low-density-lipoprotein particle number, LDL sub-class (small and medium), large HDL subclass, LDL peak size, and LDL pattern (deﬁned by predominant LDL particle size) were determined by ion mobility analysis as previously described . Carotid intima-media thickness (cIMT) and carotid plaque burden were measured using B-mode ultrasonography at enddiastole in the far wall of the bilateral distal common carotid arteries according to a standard protocol . Mean left and right cIMT from anterior, lateral, and posterior approaches were averaged and reported, along with the presence or absence of plaque. Each side was categorized as less than or greater than the seventy-ﬁfth percentile based on age and gender . 2.3. Statistical analysis Continuous variables were expressed as means ± standard deviation and discrete variables as counts and percentages. Differences between groups were assessed using the student's t-test for unpaired data, Fisher's Exact Test/Pearson Chi-Square or the Wilcoxon rank sum test, as appropriate (p b 0.05). As the current study was designed to identify potential effects of MAD use on markers of cardiovascular health, we did not use a Bonferroni correction method in order to exclude a potential type II error and consequently miss a possible real effect . All statistical analyses were performed using SPSS 11.0 for Windows (SPSS, Chicago, IL). 3. Results Ninety-seven consecutive patients were screened for potential study enrollment. Twenty-ﬁve patients were younger than 18 years of age at diet onset or on MAD for b12 months; 30 patients declined to participate or canceled, and 1 patient did not tolerate the ultrasound. Forty-one adult patients (16 men, 25 women) age 18 to 71 years were included in the study. Twenty patients were on MAD for N 12 months (mean 25 months, range 12–63 months), 4 of whom were initially on a classic KD before transitioning to MAD, and were included in the diet group. Twenty-one patients naïve to diet therapy were included in the control group (Table 1). All control group patients were deemed candidates for diet therapy and 19/21 control group patients subsequently began MAD (no control patients needed to be excluded from the study on the basis of prior cardiovascular disease, stroke, statin use, or other preexisting cardiovascular risk factors). There were no signiﬁcant differences between groups in gender, age, ethnicity, blood pressure, smoking history, personal history of diabetes, dyslipidemia, or hypertension, nor in family history of myocardial infarction or stroke (Table S1). There was also no signiﬁcant difference between groups in the number of patients receiving treatment for
Table 1 Demographic and anthropometric patient characteristics.
Male gender Age (at study onset) Age (at diet onset) Ethnicity Caucasian African American Other (Indian, Hispanic) Number of medications Prior to diet onset At time of diet onset Height (m) Weight (kg) Body mass index (kg/m2) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Waist circumference (cm) Hip circumference (cm)d Percent body fat
Control (n = 21)
Diet (n = 20)
6 (29%) 34.6 ± 12.1 34.6 ± 12.1
10 (50%) 35.7 ± 14.7 33.5 ± 14.9
0.208b 0.865 0.592 0.322c
14 (66.7%) 3 (14.3%) 4 (19%)
17 (85%) 2 (10%) 1 (5%)
8.7 ± 3.4 2.6 ± 1.0 1.69 ± 0.10 81.4 ± 20.2 28.6 ± 7.2 112.7 ± 12.0 72.8 ± 7.3 96.2 ± 14.5 109.9 ± 11.6 32.1 ± 8.5
4.6 ± 3.3 2.0 ± 1.7 1.71 ± 0.11 68.8 ± 12.9 23.5 ± 3.8 115.2 ± 13.6 74.6 ± 9.3 87.2 ± 11.0 99.1 ± 7.6 25.9 ± 9.3
0.000 0.125 0.559 0.023 0.009 0.538 0.496 0.031 0.002 0.033
Results presented as mean ± standard deviation or number (%). P values b 0.05 are indicated in bold. a Student t-test. b Fisher's Exact test. c Pearson Chi-square test. d 1 patient from each group did not complete hip circumference assessment.
