Bilal Khokhar Department of Pharmaceutical Health Services

Management of Traumatic Brain Injury with Statins among Older Medicare Beneficiaries

Item Type dissertation

Authors Khokhar, Bilal

Publication Date 2016

Abstract Background: Traumatic brain injury (TBI) is a major health concern for older adults aged 65 and older. Older TBI patients are at increased risk of primary injury (in-hospital and all-cause mortality) and secondary injury (stroke, depression, and Alzh...

Keywords older adults; secondary injury; statins; Older people; Brain Injuries, Traumatic; Medicare; Mortality

Download date 25/09/2021 02:50:19

Link to Item http://hdl.handle.net/10713/6289 Bilal Khokhar Department of Pharmaceutical Health Services Research University of Maryland School of Pharmacy 220 Arch Street, Saratoga Offices 12th Floor, Baltimore MD, 21201, USA Email: [email protected]

EDUCATION

PhD University of Maryland Baltimore December, 2016 Pharmaceutical Health Services Research Dissertation: Management of Traumatic Brain Injury in Older Medicare Beneficiaries with Statins Committee: Linda Simoni-Wastila (chair), Eleanor Perfetto, Julia Slejko, Gordon Smith, Min Zhan

MS University of Maryland Baltimore December, 2016 Epidemiology

Certificate of University of Maryland Baltimore August, 2016 Completion Epidemiology of Aging

MA University of Maryland Baltimore County December 1, 2010 Sociology

BS Towson University May 1, 2008 Sociology-Anthropology

TEACHING EXPERIENCE

Guest Lecturer, University of Maryland Baltimore "Neurology Lectures: Traumatic Brain Injury (TBI)" Applied Spring 2016 Sciences and Therapeutics 8

 Epidemiology and public health consequences of TBI  Management of TBI  Unmet needs and direction of future research

"Pharmacovigilance and regulatory science" Population Based Fall 2013 Strategies I: Pharmacoepidemiology and Pharmacoeconomics

 Define science of drug safety detection and assessment  Identify regulatory tools to examine efficacy and quality of FDA-regulated products

Teaching Assistant, University of Maryland Baltimore Population Based Strategies I: Pharmacoepidemiology and September 2011 - Pharmacoeconomics December 2015  Develop and lead live case studies and review sessions  Develop and proctor quizzes and exams  Manage course on Blackboard

Population Based Strategies II: Public Health Pharmacy January 2012 - June 2013  Lead five large group discussions and several review sessions  Develop and proctor quizzes and exams  Manage course on Blackboard

Teaching Interests Medical sociology, health services research, health policy analysis, health equity, health behavior, epidemiology, trauma epidemiology, statistics, research methods

PROFESSIONAL EXPERIENCE

Pre-doctoral Fellow, University of Maryland Baltimore School of September 2013 - Pharmacy present  Conduct administrative claims research on a variety of topics including traumatic brain injury, depression, and chronic obstructive pulmonary disease  Conduct research on relevant health outcomes including mortality, hospitalization, and drug adherence  Conduct focus groups and survey research  Conduct FDA and private industry sponsored research  Help prepare grants for future research  Work on a collaborative team of health services researchers, epidemiologists, pharmacists, and physicians

Graduate Research Assistant, University of Maryland Baltimore November 2009 - County & University of Maryland Baltimore August 2013  Conduct primary and secondary data analysis  Screen Maryland Violent Death Reporting Systems (MVDRS) data

 Review violent death cases from the Office of Medical Examiner (OME)  Code circumstances for deaths including alcohol/drug overdoses, depression, and other mental illnesses  Work as a team with the OME and various law enforcement agencies

Family Services Caseworker, Baltimore City Department of March 2009 - Social Services October 2009  Provide Case management and direct care to Baltimore City youth  Ensure children in care receive sufficient medical care including mental health care  Mentor and provide support and resources and services for struggling youth in care

HONORS & AWARDS

Pre-doctoral Research Fellowship, Epidemiology of Aging September 2013 - Training Program, Department of Epidemiology and Preventive present Medicine, School of Medicine, University of Maryland, Baltimore (T32 AG000262-14)

Winner of student travel award to present “In-hospital mortality May, 2016 following traumatic brain injury in older adult statin users.” Graduate School Association, University of Maryland, Baltimore

Winner of the M-CERSI Talent Competition Session, "Mobile May, 2015 Application Conveying Safety Information to Communicate Risks for Medications and Devices". University of Maryland Baltimore School of Pharmacy

Best Student Abstract, “The risk of stroke among elderly June, 2014 antidepressant users who have suffered a traumatic brain injury: a nested-cohort study” Academy Health Annual Research Meeting, San Diego

Selected to participate in AcademyHealth’s inaugural Pecha June, 2014 Kucha presentation competition for research “The risk of stroke among elderly antidepressant users who have suffered a traumatic brain injury: a nested-cohort study”

Winner, Geriatrics and Gerontology Education and Research March, 2014 (GGEAR) Award for Excellence in Research in the Field of Aging. Graduate Research Conference, University of Maryland Baltimore

Winner, poster session B “The risk of stroke among elderly March, 2014 antidepressant users with incident depression following traumatic brain injury.” Graduate Research Conference, University of Maryland Baltimore

Graduate Student of the Month, University of Maryland October, 2012 Baltimore School of Pharmacy

GRANTS

Active Pre-doctoral trainee, "Research Training in the Epidemiology of September 2013 - Aging" National Institutes of Health (T32AG000262-14). PI: Jay August 2016 Magaziner

Submitted Principal Investigator, "Management of Traumatic Brain Injury among Older Medicare Beneficiaries with statins" National Institutes of Health, unfunded

PUBLISHED MANUSCRIPTS

Khokhar B, Simoni-Wastila L, Albrecht JS. Risk of stroke among older Medicare antidepressant users with traumatic brain injury. J Head Trauma Rehabil. 2016. [Epub ahead of print].

Albrecht JS, Park Y, Hur P, Huang TY, Harris I, Netzer G, Lehmann SW, Langenberg P, Khokhar B, Wei YJ, Moyo P, Simoni-Wastila L. Adherence to Maintenance Medications among Older Adults with Chronic Obstructive Pulmonary Disease: The Role of Depression. Ann Am Thorac Soc, 2016. [Epub ahead of print].

Albrecht JS, Huang TY, Park Y, Langenberg P, Harris I, Netzer G, Lehmann SW, Khokhar B, Simoni-Wastila L. New Episodes of Depression among Medicare Beneficiaries with Chronic Obstructive Pulmonary Disease. Int J Geriatr Psychiatry. 2015; 31(5): 441-449.

Albrecht JS, Khokhar B, Pradel F, Campbell M, Palmer J, Harris I and Palumbo F. Perceptions of patient provider agreements. J Pharm Health Serv Res. 2015; 6(3), 139- 144.

Quinn C, Khokhar B, Weed K, Barr E, & Gruber-Baldini AL Older Adult Self-Efficacy Study of Mobile Phone Diabetes Management. Diabetes technol ther. 2015; 17(7), 455- 461.

Albrecht JS, Kiptanui Z, Tsang Y, Khokhar B, Smith GS, Zuckerman IH, Simoni-Wastila L. Patterns of Depression Treatment in Medicare Beneficiaries with Depression Following Traumatic Brain Injury. J Neurotrauma. 2014; 32(16), 1223-1229.

Albrecht JS, Kiptanui Z, Tsang Y, Khokhar B, Liu X, Simoni-Wastila L, Zuckerman IH. Depression among Older Adults Following Traumatic Brain Injury: A National Analysis. Am J of Geriatr Psychiatry. 2014; 23(6), 607-614.

Shen X, Dutcher SK, Palmer JB, Liu X, Kiptanui Z, Khokhar B, Al-Jawadi MH, Zhu Y, Zuckerman IH. A Systematic Review of Benefits and Risks of Anticoagulation Following Traumatic Brain Injury. J Head Trauma Rehabil. 2014; 30(4), E29-E37.

Under Review:

Albrecht JS, Khokhar B, Huang TY, Wei YJ, Harris I, Moyo P, Hur P, Lehmann SW, Netzer G, Simoni-Wastila L. Medication Adherence and Healthcare Utilization among Older Adults with Chronic Obstructive Pulmonary Disease and Depression. J Am Geriatr Soc.

Wei YJ, Simoni-Wastila L, Albrecht JS, Huang TY, Moyo P, Khokhar B, Harris I, Langenberg P, Netzer G, Lehmann SW. Influence of antidepressant treatment on COPD maintenance medication use and adherence in a comorbid Medicare population: A longitudinal cohort study. Med Care

Working Manuscripts:

Khokhar B, Simoni-Wastila L, Slejko JF, Perfetto E, Zhan M, Smith GS. Statin use in older Medicare beneficiaries with traumatic brain injury.

Khokhar B, Simoni-Wastila L, Slejko JF, Perfetto E, Zhan M, Smith GS. In-hospital Mortality Following Traumatic Brain Injury in Older Adult Statin Users.

Khokhar B, Simoni-Wastila L, Slejko JF, Perfetto E, Zhan M, Smith GS. Mortality and Secondary Injury Following Traumatic Brain Injury in Older Medicare Statin Users.

PODIUM PRESENTATIONS

Khokhar B, Slejko J, Perfetto E, Smith GS, Zhan M, Simoni-Wastila L. In-hospital mortality following traumatic brain injury in older adult statin users. World Congress on Brain Injury, The Hague, Netherlands, March 2016.

Dutcher S, Franey C, Harris I, Khokhar B, Kiptanui Z, Palumbo F, Park J, Polli J, Pradel F. Post-marketing Surveillance of Generic Drug Usage and Substitution Patterns. U.S. Food and Drug Administration. Office of Generic Drugs, Silver Spring, Maryland, October 2015.

Khokhar B, Sieluk S, Gaitonde P, Oehrlein E, Hanna M. Mobile Application Conveying Safety Information to Communicate Risks of Medications and Devices. 2015 ORSI Science Symposium. U.S. Food and Drug Administration, Silver Spring, Maryland, April 2015.

POSTER PRESENTATIONS

Hur P, Albrecht JS, Huang TY, Moyo P, Khokhar B, Wei YJ, Harris I, Simoni-Wastila L. Relationship Between COPD Severity and Acute Inhaler Use in Chronic Obstructive Pulmonary Disease. ISPOR 21st International Meeting, Washington DC, May 2016.

Hur P, Albrecht JS, Simoni-Wastila L., Huang TY, Moyo P, Wei YJ, Harris I, Khokhar B, Severity and Acute Inhaler Use in Chronic Obstructive Pulmonary Disease. AMCP Managed Care & Specialty Pharmacy Annual Meeting, San Francisco, CA, April 2016.

Khokhar B, Park Y, Pradel F, Palumbo F, Kiptanui Z, Dutcher S, Jiang W, Harris I. Patients’ awareness of bioequivalence study methods supporting generic venlafaxine extended release (ER) tablet approval. International Conference of Pharmacoepidemiology & Therapeutic Risk Management, Boston, Massachusetts, August 2015.

Khokhar, B, Albrecht JS, Liu X, Tsang Y, Kiptanui Z, Simoni-Wastila L, Zuckerman IH. The risk of stroke among elderly antidepressant users who have suffered a traumatic brain injury: a nested-cohort study Academy Health Annual Research Meeting, San Diego, California, June 2014.

Khokhar B, Albrecht JS, Tsang Y, Kiptanui Z, Simoni-Wastila L, Zuckerman IH. The risk of stroke among elderly antidepressant users with incident depression following traumatic brain injury. Graduate Research Conference, University of Maryland Baltimore, Baltimore, Maryland, 2014.

Khokhar B, Albrecht JS, Tsang Y, Kiptanui Z, Simoni-Wastila L, Zuckerman IH. The Effect of Antidepressants on the Risk of Stroke among Older Adults Who Develop Depression after a Traumatic Brain Injury. The Gerontological Society of America's Annual Meeting, Washington, DC, November 2014.

Albrecht JS, Kiptanui Z, Tsang Y, Khokhar B, Liu X, Simoni-Wastila L, Zuckerman IH. Depression among Older Adults Following Traumatic Brain Injury. The Gerontological Society of America's Annual Meeting, Washington, DC, November 2014.

Liu X, Baumgarten M, Smith GS, Gottlieb S, Franey C, Khokhar B, Rattinger G, Zuckerman I Warfarin usage in elderly atrial fibrillation patients who experienced traumatic brain injury. Poster session presented at: GSA Annual Scientific Meeting. 66th Annual Conference of the Gerontological Society of America; 2013 Nov 20-24; New Orleans, Louisiana.

Shen X, Dutcher S, Palmer J, Liu X, Al-Jawadi M, Khokhar B, Kiptanui Z, Zhu Y. A systematic review on benefits and risks of anticoagulant use post traumatic brain injury. Poster session presented at: GSA Annual Scientific Meeting. 66th Annual Conference of the Gerontological Society of America; 2013 Nov 20-24; New Orleans, Louisiana.

Khokhar B, Zuckerman I, Stuart B. Preventable hospitalization rates among Spanish- speaking and English-speaking Hispanics in Medicare: Analysis using the Medicare Current Beneficiary Survey. Poster session presented at: Academy Health Annual Research Meeting. 30th Annual Conference of Academy Health; 2013 Jun 23-25; Baltimore, Maryland.

INSTITUTIONAL SERVICE

Cultural Secretary, Indian Association February 2016 - present  Promote cultural awareness through campus  Organize and promote university wide cultural events

Project Jump Start August 2015 - present  Prepare meals for the homeless of Baltimore  Distribute prepared meals to homeless in various inner-city locations

Strategic planning committee member, The Writing Center July 2015 - present  Brainstorm avenues to engage university students  Conceptualize key aspects of the writing center, including mission, budget, communication, resources, and evaluation

Muslim Student and Scholars Association August 2012 - present  Assist in association's fundraising efforts  Promote cultural awareness throughout university community

Department of Pharmaceutical Health Services Research July 2015 - June 2016 Student Representative, University of Maryland Baltimore Graduate Student Association

Epidemiology of Aging Trainee Seminar Coordinator June 2014 - June 2015  Invite speakers to present pertinent aging-related topics to students  Organize student networking sessions with professionals conducting aging research

PROFESSIONAL MEMBERSHIPS

AcademyHealth, student member Gerontological Society of America, student member International Society for Pharmacoepidemiology, student member Secretary, student chapter, University of Maryland Baltimore, May 2013 - May 2014 International Society for Pharmacoeconomics and Outcomes Research student member

SKILLS

Experience in study design and analysis of large claims datasets Experience with primary data collection and analysis Proficient in SAS, STATA, SPSS, & Microsoft Office Fluent in three languages (English, Urdu, Punjabi)

ABSTRACT

Title of Dissertation: MANAGEMENT OF TRAUMATIC BRAIN INJURY WITH STATINS AMONG OLDER MEDICARE BENEFICIARIES

Bilal Khokhar, Doctor of Philosophy, 2016

Dissertation directed by: Linda Simoni-Wastila, PhD, Professor Department of Pharmaceutical Health Services Research

Background: Traumatic brain injury (TBI) is a major health concern for older adults aged 65 and older. Older TBI patients are at increased risk of primary injury (in-hospital and all-cause mortality) and secondary injury (stroke, depression, and Alzheimer’s disease and related dementias (ADRD)). There is limited research regarding optimal pharmacotherapeutic options and management of TBI patients; however, several studies have highlighted statins, used to treat hyperlipidemia, as potential pharmacologic agents to reduce inflammation and improve impaired cerebral blood flow associated with primary and secondary injury. The objectives of the study are to: 1) quantify statin utilization, and 2) determine the associations between statin use and primary and secondary injury among TBI patients. Methods: Statin use (atorvastatin, fluvastatin, lovastatin, pravastatin, rosuvastatin, and simvastatin), primary injury, and secondary injury were examined among Medicare beneficiaries hospitalized with a TBI between

2006 and 2010. Logistic regression was used to investigate the relationship between pre-

TBI statin use and in-hospital mortality, while discrete time analysis was used to investigate the relationship between statin use following TBI and all-cause mortality and

secondary injury. Results: Among the 75,698 beneficiaries who met study criteria,

37,874 (50.0%) beneficiaries used a statin at least once during the study period. The most common statin used was simvastatin, followed by atorvastatin. Fluvastatin was the least used statin. Pre-TBI use of atorvastatin (odds ratio (OR) 0.88; 95% confidence interval (CI) 0.82, 0.96), simvastatin (OR 0.84; 95% CI 0.79, 0.91), and rosuvastatin (OR

0.79; 95% CI 0.67, 0.94) were associated with significant decreases in the risk of in- hospital mortality. Any statin use was associated with reduced all-cause mortality following TBI-hospitalization discharge. Atorvastatin and simvastatin use also were associated with reductions in all secondary injury outcomes. Conclusion: Tens of thousands of older adults are hospitalized annually with TBI and experience disabling primary and secondary injury; findings from these analyses have salient implications for reducing the risk of TBI complications among older adults. The evidence generated suggests that preemptive use of statins may decrease the risk of in-hospital and all-cause mortality, as well as reduce the likelihood of stroke, depression, and ADRD.

Management of Traumatic Brain Injury with Statins among Older Medicare Beneficiaries

by Bilal Khokhar

Dissertation submitted to the Faculty of the Graduate School of the University of Maryland, Baltimore in partial fulfillment of the requirements for the degree of Doctor of Philosophy 2016

Everything that we do must be done with the proper intentions. This dissertation is dedicated to anyone with a desire to learn about the scientific method and traumatic brain injury.

iii

Acknowledgements

I have been blessed to have not just one, but three incredible mentors. As a fresh PhD student just over five years ago, Dr. Françoise Pradel taught me about professionalism and how to navigate life as a PhD student. Later, Dr. Ilene Harris took me under her wing and gave me the opportunities to expand my boundaries as a scientist, allowing me to become a more rounded researcher. Finally, Dr. Linda Simoni-Wastila is more than just a mentor; she is a friend and confidant. She has provided me with autonomy to complete my dissertation while also giving me valuable experiences with claims data analyses. Furthermore, she has provided cherished mentorship regarding leadership and the pursuit of career goals. Having the benefit of three wonderful mentors, I was able to complete my dissertation and grow professionally and personally. I owe gratitude to my dissertation committee: Dr. Simoni-Wastila for advising me on pharmacy related concerns, improving my writing skills, and on becoming a story-teller; Dr. Eleanor Perfetto for being instrumental in conceptualizing my study; Dr. Julia Slejko for helping me model statin use; Dr. Min Zhan for providing statistical and methodological guidance; and Dr. Gordon Smith for always providing thought-provoking comments and trauma expertise. In addition to mentors and committee members, I must acknowledge the PHSR administrative family for being the vital cogs to keep the department functioning. Namely, I owe gratitude to Colleen Day and Kiana Harvey, for always having their doors opens for repetitive, pestering questions and for providing a sense of calm to anxious graduate students. Furthermore, I thank the delightful students of PHSR. Without you, climbing this PhD mountain would have much more dreadful and much less exciting and entertaining. I must recognize the importance of my close group of friends for providing me with much needed work and social life balance. These life-long friends have given me respite when stressed, laughs when sad, and joined in my happiness. They have been an abundant source of joy. Among this close group of friends, I must individually express gratitude to my fellow PhD mountain climber and my little, big sister, Melissa Ross, for being my most honest friend, ever. She never said what I wanted to hear, but always said what I needed to hear. Lastly, but most importantly, I must acknowledge my mother, Shagufta Khokhar, for always being encouraging of my pursuits, for the known and unknown sacrifices, for the prayers, for the hugs, for the kisses, for the unconditional love.

