Aging.wisc.edu

Journal of Alzheimer’s Disease 13 (2008) 187–197 Effects of Simvastatin on Cerebrospinal FluidBiomarkers and Cognition in Middle-AgedAdults at Risk for Alzheimer’s Disease Cynthia M. Carlssona,e,∗, Carey E. Gleasona,e, Timothy M. Hessb, Kimberly A. Morelanda,Hanna M. Blazela, Rebecca L. Koscikb, Nathan T.N. Schreibera, Sterling C. Johnsona,e,Craig S. Atwooda,e, Luigi Pugliellia,e, Bruce P. Hermannc, Patrick E. McBrided, James H. Steind,Mark A. Sagera and Sanjay Asthanaa,eaDepartment of Medicine, Section of Geriatrics and Gerontology, University of Wisconsin School of Medicine andPublic Health, Madison, WI 53705, USAbDepartment of Biostatistics and Medical Informatics, University of Wisconsin School of Medicine and PublicHealth, Madison, WI 53705, USAcDepartment of Neurology, University of Wisconsin School of Medicine and Public Health, Madison, WI 53705,USAdDivision of Cardiovascular Medicine, Department of Medicine, University of Wisconsin School of Medicine andPublic Health, Madison, WI 53705, USAeDepartment of Veterans Affairs (VA) Geriatric Research, Education and Clinical Center (GRECC), Madison, WI,USA Abstract. Background: Statins reduce amyloid-β (Aβ) levels in the brain and cerebrospinal fluid (CSF) in animals and may
thereby favorably alter the pathobiology of AD. It is unclear if statins modify Aβ metabolism or improve cognition in asymptomatic
middle-aged adults at increased risk for AD.
Methods: In a 4-month randomized, double-blind, controlled study, we evaluated the effects of simvastatin 40 mg daily vs. placebo
on CSF Aβ42 levels and cognition in 57 asymptomatic middle-aged adult children of persons with AD.
Results: Compared to placebo, individuals randomized to simvastatin for 4 months had similar changes in CSF Aβ42 (p = 0.344)
and total tau levels (p = 0.226), yet greater improvements in some measures of verbal fluency (p = 0.024) and working memory
(p = 0.015). APOE4 genotype, gender, and vascular risk factors were associated with CSF biomarker levels, but did not modify
treatment effects.
Conclusion: In asymptomatic middle-aged adults at increased risk for AD, simvastatin use improved selected measures of
cognitive function without significantly changing CSF Aβ42 or total tau levels. Further studies are needed to clarify the impact
of higher dose and/or longer duration statin therapy on not only Aβ metabolism, but also other preclinical processes related to
the development of AD.
Keywords: Alzheimer’s disease, amyloid β-protein, biological markers, cerebrospinal fluid, genetic predisposition to disease,hydroxymethylglutaryl-CoA reductase inhibitors, neuropsychological tests, simvastatin, tau proteins ∗Corresponding author: Cynthia M. Carlsson, MD, MS, Assistant Professor of Medicine, Section of Geriatrics and Gerontology, Uni- cation and Clinical Center (GRECC) (11G), 2500 Overlook Terrace, versity of Wisconsin School of Medicine and Public Health, William Madison, WI 53705, USA. Tel.: +1 608 280 7000; Fax: +1 608 280 S. Middleton Memorial Veterans Hospital Geriatric Research, Edu- 7165; E-mail: cmc@medicine.wisc.edu.
ISSN 1387-2877/08/$17.00  2008 – IOS Press and the authors. All rights reserved C.M. Carlsson et al. / Effects of Simvastatin on Cerebrospinal Fluid Biomarkers and Cognition in Middle-Aged Adults INTRODUCTION
Alzheimer’s disease (AD) is a devastating illness that will affect an increasing number of older adults in thecoming decades unless effective preventive strategies Fifty-seven men and women without cognitive im- are developed. Hypercholesterolemia in midlife leads pairment were recruited from the Wisconsin Registry to an increased risk of developing AD later in life and, for Alzheimer’s Prevention (WRAP), a statewide reg- therefore, has been targeted as a potentially modifiable istry of adult children (ages 40–65 years) of persons risk factor for AD [14]. While some observational stud- with late-onset AD [30]. Prior to enrollment in WRAP, ies show that use of lipid-lowering agents, specifical- a diagnosis of definite or probable AD in one or both ly statins, is associated with a reduced incidence and parents was confirmed using the National Institute of prevalence of AD [10,29,41], not all studies support Neurological and Communicative Disorders and Stroke these findings [25,42]. In animal models, hypercholes- and Alzheimer’s Disease and Related Disorders Asso- terolemia stimulates amyloid-β (Aβ) deposition in the ciation (NINCDS-ADRDA) criteria [19] through clin- brain [26], a pathologic hallmark of AD. Statins reduce ical evaluation and/or chart review by physicians and CSF and brain Aβ42 levels in animals by 50%, sug- neuropsychologists with expertise in the diagnosis of gesting that statins may modify AD risk by affecting amyloid-β protein precursor (AβPP) processing [7]. It WRAP registrants were eligible for participation if remains unclear, however, whether statins will effec- they did not meet any of the following exclusion cri- tively lower CSF and brain Aβ levels in humans [12, teria: current use of cholesterol-lowering medications 28,31,32]. Furthermore, the effects of such changes in or medications known to interact with simvastatin, liv- Aβ levels on cognition are unknown.