diabetes, dyslipidemia, or hypertension. Only 1 patient in the study, in the diet group, was on lipid-lowering medication at the time of study enrollment (Table S1). Patients treated with MAD had signiﬁcantly lower weight (95% conﬁdence interval [CI] +1.8 to + 23.3, p = 0.023), BMI (95% CI +1.4 to +8.7, p = 0.009), waist circumference (95% CI +0.9 to +17.2, p = 0.031), hip circumference (95% CI +4.4 to +17.2, p = 0.002), and percent body fat (95% CI +0.5 to +11.8, p = 0.033) when compared with controls (Table 1). Patients in the control group were on one or more ASDs that included carbamazepine (1 patient), clobazam (6), clonazepam (3), diazepam (1), eslicarbazepine (1), lacosamide (8), lamotrigine (6), levetiracetam (10), lorazepam (2), oxcarbazepine (1), perampanel (3), pregabalin (1), ruﬁnamide (1), topiramate (4), valproic acid (2), and zonisamide (4). Patients in the diet group were on ASDs that included clobazam (3 patients), clonazepam (3), lacosamide (6), lamotrigine (3), levetiracetam (10), lorazepam (3), phenobarbital (2), phenytoin (3), valproic acid (5), and zonisamide (2) and 4 patients in the diet group were on no ASDs by choice. Although MAD patients tried signiﬁcantly fewer ASDs prior to diet onset (95% CI + 2.0 to + 6.2, p b 0.001), there was no signiﬁcant difference between groups in the number of patients on traditional strong enzyme-inducing ASDs including carbamazepine, phenytoin, and phenobarbital, or in the number of patients on valproic acid at study assessment (Table S2). Patients treated with MAD had signiﬁcantly lower triglyceride levels (p = 0.024) but increased levels of small LDL particles (p = 0.023) and were signiﬁcantly more likely to have LDL pattern B in which small LDL particles predominate when compared with controls (p = 0.032; Fig. 1). These ﬁndings remained signiﬁcant (triglycerides, p = 0.015 and LDL small, p = 0.011) or showed a trend for signiﬁcance (LDL pattern, p = 0.055) when the 4 patients on classic KD prior to MAD were excluded from the analysis (Table S3) and small LDL level did not correlate with diet duration (p = 0.192) or age (p = 0.454; Fig. S1). In addition, older patients did not appear to account for the difference in small LDL level between groups as when stratiﬁed by age (age 40 chosen as the ﬁrst time an estimated 10-year atherosclerotic cardiovascular disease (ASCVD) risk should be calculated according to American College of Cardiology/American Heart Association guidelines ) no signiﬁcant difference was found between groups for patients ≥ 40 (LDL-small: 226.3 ± 59.8 MAD vs 268.8 ± 158.2 Control, p = 0.779) whereas MAD patients b 40 years old had signiﬁcantly greater serum level of small LDL compared with
T.J.W. McDonald et al. / Epilepsy & Behavior 79 (2018) 82–86
Goal < 200
Total Cholesterol (mg/dL)
Goal < 150 100
Goal > 40
LDL Cholesterol (mg/dL)
HDL Cholesterol (mg/dL)
Goal < 130 100
0 Control (n=20)
Fig. 1. Serum lipids and carotid assessments. (A) Average fasting serum total cholesterol, triglycerides, HDL and LDL levels in control and diet patients. **p = 0.024. (B) Fasting apolipoprotein, lipoprotein sub-class, cIMT and plaque assessment results by study group. Results presented as mean ± standard deviation or number (%). aStudent t test. bWilcoxon rank sum test. cFisher's Exact Test. dn = 20.