TABLE OF CONTENTS

LIST OF TABLES ...... VIII

LIST OF FIGURES ...... XI

LIST OF ABBREVIATIONS ...... XII

CHAPTER I: INTRODUCTION ...... 1

PROBLEM STATEMENT ...... 1 GOAL, PURPOSE, OBJECTIVES AND AIMS OF STUDY...... 2 CONCEPTUAL FRAMEWORK...... 4 RATIONALE AND INNOVATION OF CURRENT STUDY...... 6 CHAPTER II: BACKGROUND ...... 10

TRAUMATIC BRAIN INJURY ...... 10 PHARMACOLOGICAL MANAGEMENT OF TBI ...... 11 STATINS ...... 12 STATINS AND TRAUMA...... 14 Animal Studies ...... 14 Human Studies ...... 15 Conclusion ...... 18 CHAPTER III: DATA & SAMPLE...... 19

DATA ...... 19 Centers of Medicare & Medicaid Services ...... 19 Area Health Resource File ...... 20 STUDY SAMPLE ...... 21 CHAPTER IV: DESIGN & MEASURES ...... 23

STUDY DESIGN ...... 23 MEASURES ...... 23 Independent Measure...... 23 Dependent Measures ...... 24 CHAPTER V: TBI SEVERITY ...... 26

CHAPTER VI: STATIN USE IN OLDER MEDICARE BENEFICIARIES WITH TBI...... 32

ABSTRACT:...... 32 INTRODUCTION...... 34 METHODS ...... 36

Data and sample ...... 36 Measures ...... 36 Data analysis ...... 38 RESULTS...... 38 DISCUSSION...... 46 CHAPTER VII: IN-HOSPITAL MORTALITY FOLLOWING TBI IN OLDER MEDICARE STATIN USERS...... 50

ABSTRACT:...... 50 INTRODUCTION...... 52 METHODS ...... 54 Data and sample ...... 54 Exposure ...... 54 Outcome ...... 55 Covariates ...... 55 Data analysis ...... 57 RESULTS...... 58 DISCUSSION...... 68 CHAPTER VIII: MORTALITY AND SECONDARY INJURY FOLLOWING TBI IN OLDER MEDICARE STATIN USERS ...... 75

ABSTRACT:...... 75 INTRODUCTION...... 77 METHODS ...... 79 Data and sample ...... 79 Exposure ...... 79 Outcomes ...... 80 Covariates ...... 81 Data Analysis ...... 82 RESULTS...... 85 DISCUSSION...... 94 CHAPTER IX: SUMMARY, IMPLICATIONS, & CONCLUSION ...... 99

SUMMARY ...... 99 Aim 1: Statin use in older Medicare beneficiaries with TBI ...... 100 Aim 2: In-hospital mortality following TBI in older Medicare statin users ...... 101 Aim 3: Mortality and secondary injury following TBI in older Medicare statin users ...... 102 Strengths and limitations ...... 103 IMPLICATIONS ...... 105 Clinical ...... 105 Research ...... 107

Societal ...... 108 CONCLUSION ...... 109 CHAPTER II APPENDIX ...... 111

CHAPTER VI APPENDICES ...... 112

CHAPTER VII APPENDICES ...... 115

CHAPTER VIII APPENDICES ...... 123

REFERENCES...... 127

vii

LIST OF TABLES

Table 3.1: ICD-9-CM codes to define head trauma...... 22

Table 5.1: Area under the curve statistics for in-hospital mortality, comparing SRR, SRRi, and TMPM...... 29

Table 5.2: Area under the curve statistics for in-hospital mortality, comparing SRR, SRRi, and TMPM...... 31

Table 6.1: Descriptive characteristics of statin users and non-users among Medicare beneficiaries with TBI ...... 40

Table 6.2: Statin use characteristics among statin users...... 41

Table 6.3: Odds ratios (95% CI) of factors associated with statin use among older Medicare Beneficiaries with TBI ...... 45

Table 7.1: Descriptive characteristics of statin users and non-users among Medicare beneficiaries with TBI ...... 60

Table 7.2: Statin use among Medicare beneficiaries prior to TBI ...... 62

Table 7.3: Odds ratios (95% CI) of in-hospital mortality following TBI comparing statin users and non-users ...... 63

Table 7.4: Odds ratios (95% CI) of in-hospital mortality following TBI comparing statin users and non-users by individual months prior to TBI ...... 65

Table 7.5: Odds ratios (95% CI) of in-hospital mortality following TBI comparing statin users and non-users, by consecutive and total use prior to TBI ...... 67

Table 8.1: Descriptive characteristics of statin users and non-users among Medicare beneficiaries with TBI ...... 86

Table 8.2: Relative risks (95% CI) of mortality following TBI among Medicare beneficiaries, comparing statin use to non-use ...... 90

Table 8.3: Relative risks (95% CI) of secondary injury following TBI among Medicare beneficiaries, comparing statin use to non-use ...... 92

Table 8.4: Relative risks (95% CI) of secondary injury following TBI among Medicare beneficiaries, comparing statin use to non-use among beneficiaries with at least 12 months of observation time following TBI ...... 93

viii

Table A.2.1: Animal studies assessing the impact of statins on TBI A.6.1— List of CCW chronic conditions excluding CVD conditions and ADRD...... 111

Table A.6.1: List of CCW chronic conditions excluding CVD conditions and ADRD...112

Table A.6.2: Odds ratios (95% CI) of factors associated with statin users only prior to TBI among older Medicare Beneficiaries ...... 113

Table A.6.3: Odds ratios (95% CI) of factors associated with statin users only following TBI among older Medicare Beneficiaries ...... 114

Table A.7.1: Correlation matrix between AHRF variables, Aim 2 ...... 115

Table A.7.2: Beneficiaries with monthly statin use prior to TBI ...... 116

Table A.7.3: Frequency of atorvastatin, rosuvastatin, and simvastatin doses per month, two months prior to TBI ...... 117

Table A.7.4: Odds ratios (95% CI) of in-hospital mortality following TBI comparing atorvastatin, rosuvastatin, and simvastatin doses two months prior to TBI ...... 118

Table A.7.5: Odds ratios (95% CI) of in-hospital mortality following TBI comparing statin users and non-users among beneficiaries that died within 24 hours of hospitalization ...... 119

Table A.7.6: Odds ratios (95% CI) of in-hospital mortality following TBI comparing statin users and non-users, by individual months prior to TBI among beneficiaries who did not die within 24 hours of hospitalization...... 120

Table A.7.7: Odds ratios (95% CI) of in-hospital mortality following TBI comparing statin users and non-users, by consecutive and total use prior to TBI among beneficiaries that died within 24 hours of hospitalization ...... 121

Table A.7.8: Odds ratios (95% CI) of in-hospital mortality following TBI comparing statin doses among atorvastatin, rosuvastatin, and simvastatin users, among beneficiaries that died within 24 hours of hospitalization ...... 122

Table A.8.1: Correlation matrix between AHRF variables, Aim 3 ...... 123

Table A.8.2: Relative risks (95% CI) of Mortality following TBI among Medicare beneficiaries, comparing statin use to non-use, IPTW adjusted...... 124

Table A.8.3: Relative risks (95% CI) of secondary injury following TBI among Medicare beneficiaries, comparing statin use to non-use, IPTW adjusted...... 125

Table A.8.4: Relative risks (95% CI) of secondary injury following TBI among Medicare beneficiaries, comparing statin use to non-use among beneficiaries with at least 12 months of observation time following TBI, IPTW adjusted ...... 126

LIST OF FIGURES

Figure 1.1: Conceptual framework highlighting the relationship between environmental, population, and health behavior characteristics with mortality and secondary injury ...... 5

Figure 5.1: Area under the curve for in-hospital mortality, comparing SRR, SRRi, and TMPM ...... 28

Figure 5.2: Area under the curve for in-hospital mortality, comparing SRR, SRRi, and TMPM ...... 30

Figure 6.1: Statin use per month relative to TBI among older Medicare beneficiaries ..... 42

Figure 6.2: Statin dose changes per month relative to TBI among older Medicare beneficiaries ...... 43

Figure 6.3: Statin use per month relative to TBI among older Medicare beneficiaries, by statin lipophilicity...... 44

Figure 8.1: Statin use per month following TBI among older Medicare beneficiaries...... 88

Figure 8.2: Statin use per month following TBI among older Medicare beneficiaries, by statin lipophilicity...... 89

LIST OF ABBREVIATIONS

ADRD—Alzheimer’s disease and Related Dementias

AHRF—Area Health Resource File

AIS—Abbreviated Injury Score

AMI—Acute Myocardial Infarction

AUC—Area Under the Curve

BBB—Blood Brain Barrier

CBF—Cerebral Blood Flow

CCW—Chronic Condition Data Warehouse

CHF—Congestive Heart Failure

CI—Confidence Interval

CMS—Centers for Medicare and Medicaid Services

CRASH—Corticosteroid Randomization after Significant Head Injury

CVD—Cardiovascular Disease

DME—Durable Medical Equipment

ED—Emergency Department

GEE—Generalized Estimating Equations

ICD-9-CM—International Classification of Diseases, 9th edition, Clinical Modification

ICISS—ICD-9 based Injury Severity Score

IHD—Ischemic Heart Disease

IPTW—Inverse Probability Treatment Weights

ISS—Injury Severity Score

LIS—Low Income Subsidy

LOS—Length of hospital Stay

MBSF—Master Beneficiary Summary File

xii

NSCOT—National Study on the Costs and Outcomes of Trauma

NSAID—Non-Steroidal Anti-Inflammatory Drug

OBTT—Operation Brain Trauma Therapy

OR—Odds Ratio

PDC—Proportion of Days Covered

PDE—Prescription Drug Event

QOL—Quality of Life

RCT—Randomized Clinical Trial

ROC—Receiver Operating Characteristic

ROS—Reactive Oxygen Species

RR—Relative Risk

SD—Standard Deviation

SE—Standard Error

SNF—Skilled Nursing Facility

SRR—Survival Risk Ratio

SRRi—Independent Survival Risk Ratio

TBI—Traumatic Brain Injury

TIA—Transient Ischemic Stroke

TMPM—Trauma Mortality Prediction Model

VHD—Valvular Heart Disease

xiii

CHAPTER I: INTRODUCTION

Problem Statement

Traumatic brain injury (TBI) is a major health concern for Americans.1 While

TBI impacts all age groups, the impact is greatest among adults aged 65 and older.2-5 In this population, TBI is associated with increased mortality and decreased life expectancy.3,6 Older TBI patients also frequently experience secondary injury following

TBI, such as stroke, neurological disorders, and mood disorders.7

TBI is characterized by damage to brain tissue, which can lead to inflammation and abnormal cerebral blood flow (CBF).8 TBI patients also may experience sustained inflammation.7 Inflammation and abnormal CBF can lead to mortality, and sustained inflammation puts older adults at risk for secondary injury, which may occur in the months and years following TBI.9-11 Secondary injury includes stroke, Alzheimer’s disease and related dementias (ADRD), and depression.10,12,13

There is limited research regarding ideal pharmacological treatment for TBI patients to reduce the impact of primary injury or to reduce the likelihood of secondary injury, although several studies have suggested new uses of pharmacological agents.14-17

Since TBI is characterized by inflammation and abnormal CBF, an optimal pharmacological agent would be multifunctional; that is, it would target both inflammation and CBF associated with TBI.9,10 One such multifunctional drug class is 3- hydroxy-3methyl-glutaryl-CoA reductase inhibitors, or statins. Statins possess anti-

1 inflammatory and vasoprotective properties, which may help improve survival rates among TBI patients and help prevent secondary injury following TBI.8-11

Goal, Purpose, Objectives and Aims of Study

The long-term goals of this research are to: 1) provide a better understanding of the potential pharmacological management for older TBI patients; and 2) improve the quality of life (QOL) and health outcomes among TBI patients. The purpose of this research is to assess the potential beneficial impact of statin therapy in older adults prior to and after TBI. The study objectives of the study are to quantify statin utilization and to determine the associations between statin use and mortality and secondary injury among

TBI patients. To achieve these objectives, this study used a large sample of Medicare beneficiaries aged 65 and older suffering a TBI to address three specific aims:

Aim 1: To characterize statin utilization patterns among older adults prior to and

following TBI.

Aim 1a: To quantify statin use pre-and-post-TBI; and

Aim 1b: To identify factors associated with statin use pre-and-post-TBI.

Aim 2: To assess TBI-specific inpatient mortality rates, comparing statin users

with non-users prior to TBI.

Aim 3: To assess the incidence of mortality rates following TBI hospitalization

discharge, comparing statin users with non-users BI.

Aim 3a: To assess the incidence of secondary injury, including stroke,

ADRD, and depression following TBI hospitalization discharge,

comparing statin users with non-users.

There is little research regarding statin utilization patterns among older Medicare beneficiaries with TBI. Therefore, findings from Aim 1 characterize statin utilization patterns prior to and after TBI. Aim 1 identifies and compares the disease and sociodemographic characteristics of statin users with non-users. Results from Aim 1 provide an essential understanding of statin utilization in older Medicare beneficiaries with TBI.

The results from Aim 2 examine the impact of statin use and non-use in older TBI patients. As previously indicated, statin use prior to TBI may help increase survival.

Therefore, findings from Aim 2 provide an understanding of the impact of statin use in a large sample of Medicare beneficiaries with TBI. Thus, the hypothesis for Aim 2 is:

H1: TBI-specific inpatient mortality rates will be greater in patients without prior

statin utilization, relative to patients with prior statin utilization.

The results from Aim 3 assess the impact of statin use prior to and following TBI by comparing statin users to non-users. As previously stated, long-term statin use following TBI may increase survival while statin discontinuation following TBI may decrease survival rates. Therefore, findings from Aim 3 explore the potential benefits of long-term statin use in TBI patients regarding mortality and secondary injury. Results from Aim 3 help provide a fundamental understanding for the novel application of statins. The specific hypotheses for Aim 3 are:

H1: Mortality rates following TBI hospitalization discharge will be greater in

patients without statin use following TBI, relative to patients with statin use

following TBI, while controlling for pre-TBI statin use; and

H2: Secondary injury rates following TBI hospitalization discharge will be greater

in patients without statin use following TBI, relative to patients with statin use

following TBI, while controlling for pre-TBI statin use.

Conceptual Framework

The conceptual framework of the study is based on the Anderson Model of Health

Care Utilization (Figure 1.1). The Anderson Model of Health Care Utilization model incorporates population, environmental, and health behavior characteristics of health services research.18 The framework illustrates the relationships between environmental, population, and health behavior characteristics with mortality and secondary injury in older TBI patients. This study explores the potential of statins for the pharmacological management of TBI in older adults by assessing pre-and post-TBI statin therapy and its impact on post-TBI mortality and secondary injury. Aim 1 explores pre-and post-TBI statin utilization patterns, and describes the demographic and disease characteristics of statin users and non-users. Additionally, Aim 1 investigates factors associated with statin use pre-and post-TBI.

Figure 1.1: Conceptual framework highlighting the relationship between environmental, population, and health behavior characteristics with mortality and secondary injury

Aim 2 investigates in-hospital mortality following TBI, comparing statin users and non-users because TBI is associated with an increase in in-hospital mortality.3

Approximately 9% of Medicare beneficiaries hospitalized with a TBI die during TBI hospitalization.19 Pre-TBI statin use may lessen the impact of the primary injury by reducing in-hospital mortality.20,21

Aim 3 assesses mortality following TBI-hospitalization discharge due to the continued increase in risks of mortality and secondary injury following hospital discharge. Among Medicare beneficiaries, 23%, 32%, and 46% of adults aged 65-74, 75-

84, and 85 and older, respectively, die within one year of head injury.22 TBI also is associated with secondary injury in the months and years following TBI.9-11 Thus, Aim 3 also assesses the incidence of secondary injury following TBI, comparing statin users and non-users.

Rationale and Innovation of Current Study

TBI is characterized by inflammation and cerebral damage.23-25 The initial insult to the brain, referred to as primary injury, initiates a sequence of events including inflammation and impaired CBF.24 This is due to the breakdown of the blood brain barrier (BBB) and the increase in reactive oxygen species (ROS) following injury.23-26

Immediately following injury, pro-inflammatory cytokines are released to repair the injured brain tissue; however, an unintentional consequence of this release is the increase in inflammation.23 Additionally, due to the initial injury, the BBB becomes more permeable, allowing toxins and pro-inflammatory cytokines to enter the brain, further causing inflammation.23,26,27 The structural damage from TBI can cause cerebral and microvascular damage.25 Furthermore, the increase of ROS can lead to lipid peroxidation and the lack of oxygenated cells in the brain can impair CBF and increase the risk of ischemia.23,24 Inflammation and impaired CBF caused by the primary injury often lead to mortality, either immediately or within a few days of the initial insult.20,21,28

If they survive primary injury, TBI patients are then at risk for secondary injury, which is the delayed impact caused by the neuroinflammatory cascade following the initial injury.9,10,29 After the primary injury, microglia cells are activated. This activation causes inflammation and releases pro-inflammatory cytokines. Following injury, the microglia cells may not deactivate, causing the chronic release of pro-inflammatory cytokines and inflammation.10 Furthermore, the chronic inflammation is more prominent in older adults as natural aging can also cause microglia cell activation. Therefore, an aging brain with trauma is more susceptible to sustained microglia activation and ensuing inflammation.10

Secondary injury includes a variety of events and illnesses such as stroke, neurological disorders, and mood disorders.10,12 TBI patients are at an increased risk of stroke in the year following TBI, and this may be linked to a damaged cerebrovascular structure.10,30 Neurological disorders caused by chronic neuroinflammation include dementias such as ADRD.10,12,31 Moreover, TBI also has been linked to mood disorders including depression, possibly due to chronic neuroinflammation.12,32

This research is innovative because it investigates long-term treatment following

TBI. Current pharmacological guidelines for TBI focus on either acute care to treat the primary injury or long-term symptomatic care for secondary injury following TBI.33,34

Acute pharmacological treatment includes the administration of analgesics, sedatives, barbiturates, and steroids.35,36 Unfortunately, due to an increased risk of mortality and bleeding, both steroidal and non-steroidal anti-inflammatory drug (NSAID) treatments are considered less than satisfactory in reducing inflammation following TBI.35-38 Due to secondary injury, long-term treatment has focused on alleviating symptoms from the

7 conditions that follow TBI.33,34 While acute treatment may not adequately treat primary injury and symptomatic treatment can be used after developing secondary injury, there is a gap in research exploring the effectiveness of a pharmacological agent that may both, reduce mortality after primary injury and help prevent, rather than treat, secondary injury following TBI.

Statin therapy following TBI may prove to be a promising option for the pharmacological treatment of TBI.8 The rationale of the study is based on the multiple properties of statins. While statins are recommended for the treatment of hyperlipidemia, statins also may help TBI patients because of their impact on the reduction of inflammatory cytokines and oxidative stress.16 Statins also may decrease platelet aggregation, thereby reducing the likelihood of thrombosis following TBI.17 With their vasoprotective and anti-inflammatory properties, statins are multifunctional drugs that may reduce the impact of primary injury and prevent secondary injury.9,10

There is limited research conducted in older TBI patients investigating statin utilization patterns, and the relationships between statin use and mortality and secondary injury.20,21,39 However, several animal studies indicate the potential of statins to treat TBI that this study will build upon.40-46 The proposed research will use Medicare administrative data from 2006 to 2010 to identify statin utilization patterns pre-and post-

TBI and assess the relationships between statin use and mortality and secondary injury following TBI in older adults.

This research produces the first study in older adults examining the potential beneficial impact of statins following TBI in a real world setting, guiding future research and practice. The findings provide a foundation needed to enhance care for TBI patients

8 and will help foster future research that can be used to develop evidence-based guidelines to treat older TBI patients.

CHAPTER II: BACKGROUND

Traumatic Brain Injury

There are over a million cases of TBI in the US every year.47 In 2010, the rate of

TBI-related emergency department (ED) visits, hospitalizations, or deaths was 823.7 per

100,000 people.48 The rate of ED visits, hospitalizations, and deaths is greater among males than females (932.1 per 100,000 males compared to 720.3 per 100,000 females).49

Most TBI hospitalizations are among whites.2,50 However, TBI-related death rates are greatest among ethnic and racial minorities.3

TBI disproportionately impacts older adults.1,2,50 More than half of TBIs in the US are in adults 50 or older.50 TBI causes more than 80,000 emergency visits every year for adults over 65.4 The leading cause of TBI among the elderly is falls.4 TBI among older adults is more prevalent in males and white Americans when compared to females and other races/ethnicities, respectively.4,5

Mortality rates for TBI are greater across all ages compared to the general population without TBI.51,52 TBI among adults 65 and older decreases life expectancy by approximately 3-7 years. For adults 85 and older, life expectancy following TBI is three years.6,52,53 The increased mortality decreased life expectancy highlights the need to explore pharmacological treatment options for older TBI patients. Considering the vasoprotective and anti-inflammatory properties of statins, statin use may reduce mortality and secondary injury following TBI.

Pharmacological Management of TBI

There are no standard pharmacological treatment guidelines following TBI.33,54,55

Pharmacotherapy guidelines largely focus on acute care, followed by secondary injury.35

Primary acute care includes the administration of corticosteroids to reduce inflammation and intracranial pressure.35 Trials assessing the impact of corticosteroids in TBI patients have been less than satisfactory.23,54,56 The large scale Corticosteroid Randomization after

Significant Head Injury (CRASH) trial indicated that corticosteroids were associated with an increased risk of mortality following TBI.36 The CRASH trial investigated the efficacy of methylprednisolone following head injury but had to be discontinued after 10,000 patients were enrolled because mortality rates were higher in the methylprednisolone group (21.1%) when compared to the placebo group (17.9%) (p < 0.0001).57 Although, the mechanisms that caused an increase in mortality remain unclear, the increase is unlikely due to gastrointestinal bleeding or infections.35,58 The Brain Trauma

Foundation’s “Guidelines for the Management of Severe Traumatic Brain Injury,” indicates that some studies have suggested that treatment with steroids have shown no significant benefit while others have suggested that steroid treatment has been deleterious.35 It should be noted that while corticosteroids are not recommended to treat

TBI, literature suggests that progesterone, a neurosteroid, may improve survival among

TBI patients.36,59 Progesterone has neuroprotective properties and may also decrease cerebral edema.36

NSAID treatment also is considered suboptimal, as NSAIDs may increase the risk of adverse events, such as hemorrhage.37,38 The use of over-the-counter NSAIDs such as ibuprofen and aspirin have shown little efficacy in treating TBI.27,37,38

Since TBI is typically followed by secondary injury, pharmacological treatment also may be geared toward alleviating symptoms of secondary injury.33,34 For instance, older adults face an increased risk of depression following TBI; therefore, treatment may be focused on alleviating depressive symptoms.33,60 However, symptomatic treatment does not prevent the secondary injury. In addition to symptomatic treatment, long-term use of beta-blockers, which are used to treat hypertension, have been suggested to improve survival following TBI.8,61,62 This is because TBI patients may develop hypertension following TBI, which may lead to cardiovascular events and mortality.61,63

Because TBI is a complex event that has multiple targets, including inflammation and CBF,14 treatment for TBI should both inflammation and CBF.9,10 Statins have been highlighted as multifunctional drugs that target multiple pathways and have the potential to decrease mortality and secondary injury following TBI.8-10

Statins

Statins are a candidate to decrease the burden of primary injury and to reduce the likelihood of secondary injury associated with TBI.7,8,14,16,17,55,64-66 In addition to reducing serum cholesterol, statins decrease inflammation, ROS, and improve CBF, which may decrease mortality.20,21

Statin therapy has been approved for the treatment of hyperlipidemia and is used for the primary and secondary prevention of cardiovascular events among older adults with cardiovascular disease (CVD).67-69 Currently, the statins approved for use in the US include Lipitor (atorvastatin), Lescol(fluvastatin), Mevacor and Altoprev (lovastatin),

Livalo (pitavastatin), Pravachol (pravastatin), Crestor (rosuvastatin), and Zocor

(simvastatin).70 While all statins help reduce cholesterol, each statin varies with regards to pharmacology.71 For instance, lovastatin, simvastatin, fluvastatin, and atorvastatin have greater potential to cross the BBB while pravastatin, pitavastatin, and rosuvastatin, and are less likely to enter non-hepatic cells than lipophilic statins.71-73 Though there are variations in the pharmacology of statins, statins can be titrated and/or switched in order to lower cholesterol.74

Statins are associated with few adverse events.75-78 The most common adverse events associated with statin use are myopathy and diabetes.75,77 Myopathy is defined as unexplained muscle pain and weakness that may cause muscle damage.77 Fortunately, myopathy is a rare adverse event.77 Patients who develop diabetes following statin use are often at a higher risk of developing diabetes due to other pre-existing risk factors.75,77

When considering older adults, statin therapy offers great benefits for patients with and without CVD, because the risk of adverse events is low.76,78

Considering the protective impact of statins among patients with CVD or CVD risk factors, only two-thirds of adults take statins.79 Among elderly Medicare beneficiaries, the implementation of Medicare Part D has increased the use of statins; however adherence to statins is suboptimal among patients with CVD or recovering from cardiovascular events.80-83

Statins and Trauma

Animal Studies

Several animal studies have examined the impact of statin use prior to and after

TBI. These studies simulated TBI on rats and mice. The animals in these placebo- controlled studies were exposed to simvastatin, lovastatin, rosuvastatin, or atorvastatin, in order to examine the impact of statins on inflammation,42,44-46,84,85 CBF,46,86 and neurological function (Chapter II Appendix, Table A.2.1).87,88

Six studies have assessed the impact of statin therapy on inflammation following

TBI, of which one study assessed the impact of pre-TBI statin therapy and five studies investigated the post-TBI impact of statin therapy.42,44-46,84,85 Results from these studies found that statin use was associated with a significant decrease in inflammation following

TBI.