er disease, past adverse reaction to a statin drug, or Previous investigations support the notion that statins elevated creatine kinase (CK). CSF samples were not may have the greatest benefit when they are target- collected in persons with a contraindication to lumbar ed earlier in the disease course in high risk popula- puncture or who declined the procedure. This study tions. Clinical studies to date show that short-term was approved by the University of Wisconsin Human statin therapy (3 to 12 months) may beneficially modify Subjects Committee and was conducted in accord with CSF Aβ levels [12,31,32] and cognition [33] in persons the Helsinki Declaration of 1975. Informed consent with mild AD, but are less effective in persons with more advanced disease [31,33]. Persons carrying theapolipoprotein E ε4 (APOE4) allele, the main genet- ic risk factor for late-onset AD, may also have greatercognitive benefit from statin therapy [33]. These find- All data were collected at the University of Wiscon- ings, together with the fact that it is midlife hyper- sin General Clinical Research Center. Participants at- cholesterolemia that contributes to AD risk decades lat- tended three visits during the 4-month study (baseline, er, suggest that an optimal prevention strategy may be month 1, and month 4). At each visit, subjects had to target statin therapy in a middle-aged, preclinical, fasting blood tests and completed a two-hour cognitive at-risk population with high APOE4 allele prevalence.
battery measuring memory, language fluency, visual- Asymptomatic middle-aged adult children of persons motor function, and executive function. At baseline with late-onset AD are ideal to study as their risk for and month 4, a lumbar puncture was performed for CSF AD is increased not only by their family history (a po- tential marker of unidentified genetic and/or environ- At baseline, participants were randomized in a 1:1 mental AD risk factors) [13], but also by their high ratio to receive simvastatin 40 mg nightly (n = 29) prevalence of APOE4 allele [30].
vs. matching placebo (n = 28). Randomization was Our study is a randomized, placebo-controlled, conducted by the University of Wisconsin Pharmaceuti- double-blind, parallel-group design clinical trial evalu- cal Research Center and was stratified based on APOE4 ating the effects of four months of simvastatin therapy allele status (positive or negative) using a forced block on CSF Aβ42 levels and cognition in asymptomatic design. Subjects and investigators were blinded to the middle-aged adults with a parental history of AD and identity of the study medication and to cholesterol- high prevalence of APOE4 genotype.
lowering effects during the course of the study.
C.M. Carlsson et al. / Effects of Simvastatin on Cerebrospinal Fluid Biomarkers and Cognition in Middle-Aged Adults Systolic blood pressure, mean (SD), mm Hg Current use of antihypertensive medications, No. (%) Physical exercise per week (30+ min), median Serum total cholesterol, mean (SD), mg/dL History of treated thyroid disease, No. (%) CSF biomarker criteria consistent with AD (despite normal cognitive testing), No. (%)* * Baseline CSF data based on n = 50 (simvastatin group: n = 25; placebo group: n = 25).
Values are means (SD) unless otherwise indicated. To convert from mg/dL to mmol/L multiply by 0.0259.
SSRI = selective serotonin reuptake inhibitor; HDL-C = high density lipoprotein cholesterol; LDL-C =low-density lipoprotein cholesterol; Aβ42 = amyloid-β 42; AD = Alzheimer’s disease.
zymatic precipitation techniques. Fasting plasma sam-ples were sent for NMR spectroscopy lipoprotein sub- CSF samples were collected at the same time in the class analysis (NMR LipoProfile , LipoScience, Inc., morning at both the baseline and month 4 visits. Sub- Raleigh, NC). APOE4 allele was measured using stan- jects were placed in a sitting or lateral decubitus posi- dard PCR and DNA sequencing techniques. CSF sam- tion and a Sprotte 20- or 22-gauge spinal needle was in- ples were shipped to Athena Diagnostics (Worcester, serted into the L3/L4 or L4/L5 interspace using sterile technique. Approximately 10 mL of CSF were collect- ysis and Interpretation using capture ELISA.
ed, aliquoted, and transported immediately for process-ing. CSF samples were refrigerated and shipped within 24 hours to Athena Diagnostics (Worcester, MA) foranalysis of CSF Aβ42 and T-tau levels.
Prior to lumbar puncture, cognitive testing was ad- ministered by a trained technician according to pro- tocols established for each standardized test. Sever-al cognitive domains were assessed, including: ver- All blood tests were collected following a 12-hour bal and nonverbal memory (California Verbal Learning fast. Glucose, CK, aspartate aminotransferase, and ala- Test – Second Edition [CVLT-II] [5], Wechsler Mem- nine aminotransferase levels were measured using en- ory Scale – Third Edition [WMS-III] Logical Memory C.M. Carlsson et al. / Effects of Simvastatin on Cerebrospinal Fluid Biomarkers and Cognition in Middle-Aged Adults 2 withdrew from study
3 withdrew from study
1 with scheduling conflicts1 needed to begin medication I and II [40], Visual Spatial Learning Test [VSLT] [17, 18], WMS-III Visual Reproduction I and II [40]); lan-guage fluency (Controlled Oral Word Association Test [COWAT] – CFL, category fluency [2,20]); visual- motor function (Rey Complex Figure Test – Copy [27], Grooved Peg Board [15]); and measures of attention, concentration, working memory, and executive func-tioning (Bells Visual Neglect Test [9], Wechsler Adult Intelligence Scale [WAIS]-III Symbol Search and Dig- it Symbol Substitution Subtests [39], WMS-III Mental Control, Letter-Number Sequencing, and Spatial Span subtests [40], and Trailmaking Tests A and B [34]).