T.J.W. McDonald et al. / Epilepsy & Behavior 79 (2018) 82–86
control patients b40 years old (LDL-small: 249.9 ± 86.3 vs 150.0 ± 31.4, p = 0.001). There were no signiﬁcant differences between groups in levels of TC, HDL, LDL, apoB/A1 ratio, or total number of LDL particles. There were also no signiﬁcant differences between groups in cIMT measurement, proportion of patients with cIMT greater than 75th percentile, or plaque presence (Fig. 1). 4. Discussion 4.1. Results Herein, we demonstrate that adults on MAD for over 12 months have lower weight, BMI, waist and hip circumference, percent body fat, and serum triglycerides when compared with controls. These results are consistent with prior reports in adults demonstrating reductions in weight and triglycerides  and that total cholesterol and LDL levels, if transiently elevated following MAD initiation, normalize after 1 year of therapy . However, while prior studies of carbohydraterestricted diets showed a shift from small LDL particles to large LDL particles , we found that MAD patients had increased small LDL particles (pattern B) when compared with controls. Small LDL particles cross the arterial wall more readily, bind more avidly to proteoglycans in the intimal matrix, are more easily oxidized, and promote endothelial cell dysfunction, characteristics potentially conducive to atherogenesis and intima-layer thickening . Notably, increased concentration of small dense LDL appears to be one of the most frequent dyslipidemias found in patients with premature coronary artery disease [14,15]. In addition, although not always observed, smaller peak LDL particle size and small LDL particle predominance have been shown to be predictive of increased risk for myocardial infarction and related to annual angiographic progression of coronary artery disease [16,17]. The precise role of each lipoprotein subfraction in cardiovascular risk continues to be investigated, with some studies suggesting that LDL particle number, both small and large, rather than absolute particle size correlates with carotid IMT  and other studies suggesting that both LDL particle number and small size are positively associated with coronary artery calciﬁcation . Measurements of LDL size and particle number have the potential to improve cardiovascular disease risk assessment as well as decisions about LDL treatment intensity and diet modiﬁcation, as they account for aspects of lipoprotein atherogenicity that are incompletely reﬂected by traditional LDL values . In the current study, despite increased small LDL particles in patients on MAD for ≥1 year, we did not ﬁnd a signiﬁcant difference in cIMT or plaque presence between groups. Further long-term monitoring of MAD effects on coronary and carotid vasculature, using carotid ultrasound, echocardiography, and potentially serial electrocardiogram, is warranted to explore whether length of diet therapy effects cardiovascular risk in adults. While we already recommend adherence to published Heart Healthy recommendations for MAD in patients with persistent dyslipidemia , we have adjusted current clinical practice in response to study results. Two MAD patients in the study had an estimated 10-year ASCVD risk of 5% or higher, LDL pattern B and carotid plaque. One patient with 13.1% risk was already on moderate-intensity statin therapy for dyslipidemia preceding diet onset (this was the only patient in the study in either group on a lipid-lowering medication and with a personal history of stroke) and was advised to switch to a high-intensity statin. The other patient with 7.1% risk was not on statin therapy because of low LDL and total cholesterol levels and was advised to begin a moderateintensity statin. Of note, this patient was initially on the classic ketogenic diet prior to transitioning to MAD and further investigation is warranted to determine whether this worsens risk of future cardiovascular disease.
that one or more unmeasured baseline clinical features distinct between the control and diet group participants confounded the ﬁndings, and we cannot make any conclusions regarding changes within subjects with treatment. However, we found no signiﬁcant differences between groups regarding the known personal cardiovascular risk factors of gender, age, ethnicity, smoking, diabetes, dyslipidemia, hypertension, or family history. There was also no signiﬁcant difference between groups in the number of participants receiving treatment for diabetes, dyslipidemia, or hypertension. Self-selection bias may have entered into the study as patients with a history of stroke, myocardial infarction, or other cardiovascular disease may have been less likely to seek diet treatment and would therefore not have been asked to enroll in the study. The majority of control patients (16/21) who presented to the AEDC to initiate MAD reported sedentary or low activity levels to the research dietitian for nutritional caloric calculations. We did not perform systematic assessment of physical activity level in patients already on MAD which raises the possibility that exercise, an encouraged management strategy for dyslipidemia, may have confounded results. However, as MAD patients showed improvements in weight, percent body fat and BMI relative to control patients—markers that typically improve with physical activity level, we would have expected exercise to also result in an improved LDL sub-class proﬁle in patients on MAD rather than what we found. Previous studies suggest that the older-generation ASDs including phenytoin, carbamazepine, phenobarbital, and valproic acid, exert prominent effects on the hepatic enzyme system and may alter metabolic pathways related to dyslipidemia and increased vascular risk , although ﬁndings related to speciﬁc ASDs have been inconsistent and studies may not have adjusted for confounding factors. Liver enzymeinducing drugs like carbamazepine are thought to inﬂuence cholesterol levels by enhancing the catalytic effect of cytochrome P450 in the conversion of lanosterol into cholesterol intermediates. A recent systemic review evaluating the association between ASDs and dyslipidemia found small effect sizes with the highest difference observed for plasma cholesterol in carbamazepine users. In addition, carbamazepine use was associated with increased levels of HDL, LDL, and total cholesterol . Thus, certain ASDs could inﬂuence lipid levels in the current study. While we did ﬁnd that patients on MAD were on fewer ASDs prior to diet onset, there was no signiﬁcant difference between groups in the number of patients on enzyme-inducing ASDs at the time of study measures. Furthermore, we found no differences in levels of HDL, LDL, total cholesterol, or apolipoproteins between groups. A prospective study with repeated assessments of the biochemical and vascular markers investigated in the current study as well as other potential atherogenic inﬂammatory biomarkers including c-reactive protein, homocysteine, and lipoprotein A at baseline and beyond 12 months in the same patient is warranted to exclude other potential confounding variables. 4.3. Conclusion This study provides evidence demonstrating the cardiovascular safety of a high-fat, low-carbohydrate diet used in adults with epilepsy for at least 12 months. The study also highlights a marker of cardiovascular risk – small dense LDL particles – that should be closely monitored in adult patients with epilepsy treated with diet therapy long-term. Acknowledgments This study was funded by the generous philanthropic support from Chris Garrod and Dawn Grifﬁths. Ethical publication statement
4.2. Methodological aspects and limitations The study's main limitation is that MAD patients are compared with a control population rather than their prediet baseline. Thus, it is possible
We conﬁrm that we have read the journal's position on issues involved in ethical publication and afﬁrm that this report is consistent with those guidelines.