Two studies investigated the impact of statin use on CBF.46,86 While one study found that pretreatment with simvastatin increased CBF following injury,46 another study indicated that rosuvastatin treatment following injury did not significantly impact blood flow following TBI.86 The authors of the rosuvastatin study suggest that the findings of their study were different from previous studies assessing the impact of statins on TBI because the vascular effects of statins may depend on the dissolvability of the drugs. The authors suggest that rosuvastatin may be less amenable to BBB permeability than other statins such as simvastatin and atorvastatin.86

Two studies investigated the impact of long-term statin use on neurological function following TBI.87,88 Both studies indicated that simvastatin treatment

14 significantly improved neurological function up to three months following TBI.87,88

Neurological deficits are considered secondary injury following TBI and are caused by chronic inflammation.9,10 These studies suggest that the multifunctional properties of statins may be neuroprotective in the long-term.87,88

These animal studies indicate that statins may help decrease inflammation and improve CBF following TBI, which may help reduce the burden of primary injury and reduce secondary injury.42,44-46,84-86 The encouraging results from animal studies highlight the need for translational research in human subjects.

Human Studies

Human Studies Assessing Statin Use Prior to TBI

Three studies have assessed the relationship between pre-injury statin use and in- hospital survival among older adults. Two of these studies utilized the National Study on the Costs and Outcomes of Trauma (NSCOT) data from July 2001 to November 2002, and excluded patients younger than 65 and those who died within 24 hours of hospitalization.20,21 The NSCOT data provides pre-injury statin utilization through medical charts; however, the NSCOT dataset does not collect medication utilization information during follow-up, so long-term statin use cannot be ascertained.89

The first study using NSCOT data statin use analyzed 1,224 TBI patients, of whom 22% used statins prior to injury.21 The authors stratified their analysis by CVD status in order to understand the role of CVD between the relationship of statin use and mortality. Approximately 44% of patients had CVD.21 Stratified analysis based on CVD suggested that among patients without CVD, pre-injury statin use was associated with a

7% absolute risk reduction in mortality, relative to non-users. However, among TBI patients with CVD, the relationship between statin use and mortality was not significant.21

The second study using NSCOT data analyzed 523 TBI patients and found 25% had used statins prior to injury.20 Statin use was significantly associated with decreased in-hospital mortality (relative risk (RR): 0.24, 95% confidence interval (CI): 0.08-0.69), relative to non-users.20 The authors also conducted additional analyses stratified by CVD status. Similar to the Efron et al (2008) study,21 this study suggested the relationship between statin use and mortality was modified by CVD. Stratified analysis among TBI patients with CVD indicated that statin use was associated with a RR of 0.87 (95% CI:

0.50-1.50) for in-hospital mortality. Among TBI patients without CVD, statin use was associated with a RR of 0.17 (95% CI: 0.05-0.63) for in-hospital mortality.20

A third study assessed pre-TBI statin use and 14 day mortality following TBI among severe trauma patients in Singapore.90 This study included 59 statin users matched with 59 non-users and found that pre-injury statin use was not associated with reduced

14-day mortality.90

Two of the three studies suggest statin use may be associated with decreased mortality; however, the relationship may be modified by the presence of CVD.20,21 A major limitation of the three studies is that pre-injury statin use was not quantified.

Additionally, the type of statin use was not specified. The current research quantified statin use, as well as specified specific statin(s) taken prior to TBI.

Human Studies Assessing Statin Use Following TBI

One study assessed the relationship between statin discontinuation during TBI- hospitalization and in-hospital mortality.39 Orlando et al (2013) conducted a retrospective study of data collected from two Level I Trauma Centers during 2007 to

2009 which included patients 55 and older with pre-injury statin use.39 Patients who were not administered statins within 48 hours of hospitalization were considered to be discontinued from therapy.39 Only 61 patients were included in the study, of whom 46 continued statin therapy during hospitalization. Mortality was greater among patients who discontinued; however, this relationship was not statistically significant (p =

0.055).39

A clinical trial focused on the effect of rosuvastatin on inflammatory cytokines after TBI; however, the study did not restrict the patient population to older adults.91 The objective of this placebo-controlled, double blind randomized clinical trial (RCT) was to examine whether rosuvastatin reduced inflammatory cytokine plasma levels following

TBI.91 The study recruited 39 patients from a hospital in Mexico, where 19 patients received 20 mg rosuvastatin for ten days following TBI. Inflammatory cytokines were measured on the day of TBI and three days following TBI. The study found that rosuvastatin reduced inflammation following TBI.91

A major limitation of both studies was the small sample size of patients whose type of pre-injury statin use was not identified. Additionally, adherence and persistence to statins, and long-term statin use following TBI were not addressed. Furthermore, the clinical trial was limited because of the omission of other factors that could reduce cytokine plasma levels, such as low hemoglobin.91,92

Conclusion

The scarcity and major limitations of human studies examining the relationship between pre-and post-TBI statin use and mortality highlights the need for observational research. This study addresses the limitations of previous research by examining characterizing statin use pre-and post-TBI, (Aim 1). After characterizing statin use, this study assesses the relationship between pre-injury statin use and hospital mortality (Aim

2); and pre-and post-TBI statin use and mortality and secondary injury following TBI hospitalization discharge (Aim 3). The aims of this study are fulfilled by using Medicare administrative data, a rich dataset that includes multiple covariates and confounders, monthly statin use, hospitalizations, and mortality.

CHAPTER III: DATA & SAMPLE

Data

Centers of Medicare & Medicaid Services

The Centers of Medicare & Medicaid Services (CMS) data was used to investigate statin utilization following TBI and the impact of pre-and post-TBI statin use on mortality and secondary injury. This data contains 100% of Medicare beneficiaries who were hospitalized with a TBI during 2006 to 2010 and who had Medicare Parts A, B, and D. This study will be able to assess statin utilization patterns pre-and post-TBI (Aim

1) through Medicare Part D claims, which will allow the assessment of monthly statin utilization. Medicare administrative data also was used to investigate the relationship between in-hospital mortality following TBI and pre-injury statin use (Aim 2), and the relationship between pre-and post-TBI statin use and mortality and secondary injury

(Aim 3).

The data includes demographic, healthcare utilization (inpatient and outpatient claims), and prescription drug information.93 The Medicare administrative data includes the master beneficiary summary file (MBSF), chronic conditions, inpatient, outpatient, skilled nursing facility (SNF), hospice, home health care, carrier, durable medical equipment (DME), and Medicare Part D event (PDE) files.

The MBSF will be used to create monthly Medicare coverage analytic files. These files will also be used to capture date of death and demographic characteristics including age, sex, and race. The chronic conditions data warehouse (CCW) file identifies 27

19 chronic conditions that are identified through inpatient claims using International

Classification of Diseases, Ninth Revision-Clinical Modification (ICD-9-CM) diagnosis codes. Each inpatient claim contains up to ten ICD-9-CM diagnosis codes.93 A chronic condition is identified through a mid-year flag, yearly flag, and a flag for the first occurrence of the said condition.93 This file was used to capture comorbidities and also was used to define overall beneficiary health and disease burden.

The inpatient, outpatient, SNF, hospice, home health care, carrier, and DME files were used to capture information regarding health status, outcomes, and confounders, such as inpatient claims for diseases, adverse events, or discharge to a SNF or hospice.

The PDE data file includes prescription claims for all Medicare beneficiaries enrolled in Medicare Part D. The prescription claims are submitted by pharmacies at the point of service; this means that once a prescription is filled, the claim will be submitted to a Medicare Part D health plan.94 Thus, the PDE file does not contain prescribing information; rather it contains filled prescription information. The PDE file was used to determine Medicare Part D coverage status and to create monthly analytic files of statin and other medication usage.

Area Health Resource File

Data from the 2008 Area Health Resource File (AHRF) was used to augment

CMS data. Data from the 2008 file was selected as it is the center point of the study

(2006 – 2010). This dataset is a compendium of sociodemographic and healthcare variables.95 The dataset comprises over 6,000 variables from multiple sources including the census and the American Health Association.95 The dataset includes demographic

20 variables such as education and income levels, healthcare profession variables such as number of physicians, and healthcare facility variables such as number of hospitals and beds, among a variety of other variables.95,96

A selection of sociodemographic, health professional and health facility variables were extracted from the AHRF dataset to supplement the CMS data. These data were then linked to CMS data through beneficiaries’ Social Security Administration county codes. Therefore, the CMS data was augmented by county level sociodemographic, health professional and health facility information.

Study Sample

The cohort for the three aims will be selected from Medicare beneficiaries with claims for a TBI between 01/01/2006 to 12/31/2010. Table 3.1 lists the ICD-9-CM codes that used to define head trauma. These codes have been used in prior research, and the

CDC also defines adult TBI through these codes.30,97,98 Multiple TBI admissions occurring within 14 days were counted as a single TBI, with the admission date for the first TBI being the index date and index TBI. This approach has been previously used to account for the likelihood of multiple TBI claims being made for the care of a single

TBI.19,30,99 From this sample of TBI patients, the cohorts for the three aims were created.

Table 3.1: ICD-9-CM codes to define head trauma

CHAPTER IV: DESIGN & MEASURES

Study Design

The current research is a longitudinal, retrospective study of Medicare beneficiaries hospitalized with a TBI between 2006 and 2010. The longitudinal component of the study permits multiple observations per beneficiary. Beneficiaries’ exposure, and certain covariates, and confounders are observed per 30-day period pre-and post-TBI. The observation for the outcome begins at TBI-hospital admission for Aim 2 and at TBI-hospital discharge for Aim 3.

Measures

Measures for this study include the independent (exposure) and dependent

(outcome) measures of interest. Since the proposed study is longitudinal, some measures are observed multiple times per beneficiary and are time-varying.

Independent Measure

The exposure of interest for Aims 1, 2, and 3 is statin therapy. Specifically, Aims

1 and 3 observes statin use prior to and following TBI, while Aim 2 observes statin use only prior to TBI. Statin use was defined if a beneficiary has a PDE claim for any of the following statins: atorvastatin, fluvastatin, lovastatin, pravastatin, rosuvastatin, and simvastatin. These statins are selected based on guidelines for the treatment of blood

23 cholesterol, prior research, and availability of statins during the study period.82,100-102

Combination drugs that contained a statin were classified as that individual statin; for instance, beneficiaries using ezetimibe/simvastatin were classified as using simvastatin.

Exposure to these statins are ascertained per 30-day period (henceforth referred to as month) relative to the TBI. For Aims 1 and 3, the exposure is time-varying as it is ascertained per month following TBI.

Proportion of days covered (PDC) was used to determine monthly statin adherence. The PDC is a measure that assesses medication adherence for a specified time period.103 PDC is calculated by dividing the number of days the drug is available (days of use covered per prescriptions filled) by the total number of days in a specified time period.103 The specified time period for the PDC calculation is 30-days. PDC is calculated at both the specific drug level and class level. The PDC measure of adherence is a widely used method for studying adherence in all age groups.103,104

Beneficiaries were flagged as having statin use in a period if there was a fill for a statin in the period or a PDC greater than zero. Both prescriptions filled and PDC information were used to classify statin use because of the possibility of beneficiaries being prescribed multiple months’ worth of statins in one prescription, which will cause beneficiaries to have a prescription filled in only one period, while a PDC lasting multiple months.19

Dependent Measures

Aim 1 is descriptive and assesses statin utilization prior to and following TBI. The outcome for Aim 2 is primary injury (in-hospital mortality). The outcomes for Aim 3 are

24 mortality following hospitalization discharge and secondary injury including stroke,

ADRD and depression.

The CCW inpatient claims file includes beneficiary date of death.93 Aim 2 observes TBI-specific mortality by observing the discharge status following the index

TBI. If a beneficiary is discharged dead following TBI hospitalization, then it is assumed that the mortality is TBI-specific. For Aim 2, beneficiaries are observed until they are discharged from the hospital. Beneficiaries may be discharged either alive or dead and are censored following discharge.

Aim 3 observes mortality following hospital discharge, making the outcome all- cause mortality rather TBI-specific, as cause of death in settings other than inpatient settings cannot be assessed in the CCW data. For Aim 3, beneficiaries are followed until they experience the outcome or until the end of the study period, whichever comes first.

Aim 3 also observes secondary injury following TBI-hospitalization discharge.

Stroke will be defined by inpatient claims using ICD-9-CM codes 430.xx-432.xx

(hemorrhagic stroke) and 433.xx, 434.xx, 435.xx, 437.0x, 437.1x (ischemic stroke).

ADRD and depression were defined through the CCW flags of ADRD and depression, respectively. The codes for stroke and the CCW flags for ADRD and depression have been used in prior claims data analysis.19,60,105,106

CHAPTER V: TBI SEVERITY

A major confounder between the relationship of statin use and mortality is TBI severity. TBI severity is measured through varying scales such as the Abbreviated Injury

Score (AIS) and the Injury Severity Score (ISS). The ISS is derived from the AIS.33 The

AIS and ISS are considered the gold-standards in measuring injury severity.107

Unfortunately, Medicare administrative data does not provide this information. However, several proxy injury severity measures have been developed to determine injury severity using ICD-9-CM codes available in Medicare administrative data108 Measures of injury severity based on ICD-9-CM codes have demonstrated to be valuable in administrative claims research.108

One such measure is the ICD-9-CM -based ISS (ICISS). The ICISS can measure injury severity through two multiplicative models. The first multiplicative model calculates the survival risk ratios (SRR) for each injury and then multiplies the SRRs together to obtain an ISS.109 SRR are calculated by dividing the number of survivors in each injury by the total number of patients with the same injury.109,110 The second multiplicative model calculates independent survival risk ratios (SRRi), which calculates the SRR for each individual ICD-9-CM code in an injury event.111

Another measure that predicts mortality following injury is the trauma mortality prediction model based on ICD-9-CM codes (TMPM).112 This model utilizes a two-stage logistic regression to determine injury severity. The model first predicts mortality based on ICD-9-CM codes and then estimates injury severity for each injury.112 When

26 compared to the ICISS, the TMPM model has a better predictive value in predicting mortality in injury patients.112

Assessing TBI severity is vital for the aims of this study; therefore, prior to investigating the aims of this study, this study first examined the appropriate measure to utilize to determine TBI severity. This was done by examining the receiver operating characteristic (ROC) curves for all three: SRR, SRRi, and TMPM. The ROC curves help assess the predictive probability of a binary event, in this case mortality, by plotting the sensitivity and specificity of a measure.113,114 The area under the curve (AUC) represents the amount of the sample for which mortality was correctly predictive; therefore, a higher

AUC is indicative of a higher predictive capability of a measure.113,114 A chi-square p- value is given to determine if differences in AUC are different for the TBI severity measures.113,114

CCW study data (refer to Chapter III) was used to examine the AUC for the TBI severity measures. As done with the aims of the study, all TBIs occurring within 14 days of the first TBI were collapsed and classified as a single TBI event. However, for the purpose of identifying the best TBI severity measure, beneficiaries were allowed to have multiple TBIs as long as they were more than 14 days apart. Therefore, 123,691 TBIs were observed among 116,193 Medicare beneficiaries during 2006 to 2010.

Since Aim 2 examines in-hospital mortality and Aim 3 examines post- hospitalization mortality, SRR, SRRi, and TMPM were calculated for both in-hospital mortality and post-hospitalization mortality. Then the AUC for the TBI severity measures were compared for both in-hospital and post-hospitalization mortality.

A total of 10,304 beneficiaries died during TBI-hospitalization and 43,279 died post-hospitalization. Figure 5.1 displays the ROC curves for in-hospital mortality while

Table 5.1 provides corresponding ROC statistics. The SRRi has a significantly better predictive capability (higher AUC) than the SRR and TMPM.

Figure 5.1: Area under the curve for in-hospital mortality, comparing SRR, SRRi, and

TMPM

Table 5.1: Area under the curve statistics for in-hospital mortality, comparing SRR,

SRRi, and TMPM

a Chi-square Abbreviations: AUC, area under the curve; ROC, receiver operating characteristic; CI, confidence interval; SE, standard error

Figure 5.2 displays the ROC curves for post-hospitalization mortality while Table

5.2 provides corresponding ROC statistics. Similarly, to inpatient mortality, SRRi also has significantly greater predictive capabilities than the SRR and TMPM for post- hospitalization mortality.

Figure 5.2: Area under the curve for in-hospital mortality, comparing SRR, SRRi, and

TMPM

Table 5.2: Area under the curve statistics for in-hospital mortality, comparing SRR,

SRRi, and TMPM

a Chi-square Abbreviations: AUC, area under the curve; ROC, receiver operating characteristic; CI, confidence interval; SE, standard error

Based on the AUC, the SRRi model of the ICISS performed better in predicting mortality than the SRR model of ICISS and the TMPM. While the differences between the AUCs were small for both in-hospital and post-hospitalization mortality, these differences were statistically significant. Considering SRRi’s better predictive probability to predict mortality over SRR and TMPM, the SRRi measure was used as a TBI severity proxy for the study aims.

CHAPTER VI: STATIN USE IN OLDER MEDICARE BENEFICIARIES WITH TBI

Abstract

Background: There are approximately 30 million statin users in the US. While the lipid lowering effects of statin therapy for older adults with and without CVD is established, statins also may be beneficial for older adults sustaining a TBI. There is limited research regarding statin use among older adults, especially among TBI patients.

Objective: To characterize and investigate factors associated with statin use among older adults with TBI.

Methods: A retrospective drug utilization study was conducted among 75,698 Medicare beneficiaries 65 and older with continuous Medicare Parts A, B, and D coverage six months prior and 12 months following TBI among beneficiaries hospitalized with a TBI during 2006 to 2010. The exposure of interest was statin use prior to and following TBI.

Logistic regression was used to investigate the factors associated with statin use.

Results: 37,874 (50%) of beneficiaries used a statin at least once during the study period.

The most common statin used was simvastatin, followed by atorvastatin. Fluvastatin was the least used statin, with approximately 1% of beneficiaries using fluvastatin per 30-day period throughout the study period. Statin users were more likely to have cardiovascular diseases when compared to non-users. Hyperlipidemia was a major factor associated with statin use and had the greatest impact on statin use compared to non-use (odds ratio (OR)

9.54; 95% confidence interval (CI) 9.07, 10.03).

Discussion: This study provides valuable information on the extent of statin use in older

TBI patients and future studies will examine the association between statin use and mortality.

Introduction

There are approximately 200 million statin users worldwide, 30 million of whom live in the US.115 Statins are FDA-approved therapy for the treatment of hyperlipidemia and are used for the primary and secondary prevention of cardiovascular events among older adults with CVD.67-69 Also, statins offer a favorable safety profile as the risk of adverse events associated with use is low.75-78

While the lipid lowering effects of statin therapy for older adults with and without

CVD is established,76,78 it has been suggested that statins also may be beneficial for older adults sustaining a TBI.7,64 Among older adults, TBI is associated with increased mortality and decreased life expectancy.3,6 Older patients with TBI also frequently experience secondary injury following TBI, such as stroke, neurological disorders, and mood disorders.7 Currently, there is no ideal pharmacological treatment for older adults with TBI, as most treatments are for symptoms of secondary injury occurring after TBI.35

However, statin use prior to trauma may decrease in-hospital mortality in the elderly following injury.20,21 Additionally, statin use following TBI also has been suggested to decrease secondary injury, possibly due to statins’ anti-inflammatory properties.8-11,116

Despite statins’ potential to reduce in-hospital mortality if taken prior to TBI and their promise as an option for the pharmacological treatment following TBI in older adults,8,116 there is limited research regarding statin use among older adults. Among elderly Medicare beneficiaries, the implementation of the Medicare Part D drug benefit program has increased the use of statins; however, adherence to statins is suboptimal among many patients with CVD or recovering from cardiovascular events.80-83 With

34 regards to TBI, only a few observational studies have examined statin use in trauma patients; of these, three assessed statin use prior to trauma20,21,90 and a fourth assessed statin continuation following trauma.39 Another study examined the impact of hyperlipidemia on depression following TBI, in which secondary analyses examined the role of statins.117 Regardless, these studies were limited due to small sample sizes and poor identification of statin use, as the statin used was not specified. Furthermore, none of these studies assessed persistence and factors associated with statin use. Lastly, these studies did not seek to characterize statin use prior to and following TBI; rather, they sought to examine the beneficial impact of statin therapy on trauma.

Given the potential protective effect statins may have in older patients with TBI, it is important to understand the extent of statin use in this population. Yet, the data are relatively scant despite myriad data on statins. Therefore, it is important to characterize the extent of statin use in older adults with TBI prior to examining the potential beneficial impact of statins on in-hospital mortality and secondary injury following TBI; this study would provide important information that will lay the foundation for studies examining their protective effect. The objective of this study is to characterize statin use before and after TBI in a nationally representative sample of older Medicare beneficiaries and to identify the factors associated with that use.

Methods

Data and sample

This study used Medicare administrative claims data from the CMS. These data include all Medicare beneficiaries 65 years and older who suffered a TBI between 2006 and 2010. Codes used to define TBI have been explained in Chapter III.

Beneficiaries were required to have Medicare parts A, B, and D coverage six months prior to TBI and 12 months following TBI. These data include inpatient, outpatient, and prescription drug claims.93 Beneficiaries with Medicare Part C coverage were excluded because prescription drug claims were not available for those beneficiaries.

Measures

The exposure of interest for this drug utilization study was statin use. Statins included in the study were atorvastatin, fluvastatin, lovastatin, pravastatin, rosuvastatin, and simvastatin. Exposure to statins was determined using Medicare Part D prescription drug claims.

Detailed information regarding exposure can be found under ‘Measures’ in Chapter IV.

Briefly, both prescriptions filled and PDC information was used to flag statin use in periods.