All study personnel were blinded to study drug assign- ment and APOE4 allele status when administering andscoring cognitive tests.
Fig. 2. CSF Aβ42 vs. CSF T-tau at baseline. The figure displaysthe raw values and predicted relationship after adjusting for gender,APOE4 allele status, age, BMI, and lipoprotein values using Ro- bust-Estimation Wilcoxon-Score General Linear Modeling. A vari-ety of models were analyzed and all showed a statistically significant Laboratory parameters and raw cognitive test scores positive relationship between baseline CSF Aβ42 and CSF T-tau.
were described by means and standard deviationsalong with their corresponding change from baseline.
38], a total sample size of 50 subjects should provide Changes within study groups were assessed by Wilcox- greater than 90% power to detect a 10% change in on paired signed-rank tests. Between group compar- CSF Aβ42 in the simvastatin arm compared to placebo isons of values and changes for CSF biomarkers, the primary outcome, and other laboratory parameters usedWilcoxon signed-rank tests. Between group compar-isons of values and changes in the cognitive tests scores were assessed using Robust Estimation Wilcoxon-scoreanalysis of covariance (ANCOVA), adjusting for age Baseline participant characteristics and education. Analyses were performed using R sta-tistical computing environment [37]. Using data from Baseline participant characteristics (n = 57) are pre- Simons et al. applied to a preclinical population [31, sented in Table 1. CSF was obtained in 50 (88%) partic- C.M. Carlsson et al. / Effects of Simvastatin on Cerebrospinal Fluid Biomarkers and Cognition in Middle-Aged Adults Effects of simvastatin vs. placebo on blood and CSF biomarkers *All values are mean (SD). To convert from mg/dL to mmol/L multiply by 0.0259. HDL-C =high density lipoprotein cholesterol; LDL-C = low-density lipoprotein cholesterol; Aβ42 =amyloid-β 42.
Fig. 3. Boxplots of baseline CSF Aβ42 (A) and CSF T-tau (B) by APOE4 allele status. Whiskers represent the minimum and maximum values.
Boxes represent the 25th and 75th percentiles, while the center line represents the median. Notches signify the 95% confidence interval of themedian.
ipants. Twenty-four (42%) participants were APOE4 oratory abnormalities or unexpected adverse outcomes.
allele carriers (ε2/ε4 n = 2 [8%]; ε3/ε4 n = 21 [88%]; Post-lumbar puncture headache rate was 7%.
and ε4/ε4 n = 1 [4%]). At baseline, all participantshad normal cognitive function, yet 34 (68%) of par- ticipants met CSF biomarker criteria for AD accord-ing to previously established criteria [8,21] (Table 1).
CSF Aβ42 and T-tau levels at baseline were directly At baseline, more subjects in the placebo group were correlated (F [1,36] = 6.10, p = 0.018) in both unad-justed and models adjusted for APOE4, vascular risk using selective serotonin reuptake inhibitors (SSRIs) factors, and gender (Fig. 2). At baseline, APOE4 al- than those in the simvastatin group (simvastatin 3.4% lele carriers had similar demographic and vascular risk vs. placebo 21.4%, p = 0.039). Only one partici- factor characteristics to non-carriers (data not shown, pant (placebo group) was currently clinically depressed all p-values > 0.2). Compared to non-carriers, howev- (Center for Epidemiological Studies Depression Scale er, APOE4 carriers had significantly lower mean val- [CESD] [24] score > 16 points). All other subject char- ues of both CSF Aβ42 (APOE4 non-carriers 945.4 ± acteristics, significant medical history, and medication 255.2 pg/mL; carriers 727.4 ± 184.5 pg/mL, p = 0.002) use were similar in both treatment arms (all p-values > and T-tau (APOE4 non-carriers 620.7 ± 288.7 pg/mL; carriers 434.0 ± 246.9 pg/mL, p = 0.006) (Fig. 3).
Fifty-three subjects (93%) (46 with CSF) completed Age, education, gender, and individual vascular risk the study. Reasons for withdrawal are outlined in Fig. 1.
factors did not correlate with CSF biomarker levels.
None of the participants developed any significant lab- However, gender interacted with body mass index C.M. Carlsson et al. / Effects of Simvastatin on Cerebrospinal Fluid Biomarkers and Cognition in Middle-Aged Adults Effects of simvastatin vs. placebo on selective cognitive measures Visual Reproduction II Delayed Recall [40] Bells Visual Neglect Test (Time, Sec) [9] *All values are mean (SD) changes in raw scores (unless otherwise indicated), adjusted for age and education. CVLT-II = California VerbalLearning Test – 2nd Edition; VSLT = Visual Spatial Learning Test.
Fig. 4. Baseline CSF Aβ42 (A) and T-tau (B) vs. BMI by gender. Figures display raw values (points) and predicted values for women and men(lines) using Robust-Estimation Wilcoxon-Score General Linear Modeling.