T.J.W. McDonald et al. / Epilepsy & Behavior 79 (2018) 82–86
Disclosures of conﬂicts of interest Dr. McDonald has no conﬂicts of interest to disclose. Dr. Ratchford's spouse is employed by MedImmune and thus receives stock in AstraZeneca. Ms. Henry-Barron receives grants from Johns Hopkins Institute for Clinical and Translational Research (ICTR) which is funded in part by Grant Number UL1 TR 001079 from the National Center for Advancing Translational Sciences (NCATS) a component of the National Institutes of Health (NIH), and NIH Roadmap for Medical Research, Nutricia and Vitaﬂo. Dr. Kossoff receives a grant from Nutricia and provides consulting services to Atkins Nutritionals, Inc., Nutricia, and NeuroPace. Dr. Cervenka receives grants from Nutricia, Vitaﬂo, BrightFocus Foundation, and Army Research Laboratory as well as honoraria from American Epilepsy Society, New York University, The Neurology Center, Nutricia, and LivaNova. She provides consulting services for Nutricia and Sage Therapeutics. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.yebeh.2017.10.035. References  Paoli A, Rubini A, Volek JS, Grimaldi KA. Beyond weight loss: a review of the therapeutic uses of very-low-carbohydrate (ketogenic) diets. Eur J Clin Nutr 2013; 67:789–96. https://doi.org/10.1038/ejcn.2013.116.  Klein P, Tyrlikova I, Mathews GC. Dietary treatment in adults with refractory epilepsy: a review. Neurology 2014;83:1978–85. https://doi.org/10.1212/WNL. 0000000000001004.  Cervenka MC, Henry BJ, Felton EA, Patton K, Kossoff EH. Establishing an adult epilepsy diet center: experience, efﬁcacy and challenges. Epilepsy Behav 2016;58: 61–8. https://doi.org/10.1016/j.yebeh.2016.02.038.  Volek JS, Phinney SD, Forsythe CE, Quann EE, Wood RJ, Puglisi MJ, et al. Carbohydrate restriction has a more favorable impact on the metabolic syndrome than a low fat diet. Lipids 2009;44:297–309. https://doi.org/10.1007/s11745-008-3274-2.  Özdemir R, Güzel O, Küçük M, Karadeniz C, Katipoglu N, Yilmaz Ü, et al. The effect of the ketogenic diet on the vascular structure and functions in children with intractable epilepsy. Pediatr Neurol 2016;56:30–4. https://doi.org/10.1016/j.pediatrneurol. 2015.10.017.  Coppola G, Natale F, Torino A, Capasso R, D'Aniello A, Pironti E, et al. The impact of the ketogenic diet on arterial morphology and endothelial function in children and young adults with epilepsy: a case-control study. Seizure 2014;23:260–5. https:// doi.org/10.1016/j.seizure.2013.12.002.  Durnin JVGA, Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Br J Nutr 1973;32:77–97. https://doi.org/10.1079/BJN19740060.