Beneficiaries had six periods of coverage to examine exposure prior to the index

TBI and 12 periods of coverage to examine exposure following the index TBI. Therefore, proportion of statin use by drug class and individual statins was calculated per 30-day

36 period six periods prior to and 12 periods following TBI. Average duration of use, and average number of and duration of gaps among statin users were calculated. A gap in use was defined as a period without a prescription fill of a statin and a PDC of 0. Average statin dose per 30-day period also was calculated for individual statins. Lastly, statin use by lipophilicity and hydrophilicity per 30-day period also was calculated. This is because while all statins help reduce hyperlipidemia, their pharmacology differs with regards to lipophilicity and hydrophilicity. Atorvastatin, lovastatin, fluvastatin and simvastatin are considered lipophilic statins and have greater potential to cross the BBB, which may be related to a statins ability to reduce cerebral inflammation.71-73

Covariates examined included demographic characteristics (age, race, and sex), and conditions associated with CVD (hypertension, hyperlipidemia, ischemic heart disease (IHD), congestive heart failure (CHF), acute myocardial infarction (AMI), valvular heart disease (VHD), and history of stroke/transient ischemic attack (TIA)).

ADRD, diabetes, and a count of CMS’s CCW chronic conditions, excluding the above mentioned cardiovascular and chronic conditions (Chapter VI Appendices, Table A.6.1).

Lastly, covariates associated with injury severity, including length of hospital stay (LOS) and SRRi were examined. SRRi is part of the ICISS, which is used to assess injury severity based on ICD-9-CM diagnosis codes.109,110 More information on SRRi can be found in Chapter V. It should be noted that injury severity covariates were only examined for beneficiaries that used statins only following TBI as injury severity cannot be related to statin use prior to injury.

Data analysis

Associations with covariates between statin users and non-users were tested using

Chi-square test for categorical variables and Student’s t-tests for continuous variables.

Additionally, the Wilcoxon rank sum test was used to test differences between two medians. Logistic regression was used to identify factors associated with any statin use.

Statin use was first categorized as use or non-use, and then as no use, less than or equal to

50% use throughout study period, and greater than 50% use throughout study period.

Additional sensitivity analyses were conducted by stratifying factors associated with statin use among beneficiaries that only used statins prior to TBI and beneficiaries that only used statins following TBI. Odds ratios (OR) and 95% CI were reported.

All analyses were performed with SAS version 9.2 (Cary, NC). This study was approved by the University of Maryland Baltimore’s Institutional Review Board.

Results

A total of 116,170 Medicare beneficiaries 65 years and older without Medicare

Part C were hospitalized with a TBI during 2006 to 2010. Of these beneficiaries, 83,461 had continuous Medicare Parts A, B, and D coverage six months prior to TBI. The final sample included 75,698 beneficiaries that also had continuous Medicare Parts A, B, and

D coverage 12 months following TBI. Of the 75,698 beneficiaries in the study sample,

37,874 (50%) used at least one statin during the study period.

The study population was predominately white (87%) and female (67%). The mean age of statin users was less than that of non-users (81.3 years compared to 83.3 years, p <0.001). Statin users were more likely to have a diagnosis of hyperlipidemia

38 recorded when compared to non-users (94% vs 59%, p <0.001). Additionally, statin users were more likely to have CVD, including AMI, IHD, CHF, and VHD (All p <0.001)

(Table 6.1).

Table 6.1: Descriptive characteristics of statin users and non-users among Medicare beneficiaries with TBI (n=75,698)

40 d Chronic conditions present at time of traumatic brain injury e Count of CCW chronic conditions excluding cardiovascular conditions Abbreviations: TBI, traumatic brain injury; SD, standard deviation; SNF, skilled nursing facility; ADRD, Alzheimer’s disease and related dementias

Of the 37,874 statin users, there were 30,378 (80%) beneficiaries who used statins prior to and following TBI; 1,969 (5%) only used statins prior to TBI and 5,527 (15%) used statins only following TBI. Among statin users, the average duration of use was

12.4 months. There were 14,475 (38.2%) users with at least one gap in use. Among these users, the average number of gaps was 1.5 and the average gap duration was 3.3 months

(Table 6.2).

Table 6.2: Statin use characteristics among statin users (n=37,874)

a Number of 30-day periods Abbreviations: TBI, traumatic brain injury; SD, standard deviation

Figure 6.1 illustrates trends in statin use throughout the study period and the dip associated with TBI hospitalization (a period when statin use would not be recorded by

Medicare Part D claims). Statin use per 30-day period ranged from 35% to 40%, prior to and following TBI with no sudden increase post discharge. The most common statin used was simvastatin, followed by atorvastatin. Fluvastatin was the least used statin, with approximately one percent of the sample using fluvastatin per 30-day period throughout the study duration.

Figure 6.1: Statin use per month relative to TBI among older Medicare beneficiaries

(n=75,698)

Statin use 12 months following TBI (n=75,698) 40 35

30 Statin Use 25 Atorvastatin use 20 Fluvastatin Use

15 Lovastatin Use Percent Use 10 Pravastatin Use 5 Rosuvastatin Use Simvastatin Use 0 -6 -3 1 4 7 10 30-Day Periods Post-TBI

The average dose of statins per 30-day period fluctuated 1 to 4mg throughout the study duration (Figure 6.2).

Figure 6.2: Statin dose changes per month relative to TBI among older Medicare beneficiaries (n=75,698)

Average statin dose 12 months following TBI (n=75,698) 70

50 Average Atorvastatin Dose 40 Average Fluvastatin Dose

30 Average Lovastatin Dose Average Pravastatin Dose

Average Dose (mg) 20 Average Rosuvastatin Dose 10 Average SimvastatinDose

0 -6 -3 1 4 7 10 30-Day Periods relative to TBI

When stratified by statin lipophilicity and hydrophilicity, 30% to 35% of beneficiaries used lipophilic statins prior to and following TBI, while approximately 5% of beneficiaries used hydrophilic statins prior to and following TBI (Figure 6.3).

Figure 6.3: Statin use per month relative to TBI among older Medicare beneficiaries, by statin lipophilicity (n=75,698)

Statin use 12 months following TBI by statin philicitya (n=75,698) 35 30 25 20 15 Lipophilic Statin Use

Percent Use 10 Hydrophilic Statin Use 5 0 -6 -3 1 4 7 10 30-Day Periods Post-TBI a Lipophilic statins: Atorvastatin, lovastatin, fluvastatin and simvastatin; hydrophilic statins: pravastatin, rosuvastatin

When comparing any statin use to non-use, increasing age was associated with decreased use (OR 0.52; 95% CI 0.50, 0.55 for beneficiaries 85 years or older, compared with beneficiaries aged 65-75 years). Additionally, female sex was associated with lower odds of using statins (OR 0.91; 95% CI 0.88, 0.94) than males. Among chronic conditions, hyperlipidemia had the greatest association with statin use (OR 9.54; 95% CI

9.07, 10.03). Hypertension, IHD, AMI, VHD, diabetes and prior stroke/TIA also were significantly associated with statin use (Table 6.3). Logistic regression comparing less

44 than 50% statin use to no use, and comparing greater than 50% statin use to no use showed similar associations (Table 6.3).

Table 6.3: Odds ratios (95% CI) of factors associated with statin use among older

Medicare Beneficiaries with TBI (n=75,813)

a Other races include Native American, other and unknown race b Count of CCW chronic conditions excluding cardiovascular conditions

Sensitivity analyses among beneficiaries with statin use only prior to TBI indicated similar results as the primary analysis. However, when examining beneficiaries

45 who only used statins following TBI, the association with hyperlipidemia lessened when compared to the primary analysis for beneficiaries that used statins prior to and following

TBI. While still a significant factor associated with statin use, the OR for hyperlipidemia was 2.39 (95% CI 2.22, 2.53), comparing any statin use to no use; 2.27 (95% CI 2.04,

2.53) comparing 50% use to no use; and 2.48 (95% CI 2.26, 2.74) comparing greater than

50% use to no use (Chapter VI Appendices, Tables A.6.2, A.6.3).

Discussion

This national sample of older adults with TBI suggests that statins are commonly used. Among the 75,698 beneficiaries in the study sample, 50% of beneficiaries (37,874) used statins at some point during the study period. The data suggest that among users, the average statin use duration was 12.4 months (69% of study period), and users had an average of 0.9 gaps in statin treatment. Of statin users, 85% (32,347) of beneficiaries used statins prior to TBI. Of these, 6% (1,969) of beneficiaries who used statins prior to

TBI stopped using statins following injury, suggesting that most statin users continued statin therapy following TBI.

The three observational studies that assessed statin use prior to trauma found that

20% - 22% of the participants took statins prior to injury.20,21,90 It should be noted that one study examining pre trauma statin use did not limit their study population to older adults and included all trauma patients.21 One study assessed statin continuation following trauma found that among adults 55 years or older, approximately 50% continued to receive statins during trauma hospitalization.39 Our results suggest greater

46 statin use prior to injury among older Medicare beneficiaries with TBI. This may be because our study includes Medicare beneficiaries with Medicare Part D and availability of prescription drug insurance is associated with statin use.118 The previous studies did not examine the drug insurance status of trauma patients. Nonetheless, no previous study has examined trends in statin utilization prior to and following TBI.

Findings indicate that 76% of the study population had hyperlipidemia, of which

62% (35,667) used statins at some point in the study period. Previous research utilizing

National Ambulatory Medical Care Survey and the National Hospital Ambulatory

Medical Care Survey suggests that 41% of participants with hyperlipidemia use statins.119

Furthermore, previous research also has indicated that two-thirds of patients with CVD or

CVD risk factors take statins.79

Results also suggest that older Medicare beneficiaries use lipophilic statins more than hydrophilic statins. Lipophilic statins are more likely to enter the BBB, which may help decrease cerebral inflammation following TBI.71-73 This characteristic may lead to decreased mortality following TBI.

The biggest factor associated with statin use was hyperlipidemia. Comparing any statin use to non-use, beneficiaries with hyperlipidemia were nine times more likely to take statins. When comparing statin use for greater than 50% of the study period to no use, beneficiaries with hyperlipidemia were more than 13 times more likely to use statins.

CVD such as IHD and AMI also were significantly associated with statins. This data supports previous research indicating that hyperlipidemia and CVD are major factors associated with statin use.119 However, no previous study has examined factors associated with statin use in an older Medicare population with TBI.

Among beneficiaries that only used statins following TBI (n=5,527), beneficiaries with hyperlipidemia were approximately 2.5 times more likely to use statins. The low impact of hyperlipidemia among beneficiaries that only took statins following TBI, when compared to beneficiaries that took statins prior to and following TBI may suggest that beneficiaries may have been prescribed statins following injury in order to protect against adverse events associated with TBI, rather than for hyperlipidemia. Beneficiaries that started using statins following TBI were significantly younger and had fewer cardiovascular conditions at the time of injury than non-statin users. However, new-users following TBI had longer LOS and greater injury severity than non-users (Chapter VI

Appendices, Table A.6.3). Several animal studies have indicated that statin use following trauma may help reduce physiological changes associated with secondary injury.44,84,87,88

Therefore, it is possible that clinicians prescribed statins following injury to protect against secondary injury associated with TBI.

While this is the first study to examine statin use in older Medicare beneficiaries with TBI, there are limitations to this study. This study uses prescription fill and PDC information to assess statin use. However, given the data, it is not possible to assess statin consumption by beneficiaries. Furthermore, the indication for statin use cannot be assessed in this data. Indications of statin use can help in better understanding the factors associated with statin use. Another limitation of the study is that this study excluded beneficiaries with Medicare Part C because their prescription drug information was not available; thus, study findings are only generalizable to Medicare beneficiaries with fee- for-service coverage.

Strengths of this study include a large sample of older Medicare beneficiaries with

TBI. This large sample allowed researchers to assess statin utilization by class and by individual statins. This is the first study to examine statin use by lipophilicity and hydrophilicity, as well as statin dose changes over time in older adults with TBI. Another strength of this study is that statin use both prior to and following to TBI was examined.

This allowed researchers to examine average duration and gaps in statin use prior to and following TBI.

Results of this study provide valuable information regarding statin use in a vulnerable population of older adults with TBI. The current study indicated that more than 76% of the study sample has hyperlipidemia; however, 50% of the sample took statins. Additionally, 15% of users started statin therapy following TBI, which may potentially have been given to reduce secondary injury following TBI. Given the potential protective impact of statins in reducing primary and secondary injury, clinicians can look to prescribe and encourage statin use among vulnerable older adults. Clinicians also can look to prescribe statins to patients with a history of or risk of falls as more than

80% of TBIs among adults 65 and older have TBIs due falls.120 While this study provides understanding on the extent of statin use, future studies will next examine the impact of statin use on mortality and secondary injury in order to shape pharmacological therapy guidelines following TBI.

CHAPTER VII: IN-HOSPITAL MORTALITY FOLLOWING TBI IN OLDER MEDICARE STATIN USERS

Abstract

Background: TBI disproportionately impacts older adults as hospitalization rates in the

US are more than three times greater in adults over 65, compared to adults aged 45 to 64.

Mortality rates for adults 65 and older from 2009 to 2010 was 45.2 per 100,000 people.

There are few pharmacotherapeutic options to decrease in-hospital TBI mortality. Statins have been highlighted as a potential option to decrease in-hospital mortality following

TBI.

Objective: To assess the relationship between pre-TBI statin uses and non-users and subsequent in-hospital mortality.

Methods: Atorvastatin, fluvastatin, lovastatin, pravastatin, rosuvastatin, and simvastatin use for Medicare beneficiaries 65 and older hospitalized with a TBI during 2006 through

2010 was assessed six months prior to TBI. Beneficiaries were classified as current, recent, past, and non-users of statins prior to TBI. The outcome of interest was in-hospital mortality. Logistic regression was used to obtain odds ratios (OR) and 95% confidence interval (CI) for mortality, comparing current, recent, and prior statin use to non-use.

Results: Current atorvastatin (OR 0.88; 95% CI 0.82, 0.96), simvastatin (OR 0.84; 95%

CI 0.79, 0.91) and rosuvastatin (OR 0.79; 95% CI 0.67, 0.94) use were associated with a significant decrease in the risk of in-hospital mortality following TBI.

Discussion: In addition to being the most used statins, current use of atorvastatin, rosuvastatin, and simvastatin was associated with at least a 10% decrease in-hospital

50 mortality following TBI. Clinicians can look to prescribe statins, when appropriate to older adults at risk of TBIs. These statins may help prolong survival following TBI.

Introduction

In 2010, TBI caused more than 50,000 deaths in the US.120 While a concern for all

Americans, TBI disproportionately impacts older adults due to their higher risk of falls.1,2,50,120 The rate of TBI hospitalization in the US is more than three times greater in adults over 65, compared to adults aged 45 to 64.121 Among older adults 65 and older,

TBI mortality rates from 2009 to 2010 was 45.2 per 100,000 people.121 Furthermore, mortality rates for TBI in older adults also increase with age.3

TBI is characterized by damage to brain tissue, which can lead to inflammation and abnormal CBF.8 The inflammation and abnormal CBF from the primary injury, the initial insult, can lead to mortality.9-11 Currently, there are no universally accepted pharmacological treatment guidelines to decrease mortality following TBI.33,54,55 The lack of treatment options for TBI led to the creation of the Operation Brain Trauma Therapy

(OBTT), a research consortium created to screen and advance novel biomarkers and therapies for TBI, from animal studies to human clinical trials.122 The OBTT has highlighted several therapies that can potentially be used for TBI patients, including statins.122

In addition to their primary utility in reducing serum cholesterol, statins decrease inflammation, and increase CBF.16 Statins accomplish this by limiting the production of inflammatory cytokines and promoting BBB integrity.14,16,55 Furthermore, lipophilic statins may have a greater potential to cross the BBB than hydrophilic statins, which in turn may potentially have a greater impact in reducing mortality.71-73 These properties

52 may help pre-injury statin use to decrease in-hospital mortality and a few studies have suggested this relationship.20,21

Currently, three small-scale human studies have assessed the relationship between pre-injury statin use and mortality survival among older adults with TBI, two conducted in the US and one in Singapore.20,21,90 Both US studies found a significant decrease of in- hospital mortality comparing pre-injury statin users to non-users. It should be noted that these studies also stratified patients by CVD and found that the protective effect of statins was only statistically significant among TBI patients without CVD.20,21 The third study, among Asian TBI patients, examined pre-injury statin use and 14-day mortality and did not find any association between statin use mortality.90 All three of these studies were limited due to small sample sizes and non-differentiation of statins.20,21,90 In addition to the two human studies indicating a protective effect of statins, one animal study has shown that prior statin use reduced inflammation following simulated TBI on rats and mice, postulating that statin use may lead to decreased mortality.84

Given the potential of statins to reduce mortality indicated in prior small-scale human and animal research, the objective of this study was to assess the relationship between pre-TBI statin uses and non-users who suffer a TBI and subsequent in-hospital mortality in a large-scale population of older Medicare beneficiaries.

Methods

Data and sample

CMS data was used to investigate pre-injury statin utilization and in-hospital mortality following hospital admission for TBI. Detailed information on the data and the

ICD-9-CM codes used to define TBI is available in Chapter III.

Beneficiaries were required to be at least 65 at the time of their first TBI and have at least six months of continuous Medicare Parts A, B, and D coverage prior to TBI.

Medicare Parts A, B, and D include inpatient, outpatient, and prescription drug claims, respectively.93 Beneficiaries with Medicare Part C are part of managed care programs and their medical claims are not available; therefore, these beneficiaries were excluded.

Exposure

The exposure of interest was statins, which comprised atorvastatin, fluvastatin, lovastatin, pravastatin, rosuvastatin, and simvastatin. These statins were approved and available for use in the US during the study period.70 More information on the exposure of interest can be found in Chapter IV. Since beneficiaries were required to have six months of coverage prior to TBI, each beneficiary had six periods (hereafter referred to as months) of statin use information prior to TBI. Statin use was examined by class and by individual statins.

Statin users (by class and by individual statins) and non-users was classified into four mutually exclusive categories: current use (use in any of the two months immediately prior to TBI); recent use (use in three or four months and no use in one or

54 two months prior to TBI); prior use (use in five or six months and no use in one to four months prior to TBI); and no use in any month prior to TBI. Duration of statin use and total statin use was measured in order to obtain a duration/total use effect of statins. Statin duration was measured by examining consecutive use of statins in the months prior to

TBI. Statin duration was categorized if beneficiaries used statins consecutively for 1-2 months, 3-4 months, and 5-6 months immediately prior to TBI. Similarly, total use was categorized if beneficiaries used statins for a total of 1-2 months, 3-4 months, and 5-6 months prior to TBI. Statin use also was classified per individual months prior to TBI.

Lastly, statin use was classified by lipophilicity and hydrophilicity. Lipophilic statins included atorvastatin, lovastatin, fluvastatin and simvastatin while hydrophilic statins included pravastatin and rosuvastatin.

Outcome

The outcome of interest was in-hospital mortality, which was observed through inpatient Medicare claims data. In-hospital mortality was flagged if beneficiaries were discharged dead following index TBI. Beneficiaries were censored following hospitalization discharge. Mortality was assessed by any statin use, individual statins, and statins classified as either lipophilic or hydrophilic.

Covariates

Covariates included demographic characteristics (age, race, and sex) and chronic conditions, which were obtained through the CCW. More information on CCW can be found in Chapter III. Overall health was determined by including a comorbidity count

55 excluding CVD (hypertension, hyperlipidemia, IHD, CHF, AMI, VHD, and diabetes), and ADRD, as these conditions were included separately. CVD and ADRD were added separately because of their association with statin use and mortality.123,124 A SNF stay in the month prior to TBI also was included to determine overall health as a SNF stay may be indicative of poor health.19,99

Medication use that may help improve survival following TBI, such as antiplatelets, anticoagulants, and beta-blockers also were included as covariates.

Additionally, other statin use also was included. For instance, when assessing the impact of atorvastatin on in-hospital mortality, other statins used also were included as covariates.

TBI severity was measured using SRRi. More information on SRRi can be found in Chapter V. Additionally, LOS also was included as a TBI severity proxy measure.

Lastly, AHRF was used to include demographic and healthcare related geographical characteristics. The AHRF database contains over 6,000 variables from over

50 sources including population, healthcare professional and healthcare facility data.95,96

Given the amount of variables, correlations between selected geographical variables were calculated and variables correlated 0.80 or more were excluded (Chapter VII Appendices,

Table A.7.1).125 AHRF data was linked to CMS data through Social Security

Administration’s county codes. Therefore, county-level demographic covariates (median income and geographical region), and provider and hospital characteristics per 100,000 people (total physicians, hospitals, hospitals with trauma services, hospitals with neurological services, and beds) were included.

Data analysis

Bivariate analysis was conducted to compare the baseline descriptive and sociodemographic characteristics of statin users and non-users. Chi-square and t-tests were used to assess significant differences for categorical and continuous variables, respectively. Additionally, a Wilcoxon signed rank sum test was used to assess differences in medians between statin users and non-users. Logistic regression was used to obtain OR and 95% CI for mortality, comparing current, recent, and prior statin use to non-use. Furthermore, additional analysis was conducted comparing statin use per each individual month prior to TBI with non-use, rather than comparing the categorical use variables of current use, recent use, and prior use with non-use.

To better understand the impact of statin use and in-hospital mortality, the relationship between different statin doses and in-hospital mortality was examined among statin users for the statins that indicated a significant protective effect against in-hospital mortality. This analysis was conducted by comparing statin doses at the monthly level among statin users prior to TBI for only the months that statin use indicated a protective effect against in-hospital mortality. Statin doses for comparison were selected by examining the distribution of statin doses.

Secondary analysis was conducted to examine the duration-effect relationship between statin use and mortality; that is to assess if consecutive use and total use prior to

TBI is associated with reduced in-hospital mortality. Therefore, analysis was conducted to examine if: 1) consecutive statin use for 1-2 months, 3-4 months, and 5-6 months prior to TBI; and 2) 1-2 months, 3-4 months, and 5-6 of total statin use prior to TBI was associated with reduced in-hospital mortality.

Lastly, a sensitivity analysis was conducted to assess robustness of results by conducted all analyses after excluding beneficiaries who died within 24 hours of TBI hospitalization. This approach has been used in prior studies in order to exclude the most severe TBIs.20,21

All analyses were performed with SAS version 9.2 (Cary, NC). This study was approved by the University of Maryland Baltimore’s Institutional Review Board.