(BMI) (Aβ42: F [1, 42] = 7.20, p = 0.010; T-tau: (−41%), and triglyceride levels (−27%) compared to F [1, 42] = 4.47, p = 0.041) and total cholesterol (TC) placebo (Table 2). In unadjusted models and models levels (Aβ42: F [1, 42] = 5.23, p = 0.027) to predict adjusted for APOE4, vascular risk factors, gender, and baseline CSF biomarker values (Figs 4 and 5). Thus, CSF biomarker status (“consistent with AD” or not), men with higher BMI had lower CSF Aβ42 and T-tau changes in CSF Aβ42 and T-tau levels in subjects as- levels compared to men with lower BMI. BMI had lit- signed simvastatin therapy were similar to those seen in tle relationship to CSF biomarkers in women (Fig. 4).
subjects on placebo (Table 2). APOE4 genotype, CSF Men with higher serum TC levels had lower CSF Aβ42 biomarker status (“consistent with AD” or not), and levels than men with normal serum TC levels, while other vascular risk factors did not influence response to in women this relationship was reversed (Fig. 5). No other relationships between vascular risk factors andCSF T-tau levels were observed.
Simvastatin effects on blood and CSF biomarkers At baseline, none of the participants had cognitive As expected, simvastatin significantly lowered TC impairment as demonstrated by a pattern of test scores (−27%), low-density lipoprotein cholesterol (LDL-C) below the 9th percentile rank. Cognitive measures (Ta- C.M. Carlsson et al. / Effects of Simvastatin on Cerebrospinal Fluid Biomarkers and Cognition in Middle-Aged Adults ble 3) were similar in the two treatment groups (all p > 0.1) at baseline, with the exception of the baselinescore on a WMS-III Mental Control subtest [40], on which subjects assigned to placebo outperformed those assigned to simvastatin (F [1,44] = 5.10, p = 0.029).
Subjects assigned simvastatin showed greater improve- ments in tests of verbal fluency and working memory compared to placebo, with trends toward improvementin verbal delayed free recall and construction tasks (Ta- ble 3). At month 4, subjects on placebo were faster on a visual cancellation test compared to those on sim-vastatin (Table 3). Adjusting for APOE4 genotype and baseline differences in SSRI use did not alter cognitive Fig. 5. Baseline CSF Aβ42 vs. TC by gender with predicted valuesfor women and men (lines) using Robust-Estimation Wilcoxon-ScoreGeneral Linear Modeling.
DISCUSSION
differs from previous investigations in that it is larg- In this randomized controlled study of asymptomat- er [12,28,31,32], used a randomized placebo-controlled ic middle-aged adults at increased risk for AD, we design [12,28,32], measured domain-specific cogni- found that subjects treated for four months with sim- tive outcomes [12,28,31,32], and focused on a pre- vastatin 40 mg daily demonstrated similar changes in clinical, middle-aged population with increased risk CSF Aβ42 and T-tau levels compared to those taking for AD (family history of AD and high APOE4 allele placebo, yet greater improvement on some cognitive prevalence) [12,28,31,32]. In an uncontrolled study of measures of verbal fluency, working memory, verbal 19 persons with AD, use of simvastatin 20 mg daily delayed free recall, and construction tasks. Baseline for 3 months was associated with reduced concentra- CSF Aβ42 and T-tau values were lower in APOE4 car- tions of CSF α- and β-cleaved soluble AβPP (αsAβPP, riers compared to non-carriers, but APOE4 genotype βsAβPP), but not with changes in CSF Aβ42, T-tau, did not affect changes in CSF biomarkers or response or phosphorylated-tau (p-tau) [32]. In contrast, a 12- to therapy. While individual vascular risk factors did month follow-up trial in 19 AD patients showed an not significantly influence CSF Aβ42 or T-tau levels increase in both CSF αsAβPP and p-tau following in this study, gender interacted with BMI and TC to simvastatin therapy, but no effect on βsAβPP [12] As predict baseline CSF biomarker levels and changes in this open trial did not have a placebo group, it is un- biomarkers, independent of therapeutic effects. These clear how these statin-associated changes compared novel, preliminary data in a unique group of at-risk to CSF changes noted with natural disease progres- adults may help guide future research investigating the sion. In a 26-week randomized controlled clinical trial effects of vascular risk factor modification on preclini- of 44 persons with AD, simvastatin 80 mg daily re- cal processes related to the risk of developing AD.