 Caulﬁeld MP, Li S, Lee G, Blanche PJ, Salameh WA, Benner WH, et al. Direct determination of lipoprotein particle sizes and concentrations by ion mobility analysis. Clin Chem 2008;54:1307–16. https://doi.org/10.1373/clinchem.2007.100586.  Stein JH, Korcarz CE, Hurst RT, Lonn E, Kendall CB, Mohler ER, et al. Use of carotid ultrasound to identify subclinical vascular disease and evaluate cardiovascular disease risk: a consensus statement from the American Society of Echocardiography carotid intima-media thickness task force endorsed by the Society for Vascular. J Am Soc Echocardiogr 2008;21:93–111. https://doi.org/10.1016/j.echo.2007.11.011.  Denarié N, Gariepy J, Chironi G, Massonneau M, Laskri F, Salomon J, et al. Distribution of ultrasonographically-assessed dimensions of common carotid arteries in healthy adults of both sexes. Atherosclerosis 2000;148:297–302. https://doi.org/10.1016/ S0021-9150(99)00276-2.  Armstrong RA. When to use the Bonferroni correction. Ophthalmic Physiol Opt 2014;34:502–8. https://doi.org/10.1111/opo.12131.  Goff DC, Lloyd-Jones DM, Bennett G, Coady S, D'Agostino RB, Gibbons R, et al. ACC/ AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association task force on practice guidelines. Circulation 2013;2014:129. https://doi.org/10.1161/01.cir.0000437741.48606.98.  Cervenka MC, Patton K, Eloyan A, Henry B, Kossoff EH. The impact of the modiﬁed Atkins diet on lipid proﬁles in adults with epilepsy. Nutr Neurosci 2016;19:131–7. https://doi.org/10.1179/1476830514Y.0000000162.  Mudd JO, Borlaug BA, Johnston PV, Kral BG, Rouf R, Blumenthal RS, et al. Beyond low-density lipoprotein cholesterol. Deﬁning the role of low-density lipoprotein heterogeneity in coronary artery disease. J Am Coll Cardiol 2007;50:1735–41. https://doi.org/10.1016/j.jacc.2007.07.045.  Blake GJ, Otvos JD, Rifai N, Ridker PM. Low-density lipoprotein particle concentration and size as determined by nuclear magnetic resonance spectroscopy as predictors of cardiovascular disease in women. Circulation 2002;106:1930–7. https://doi.org/10. 1161/01.CIR.0000033222.75187.B9.  Austin MA, Breslow JL, Hennekens CH, Buring JE, Willett WC, Krauss RM. Lowdensity lipoprotein subclass patterns and risk of myocardial infarction. JAMA 1988;260:1917–21. https://doi.org/10.1001/jama.1988.03410130125037.  Williams PT, Superko HR, Haskell WL, Alderman EL, Blanche PJ, Holl LG, et al. Smallest LDL particles are most strongly related to coronary disease progression in men. Arterioscler Thromb Vasc Biol 2003;23:314–21. https://doi.org/10.1161/01. ATV.0000053385.64132.2D.  Mora S, Szklo M, Otvos JD, Greenland P, Psaty BM, Goff DC, et al. LDL particle subclasses, LDL particle size, and carotid atherosclerosis in the multi-ethnic study of atherosclerosis (MESA). Atherosclerosis 2007;192:211–7. https://doi.org/10. 1016/j.atherosclerosis.2006.05.007.  Mackey RH, Kuller LH, Sutton-Tyrrell K, Evans RW, Holubkov R, Matthews KA. Lipoprotein subclasses and coronary artery calcium in postmenopausal women from the healthy women study. Am J Cardiol 2002;90:71i–6i. https://doi.org/10. 1016/S0002-9149(02)02636-X.  Superko HR, Gadesam RR. Is it LDL particle size or number that correlates with risk for cardiovascular disease? Curr Atheroscler Rep 2008;10:377–85. https://doi.org/ 10.1007/s11883-008-0059-2.  Chuang Y-C, Chuang H-Y, Lin T-K, Chang C-C, C-H Lu, Chang W-N, et al. Effects of long-term antiepileptic drug monotherapy on vascular risk factors and atherosclerosis. Epilepsia 2012;53:120–8. https://doi.org/10.1111/j.1528-1167.2011.03316.x.  Vyas MV, Davidson BA, Escalaya L, Costella J, Saposnik G, Burneo JG. Antiepileptic drug use for treatment of epilepsy and dyslipidemia: systematic review. Epilepsy Res 2015;113:44–67. https://doi.org/10.1016/j.eplepsyres.2015.03.002.