Results

A total of 116,193 Medicare beneficiaries had a TBI during 2006 to 2010; of these individuals, 116,170 were at least 65 year old at the time of their TBI. Among those

65 and older, 112,716 beneficiaries had continuous coverage of Medicare Parts A, B and

D six months prior to their index TBI. An additional 607 beneficiaries were excluded due to missing prescription drug information because these beneficiaries had a hospital or

SNF stays throughout the study month and prescription drug claims are not available during facility stays. Thus, the final sample included 112,109 beneficiaries.

Of the 112,109 beneficiaries in the study sample, 45,752 (41%) used statins at some point prior to TBI. Overall, the majority of the sample was white (87%) and female

(64%). Statin users were younger (80 years compared to 82 years, p <0.0001) and had slightly higher median income ($50,660 compared to $49,518, p <0.0001). When comparing provider characteristics, statin users lived in areas with a greater density of physicians per 100,000 people (239 compared 231, p <0.0001) but a lesser density of hospital beds per 100,000 population (301 compared to 304, p <0.0001). The majority of

58 the sample had hypertension (91%) and hyperlipidemia (75%). Statin users were significantly more likely to have cardiovascular conditions including IHD and CHF.

While 97% of statin users had hyperlipidemia, the majority (60%) of non-users also had hyperlipidemia. There were no significant differences in the TBI severity measures

(Table 7.1).

Table 7.1: Descriptive characteristics of statin users and non-users among Medicare beneficiaries with TBI (n=112,109)

Table 7.1 (continued): Descriptive characteristics of statin users and non-users among

Medicare beneficiaries with TBI (n=112,109)

a Beneficiaries were categorized as non-users if they did not use statins at any time during study period, while beneficiaries were categorized as users if they had any statin use at any time during study period. b P-value from chi-square for categorical variables, Student’s t-test for continuous variables and Wilcoxon signed rank sum test to test differences between medians reflects comparison between statin users and non - users. c Other races include Native American, other and unknown race d SNF stay in the month immediately preceding TBI e Count of CCW chronic conditions excluding cardiovascular conditions (hypertension, hyperlipidemia, ischemic heart disease, congestive heart failure, acute myocardial infarction, valvular heard disease) and diabetes and ADRD Abbreviations: TBI, traumatic brain injury; SD, standard deviation; SNF, skilled nursing facility; ADRD, Alzheimer’s disease and related dementias

Most statin users were classified as current users 41,174 (90%), while 2,852 (8%) were classified as recent users and 930 (2%) were classified as past users. The most commonly used statins were simvastatin and atorvastatin (Table 7.2).

Table 7.2: Statin use among Medicare beneficiaries prior to TBI

a Lipophilic statins: Atorvastatin, lovastatin, fluvastatin and simvastatin; hydrophilic statins: pravastatin, rosuvastatin

When examining statin use per prior month relative to TBI, most statin users

(84%) used statins in the month immediately prior to TBI (Chapter VII Appendices,

Table A.7.2). With regard to the outcome of interest, a total of 9,810 (9%) of beneficiaries were discharged dead following TBI hospitalization.

In adjusted analysis, current use of any statins was associated with decreased in- hospital mortality (OR 0.87; 95% CI 0.82, 0.92). When assessing individual statins, current use of the lipophilic statins atorvastatin (OR 0.88; 95% CI 0.82, 0.96), and simvastatin (OR 0.84; 95% CI 0.79, 0.91) and hydrophilic rosuvastatin (OR 0.79; 95% CI

0.67, 0.94), were associated with a significant decrease in the risk of in-hospital mortality following TBI. Recent use and past use of any statin was not associated with a decreased risk of in-hospital mortality. Current use of both lipophilic (OR 0.87; 95% CI 0.83, 0.93) and hydrophilic (OR 0.87; 95% CI 0.78, 0.98) statins was associated with a significant decrease in in-hospital mortality (Table 7.3).

Table 7.3: Odds ratios (95% CI) of in-hospital mortality following TBI comparing statin users and non-users

a Lipophilic statins: Atorvastatin, lovastatin, fluvastatin and simvastatin; hydrophilic statins: pravastatin, rosuvastatin b Adjusted for: demographic characteristics (age, sex, race, income, region); medications (antiplatelets, anticoagulants, beta blockers, other statins); cardiovascular disease (CHF, AMI, IHD, VHD, hyperlipidemia, hypertension, prior stroke); diabetes; Alzheimer’s disease and related dementia; count of CCW chronic conditions excluding cardiovascular conditions (hypertension, hyperlipidemia, ischemic heart disease, congestive heart failure, acute myocardial infarction, valvular heard disease) and diabetes and ADRD; SNF stay in month prior to TBI; injury severity (SRRi, length of hospital stay); provider and hospital characteristics per 100,000 population (physicians, hospitals, hospitals with trauma services, hospitals with neurological services, beds)

Analysis assessing statin use per month prior to TBI indicated that any statin use, atorvastatin use, rosuvastatin use, and simvastatin use in the month immediately prior to

TBI was associated with a decreased risk of in-hospital mortality when compared to non- use (Table 7.4).

The relationship between statin dose in the two months prior to TBI and in- hospital mortality was assessed for these three statins. Statin doses were categorized by examining the most prominent doses among beneficiaries (Chapter VII Appendices,

Table A.7.3). This analysis, conducted only among statin users, found no relationship between atorvastatin and rosuvastatin dose and mortality. However, higher doses of simvastatin were significantly associated with a decrease in in-hospital mortality (Chapter

VII Appendices, Table A.7.4).

Table 7.4: Odds ratios (95% CI) of in-hospital mortality following TBI comparing statin users and non-users by individual months prior to TBI

The findings indicated that 1-2 months of consecutive use of rosuvastatin (OR

0.57; 95% CI 0.35, 0.91) and simvastatin (OR 0.84; 95% CI 0.71, 0.98) immediately prior to TBI was associated with a significant decrease in in-hospital mortality.

Additionally, 5-6 months of consecutive use of these statins immediately prior to TBI also was associated with a significant reduction of in-hospital mortality. Similarly, 5-6 months of total use of rosuvastatin and simvastatin prior to TBI was significantly associated with a decrease in in-hospital mortality (Table 7.5).

Sensitivity analysis restricted to beneficiaries who did not die with 24 hours of hospitalization showed similar results as the primary analysis (Chapter VII Appendices,

Table A.7.5 – A.7.8).

Table 7.5: Odds ratios (95% CI) of in-hospital mortality following TBI comparing statin users and non-users, by consecutive and total use prior to TBI

Discussion

Results from this national sample of Medicare beneficiaries indicate that statin use in the months prior to TBI is associated with reduced in-hospital TBI mortality.

Specifically, current statin use, defined as statin use in any of two months immediately preceding TBI, was linked to decreased in-hospital mortality. The decrease in mortality appears to be driven by atorvastatin, rosuvastatin, and simvastatin use, all of which are associated with a greater than 10% reduction in mortality. The greatest reduction in mortality is associated with current use of rosuvastatin (21%), followed by simvastatin

(16%) and atorvastatin (12%). Fluvastatin, lovastatin and pravastatin also are associated with decreased mortality; however, these relationships were not statistically significant.

This may be due to insufficient number of beneficiaries taking these statins or channeling bias, where different treatments are given for patients with difference prognoses.126 More research is required, particularly clinical trials, to investigate the differences in mortality among statin users.

Atorvastatin, rosuvastatin, and simvastatin use in the month immediately prior to

TBI was associated with decreased in-hospital mortality. This may because of the lasting impact of reduced inflammation associated with statin use. While statins’ impact on lipids is apparent four to six weeks following use, there is limited research regarding statins’ duration of impact on inflammation.72 No study has examined the length of time it takes for statins to impact inflammation following TBI; however, some studies have examined secondary endpoints and inflammation associated with other chronic conditions.71,127 One clinical trial among atherosclerosis patients found that atorvastatin

68 treatment reduced atherosclerotic plaque inflammation starting at four weeks after treatment.127 Another clinical trial found that simvastatin use reduced oxidative stress, which is associated with inflammation, four weeks following treatment among heart failure patients.71 These studies indicate that, similar to its impact on lipids, statin therapy may impact inflammation at least four weeks after initiation. Therefore, the reduced in- hospital mortality associated with statin therapy may be due to the lasting anti- inflammatory impact of statins as the reduction of in-hospital mortality is mainly witnessed among patients with statin use in the month immediately prior to TBI, rather than several months prior to TBI.

The results also indicated that in addition to being protective against in-hospital mortality, atorvastatin and simvastatin are the most commonly used statins. A total of

21,598 beneficiaries used simvastatin, of which 19,194 were classified as current users, and 15,597 beneficiaries used atorvastatin, of which 13,156 were classified as current users (Chapter VII Appendices, Table A.7.2).

Three previous human studies assessing pre-TBI statin use and in-hospital mortality; however, these studies did not distinguish between statins.20,21 Although these studies were limited due to small sample size and poor determination of statin use, both studies among US TBI patients did indicate the protective effect of statins.20,21 While a few animal studies have assessed the impact of specific statins, most studies administered statins following simulated TBI.42,44-46,85 The results of these animal studies significantly demonstrated that atorvastatin, rosuvastatin, and simvastatin reduced cerebral inflammation and improved CBF following TBI.42,44-46,85

In secondary analysis examining the impact of duration of statin use, the greatest reduction of in-hospital mortality (43%) was associated among beneficiaries that consecutively used rosuvastatin 1-2 months immediately prior to TBI. Additionally, there was a 31% reduction of in-hospital mortality among beneficiaries that consecutively used rosuvastatin 5-6 months immediately prior to TBI. The results are similar for analysis considering total month use prior to TBI. These results indicate a potential duration-effect of statin use on mortality, with beneficiaries using statins consecutively and more often witnessing the greatest reduction of in-hospital mortality. No previous human study has examined a duration effect of statin use on in-hospital mortality among

TBI patients. However, a few meta-analyses have examined the impact of statin duration on cardiovascular events.128,129 One meta-analysis found that longer duration of statin use was associated with greater reduction in IHD, while another found that CVD patients that used statins for longer durations had lower risks of atrial fibrillation following cardiac surgery.128,129 More research is required examining the impact of statin duration in older

TBI patients to corroborate the results of this study.

Our results also indicate that current use of both lipophilic and hydrophilic statins reduce in-hospital mortality following TBI by 13%. A total of 39,812 beneficiaries used lipophilic statins and 6,831 used hydrophilic statins. Of the statins studied, atorvastatin and simvastatin are the most lipophilic statins, making them most susceptible to cross the

BBB.130 While it is expected that lipophilic statins may have a greater impact in reducing mortality due to their ability to cross the BBB, there are multiple mechanisms that make statins neuroprotective following TBI, including lowering brain cholesterol and reducing neuronal cell death.130 Rosuvastatin helps lower brain cholesterol and reduce neuronal

70 cell death, which may be why rosuvastatin, a hydrophilic statin, reduced in-hospital mortality along with atorvastatin and simvastatin.130

There are several limitations to consider when interpreting the results of this study. First, this study excluded Medicare Part C beneficiaries. Therefore, the study results cannot be generalized to the entire Medicare population. Furthermore, the administrative nature of the data made it impossible to capture injury severity, which could have been possible with medical chart reviews. However, injury severity was ascertained through the ICD-9-CM code-based SRRi. Measures of injury severity based on ICD-9 codes, such as the SRRi, have been demonstrated to be valuable in administrative claims research.108 Another limitation is that the data only includes patients hospitalized with a TBI and excludes patients with less severe TBIs who did not experience a hospitalization. While this limitation cannot be overcome within the data, sensitivity analysis was conducted by excluding beneficiaries who died within 24 hours of hospitalization, as this is indicative of most severe TBI.20,21

Lastly, the data does not capture medication usage during TBI hospitalization.

This is a major limitation as other medications given in-hospital that may improve survival could not be adjusted in the analysis. Furthermore, it is not known whether statins were continued during hospitalization. However, one study examined statin discontinuation during TBI hospitalization and found that 39% of pre-TBI statin users did not receive statins within 48 hours of hospitalization.39 Of those that discontinued, 58% discontinued due to severity of injury.39 It is unclear why non-severe patients discontinued statins; however, it may be due to medical oversight.39 Future studies can

71 further investigate statin discontinuation rates while hospitalized and explore potential reasons for discontinuation.

Despite the limitations, the study found that atorvastatin, rosuvastatin, and simvastatin use, possibly due to their lasting anti-inflammatory effects, in one or two months immediately prior to TBI reduced in-hospital mortality while consecutive use of rosuvastatin and simvastatin in the two months immediately prior to TBI had the greatest impact in reducing in-hospital mortality.

There also are numerous strengths of this study. This is the first large-scale observational study assessing pre-TBI statin use and mortality. This is the first study to distinguish between individual and lipophilic and hydrophilic statins. Furthermore, this study is the first to examine the impact of the duration of statin therapy on in-hospital mortality. This study also was able to include a myriad of covariates from CCW and

AHRF data to reduce bias. The CCW dataset included health-related characteristics such as comorbidities and medication use. The AHRF dataset included sociodemographic and geographical characteristics including income, and region. This study also was able to control for healthcare characteristics including the density of physicians, hospitals, and hospital beds in the beneficiaries’ geographic area. Adjusting for these covariates helped reduce confounding associated with beneficiary health, sociodemographic, and geographical characteristics. Additionally, a strength of this study is the large sample that allowed sensitivity analysis to assess the robustness of the results by excluding beneficiaries that died within 24 hours of hospitalization. The results from the sensitivity analyses supported primary analyses indicating that current use of atorvastatin, rosuvastatin, and simvastatin and use in the month immediately prior to TBI significantly

72 reduced in-hospital TBI mortality. Furthermore, sensitivity analysis regarding the impact of statin dose on mortality showed similar results.

More research is required examining the impact of statin duration in older TBI patients to corroborate the results of this study. However, the results of this study make an important contribution for clinicians treating older adults with TBI. They indicate that with atorvastatin, rosuvastatin, or simvastatin use, there is the potential to reduce the death rate among older adults that suffer a TBI. Furthermore, the results also indicate that consecutive use of rosuvastatin and simvastatin in the two months immediately prior to

TBI had the greatest impact in reducing in-hospital mortality.

The current translational research has serious implications for healthcare providers. If appropriate, clinicians could prescribe statin therapy, especially among those with hyperlipidemia, to not only treat hyperlipidemia but to potentially reduce the risk of mortality in the event of a TBI. Furthermore, among patients with pre-injury statin use, healthcare providers can look to continue statin use during hospitalization. Discontinuing statins in hospital after injury may have deleterious effects on mortality.39 Statins’ anti- inflammatory benefits abrogate within 72 hours of statin discontinuation, which may increase mortality.131,132 Additionally, several animal studies also have shown that statins administered in the hours immediately following TBI significantly reduce inflammation.42,44-46,85 The quick reversal of the anti-inflammatory properties of statins following discontinuation and animal studies suggesting the immediate benefit of statins following injury highlight the need for quality clinical trials investigating the impact of statin use in-hospitalization and mortality. However, given available evidence and statins’ positive safety profile, clinicians can look to administer statins in-hospital for all older

TBI patients. Such preventative measures can potentially reduce the death rate due to TBI in older adults. This study provides valuable information to clinicians providing care for older adults and its results can be used to help patients and families make informed healthcare decisions.

CHAPTER VIII: MORTALITY AND SECONDARY INJURY FOLLOWING TBI IN OLDER MEDICARE STATIN USERS

Abstract

Background: TBI is a major health concern, especially older adults. TBI among older adults is associated with an increase in mortality and secondary injury, including stroke, depression, and ADRD following TBI. There are no universally accepted pharmacological treatment guidelines following TBI. Statins are highlighted as a potential therapy for primary and secondary injury associated with TBI.

Objective: To assess the relationship between post-TBI statin use and: 1) mortality; and

2) the incidence of secondary injury.

Methods: Atorvastatin, fluvastatin, lovastatin, pravastatin, rosuvastatin, and simvastatin use was assessed following TBI among for Medicare beneficiaries 65 and older hospitalized with a TBI during 2006 through 2010. Outcomes of interest included mortality, stroke, depression, and ADRD. RR were obtained using discrete time analysis and generalized estimating equations.

Results: Statin use of any kind was associated with decreased mortality following TBI- hospitalization discharge. Any statin use also was associated with a decrease in any stroke (RR, 0.86; 95% CI, 0.81, 0.91), depression (RR, 0.85; 95% CI, 0.79, 0.90), and

ADRD (RR, 0.77; 95% CI, 0.73, 0.81).

Discussion: While all statins were associated with a decrease in mortality, only atorvastatin and simvastatin use was associated with a decrease in all secondary injury outcomes. Both atorvastatin and simvastatin also were the most commonly used statins.

These findings provide valuable information for clinicians treating older adults with TBI as clinicians can prescribe atorvastatin and simvastatin to all beneficiaries with TBI in order to decrease primary and secondary injury.

Introduction

TBI impacts millions of adults in the US and accounts for almost one-third of all injury-related deaths.120 The rates of TBI-related hospitalizations and deaths is highest among adults 65 years of age and older, and TBI mortality rates are estimated at 24.5,

51.4, and 103.8 per 100,000 for adults aged 65-74, 75-84, and 85 and older, respectively.3,120

In addition to an increase in mortality for older adults, there also is a chronic neuroinflammation following injury. This sustained inflammatory cascade can lead to several unfavorable outcomes including stroke, depression, and ADRD.10,12,13,30,60,133

These events are commonly referred to as secondary injury and can occur in the days, months, and years following TBI.9,10,29 Furthermore, natural aging also contributes to chronic neuroinflammation; therefore, older adults with TBI are particularly susceptible to both chronic neuroinflammation and ensuing secondary injury.10

There are no universally accepted pharmacological treatment guidelines following

TBI.33,54,55 Pharmacotherapy guidelines have focused largely on treating the primary or acute injury.35 These pharmacotherapies include corticosteroids, analgesics, and sedatives and barbiturates, which are used to either reduce intracranial pressure or to treat pain and agitation.35,36 Since TBI is typically followed by secondary injury, pharmacological treatment also may be geared toward alleviating symptoms of secondary injury.33,34 For instance, older adults face an increased risk of depression following TBI; therefore, treatment may be focused on alleviating depressive symptoms.33,60,134

The OBTT consortium was established to highlight acute and long-term potential pharmacotherapies for TBI.122 The consortium has suggested statins, typically used to treat CVD, as a potential therapy for primary and secondary injury associated with TBI because animal studies have suggested that statins may help to reduce chronic neuroinflammation and increase CBF following TBI.122,135 Furthermore, lipophilic statins may have greater potential as they are more likely to cross the BBB.71-73

There are currently only two human studies investigating the impact of statin use following TBI on either inflammation or mortality.39,91 One study consisted of 39 TBI patients, of which 19 received statins for only ten days following TBI. The results of this study indicated that statin use decreased inflammation following TBI.91 Another study examined statin discontinuation among 61 TBI patients that used statins prior to TBI and found that statin discontinuation was not significantly associated with increased mortality.39 It should be noted that multiple animal studies have examined inflammatory biomarkers after TBI and have indicated that statin use may decrease inflammation and improved CBF following TBI.42,44-46,85

No study has investigated the relationship between statin use following TBI and mortality or secondary injury. This, the objective of this study is to assess the relationship between post-TBI statin use and: 1) mortality; and 2) the development of secondary injury, including stroke, depression, and ADRD among older Medicare beneficiaries

Methods

Data and sample

CCW data from CMS was used to investigate the relationship between post-TBI statin use and mortality or secondary injury following TBI. These data include all TBIs occurring among Medicare beneficiaries from 2006 through 2010. TBI was identified by inpatient claims using ICD-9-CM diagnosis codes 800.xx, 801.xx, 803.xx, 804.xx,

850.xx- 854.1x, 950.1-950.3, and 959.01. More information regarding the identification of TBI and the index date/index TBI can be found in Chapter III.

To be included in the study, beneficiaries were required to be at least 65 years of age at the time of TBI, have six months of observation time prior to TBI, and continuous coverage of Medicare Parts A, B, and D throughout the study period following TBI.

Therefore, beneficiaries could have up to 54 months of follow-up. Medicare Parts A, B, and D were used to identify inpatient, outpatient, and medication claims, respectively, throughout the study period.93 Beneficiaries with Medicare Part C (private insurance) were excluded because their inpatient, outpatient, and medication claims were not available.

Exposure

The exposure of interest was use of statins, including atorvastatin, fluvastatin, lovastatin, pravastatin, rosuvastatin, and simvastatin were identified using PDE claims.

Individual and any statin use was determined per 30-day period (henceforth referred to as

79 month) prior to and following TBI. More information on the exposure of interest can be found in Chapter IV.

Since beneficiaries had six months of coverage prior to TBI, pre-TBI statin use was divided into four categories: 1) use in the first or second month immediately prior to

TBI; 2) use that was three or four months prior to TBI; 3) use that was over four months prior to TBI; and 4) no use prior to TBI. Following TBI, beneficiaries were classified as users or non-users on a monthly basis, creating a time-varying exposure.

In addition to examining any statin and individual statin use, statin use also was categorized by lipophilicity and hydrophilicity status. Lipophilic statins included atorvastatin, lovastatin, fluvastatin, and simvastatin; hydrophilic statins included pravastatin and rosuvastatin.

Outcomes

The two outcomes of interest included primary injury (mortality) and secondary injury (stroke, depression, or ADRD). Mortality was determined through CCW data and assessed following hospital discharge, making the outcome all-cause mortality rather than

TBI-specific. Different variables were created to assess 30-day, 60-day, 90-day and mortality at any time following TBI hospitalization discharge. The 30-day, 60-day, and

90-day mortality variables were mutually exclusive in order to examine the time period for when statin use may have been protective for primary injury. For instance, beneficiaries who died within 30 days were excluded when analyzing 60-day or 90-day mortality to avoid biasing the results.

Secondary injury comprised stroke, depression, and ADRD. Stroke was categorized as any stroke, ischemic stroke, and hemorrhagic stroke. More information on the codes used to define stroke can be found in Chapter IV. Depression and ADRD were defined through the CCW flags for chronic conditions. CCW data includes the date of first diagnosis of depression and ADRD; therefore, beneficiaries with their first diagnosis of depression or ADRD following TBI were flagged as having incident depression or

ADRD, respectively.