duced CSF Aβ40 (but not Aβ42) levels in persons with Statin effects on Aβ metabolism and cognition may mild AD, but not in those with more advanced disease, be significantly influenced by the targeted study popula- suggesting greater statin effects earlier in the disease tion. Persons with preexisting AD may be too far along process [31]. In an uncontrolled study of 23 hyper- in the disease process to reverse the cascade of neuronal cholesterolemic subjects without dementia (ages 34–87 degeneration and healthy controls at average risk for years), 14 weeks of simvastatin 40 mg daily was associ- AD may not benefit from such therapies in the pres- ated with reduced CSF p-tau levels compared to prava- ence of normal AβPP processing. To date, four other statin 80 mg daily, but not with significant changes in prospective clinical trials have evaluated the effects of CSF T-tau, Aβ42, or Aβ40 [28]. Our study population statins on CSF Aβ levels [12,28,31,32]; three of these differs from the study of asymptomatic hypercholes- studies focused on persons with AD [12,31,32] and terolemic subjects in that our participants had fairly one studied cognitively healthy, hypercholesterolemic normal cholesterol levels and were more homogeneous adults at average risk for AD [28]. Of the four stud- in age and AD risk, potential confounding factors for ies, only one included a control group [31]. Our study changes in CSF biomarkers. Together with our find- C.M. Carlsson et al. / Effects of Simvastatin on Cerebrospinal Fluid Biomarkers and Cognition in Middle-Aged Adults ings, these studies suggest that future trials should use a pendently of their effects on Aβ metabolism [16], pos- randomized controlled design, target high risk asymp- sibly through modifying signaling pathways, reducing tomatic adults, and investigate the effects of higher inflammation, improving cerebral blood flow, and/or dose and/or longer duration statin therapy on not only via more acute cholinomimetic effects [4]. Thus, some Aβ metabolism, but also on other preclinical processes of the positive cognitive changes noted in our study closely related to the development of AD.
may be due to some of these other unmeasured process- Cognitive outcomes from ours and previous trials es. While some AD-related mechanisms may be more also suggest that the optimal timing for statin therapy rapidly reversed with statins, others may require longer may be in the preclinical or early disease stages, be- duration of therapy. Although statins reduce CSF and fore neurodegeneration becomes irreversible. In addi- brain Aβ levels in animals in as short as three weeks [7], tion, findings from ours and other studies support that our study may not have been long enough to fully assess domain-specific neuropsychological tests may be more simvastatin’s impact on CSF Aβ levels. Longer trials sensitive in distinguishing preclinical cognitive bene- are underway investigating pleiotropic effects of statins fits in asymptomatic persons at risk for AD compared on cognition and AD-related mechanisms of disease in to global measures of cognitive function. In the three previously discussed CSF studies in AD patients, statin Despite normal cognitive function in our partic- therapy was associated with conflicting results on glob- ipants, our findings also suggest that asymptomat- al cognitive measures (Mini Mental State Examination ic middle-aged APOE4 carriers may already exhibit [MMSE] and Alzheimer’s Disease Assessment Scale – pathobiological changes in Aβ metabolism compared Cognitive Portion [ADAS-cog]) [12,31,32]. In a 12- to APOE4 non-carriers. Persons with AD have lower month randomized controlled trial of 63 older adults CSF Aβ42 and higher CSF T-tau levels than healthy with AD, Sparks et al. demonstrated that atorvastatin controls [35], possibly from increased Aβ deposition may improve cognition in older adults with AD, with into amyloid plaques, reduced synthesis of Aβ by dy- the greatest cognitive benefits noted in APOE4 carri- ing neurons, and intracellular release of tau into the ers who are early in their disease course [33]. In a CSF. In older adults and persons with mild cognitive study by Muldoon et al., hypercholesterolemic middle- impairment (MCI), lower levels of CSF Aβ42 [1,6] and aged adults at average risk for AD were noted to have higher levels of CSF T-tau predict future risk of demen- less improvement in cognition following simvastatin tia [1]. Our findings of lower baseline CSF Aβ42 levels therapy compared to placebo [23]. In contrast, our in asymptomatic APOE4 carriers are consistent with asymptomatic at-risk participants randomized to sim- those noted in the BIOCARD study [36] and may sug- vastatin had improvements in domain-specific tests of gest increased risk for future cognitive decline. How- verbal fluency, working memory, verbal delayed free ever, we also surprisingly found lower CSF T-tau levels recall, and construction tasks. The cognitive domains in persons at increased genetic risk for AD (APOE4 which improved with statin therapy in our study are carriers) as well as in those participants with lower CSF similar to those measures noted to decline in asymp- Aβ42 levels (suggesting a possible increased risk for tomatic middle-aged APOE4 carriers [3], suggesting disease progression). Our CSF T-tau findings differ that statins may stabilize these vulnerable cognitive do- from those noted in the BIOCARD study, in which mains in this at-risk cohort with high APOE4 preva- APOE4 carriers and non-carriers had similar mean CSF lence. Our at-risk population may be more sensitive to T-tau values [36]. While the T-tau results were un- the beneficial cognitive effects of potential preventive expected, the strong association of the CSF T-tau lev- therapies; however, given the modest magnitude of the els with APOE4 status and Aβ levels support that fur- cognitive changes observed in our study, longer studies ther investigation is needed to clarify the longitudinal are necessary to establish if these differences in cog- changes in CSF Aβ42 and T-tau levels spanning the nitive test scores lead to clinically beneficial cognitive preclinical phase of the disease through the develop- changes. Given that cognitive outcomes were an ex- ment of AD. Based on previously published studies of ploratory secondary aim of this trial, we did not cor- persons with AD [8,21], many participants in our study rect for multiple comparisons. Thus, larger prospec- had combined CSF Aβ42 and T-tau biomarker levels tive controlled trials of longer duration are needed to meeting criteria for AD despite their having normal cognitive function. Further investigation is necessary Consistent with our study results, animal and clinical to evaluate if this combination of biomarkers predict studies support that statins may benefit cognition inde- future risk for cognitive decline in an asymptomatic C.M. Carlsson et al. / Effects of Simvastatin on Cerebrospinal Fluid Biomarkers and Cognition in Middle-Aged Adults middle-aged at-risk population or if such CSF measures ACKNOWLEDGMENTS
are less accurate in this age group. Findings from large,multicenter trials such as the ongoing Alzheimer’s Dis- This study was funded by the American Federation ease Neuroimaging Initiative (ADNI) will help clarify for Aging Research (AFAR) Merck/AFAR Junior In- the natural progression of CSF biomarkers throughout vestigator Award in Geriatric Clinical Pharmacology, the disease course and will establish the best techniques the University of Wisconsin (UW) General Clinical Re- for CSF biomarker collection and analysis to optimize search Center (GCRC) (NCRR grant M01 RR03186) and the UW GCRC Clinical Research Feasibility Fund We also found that increased TC levels and BMI (CReFF) Program. Dr. Carlsson was supported in were associated with lower CSF Aβ42 levels in men, part through NIA grant 1K23 AG021605-01 A1 and a suggesting possible gender differences in the effects of Beeson Career Development Award (1K23 AG026752- vascular risk factors on AD progression. In general, 01, a grant jointly funded by NIA, the John A. Hart- men have more atherogenic lipid profiles than wom- ford Foundation, Atlantic Philanthropies, and the Starr en, which could explain the gender differences noted Foundation). Merck and Co., Inc. provided all study in association with CSF Aβ42 levels; however, we did medications for this investigator-initiated trial.