Covariates

This study incorporated covariates including sociodemographic, health, drug utilization, and geographical characteristics. CCW data includes demographic characteristics such as age, sex, race, and low income subsidy (LIS) status. The data flags beneficiaries that received LIS during a month, making it a time-varying covariate. The data also identifies the first occurrence of 27 chronic conditions through inpatient claims using ICD-9-CM diagnosis codes. These chronic conditions include CVD such as IHD,

CHF, and AMI. Other health covariates included LOS, discharge status, and TBI injury severity. LOS was a continuous variable and discharge status was dichotomized to either being discharged to the beneficiaries’ home or a healthcare facility, including nursing home, SNF, or another hospital. TBI injury severity was measured through the ICD-9-

CM based SRRi measure, previously used to determine injury severity in a study of older

Medicare patients with injury, including TBI.136 More information on the SRRi measure can be found in Chapter V.

In addition to pre-TBI statin use, drug utilization covariates include anticoagulant, antiplatelet, and beta-blocker use following TBI. These pharmacotherapies were included because they may impact primary and secondary injury, or have been indicated as potential treatments for TBI.134,137-139 Post-TBI drug variables were made to be time- varying as their monthly use was assessed following TBI.

Geographic covariates were included by linking AHRF data with the CCW data by beneficiaries’ county codes. The AHRF data repository includes several thousand variables from multiple sources. These variables include demographic variables such as income and region, and healthcare provider characteristics such as total number of physicians, hospitals, and beds.95,96

Data Analysis

Bivariate analysis examining associations between statin use and demographic, health, and geographic characteristics were tested using chi-squared test of proportions for categorical variables, Student’s t-tests for continuous variables, and the Wilcoxon

Rank Sum test when comparing medians between statin users and non-users. Trends in statin use were examined by any statin use, individual statin use, and use based on statin lipophilicity status.

Given the longitudinal nature of the data, discrete time analysis was used to assess the relationship between statin use and primary and secondary injury. This relationship was assessed using generalized estimating equations (GEE) with a binomial distribution and complimentary log-log link. This modelling approach is appropriate for survival analysis given multiple observations per beneficiary and a time-varying exposure.140

Since it was important to confirm that meaningful exposure occurs prior to the outcomes, statin use was lagged one month for secondary injury outcomes to ensure the exposure preceded the outcome. This approach also has been used in longitudinal

Medicare data.19,99 Statin use was not lagged for the mortality models because observation time for beneficiaries ends at death; therefore, mortality will always follow the exposure. This is not necessarily true for beneficiaries who experience secondary injury events because beneficiaries are censored following the month an event is experienced. For instance, it is possible that a beneficiary experiences an event in the beginning of the month, and continues to receive statins in the same month following the event. In this case, the outcome precedes the exposure.

Separate GEE models were used for each outcome and each beneficiary was censored after experiencing the outcome of interest. In order to narrow down the multiple geographic AHRF variables and to build parsimonious models, correlations between geographic covariates were examined. If a variable was correlated more than 0.80 were with another variable, then one of the variables was excluded from the analysis (Chapter

VIII Appendices, Table A.8.1).125 All final models included sociodemographic variables

(age, race, sex, county-level income, LIS status, and region); comorbidities

(hyperlipidemia, hypertension, CHF, IHD, AMI, ADRD, valvular heart disease, and diabetes); count of CCW chronic conditions, excluding the above mentioned comorbidities due to their relationship with the exposure and outcomes; history of stroke; injury severity (LOS, discharge status, SRRi); healthcare provider characteristics (total physicians, hospitals, hospitals with trauma services, hospitals with neurological services, and beds per 100,000 people); and pre-TBI statin use. The mortality and stroke models

83 also included anticoagulant, antiplatelet, and beta-blocker use in the month due to their relationship with statin use and mortality and stroke.134,137-139 Additionally, the model assessing any stroke included history of any stroke in the six months prior to TBI.

Similarly, the models assessing ischemic and hemorrhagic stroke included history of ischemic and hemorrhagic stroke, (respectively), six months prior to TBI. Lastly, the depression model also included monthly beta-blocker use.

Secondary analyses on secondary injury was conducted after excluding beneficiaries with less than 12 months of observation time following TBI. This was done to allow beneficiaries greater time to experience or be diagnosed with secondary injury events. For instance, beneficiaries may develop depression and ADRD following TBI, but may experience death before diagnosis. Sensitivity analyses were conducted by including inverse probability of treatment weights (IPTW). IPTW create a pseudo population that predicts the probability of a beneficiary to receive treatment based on the beneficiary’s treatment and confounder history.141,142 The treatment weights decrease bias caused by non-random assignment of treatment.141 IPTWs were constructed by modelling any post-TBI statin use following hospitalization discharge as a function of the risk factors (baseline and time-varying) for the outcomes of interest. The risk factors included the covariates mentioned above for each model. The final sensitivity analysis models included their IPTW in addition to their individual covariates IPTW.

RR and 95% CI were reported. All analyses were performed using SAS (Cary,

North Carolina). This study was approved by the University of Maryland Baltimore’s

Institutional Review Board.

Results

A total of 116,170 Medicare beneficiaries 65 and older were hospitalized with a

TBI from 2006 through 2010. Of these, 115,334 beneficiaries had six months of observation time prior to TBI and 110,500 had continuous Medicare Parts A, B, and D coverage following TBI. After excluding beneficiaries who had missing prescription drug information for the entire follow-up period (due to facility stays) and who died during hospitalization, the final study sample comprised 100,515 beneficiaries. Approximately

50% (50,173) of these beneficiaries used statins either prior to and/or following TBI, while 41% (41,651) only used statins at any time following hospitalization discharge.

The study sample was predominately white (87%) and female (65%) and the mean age of the sample was 81 years of age (standard deviation (SD), 8.1). Statin users were more likely to have CVD (p < 0.0001). While 93% of statin users had hyperlipidemia, the majority (57%) of non-users also had hyperlipidemia (p < 0.0001).

Statin users also were more likely to have hypertension and diabetes (p < 0.0001). There were no significant differences with regards to the ICD-9-CM based injury severity measure (SRRi 0.90 for users and non-users, p= 0.72). However, while likely not clinically significant, statin users had a mean LOS of 6.4 days while non-users had a mean LOS of 6.3 days (p=0.0043) (Table 8.1).

Table 8.1: Descriptive characteristics of statin users and non-users among Medicare beneficiaries with TBI (n=100,515)

Table 8.1 (continued): Descriptive characteristics of statin users and non-users among

Medicare beneficiaries with TBI (n=100,515)

a Beneficiaries were categorized as non-users if they did not use statins at any time during study period, while beneficiaries were categorized as users if they had any statin use at any time during study period b P-value from chi-square for categorical variables, Student’s t-test for continuous variables and Wilcoxon signed rank sum test to test differences between medians reflects comparison between statin users and non - users c Other races include Native American, other and unknown race d Comorbidity burden: count of chronic conditions excluding cardiovascular conditions (hypertension, hyperlipidemia, ischemic heart disease, congestive heart failure, acute myocardial infarction, valvular heard disease) and diabetes and ADRD Abbreviations: TBI, traumatic brain injury; SD, standard deviation; SNF, skilled nursing facility; ADRD, Alzheimer’s disease and related dementias

Figures 8.1 and 8.2 display trends in post-TBI statin use by individual statins

(Figure 8.1) and by statin lipophilicity status (Figure 8.2). Statin use remained stable throughout the study period. The most commonly used statin was simvastatin, followed by atorvastatin. Fluvastatin was the least used statin (Figure 8.1). Lipophilic statins were more commonly used than hydrophilic statins, as the two most used statins (simvastatin and atorvastatin) are lipophilic statins (Figure 8.2).

Figure 8.1: Statin use per month following TBI among older Medicare beneficiaries (n=100,515)

Statin use following TBI (n=100,515) 45 40 35 Statin Use 30 Atorvastatin use 25 Fluvastatin Use 20 Lovastatin Use Percent Use 15 Pravastatin Use 10 Rosuvastatin Use 5 Simvastatin Use 0 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 30-Day Periods Post-TBI

Figure 8.2: Statin use per month following TBI among older Medicare beneficiaries, by statin lipophilicity (n=100,515)

Statin following TBI by statin philicitya (n=100,515) 35

15 Lipophilic Statin Use Percent Use 10 Hydrophilic Statin Use

0 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 30-Day Periods Post-TBI a Lipophilic statins: atorvastatin, fluvastatin, lovastatin, and simvastatin; hydrophilic statins: pravastatin and rosuvastatin

A total of 41,650 beneficiaries died following TBI hospitalization discharge, of whom 8,507 died within 30 days, 3,348 died after 30 days but within 60 days, and 2,439 died after 60 days but within 90 days of hospitalization discharge. With regard to secondary injury, 9,420 beneficiaries developed ischemic stroke, 3,841 had hemorrhagic stroke, 10,748 had incident depression and 14,907 had incident ADRD.

The adjusted GEE models showed that post-TBI statin use of any kind was associated with lower mortality following TBI-hospitalization discharge. The greatest difference is seen in 30-day mortality following TBI, with lovastatin (RR, 0.05; 95% CI,

0.01, 0.14) and rosuvastatin (RR, 0.10; 95% CI, 0.05, 0.23). Additionally, both lipophilic and hydrophilic statins were associated with lower mortality (Table 8.2).

Table 8.2: Relative risks (95% CI) of mortality following TBI among Medicare beneficiaries, comparing statin use to non-use

Adjusted analysis of secondary injury indicated that any statin use was associated with a lower risk of any stroke (RR, 0.86; 95% CI, 0.81, 0.91), ischemic stroke (RR,

0.91; 95% CI, 0.85, 0.96), hemorrhagic stroke (RR, 0.75; 95% CI, 0.67, 0.83), depression

(RR, 0.85; 95% CI, 0.79, 0.90), and ADRD (RR, 0.77; 95% CI, 0.73, 0.81). With regard to individual statins, both atorvastatin and simvastatin use was associated with a lower risk of all secondary injury outcomes (Table 8.3).

Secondary analysis among beneficiaries with at least 12 months of observation time following TBI showed similar associations as the primary analysis of secondary injury following TBI (Table 8.4).

Sensitivity analyses incorporating IPTW showed similar results for primary and secondary injury, with the exception of the relationship between hydrophilic statin use and depression not being statistically significant when incorporating IPTW (Chapter VII

Appendices, Tables A.8.2 – A.8.4).

Table 8.3: Relative risks (95% CI) of secondary injury following TBI among Medicare beneficiaries, comparing statin use to non-use

Discussion

This is the first study to investigate the relationship of statin use following TBI and mortality post TBI-hospitalization discharge. In this sample of older Medicare beneficiaries with TBI, statin use following injury was associated with major decreases in

30-day, 60-day, and 90-day mortality, as well as a reduction in ischemic and hemorrhagic stroke, incident depression, and ADRD. Specifically, atorvastatin and simvastatin were associated with significant decreases in all secondary injury outcomes. Furthermore, these statins were the most commonly used statins. These results are especially encouraging because statin users were more likely to have illnesses prior to TBI, such as history of stroke, and can be viewed as having higher risk of mortality and secondary injury. However, even with a history of stroke and other CVD, these beneficiaries witnessed lower mortality and secondary injury events.

One study previously examined the association with statin discontinuation during hospitalization and found that statin discontinuation among pre-injury statin users was associated with an approximately four-fold higher mortality than patients that continued statin use during hospitalization.39 However, this relationship was not significant

(p=0.055), potentially due to their small sample size (n=61).39

This is also the first study to investigate the relationship between statin use following TBI and secondary injury, including stroke, depression, and ADRD. The results indicate that any statin use was associated with 9% lower risk of ischemic stroke,

25% lower risk of hemorrhagic stroke, and 14% lower risk of any stroke, controlling for pre-TBI comorbidity. Atorvastatin and simvastatin were the only statins associated with lower risk of both ischemic and hemorrhagic stroke. One study among older Medicare

94 beneficiaries compared rates of ischemic and hemorrhagic prior to and following TBI and found that rate of ischemic stroke following TBI was 1.3 times greater than pre-TBI and the rate of hemorrhagic stroke following TBI was 6.5 times greater than pre-TBI.30 The greater increase in the rate of hemorrhagic stroke make the findings from this study especially encouraging as statin use is associated with significantly lower risk of hemorrhagic stroke.

All statins except fluvastatin were associated with lower incident depression and

ADRD following TBI. These statins were all associated at least 15% lower incident depression and at least 17% lower incident ADRD. One study examined the rates of depression among older Medicare beneficiaries with TBI and found the rates of depression almost doubled following TBI.60 Furthermore, moderate and severe TBI also has been linked to a 2.3 and 4.5 times increase in the risk of ADRD, respectively, as TBI can fasten the natural cognitive decline among older adults.143,144 The increase in the risk of depression and ADRD following TBI is partially due to chronic neuroinflammation, which is a common physiological consequence of TBI.10,13,32

While this is the first study to investigate the relationship between statin use following TBI and post-TBI hospitalization discharge mortality and secondary injury, it is not without limitations. Primarily, the mortality observed in this study is all-cause mortality rather than TBI-specific mortality. The CCW data includes a date of death, which indicates if a beneficiary died during the study period, but does not indicate cause of death. However, we examined multiple mortality models in which the relationship between 30-day, 60-day, and 90-day mortality and statin use was assessed. It is possible that 30-day mortality is more likely to be TBI-specific than other mortality models and it

95 is in this model that beneficiaries witnessed the greatest difference in mortality.

Additionally, it is possible that beneficiaries discontinued statin use during end-of-life care and died a few months following discontinuation. This analysis classifies such beneficiaries as non-users at the time of death, potentially artificially inflating the impact of statin use on mortality. Similarly, statins could have only been prescribed to less severe TBI patients with a greater anticipated life expectancy, as compared to more severe patients with a decreased life expectancy. This healthy user effect may have biased the results as healthy statin users were potentially compared to less healthy non-users.145

Lastly, potentially life-saving medications given during hospitalization is not available in the data, and could not be accounted for in the analysis.

Another limitation is that depression is episodic as beneficiaries may experience relapses from remission, therefore they may be misclassified as depressed or non- depressed.146 To counter this limitation, in the examination of depression as an outcome, beneficiaries with a prior CCW diagnosis of depression were excluded. This helped in creating a cleaner cohort in which beneficiaries did not have prevalent depression, and therefore mitigating possible misclassification. Similarly, in the examination of ADRD as an outcome, beneficiaries with prior ADRD were excluded from the analysis. However, it was not possible to exclude beneficiaries with undiagnosed depression and ADRD prior to TBI.

There are multiple strengths to this study. The longitudinal study design allowed for the assessment of a time-varying exposure, which mimics a “real-world” setting. This study also was able to differentiate the impact between statins and by statin lipophilicity.

Additionally, this study was able to incorporate several covariates by linking CCW and

AHRF data. Including sociodemographic, health, drug utilization, and geographical covariates helped reduce confounding. Furthermore, sensitivity analysis with IPTW was incorporated to reduce selection bias.

While this study had methodological strengths, the most salient strength of this study is that it provides precious information regarding statin use and outcomes following

TBI. No prior study has investigated the relationship between statin use following TBI and mortality and secondary injury. Previous animal studies have examined secondary endpoints such as neuroinflammation and other biomarkers and have found that statin use following TBI decreases neuroinflammation and improves neurological function following injury.42,44-46,84-88 It should be noted that some previous human clinical trials have failed after encouraging results from animal studies, potentially due to inadequate biomarkers.147 However, this translational retrospective study sets the platform for human clinical trial by examining primary endpoints. Unlike previous studies human clinical trials examining pharmacologic treatments for TBI following encouraging animal research, future clinical trials investigating the impact of statins on TBI patients can utilize the information gained from this large observational study.

This large-scale study of older Medicare beneficiaries highlights the potential of statins to treat primary and secondary injury associated with TBI. The study found that atorvastatin and simvastatin are not only the most commonly used statins, but are also the only two statins that were associated with a significant risk lowering in all outcomes observed. Findings provide valuable information for clinicians treating older adults with

TBI. As this study finds, the majority of non-statin users had also had hyperlipidemia and clinicians can encourage, when appropriate, statin use among TBI patients with and

97 without hyperlipidemia. While this study provides a foundation, more research is needed to investigate primary and secondary injury associated with TBI in order to corroborate these results.

CHAPTER IX: SUMMARY, IMPLICATIONS, & CONCLUSION

Summary

TBI is a major health concern for Americans, especially older adults.1 More than half of TBIs in the US are in adults 50 or older, and TBI causes over 80,000 hospitalizations among adults 65 and older.4,50 TBI is associated with primary injury, the initial insult, and secondary injury, including stroke, depression and ADRD.10,12,13

Currently, care is focused on primary injury (aimed to reduce inflammation, primarily with corticosteroids) followed by symptomatic treatment of secondary injury.34,35

Unfortunately, there is no pharmacotherapy to treat both primary and secondary injury associated with TBI.54,55 Statins have been highlighted as a potential multifunctional medication for the treatment of primary and secondary injury associated with TBI.8-10

These investigations of statins as a pharmacotherapeutic option provide valuable information to healthcare professionals providing care to TBI patients. This series of studies conducted found nearly 40% of study beneficiaries take statins prior to and following TBI, namely simvastatin and atorvastatin. The studies also found that statin use, particularly rosuvastatin, simvastatin, and atorvastatin use, in the one to two months immediately prior to TBI was associated with decreased in-hospital mortality following

TBI. Lastly, these studies found that any statin use following TBI-hospitalization discharge was associated with decreased all-cause mortality following TBI, and simvastatin and atorvastatin use following TBI also were associated with decreased risk of stroke, depression, and ADRD.

Aim 1: Statin use in older Medicare beneficiaries with TBI

The results from Aim 1 fill a gap in the literature regarding statin use in older adults with TBI. Prior to this study, only four observational studies assessed statin use prior to TBI, of which one also examined statin continuation during TBI hospitalization.20,21,39,90 Additionally, no study assessed statin use following TBI.

Furthermore, no study differentiated between statins, nor did they assess duration of statin use. Thus, this study provides important information regarding statin use both prior to and following TBI as it examined statin duration, individually and as a class, prior to and following TBI. Additionally, this study examined statin dose changes prior to and following TBI.

The Aim 1 study found that 35% to 40% of beneficiaries used a statin per month prior to or following TBI. Among users, simvastatin and atorvastatin were the most used statins, while fluvastatin was the least used statin. More than 76% of beneficiaries had hyperlipidemia, of which 38% did not use statins at any time during the study period. The study also found that the average statin dose remained stable throughout the study period.

Among statin users, the average duration of use was 12.4 months and 38% of users had at least a one month gap in use. In addition to examining statin use, this study also assessed factors associated with overall statin use, pre-TBI statin use, and post-TBI statin use.

Beneficiaries with CVD, namely hyperlipidemia, were more likely to use statins both pre- and post-TBI. By characterizing statin use and factors associated with statin use pre-and post-TBI, results from Aim 1 provide essential contextual information regarding statin use among older Medicare beneficiaries with TBI.

100

Aim 2: In-hospital mortality following TBI in older Medicare statin users

Results from Aim 2 provide the first large-scale findings regarding the impact of statins, as a class and individually, on in-hospital mortality. This study found that 75% of the study sample had hyperlipidemia at TBI hospitalization; however, of these beneficiaries, only 53% of beneficiaries used statins prior to TBI. Approximately 90% of statin users used statins in either the first or second month immediately prior to TBI and, similar to Aim 1 findings, simvastatin was the most commonly used statin, followed by atorvastatin. Fluvastatin was the least used statin with only 535 beneficiaries using the drug in the six months prior to TBI. While simvastatin and atorvastatin were the most commonly used statins, these statins also were associated with the second greatest and third greatest reductions, respectively, in in-hospital mortality following TBI.

Rosuvastatin was associated with the greatest reduction in in-hospital mortality.

Additionally, when examining total use of statins in the months prior to TBI, five to six months of use of rosuvastatin, simvastatin, and atorvastatin were associated with significant reductions in in-hospital mortality.

The results of this study provide valuable information regarding which statins have the greatest impact in reducing in-hospital mortality. These results suggest that clinicians can recommend, when appropriate, the use of rosuvastatin, simvastatin, and atorvastatin to older adults at high risk for TBI. For instance, since falls are the leading cause of TBI among older adults, clinicians can recommend these statins to patients with a history or risk of falls.120

101

Aim 3: Mortality and secondary injury following TBI in older Medicare statin users

Results from Aim 3 build upon Aim 2 and deliver information regarding the impact of statin use following TBI on mortality and secondary injury. No previous study has investigated long-term statin use following TBI nor its impact on mortality and secondary injury in the months and years following TBI. The study found that 41% of beneficiaries used statins at any time following TBI and similar to Aims 1 and 2, simvastatin and atorvastatin were the most commonly used statins. However, 38% of beneficiaries with hyperlipidemia did not use statins at any time during the study period.

Results indicated that all statins were significantly associated with a decreased mortality, with the greatest reductions witnessed in 30-day mortality. In addition to being the most commonly used statins, both simvastatin and atorvastatin were associated with a significant reduction in all secondary injury outcomes including stroke (ischemic, hemorrhagic, and overall), depression, and ADRD. These results provide crucial information regarding the beneficial impact of statins, namely simvastatin and atorvastatin, following TBI. Clinicians treating older TBI patients can look to statins as a potential pharmacotherapy to help reduce mortality and secondary injury associated with

TBI. However, as stated in Chapter VIII, these results should be interpreted with caution as the mortality investigated is all-cause mortality, rather than TBI-specific mortality.

Furthermore, while statin use was associated with fewer ADRD and depression events, the causality of these associations is unclear due to unknown magnitude of undiagnosed

ADRD and depression prior to and following TBI. Thus, there are unmeasured confounders that may bias the results regarding secondary injury events.

102

These series of investigations into statin use and its impact on primary and secondary injury give insight on the potential of statin use and its beneficial impact.

These studies show that while almost three-fourths of the study population has hyperlipidemia, only approximately 40% of beneficiaries take statins every month.

However, the beneficiaries that used statins were associated with increased survival and decreased stroke, depression, and ADRD.

Strengths and limitations

Secondary data analysis for a pharmacoepidemiology study has several advantages, the most important being the collection of reimbursed prescription data without the pitfalls of recall bias. This large dataset allowed for the assessment of statin utilization prior to and following TBI and the examination of the relationships between statin use and mortality and secondary injury following TBI.

A major strength of this research is the differentiation of statins. Prior research did not differentiate the impact of different statins on mortality.20,21,39,90 This is first research to examine statin use, individually and as a class, and its impact on primary and secondary injury. Furthermore, this research is first to measure statin dose changes among older TBI patients and to examine statin use by lipophilicity status. Additionally, the longitudinal nature of the data allowed for the assessment of monthly statin use.

Another strength of this current research is the inclusion of AHRF data to supplement Medicare data. By including AHRF data, it was possible to assess geographical socioeconomic, and healthcare facility and profession characteristics in relation to primary and secondary injury. Furthermore, by including multiple covariates

103 and incorporating IPTW, this research was able to decrease confounding and selection bias associated with non-random assignment of treatment (statin use).

While this research highlights the potential of statins in treating TBI, there are several limitations to address. First, TBI severity is widely measured through varying scales such as the ISS and the AIS, considered the gold-standards in measuring injury severity.107 Unfortunately, Medicare administrative data does not provide this information. However, several proxy injury severity measures have been developed to determine injury severity using ICD-9 codes available in Medicare administrative data108

These studies utilize the ICISS to measure TBI severity. When comparing the ICISS to the AIS and the ISS, the ICISS injury severity measure performed better when predicting mortality following injury.107 The ICISS has been used to determine injury severity in a study of older Medicare patients with injury, including TBI.136 In addition to using the

ICISS, LOS following TBI was used a proxy measure of TBI severity as it may be indicative of TBI severity.19

In addition to not including injury severity scales, the data also may be biased as only patients who are hospitalized with TBI are included. This may exclude patients with less severe TBIs who are not hospitalized. Therefore, the results of this study indicating the benefits of statins use to decrease mortality and secondary injury events may only apply to TBI patients severe enough to be hospitalized. While this data limitation cannot be overcome, sensitivity analyses were conducted by excluding beneficiaries who died within 24 hours of TBI-hospitalization, an indication of severe TBI.20,21

Another limitation of the research is that the data may not accurately capture

ADRD and depression. First, ADRD and depression may develop several years following

104

TBI and may not be captured in a dataset with limited observation time following injury.

Second, the data does not capture undiagnosed ADRD and depression. Thus, the incidence of ADRD and depression may be underestimated because of limited observation time following injury and due to the cases of undiagnosed ADRD and depression.

Lastly, the research findings are generalizable to Medicare fee-for-service beneficiaries who are enrolled in Medicare Part D. Due to data limitations, the findings are not generalizable to beneficiaries in Medicare Advantage plans. Lastly, medications taken during TBI-hospitalizations are not captured in the data. This is because medication use during hospitalization is enclosed inpatient coverage (Medicare Part A), rather than prescription drug coverage (Medicare Part D).93 Regardless of the limitations, these studies address a significant health problem and inform future research and pharmacological management of older adults with TBI.

Implications

Clinical

Findings from this research have significant clinical and societal implications.

Foremost, this study has striking clinical implications. This research indicates that a majority of older adults have hyperlipidemia; however, a substantial amount of such beneficiaries do not take statins. Findings from Aims 1, 2, and 3 show that 38%, 47%, and 38%, respectively, of beneficiaries that have hyperlipidemia do not take statins either prior to or following TBI. There may be several reasons why beneficiaries with

105 hyperlipidemia did not take statins. First, beneficiaries often discontinue statins following treatment initiation.148 One study among adults 65 and older found that statin therapy declined significantly six months following initiation.148 Therefore, it is possible that

Medicare beneficiaries in this research discontinued statin therapy prior to their observation start time, while still retaining their CCW flag for hyperlipidemia. Second, statin use is associated with adverse events including myopathy, which may prevent beneficiaries with hyperlipidemia to use statins.77 Lastly, beneficiaries may not receive statins due to their poor health and limited life expectancy after suffering a TBI. Statin therapy is recommended for older adults with at least one year of life-expectancy.145

Similarly, physicians may discontinue statins for end-of-life care. Consequently, it is possible that beneficiaries with hyperlipidemia may not receive statins due to their limited life-expectancy following TBI. Therefore, when appropriate for the patients, clinicians can look prescribe statins to all beneficiaries to potentially reduce the risk of in- hospital mortality in the event of a TBI. Furthermore, clinicians can look to prescribe statins to all TBI-patients following TBI to potentially lower secondary injury rates following TBI.

Findings from Aim 2 suggest that pre-TBI statin use is associated with decreased in-hospital mortality. Therefore, clinicians can proactively prescribe statins to older adults that are at a high risk of TBIs. Most TBIs among older adults are caused by falls.4

Therefore, clinicians can prescribe statins to older adults with a history of falls.

Furthermore, statins have a positive safety profile as there are a few adverse events with low risks associated with statin use.75-78 Given their protective effect against mortality following TBI and favorable safety profile, clinicians can, when appropriate, prescribe

106 statins to all older adults. Additionally, clinicians can administer statins during hospitalization to all TBI patients. Several animal studies have shown that statin administration in the hours following injury may suppress inflammation.42,44-46,85

Findings from Aim 3 suggest that statin use following TBI is associated with decrease all-cause mortality following hospitalization discharge, while also being associated with a reduction in stroke, depression, and ADRD. Therefore, clinicians can prescribe statins to older adults immediately following TBI. This may help reduce their mortality following discharge. Furthermore, such beneficiaries may be at a lower risk to experience secondary injury events.

The clinical implications of these studies are noteworthy, and healthcare professionals can look to implement these findings into their practices. This research helps define pharmacological guidelines for the treatment of TBI and improve health outcomes and QOL following TBI.

Research

These studies have paved the way for future research. While this study has shown the effectiveness of statin use in a large, real-world sample of older adults, the next step in would be to conduct a RCT or a prospective cohort study examining the relationship of statin use and primary and secondary injury. Currently, a single RCT has examined statin use.91 This study only examined the impact of rosuvastatin on inflammatory cytokines following TBI.91 Future RCTs or prospective cohort studies can assess the impact of all statins on primary and secondary injury. Future studies can also investigate the impact of different statin doses.

107

Future research can also address the limitations of the current studies. Due to data limitations, this study could not examine statin use during hospitalization; therefore, future studies can assess the impact of statin use during hospitalization. Additionally, due to data limitations, the findings are of this research are only generalizable to Medicare fee-for-service beneficiaries and not beneficiaries who are enrolled Medicare Advantage plans. Thus, future research can include persons with private health insurance.

In addition to addressing the limitations of this research, future studies can also build upon this research by exploring the impact of statins on veterans and younger populations. Furthermore, research can focus on the impact of statins on persons with multiple TBIs.

Societal

Societal implications from this study include improved QOL and the decrease of costs associated with TBI. Older adults experience poor QOL following TBI, potentially due to emotional and cognitive deficits caused by TBI.149,150 TBI can cause stroke, depression, and ADRD, all of which can decrease QOL. 13,30,60,149,151 As the evidence suggests, statin use can prevent secondary injury outcomes; therefore, potentially help improve QOL following TBI. Furthermore, it may be possible that statin use following injury can delay the onset of depression and ADRD, which can also improve QOL following TBI.

In addition to reducing QOL, TBI can cost patients thousands of dollars.152 TBI hospitalization among patients 65 and older can cost over $20,000.153 Statin use prior to

TBI is associated decreased in-hospital mortality. It is also possible that statin use may be

108 associated with other favorable TBI-hospitalization outcomes, such as a decreased LOS.

Therefore, if given prior to TBI, statin use can potentially reduce hospitalization costs by decreasing LOS.

The findings may help reduce secondary injury events, and thus decrease costs associated with secondary injuries. Care for stroke cost more than $18 billion in the US during 2008.154 The mean cost per stroke patient in the first year after stroke is over

$11,000 while lifetime costs can be over $100,000.155 Depression among older adults also is associated with increased healthcare costs, especially outpatient costs.156-158

Furthermore, general health services costs can be up to 50% higher for depressed older adults than non-depressed older adults.156,157 Lastly, the economic burden of ADRD also is significant.133 Among Medicare beneficiaries, cost of care for ADRD is over $100 billion, with an average annual healthcare costs of $22,206 per beneficiary 65 and older.133 Additionally, ADRD is associated with caregiver costs, with direct and indirect caregiver costs being over $10 billion in 2015.133 With treatment of TBI with statins following injury, these costs can significantly decrease given the reductions of stroke, depression, and ADRD associated with statin use following TBI.

Conclusion

Government and academic institutions have highlighted the need to explore pharmacological management of TBI and have suggested researching statins as a potential pharmacological agent.55,122 This research follows such recommendations and findings show simvastatin, atorvastatin and rosuvastatin use are associated with a

109 decrease in mortality and secondary injury. This research addresses a significant health problem and sets a platform for clinical practice and future research on the pharmacological management of all populations with a TBI.

Given that thousands of older adults are hospitalized with TBI every year and that

TBI is associated with increased mortality, these investigations into statin use have salient implications for older adults suffering from TBI.4,22 Almost a quarter of adults 65 and older die within one year of TBI, while almost half of adults 85 and older die within one year of TBI.22 Therefore, the evidence generated can help save lives if patients preemptively take statins to decrease in-hospital mortality and continue to take statins following TBI to decrease all-cause mortality following hospitalization discharge.

Furthermore, patients can take statins following TBI to reduce their risks of secondary injury including stroke, depression, and ADRD.

This research sought to assess the impact of statins in older TBI patients and found that statin use is protective against primary and secondary injury. Implementation of this research into clinical practice can help in accomplishing the research’s long-term goal of improving the health outcomes and lives of all populations suffering from a TBI.

110

CHAPTER II APPENDIX

Table A.2.1: Animal studies assessing the impact of statins on TBI

111

CHAPTER VI APPENDICES

Table A.6.1: List of CCW chronic conditions excluding CVD conditions and ADRD

112

Table A.6.2: Odds ratios (95% CI) of factors associated with statin users only prior to

TBI among older Medicare Beneficiaries (n=1,969)

a Other races include Native American, other and unknown race b Count of CCW chronic conditions excluding cardiovascular conditions

113

Table A.6.3: Odds ratios (95% CI) of factors associated with statin users only following

TBI among older Medicare Beneficiaries (n=5,527)

a Other races include Native American, other and unknown race b Count of CCW chronic conditions excluding cardiovascular conditions

114

CHAPTER VII APPENDICES

Table A.7.1: Correlation matrix between AHRF variables, Aim 2

115

Table A.7.2: Beneficiaries with monthly statin use prior to TBI (n=112,109)

116

Table A.7.3: Frequency of atorvastatin, rosuvastatin, and simvastatin doses per month, two months prior to TBI

a Beneficiaries only had one dose per month. Dose for beneficiaries that changed doses within a month were averaged to get a single dose.

117

Table A.7.4: Odds ratios (95% CI) of in-hospital mortality following TBI comparing atorvastatin, rosuvastatin, and simvastatin doses two months prior to TBI

a Adjusted for: demographic characteristics (age, sex, race, income, region); medications (antiplatelets, anticoagulants, beta blockers, other statins); cardiovascular disease (CHF, AMI, IHD, VHD, hyperlipidemia, hypertension, prior stroke); diabetes; Alzheimer’s disease and related dementia; count of CCW chronic conditions excluding cardiovascular conditions (hypertension, hyperlipidemia, ischemic heart disease, congestive heart failure, acute myocardial infarction, valvular heard disease) and diabetes and ADRD; SNF stay in month prior to TBI; injury severity (SRRi, length of hospital stay); provider and hospital characteristics per 100,000 population (physicians, hospitals, hospitals with trauma serv ices, hospitals with neurological services, beds)

118

Table A.7.5: Odds ratios (95% CI) of in-hospital mortality following TBI comparing statin users and non-users among beneficiaries that died within 24 hours of hospitalization (n=110,970)

119

120

121

122

CHAPTER VIII APPENDICES

Table A.8.1: Correlation matrix between AHRF variables, Aim 3

123

Table A.8.2: Relative risks (95% CI) of Mortality following TBI among Medicare beneficiaries, comparing statin use to non-use, IPTW adjusted

a Adjusted for sociodemographic characteristics (age, race, sex, county-level income, LIS status, region); CVD (CHF, AMI, IHD, valvular heart disease, hyperlipidemia, hypertension); diabetes; ADRD; count of CCW chronic conditions excluding CVD and ADRD and diabetes; injury severity (LOS, discharge status, SRRi); medication use in month (anticoagulant use, antiplatelet use, beta-blocker); healthcare provider characteristics per 100,000 population (physicians, hospitals, hospitals with trauma services, hospitals with neurological services, beds); pre-TBI statin use; IPTW b N/A—estimate is too small and unreliable c Lipophilic statins: atorvastatin, fluvastatin, lovastatin, and simvastatin; hydrophilic statins: pravastatin and rosuvastatin

124

Table A.8.3: Relative risks (95% CI) of secondary injury following TBI among Medicare beneficiaries, comparing statin use to non-use, IPTW adjusted

125

126

REFERENCES

1. Coronado VG, McGuire LC, Sarmiento K, et al. Trends in traumatic brain injury

in the US and the public health response: 1995–2009. Journal of safety research.

2012;43(4):299-307.

2. Coronado VG, Thomas KE, Sattin RW, Johnson RL. The CDC traumatic brain

injury surveillance system: characteristics of persons aged 65 years and older

hospitalized with a TBI. The Journal of head trauma rehabilitation.

2005;20(3):215-228.

3. Coronado VG, Xu L, Basavaraju SV, et al. Surveillance for traumatic brain

injury-related deaths: United States, 1997-2007: US Department of Health and

Human Services, Centers for Disease Control and Prevention Atlanta; 2011.

4. Thompson HJ, McCormick WC, Kagan SH. Traumatic brain injury in older

adults: epidemiology, outcomes, and future implications. Journal of the American

Geriatrics Society. 2006;54(10):1590-1595.

5. Thompson HJ, Dikmen S, Temkin N. Prevalence of comorbidity and its

association with traumatic brain injury and outcomes in older adults. Research in

gerontological nursing. 2012;5(1):17.

6. Ventura T, Harrison-Felix C, Carlson N, et al. Mortality after discharge from

acute care hospitalization with traumatic brain injury: a population-based study.

Archives of physical medicine and rehabilitation. 2010;91(1):20-29.

7. Wible EF, Laskowitz DT. Statins in traumatic brain injury. Neurotherapeutics.

2010;7(1):62-73.

127

8. Radosevich JJ, Patanwala AE, Erstad BL. Emerging pharmacological agents to

improve survival from traumatic brain injury. Brain Injury. 2013;27(13-14):1492-

1499.

9. Loane DJ, Faden AI. Neuroprotection for traumatic brain injury: translational

challenges and emerging therapeutic strategies. Trends in pharmacological

sciences. 2010;31(12):596-604.

10. Kumar A, Loane DJ. Neuroinflammation after traumatic brain injury:

opportunities for therapeutic intervention. Brain, behavior, and immunity.

2012;26(8):1191-1201.

11. Jansen JO, Lord JM, Thickett DR, Midwinter MJ, McAuley DF, Gao F. Clinical

review: Statins and trauma-a systematic review. Critical Care. 2013;17(3):1-10.

12. Griffen J, Hanks R. Cognitive and Behavioral Outcomes from Traumatic Brain

Injury. Handbook on the Neuropsychology of Traumatic Brain Injury: Springer;

2014:25-45.

13. Sivanandam TM, Thakur MK. Traumatic brain injury: a risk factor for

Alzheimer's disease. Neuroscience & Biobehavioral Reviews. 2012;36(5):1376-

1381.

14. Kabadi SV, Faden AI. Neuroprotective strategies for traumatic brain injury:

improving clinical translation. International journal of molecular sciences.

2014;15(1):1216-1236.

15. Rosenfeld JV, Maas AI, Bragge P, Morganti-Kossmann MC, Manley GT, Gruen

RL. Early management of severe traumatic brain injury. The Lancet.

2012;380(9847):1088-1098.

128

16. Tran LV. Understanding the pathophysiology of traumatic brain injury and the

mechanisms of action of neuroprotective interventions. Journal of Trauma

Nursing. 2014;21(1):30-35.

17. Vink R, Nimmo AJ. Multifunctional drugs for head injury. Neurotherapeutics.

2009;6(1):28-42.

18. Andersen RM. Revisiting the behavioral model and access to medical care: does it

matter? Journal of health and social behavior. 1995:1-10.

19. Albrecht JS, Liu X, Baumgarten M, et al. Benefits and Risks of Anticoagulation

Resumption Following Traumatic Brain Injury. JAMA internal medicine.

2014;174(8):1244-1251.

20. Schneider EB, Efron DT, MacKenzie EJ, Rivara FP, Nathens AB, Jurkovich GJ.

Premorbid statin use is associated with improved survival and functional

outcomes in older head-injured individuals. Journal of Trauma and Acute Care

Surgery. 2011;71(4):815-819.

21. Efron DT, Sorock G, Haut ER, et al. Preinjury statin use is associated with

improved in-hospital survival in elderly trauma patients. Journal of Trauma and

Acute Care Surgery. 2008;64(1):66-74.

22. Donohue JT, Clark DE, DeLorenzo MA. Long-term survival of Medicare patients

with head injury. Journal of Trauma-Injury, Infection, and Critical Care.

2007;62(2):419-423.

23. Royo N, Shimizu S, Schouten J, Stover J, McIntosh T. Pharmacology of traumatic

brain injury. Current opinion in pharmacology. 2003;3(1):27-32.

129

24. Greve MW, Zink BJ. Pathophysiology of traumatic brain injury. Mount Sinai

Journal of Medicine: A Journal of Translational and Personalized Medicine.

2009;76(2):97-104.

25. Werner C, Engelhard K. Pathophysiology of traumatic brain injury. British

journal of anaesthesia. 2007;99(1):4-9.

26. Chodobski A, Zink BJ, Szmydynger-Chodobska J. Blood–brain barrier

pathophysiology in traumatic brain injury. Translational stroke research.

2011;2(4):492-516.

27. Ziebell JM, Morganti-Kossmann MC. Involvement of pro-and anti-inflammatory

cytokines and chemokines in the pathophysiology of traumatic brain injury.

Neurotherapeutics. 2010;7(1):22-30.

28. Lorente L. New Prognostic Biomarkers in Patients With Traumatic Brain Injury.

Archives of Trauma Research. 2015;4(4).

29. Masel BE, DeWitt DS. Traumatic Brain Injury Disease: Long-term Consequences

of Traumatic Brain Injury. Understanding Traumatic Brain Injury: Current

Research and Future Directions. 2014:28.

30. Albrecht JS, Liu X, Smith GS, et al. Stroke Incidence Following Traumatic Brain

Injury in Older Adults. The Journal of head trauma rehabilitation. 2014.

31. Goldstein FC, Levin HS. Assessment and Treatment of Older Adults with

Traumatic Brain Injuries. Handbook on the Neuropsychology of Traumatic Brain

Injury: Springer; 2014:331-343.

32. Hurley LL, Tizabi Y. Neuroinflammation, neurodegeneration, and depression.

Neurotoxicity research. 2013;23(2):131-144.

130

33. Sherer M, Struchen MA, Yablon SA, Wang Y, Nick TG. Comparison of indices

of traumatic brain injury severity: Glasgow Coma Scale, length of coma and post-

traumatic amnesia. Journal of Neurology, Neurosurgery & Psychiatry.

2008;79(6):678-685.

34. Comper P, Bisschop S, Carnide N, Tricco A. A systematic review of treatments

for mild traumatic brain injury. Brain Injury. 2005;19(11):863-880.

35. Foundation BT. Guidelines for the Management of Severe Traumatic Brain

Injury. Journal of neurotrauma. 2007;24(1).

36. Beauchamp K, Mutlak H, Smith WR, Shohami E, Stahel PF. Pharmacology of

traumatic brain injury: where is the “golden bullet”? Molecular Medicine.

2008;14(11-12):731.

37. Zangbar B, Pandit V, Rhee P, et al. Clinical outcomes in patients on preinjury

ibuprofen with traumatic brain injury. The American Journal of Surgery. 2014.

38. Papa L, Mendes ME, Braga CF. Mild traumatic brain injury among the geriatric

population. Current translational geriatrics and experimental gerontology

reports. 2012;1(3):135-142.

39. Orlando A, Bar-Or D, Salottolo K, et al. Unintentional Discontinuation of Statins

May Increase Mortality After Traumatic Brain Injury in Elderly Patients: A

Preliminary Observation. Journal of clinical medicine research. 2013;5(3):168.

40. Wood WG, Eckert GP, Igbavboa U, Müller WE. Statins and neuroprotection.

Annals of the New York Academy of Sciences. 2010;1199(1):69-76.

41. Wood WG, Mΰller WE, Eckert GP. Statins and Neuroprotection: Basic

Pharmacology Needed. Molecular neurobiology. 2014:1-7.

131

42. Beziaud T, Chen XR, El Shafey N, et al. Simvastatin in traumatic brain injury:

effect on brain edema mechanisms. Critical care medicine. 2011;39(10):2300-

2307.

43. Belur PK, Chang JJ, He S, Emanuel BA, Mack WJ. Emerging experimental

therapies for intracerebral hemorrhage: targeting mechanisms of secondary brain

injury. Neurosurgical focus. 2013;34(5):E9.

44. Indraswari F, Wang H, Lei B, et al. Statins improve outcome in murine models of

intracranial hemorrhage and traumatic brain injury: a translational approach.

Journal of neurotrauma. 2012;29(7):1388-1400.

45. Turkoglu OF, Eroglu H, Okutan O, et al. Atorvastatin efficiency after traumatic

brain injury in rats. Surgical neurology. 2009;72(2):146-152.

46. Wang H, Lynch JR, Song P, et al. Simvastatin and atorvastatin improve

behavioral outcome, reduce hippocampal degeneration, and improve cerebral

blood flow after experimental traumatic brain injury. Experimental neurology.

2007;206(1):59-69.

47. Langlois JA, Rutland-Brown W, Wald MM. The epidemiology and impact of

traumatic brain injury: a brief overview. The Journal of head trauma

rehabilitation. 2006;21(5):375-378.

48. Prevention CfDCa. Rates of TBI-related Emergency Department Visits,

Hospitalizations, and Deaths - United States, 2001-2010. 2014;

http://www.cdc.gov/traumaticbraininjury/data/rates.html. Accessed December 15,

2014.

132

49. Prevention CfDCa. Rates of TBI-related Emergency Department Visits,

Hospitalizations, and Deaths by Sex - United States, 2001-2010. 2014;

http://www.cdc.gov/traumaticbraininjury/data/rates_bysex.html. Accessed

December 15, 2014.