not find an association between CSF Aβ42 levels and Stein has served as a consultant for LipoScience and lipoprotein markers of atherogenicity, including LDL Pfizer; received research and/or training grants from particle number, LDL size, or HDL size. Our findings Astra-Zeneca, Bristol-Myers Squibb, Kos, and Lipo- that gender and BMI interact to affect CSF biomarkers Science; and has been on the Speaker’s Bureaus for are supported by data showing a link between obesi- Kos, LipoScience, Merck, Pfizer, Schering-Plough, and ty, gender, inflammation, AβPP processing, and AD Takeda. Dr. McBride has served as a consultant for Re- risk [11]. Further investigation is needed into the im- liant and Johnson and Johnson. Dr. Asthana received pact of dyslipidemia, obesity, and gender on the neuro- grant support from Elan, Merck and Eisai. The other authors have no conflicts of interest to disclose.
This is Madison VA GRECC manuscript #2007-02.
The authors would like to sincerely thank all of our CONCLUSIONS
dedicated study participants as well as Tracy Ohrt,Sarah Meade, Neala Lane, and Angela Slattery for theirimportant contributions to this study.
In asymptomatic middle-aged adults at increased risk for AD, treatment with simvastatin led to similarchanges in CSF Aβ42 and T-tau levels compared to References
placebo, yet greater improvement in some cognitivemeasures of verbal fluency, working memory, verbaldelayed free recall, and construction tasks. APOE4 N. Andreasen, E. Vanmechelen, H. Vanderstichele, P. Davids-son and K. Blennow, Cerebrospinal fluid levels of total- genotype and the interaction of gender with BMI and tau, phospho-tau and A beta 42 predicts development of TC levels influenced CSF Aβ42 and T-tau levels in Alzheimer’s disease in patients with mild cognitive impair- this group of at-risk adults, but did not affect response ment, Acta Neurol Scand Suppl 179 (2003), 47–51.
A. Benton and K. Hamsher, Multilingual Aphasia Examina- to simvastatin. Further studies are needed to investi- tion, AJA Associates, Iowa City, IA, 1989.
gate the impact of higher dose and/or longer duration R.J. Caselli, E.M. Reiman, D. Osborne, J.G. Hentz, L.C. Bax- statin therapy on mechanisms related to the develop- ter, J.L. Hernandez and G.G. Alexander, Longitudinal changesin cognition and behavior in asymptomatic carriers of the ment of AD, including Aβ metabolism, inflammation, APOE e4 allele, Neurology 62 (2004), 1990–1995.
and cerebral perfusion. In addition, findings from the S. Darvesh, E. Martin, R. Walsh and K. Rockwood, Differen- present study emphasize the need for future investiga- tial effects of lipid-lowering agents on human cholinesterases, tions to assess the interactive effects of gender, genetic Clin Biochem 37 (2004), 42–49.
D.C. Delis, J.H. Kramer, E. Kaplan and B.A. Ober, California risk, and vascular biology on preclinical AD biomark- Verbal Learning Test-II, The Psychological Corporation, San ers. Ultimately, large randomized, controlled clinical trials are needed to best assess the effects of statins on A.M. Fagan, C.M. Roe, C. Xiong, M.A. Mintun, J.C. Morrisand D.M. Holtzman, Cerebrospinal Fluid tau/beta-Amyloid42 long-term cognitive outcomes and the development of Ratio as a Prediction of Cognitive Decline in Nondemented AD in asymptomatic adults at high risk for the disease.
Older Adults, Arch Neurol (2007).
C.M. Carlsson et al. / Effects of Simvastatin on Cerebrospinal Fluid Biomarkers and Cognition in Middle-Aged Adults K. Fassbender, M. Simons, C. Bergmann, M. Stroick, D.
cerebrospinal fluid of patients with Alzheimer’s disease, Ann Lutjohann, P. Keller, H. Runz, S. Kuhl, T. Bertsch, K. von Neurol 38 (1995), 643–648.
Bergmann, M. Hennerici, K. Beyreuther and T. Hartmann, S.G. Mueller, M.W. Weiner, L.J. Thal, R.C. Petersen, C. Jack, Simvastatin strongly reduces levels of Alzheimer’s disease be- W. Jagust, J.Q. Trojanowski, A.W. Toga and L. Beckett, The ta -amyloid peptides Abeta 42 and Abeta 40 in vitro and in Alzheimer’s disease neuroimaging initiative, Neuroimaging vivo, Proc Natl Acad Sci U S A 98 (2001), 5856–5861.
Clin N Am 15 (2005), 869–877, xi–xii.
D. Galasko, L. Chang, R. Motter, C.M. Clark, J. Kaye, D.
M.F. Muldoon, C.M. Ryan, S.M. Sereika, J.D. Flory and S.B.
Knopman, R. Thomas, D. Kholodenko, D. Schenk, I. Lieber- Manuck, Randomized trial of the effects of simvastatin on burg, B. Miller, R. Green, R. Basherad, L. Kertiles, M.A. Boss cognitive functioning in hypercholesterolemic adults, Am J and P. Seubert, High cerebrospinal fluid tau and low amyloid Med 117 (2004), 823–829.
beta42 levels in the clinical diagnosis of Alzheimer disease L.S. Radloff, The CES-D scale: A self-report depression scale and relation to apolipoprotein E genotype, Arch Neurol 55
for research in the general population., Appl Psychol Meas 1
L. Gauthier, F. Dehaut and Y. Joanette, The Bells Test: A T.D. Rea, J.C. Breitner, B.M. Psaty, A.L. Fitzpatrick, O.L.
quantitative and qualitative test for visual neglect, Int J of Clin Lopez, A.B. Newman, W.R. Hazzard, P.P. Zandi, G.L. Burke, Neurop 11 (1989), 49–54.
C.G. Lyketsos, C. Bernick and L.H. Kuller, Statin use and the I. Hajjar, J. Schumpert, V. Hirth, D. Wieland and G.P. Eleazer, risk of incident dementia: the Cardiovascular Health Study, The impact of the use of statins on the prevalence of dementia Arch Neurol 62 (2005), 1047–1051.
and the progression of cognitive impairment, J Gerontol A L.M. Refolo, B. Malester, J. LaFrancois, T. Bryant-Thomas, Biol Sci Med Sci 57 (2002), M414–M418.
R. Wang, G.S. Tint, K. Sambamurti, K. Duff and M.A. Pap- J. Harvey, Leptin: a multifaceted hormone in the central ner- polla, Hypercholesterolemia accelerates the Alzheimer’s amy- vous system, Mol Neurobiol 28 (2003), 245–258.
loid pathology in a transgenic mouse model, Neurobiol Dis 7
K. Hoglund, K.M. Thelen, S. Syversen, M. Sjogren, K. von Bergmann, A. Wallin, E. Vanmechelen, H. Vanderstichele, D.
A. Rey, Psychological examination of traumatic encephalopa- Lutjohann and K. Blennow, The effect of simvastatin treat- thy. Archives de Psychologie 28, 286–340; sections translated
ment on the amyloid precursor protein and brain cholesterol by J. Corwin, & F.W. Blysma, The Clinical Neuropsychologist metabolism in patients with Alzheimer’s disease, Demen Geri- atr Cogni Disord 19 (2005), 256–265.
R.G. Riekse, G. Li, E.C. Petrie, J.B. Leverenz, D. Vavrek, S.
S.C. Johnson, T.W. Schmitz, M.A. Trivedi, M.L. Ries, B.M.
Vuletic, J.J. Albers, T.J. Montine, V.M. Lee, M. Lee, P. Seu- Torgerson, C.M. Carlsson, S. Asthana, B.P. Hermann and bert, D. Galasko, G.D. Schellenberg, W.R. Hazzard and E.R.
M.A. Sager, The influence of Alzheimer disease family his- Peskind, Effect of statins on Alzheimer’s disease biomarkers tory and apolipoprotein E epsilon4 on mesial temporal lobe in cerebrospinal fluid, J Alzheimers Dis 10 (2006), 399–406.
activation, J Neurosci 26 (2006), 6069–6076.
K. Rockwood, S. Kirkland, D.B. Hogan, C. MacKnight, H.
M. Kivipelto, E.L. Helkala, M.P. Laakso, T. Hanninen, M. Hal- Merry, R. Verreault, C. Wolfson and I. McDowell, Use of lipid- likainen, K. Alhainen, S. Iivonen, A. Mannermaa, J. Tuomile- lowering agents, indication bias, and the risk of dementia in hto, A. Nissinen and H. Soininen, Apolipoprotein E epsilon4 community-dwelling elderly people, Arch Neurol 59 (2002),
allele, elevated midlife total cholesterol level, and high midlife systolic blood pressure are independent risk factors for late-life M.A. Sager, B. Hermann and A. La Rue, Middle-aged children Alzheimer disease, Ann Intern Med 137 (2002), 149–155.
of persons with Alzheimer’s disease: APOE genotypes and H. Kløve, ed., Clinical Neuropsychology, New York: Saun- cognitive function in the Wisconsin Registry for Alzheimer’s ders; 1963. Forster FM, ed. The medical clinics of North Prevention, J Geriatr Psychiatry Neurol 18 (2005), 245–249.