50. Cuthbert JP, Harrison-Felix C, Corrigan JD, et al. Epidemiology of adults

receiving acute inpatient rehabilitation for a primary diagnosis of traumatic brain

injury in the United States. Journal of Head Trauma Rehabilitation. 2014.

51. Harrison-Felix CL, Whiteneck GG, Jha A, DeVivo MJ, Hammond FM, Hart DM.

Mortality over four decades after traumatic brain injury rehabilitation: a

retrospective cohort study. Archives of physical medicine and rehabilitation.

2009;90(9):1506-1513.

52. Harrison-Felix CL, Pretz CR, Hammond F, et al. Life Expectancy following

Inpatient Rehabilitation for Traumatic Brain Injury in the United States. Journal

of neurotrauma. 2014(ja).

53. Harrison-Felix C, Kolakowsky-Hayner SA, Hammond FM, et al. Mortality after

surviving traumatic brain injury: risks based on age groups. The Journal of head

trauma rehabilitation. 2012;27(6):E45-E56.

54. Finfer SR, Cohen J. Severe traumatic brain injury. Resuscitation. 2001;48(1):77-

90.

55. Diaz-Arrastia R, Kochanek PM, Bergold P, et al. Pharmacotherapy of Traumatic

Brain Injury: State of the Science and the Road Forward: Report of the

Department of Defense Neurotrauma Pharmacology Workgroup. Journal of

Neurotrauma. 2014;31(2):135-158.

133

56. Alderson P, Roberts I. Corticosteroids in acute traumatic brain injury: systematic

review of randomised controlled trials. Bmj. 1997;314(7098):1855.

57. Roberts I, Yates D, Sandercock P, et al. Effect of intravenous corticosteroids on

death within 14 days in 10008 adults with clinically significant head injury (MRC

CRASH trial): randomised placebo-controlled trial. Lancet.

2004;364(9442):1321-1328.

58. Moppett I. Traumatic brain injury: assessment, resuscitation and early

management. British Journal of Anaesthesia. 2007;99(1):18-31.

59. Sellers ZP, Williams RA, Overbay JW, Cho J, Henderson M, Reed TT. Current

Therapeutic Modalities, Enzyme Kinetics, and Redox Proteomics in Traumatic

Brain Injury. 2014.

60. Albrecht JS, Kiptanui Z, Tsang Y, et al. Depression among older adults after

traumatic brain injury: a national analysis. The American Journal of Geriatric

Psychiatry. 2014.

61. Schroeppel TJ, Sharpe JP, Magnotti LJ, et al. Traumatic brain injury and β-

blockers: Not all drugs are created equal. Journal of Trauma and Acute Care

Surgery. 2014;76(2):504-509.

62. Mak CH, Wong SK, Wong GK, et al. Traumatic brain injury in the elderly: Is it as

bad as we think? Current translational geriatrics and experimental gerontology

reports. 2012;1(3):171-178.

63. Prathep S, Sharma D, Hallman M, et al. Preliminary Report on Cardiac

Dysfunction After Isolated Traumatic Brain Injury*. Critical care medicine.

2014;42(1):142-147.

134

64. Wilberger J, Jho DJ. Geriatric Neurotrauma. Geriatric Trauma and Critical Care:

Springer; 2014:237-248.

65. Xiong Y, Mahmood A, Chopp M. Emerging treatments for traumatic brain injury.

2009.

66. Zhang Y, Xiong Y, Mahmood A, Zhang ZG, Chopp M. Angiogenesis and

Functional Recovery After Traumatic Brain Injury. Vascular Mechanisms in CNS

Trauma: Springer; 2014:141-156.

67. Mills EJ, Rachlis B, Wu P, Devereaux PJ, Arora P, Perri D. Primary Prevention of

Cardiovascular Mortality and Events With Statin TreatmentsA Network Meta-

Analysis Involving More Than 65,000 Patients. Journal of the American College

of Cardiology. 2008;52(22):1769-1781.

68. Taylor F, Huffman MD, Macedo AF, et al. Statins for the primary prevention of

cardiovascular disease. Cochrane Database Syst Rev. 2013;1(1).

69. Brugts J, Yetgin T, Hoeks S, et al. The benefits of statins in people without

established cardiovascular disease but with cardiovascular risk factors: meta-

analysis of randomised controlled trials. Bmj. 2009;338.

70. Administration USFaD. Statins. 2014;

http://www.fda.gov/Drugs/DrugSafety/InformationbyDrugClass/ucm294358.htm.

Accessed May 10 2015.

71. Zhou Q, Liao JK. Statins and cardiovascular diseases: from cholesterol lowering

to pleiotropy. Current pharmaceutical design. 2009;15(5):467.

135

72. Chong PH, Seeger JD, Franklin C. Clinically relevant differences between the

statins: implications for therapeutic selection. The American journal of medicine.

2001;111(5):390-400.

73. Schachter M. Chemical, pharmacokinetic and pharmacodynamic properties of

statins: an update. Fundamental & clinical pharmacology. 2005;19(1):117-125.

74. Schuster H. Improving lipid management–to titrate, combine or switch.

International journal of clinical practice. 2004;58(7):689-694.

75. Finegold JA, Manisty CH, Goldacre B, Barron AJ, Francis DP. What proportion

of symptomatic side effects in patients taking statins are genuinely caused by the

drug? Systematic review of randomized placebo-controlled trials to aid individual

patient choice. European journal of preventive cardiology. 2014;21(4):464-474.

76. Sadhu JS, Rich MW. Efficacy and Safety of Statins in Older Adults. Current

Cardiovascular Risk Reports. 2012;6(5):372-379.

77. Desai CS, Martin SS, Blumenthal RS. Non-cardiovascular effects associated with

statins. Bmj. 2014;349:g3743.

78. Yan Y, Qiu B, Hu L, et al. Efficacy and safety evaluation of intensive statin

therapy in older patients with coronary heart disease: a systematic review and

meta-analysis. European journal of clinical pharmacology. 2013;69(12):2001-

2009.

79. O'Connor CM, Abraham WT, Albert NM, et al. Predictors of mortality after

discharge in patients hospitalized with heart failure: an analysis from the

Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with

Heart Failure (OPTIMIZE-HF). American heart journal. 2008;156(4):662-673.

136

80. Schneeweiss S, Patrick AR, Pedan A, et al. The effect of Medicare Part D

coverage on drug use and cost sharing among seniors without prior drug benefits.

Health Affairs. 2009;28(2):w305-w316.

81. Choudhry NK, Setoguchi S, Levin R, Winkelmayer WC, Shrank WH. Trends in

adherence to secondary prevention medications in elderly post‐myocardial

infarction patients. Pharmacoepidemiology and drug safety. 2008;17(12):1189-

1196.

82. Swindle JP, Potash J, Kulakodlu M, Kuznik A, Buikema A. Drug utilization

patterns and cardiovascular outcomes in elderly patients newly initiated on

atorvastatin or simvastatin. The American journal of geriatric pharmacotherapy.

2011;9(6):471-482.

83. Kulik A, Shrank WH, Levin R, Choudhry NK. Adherence to statin therapy in

elderly patients after hospitalization for coronary revascularization. The American

journal of cardiology. 2011;107(10):1409-1414.

84. Chen S-F, Hung T-H, Chen C-C, et al. Lovastatin improves histological and

functional outcomes and reduces inflammation after experimental traumatic brain

injury. Life sciences. 2007;81(4):288-298.

85. Chen G, Zhang S, Shi J, Ai J, Qi M, Hang C. Simvastatin reduces secondary brain

injury caused by cortical contusion in rats: possible involvement of TLR4/NF-κB

pathway. Experimental neurology. 2009;216(2):398-406.

86. Jungner M, Lundblad C, Bentzer P. Rosuvastatin in experimental brain trauma:

Improved capillary patency but no effect on edema or cerebral blood flow.

Microvascular research. 2013;88:48-55.

137

87. Mahmood A, Goussev A, Lu D, et al. Long-lasting benefits after treatment of

traumatic brain injury (TBI) in rats with combination therapy of marrow stromal

cells (MSCs) and simvastatin. Journal of Neurotrauma. 2008;25(12):1441-1447.

88. Mahmood A, Goussev A, Kazmi H, Qu C, Lu D, Chopp M. Long-term benefits

after treatment of traumatic brain injury with simvastatin in rats. Neurosurgery.

2009;65(1):187.

89. MacKenzie EJ, Rivara FP, Jurkovich GJ, et al. The national study on costs and

outcomes of trauma. Journal of Trauma-Injury, Infection, and Critical Care.

2007;63(6):S54-S67.

90. Neilson SJ, See AA, King NK. Effect of prior statin use on outcome after severe

traumatic brain injury in a South-East Asian population. Brain Injury. 2016:1-6.

91. Sánchez-Aguilar M, Tapia-Pérez JH, Sánchez-Rodríguez JJ, et al. Effect of

rosuvastatin on cytokines after traumatic head injury: Clinical article. Journal of

neurosurgery. 2013;118(3):669-675.

92. Helmy A, Timofeev I, Palmer CR, Gore A, Menon DK, Hutchinson PJ.

Hierarchical log linear analysis of admission blood parameters and clinical

outcome following traumatic brain injury. Acta neurochirurgica.

2010;152(6):953-957.

93. Medicare Administrative Data User Guide. 2014.

https://www.ccwdata.org/web/guest/user-documentation.

94. Medicare Part D Data User Guide. 2014.

https://www.ccwdata.org/web/guest/user-documentation.

138

95. Area Health Resources Files (AHRF) In: Services USDoHaH, ed. Rockville, MD:

Health Resources and Services Administration, Bureau of Health Workforce;

2016.

96. Stambler HV. The Area Resource File--a brief look. Public Health Reports.

1988;103(2):184.

97. Leibson CL, Brown AW, Ransom JE, et al. Incidence of traumatic brain injury

across the full disease spectrum: a population-based medical record review study.

Epidemiology (Cambridge, Mass.). 2011;22(6):836.

98. Faul M, Xu L, Wald M, Coronado VG. Traumatic brain injury in the United

States: emergency department visits, hospitalizations and deaths 2002–2006.

Atlanta, GA: Centers for Disease Control and Prevention, National Center for

Injury Prevention and Control. 2010:2-70.

99. Khokhar B, Simoni-Wastila L, Albrecht JS. Risk of Stroke Among Older

Medicare Antidepressant Users With Traumatic Brain Injury. The Journal of head

trauma rehabilitation. 2016.

100. Savarese G, Gotto AM, Paolillo S, et al. Benefits of statins in elderly subjects

without established cardiovascular disease: a meta-analysis. Journal of the

American College of Cardiology. 2013;62(22):2090-2099.

101. Stone NJ, Merz CNB, ScM F, et al. 2013 ACC/AHA Guideline on the Treatment

of Blood Cholesterol to Reduce Atherosclerotic Cardiovascular Risk in Adults.

Journal of the American College of Cardiology. 2013.

139

102. Bhattacharjee S, Findley PA, Sambamoorthi U. Understanding Gender

Differences in Statin Use among Elderly Medicare Beneficiaries. Drugs & aging.

2012;29(12):971-980.

103. Caetano PA, Lam J, Morgan SG. Toward a standard definition and measurement

of persistence with drug therapy: examples from research on statin and

antihypertensive utilization. Clinical therapeutics. 2006;28(9):1411-1424.

104. Steiner JF, Prochazka AV. The assessment of refill compliance using pharmacy

records: methods, validity, and applications. Journal of clinical epidemiology.

1997;50(1):105-116.

105. AlbrechtJennifer S, SmithGordon S, ZuckermanIlene H. Patterns of Depression

Treatment in Medicare Beneficiaries with Depression after Traumatic Brain

Injury. Journal of neurotrauma. 2015.

106. Moreno C, Laje G, Blanco C, Jiang H, Schmidt AB, Olfson M. National trends in

the outpatient diagnosis and treatment of bipolar disorder in youth. Archives of

general psychiatry. 2007;64(9):1032-1039.

107. Meredith JW, Evans G, Kilgo PD, et al. A comparison of the abilities of nine

scoring algorithms in predicting mortality. Journal of Trauma and Acute Care

Surgery. 2002;53(4):621-629.

108. Kim YJ. Injury severity scoring systems: a review of application to practice.

Nursing in critical care. 2012;17(3):138-150.

109. Osler T, Rutledge R, Deis J, Bedrick E. ICISS: an international classification of

disease-9 based injury severity score. Journal of Trauma and Acute Care Surgery.

1996;41(3):380-388.

140

110. Nakahara S, Yokota J. Revision of the International Classification of Diseases to

include standardized descriptions of multiple injuries and injury severity. Bulletin

of the World Health Organization. 2011;89(3):238-240.

111. Meredith JW, Kilgo PD, Osler TM. Independently derived survival risk ratios

yield better estimates of survival than traditional survival risk ratios when using

the ICISS. Journal of Trauma and Acute Care Surgery. 2003;55(5):933-938.

112. Glance LG, Osler TM, Mukamel DB, Meredith W, Wagner J, Dick AW. TMPM-

ICD9: a trauma mortality prediction model based on ICD-9-CM codes. Annals of

surgery. 2009;249(6):1032-1039.

113. Bewick V, Cheek L, Ball J. Statistics review 13: receiver operating characteristic

curves. Critical care. 2004;8(6):1.

114. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or

more correlated receiver operating characteristic curves: a nonparametric

approach. Biometrics. 1988:837-845.

115. Blaha MJ, Martin SS. How Do Statins Work? Changing Paradigms With

Implications for Statin Allocation∗. Journal of the American College of

Cardiology. 2013;62(25):2392-2394.

116. Loane D, Stoica B, Faden A. Neuroprotection for traumatic brain injury.

Handbook of clinical neurology. 2014;127:343-366.

117. Wee H-Y, Ho C-H, Liang F-W, et al. Increased Risk of New-Onset Depression in

Patients With Traumatic Brain Injury and Hyperlipidemia: The Important Role of

Statin Medications. The Journal of clinical psychiatry. 2016;77(4):505-511.

141

118. Mann D, Reynolds K, Smith D, Muntner P. Trends in statin use and low-density

lipoprotein cholesterol levels among US adults: impact of the 2001 National

Cholesterol Education Program guidelines. Annals of Pharmacotherapy.

2008;42(9):1208-1215.

119. Ma J, Sehgal NL, Ayanian JZ, Stafford RS. National trends in statin use by

coronary heart disease risk category. PLoS Med. 2005;2(5):e123.

120. Prevention CfDCa. TBI: Get the Facts. 2016;

http://www.cdc.gov/traumaticbraininjury/get_the_facts.html. Accessed May 2,

2016.

121. Prevention CfDCa. Rates of TBI-related Emergency Department Visits,

Hospitalizations, and Deaths by Age Group - United States, 2001-2010. 2014;

http://www.cdc.gov/traumaticbraininjury/data/rates_hosp_byage.html. Accessed

December 15, 2014.

122. Kochanek PM, Bramlett H, Dietrich WD, et al. A novel multicenter preclinical

drug screening and biomarker consortium for experimental traumatic brain injury:

operation brain trauma therapy. Journal of Trauma and Acute Care Surgery.

2011;71(1):S15-S24.

123. Association As. 2011 Alzheimer's disease facts and figures. Alzheimer's &

dementia: the journal of the Alzheimer's Association. 2011;7(2):208.

124. Newman AB, Fitzpatrick AL, Lopez O, et al. Dementia and Alzheimer's disease

incidence in relationship to cardiovascular disease in the Cardiovascular Health

Study cohort. Journal of the American Geriatrics Society. 2005;53(7):1101-1107.

125. Allison PD. Multiple regression: A primer: Pine Forge Press; 1999.

142

126. Petri H, Urquhart J. Channeling bias in the interpretation of drug effects. Statistics

in medicine. 1991;10(4):577-581.

127. Tawakol A, Fayad ZA, Mogg R, et al. Intensification of statin therapy results in a

rapid reduction in atherosclerotic inflammation: results of a multicenter

fluorodeoxyglucose-positron emission tomography/computed tomography

feasibility study. Journal of the American College of Cardiology.

2013;62(10):909-917.

128. Law MR, Wald NJ, Rudnicka A. Quantifying effect of statins on low density

lipoprotein cholesterol, ischaemic heart disease, and stroke: systematic review and

meta-analysis. Bmj. 2003;326(7404):1423.

129. Chen WT, Krishnan GM, Sood N, Kluger J, Coleman CI. Effect of statins on

atrial fibrillation after cardiac surgery: a duration-and dose-response meta-

analysis. The Journal of Thoracic and Cardiovascular Surgery. 2010;140(2):364-

372.

130. Sierra S, Ramos MC, Molina P, Esteo C, Vázquez JA, Burgos JS. Statins as

neuroprotectants: a comparative in vitro study of lipophilicity, blood-brain-barrier

penetration, lowering of brain cholesterol, and decrease of neuron cell death.

Journal of Alzheimer's Disease. 2011;23(2):307-318.

131. Li J-J, Li Y-S, Chu J-M, et al. Changes of plasma inflammatory markers after

withdrawal of statin therapy in patients with hyperlipidemia. Clinica Chimica

Acta. 2006;366(1):269-273.

132. Lai W-T, Lee K-T, Chu C-S, et al. Influence of withdrawal of statin treatment on

proinflammatory response and fibrinolytic activity in humans: an effect

143

independent on cholesterol elevation. International journal of cardiology.

2005;98(3):459-464.

133. Association As. 2016 Alzheimer's disease facts and figures. Alzheimer's &

Dementia. 2016;12(4):459-509.

134. Bhatnagar S, Iaccarino MA, Zafonte R. Pharmacotherapy in rehabilitation of post-

acute traumatic brain injury. Brain research. 2016;1640:164-179.

135. Kochanek PM, Bramlett HM, Dixon CE, et al. Approach to modeling, therapy

evaluation, drug selection, and biomarker assessments for a multicenter pre-

clinical drug screening consortium for acute therapies in severe traumatic brain

injury: operation brain trauma therapy. Journal of Neurotrauma. 2016;33(6):513-

522.

136. Fleischman RJ, Adams AL, Hedges JR, Ma OJ, Mullins RJ, Newgard CD. The

Optimum Follow‐Up Period for Assessing Mortality Outcomes in Injured Older

Adults. Journal of the American Geriatrics Society. 2010;58(10):1843-1849.

137. Andrus MR. Oral anticoagulant drug interactions with statins: case report of

fluvastatin and review of the literature. Pharmacotherapy: The Journal of Human

Pharmacology and Drug Therapy. 2004;24(2):285-290.

138. Nutescu E, Chuatrisorn I, Hellenbart E. Drug and dietary interactions of warfarin

and novel oral anticoagulants: an update. Journal of thrombosis and thrombolysis.

2011;31(3):326-343.

139. Williams D, Feely J. Pharmacokinetic-pharmacodynamic drug interactions with

HMG-CoA reductase inhibitors. Clinical pharmacokinetics. 2002;41(5):343-370.

144

140. Allison PD. Discrete-time methods for the analysis of event histories.

Sociological methodology. 1982;13(1):61-98.

141. Robins JM, Hernan MA, Brumback B. Marginal structural models and causal

inference in epidemiology. Epidemiology. 2000;11(5):550-560.

142. Cole SR, Hernán MA. Constructing inverse probability weights for marginal

structural models. American journal of epidemiology. 2008;168(6):656-664.

143. Plassman BL, Havlik R, Steffens D, et al. Documented head injury in early

adulthood and risk of Alzheimer’s disease and other dementias. Neurology.

2000;55(8):1158-1166.

144. Moretti L, Cristofori I, Weaver SM, Chau A, Portelli JN, Grafman J. Cognitive

decline in older adults with a history of traumatic brain injury. The Lancet

Neurology. 2012;11(12):1103-1112.

145. van Rein N, Cannegieter SC, le Cessie S, et al. Statins and Risk of Bleeding: An

Analysis to Evaluate Possible Bias Due to Prevalent Users and Healthy User

Aspects. American journal of epidemiology. 2016:kwv255.

146. Andrews G. Should depression be managed as a chronic disease? Bmj.

2001;322(7283):419-421.

147. Schumacher M, Denier C, Oudinet J-P, Adams D, Guennoun R. Progesterone

neuroprotection: The background of clinical trial failure. The Journal of steroid

biochemistry and molecular biology. 2016;160:53-66.

148. Benner JS, Glynn RJ, Mogun H, Neumann PJ, Weinstein MC, Avorn J. Long-

term persistence in use of statin therapy in elderly patients. Jama.

2002;288(4):455-461.

145

149. Corrigan JD, Cuthbert JP, Harrison-Felix C, et al. US Population Estimates of

Health and Social Outcomes 5 Years After Rehabilitation for Traumatic Brain

Injury. The Journal of head trauma rehabilitation. 2014.

150. Dijkers MP. Quality of life after traumatic brain injury: a review of research

approaches and findings. Archives of physical medicine and rehabilitation.

2004;85:21-35.

151. Dijkers M, Brandstater M, Horn S, Ryser D, Barrett R. Inpatient rehabilitation for

traumatic brain injury: The influence of age on treatments and outcomes.

NeuroRehabilitation. 2013;32(2):233-252.

152. Humphreys I, Wood RL, Phillips CJ, Macey S. The costs of traumatic brain

injury: a literature. ClinicoEconomics and outcomes research. 2013;5:281-287.

153. Schootman M, Buchman TG, Lewis LM. National estimates of hospitalization

charges for the acute care of traumatic brain injuries. Brain Injury.

2003;17(11):983-990.

154. Control CfD, Prevention. Prevalence of stroke--United States, 2005. MMWR.

Morbidity and mortality weekly report. 2007;56(19):469.

155. Go AS, Mozaffarian D, Roger VL, et al. Heart disease and stroke statistics--2014

update: a report from the American Heart Association. Circulation.

2014;129(3):e28.

156. Unützer J, Patrick DL, Simon G, et al. Depressive symptoms and the cost of

health services in HMO patients aged 65 years and older: a 4-year prospective

study. Jama. 1997;277(20):1618-1623.

146

157. Katon WJ, Lin E, Russo J, Unützer J. Increased medical costs of a population-

based sample of depressed elderly patients. Archives of general psychiatry.

2003;60(9):897-903.

158. Chapman DP, Perry GS. Depression as a major component of public health for

older adults. Prev Chronic Dis. 2008;5(1):A22.

147