M. Simons, F. Schwarzler, D. Lutjohann, K. von Bergmann, K.
L. Li, D. Cao, H. Kim, R. Lester and K. Fukuchi, Simvastatin Beyreuther, J. Dichgans, H. Wormstall, T. Hartmann and J.B.
enhances learning and memory independent of amyloid load Schulz, Treatment with simvastatin in normocholesterolemic in mice, Ann Neurol 60 (2006), 729–739.
patients with Alzheimer’s disease: A 26-week randomized, J. Malec, R. Ivnik, G. Smith, E. Tangalos, P. RC, E. Kokmen placebo-controlled, double-blind trial, Ann Neurol 52 (2002),
and L. Kurland, Visual spatial learning test: Normative data and further validation, Psychol Assess 4 (1997), 433–441.
M. Sjogren, K. Gustafsson, S. Syversen, A. Olsson, A. Ed- J.F. Malec, R.J. Ivnik and N.S. Hinkeldey, Visual Spatial man, P. Davidsson, A. Wallin and K. Blennow, Treatment with Learning Test, Psychol Assess 3 (1991), 82–88.
simvastatin in patients with Alzheimer’s disease lowers both G. McKhann, D. Drachman, M. Folstein, R. Katzman, D.
alpha- and beta-cleaved amyloid precursor protein, Dement Price and E.M. Stadlan, Clinical diagnosis of Alzheimer’s Geriatr Cogn Disord 16 (2003), 25–30.
disease: report of the NINCDS-ADRDA Work Group under D.L. Sparks, D.J. Connor, M.N. Sabbagh, R.B. Petersen, the auspices of Department of Health and Human Services J. Lopez and P. Browne, Circulating cholesterol levels, Task Force on Alzheimer’s Disease, Neurology 34 (1984),
apolipoprotein E genotype and dementia severity influence the benefit of atorvastatin treatment in Alzheimer’s disease: re- A.U. Monsch, M.W. Bondi, N. Butters, D.P. Salmon, R. Katz- sults of the Alzheimer’s Disease Cholesterol-Lowering Treat- man and L.J. Thal, Comparisons of verbal fluency tasks in the ment (ADCLT) trial, Acta Neurol Scand Suppl 185 (2006),
detection of dementia of the Alzheimer type, Arch Neurol 49
O. Spreen and E. Strauss, A compendium of neuropsychologi- R. Motter, C. Vigo-Pelfrey, D. Kholodenko, R. Barbour, K.
cal tests, Oxford University Press, New York, 1998.
Johnson-Wood, D. Galasko, L. Chang, B. Miller, C. Clark, T. Sunderland, G. Linker, N. Mirza, K.T. Putnam, D.L. Fried- R. Green et al., Reduction of beta-amyloid peptide42 in the man, L.H. Kimmel, J. Bergeson, G.J. Manetti, M. Zimmer-mann, B. Tang, J.J. Bartko and R.M. Cohen, Decreased beta- C.M. Carlsson et al. / Effects of Simvastatin on Cerebrospinal Fluid Biomarkers and Cognition in Middle-Aged Adults amyloid1-42 and increased tau levels in cerebrospinal fluid Alzheimer disease, Alzheimer Dis Assoc Disord 20 (2006),
of patients with Alzheimer disease, Jama 289 (2003), 2094–
D. Wechsler, WAIS-R Wechsler Adult Intelligence Scale-III, T. Sunderland, N. Mirza, K.T. Putnam, G. Linker, D. Bhupali, Psychological Corporation, New York, N.Y., 1991.
R. Durham, H. Soares, L. Kimmel, D. Friedman, J. Berge- D. Wechsler, Wechsler Memory Scales-3rd Edition, The Psy- son, G. Csako, J.A. Levy, J.J. Bartko and R.M. Cohen, Cere- chological Corporation, San Antonio, TX, 1997.
brospinal fluid beta-amyloid(1-42) and tau in control subjects B. Wolozin, W. Kellman, P. Ruosseau, G.G. Celesia and G.
at risk for Alzheimer’s disease: The effect of APOE epsilon4 Siegel, Decreased prevalence of Alzheimer disease associ- allele, Biol Psychiatry 56 (2004), 670–676.
ated with 3-hydroxy-3-methyglutaryl coenzyme A reductase R.D.C. Team, R: A Language and Environment for Statistical inhibitors, Arch Neurol 57 (2000), 1439–1443.
Computing., R Foundation for Statistical Computing, Vienna, P.P. Zandi, D.L. Sparks, A.S. Khachaturian, J. Tschanz, M.
Norton, M. Steinberg, K.A. Welsh-Bohmer and J.C. Breitner, L.J. Thal, K. Kantarci, E.M. Reiman, W.E. Klunk, M.W. Wein- Do statins reduce risk of incident dementia and Alzheimer er, H. Zetterberg, D. Galasko, D. Pratico, S. Griffin, D. Schenk disease? The Cache County Study, Arch Gen Psychiatry 62
and E. Siemers, The role of biomarkers in clinical trials for

Source: http://www.aging.wisc.edu/pdfs/30.pdf