DISEASE SEVERITY AND DISABILITY
IN PERSONS WITH
PERIPHERAL ARTERIAL DISEASE
DISSERTATION
Presented in Partial Fulfillment of the Requirements for
The Degree of Doctor of Philosophy in the Graduate School of
The Ohio State University
By
Jeanne Malcom Widener * * * * * The Ohio State University
2008
Dissertation Committee:
Professor Pamela J. Salsberry, advisor Approved by:
Professor Kathleen A. Stone
Professor Barbara Polivka ______Professor Linda K. Daley Advisor College of Nursing
Copyright by
Jeanne M. Widener
2008
ABSTRACT
Peripheral arterial disease (PAD) is a serious condition that can lead to long-term disability. Recently the National Heart, Lung and Blood Institute began a campaign to educate the public and increase awareness of PAD. The diagnosis of PAD frequently occurs late in the process. The purpose of this study was to understand the relationship between mild or severe PAD and disability (health-related quality of life) and determine which factors affect that relationship. This study explored pain, mobility and activity alterations in response to PAD. Sociodemographic, chronic diseases and biological risk factors were also examined. A cross-sectional design was used to examine
4559 adults age 40 and over from the NHANES 2001-2004 data. An ankle- brachial index (ABI) measured PAD severity and the Center for Disease Control and Prevention Health-Related Quality of Life 4 question set measured physical, mental and activity disability.
Comparisons of PAD levels: severe (ABI less than 0.7), mild (ABI 0.7-
0.9) and no disease showed that differences in pain, activity, mobility and risk factors become apparent when PAD is considered asymptomatic. Logistic regression showed physical disability was 1.7 times (95% CI 1.3, 2.2) more likely with mild PAD than no disease. No effect was found between PAD and
ii
mental or activity disability. Education, poverty income ratio and hypertension were confounding factors. Mobility difficulty and calf pain with walking were mediators between PAD and physical disability. Age, ability to do vigorous activity, total cholesterol and obesity were effect moderators. Compared with no disease, odds ratio of physical disability were 4.4 times higher (95% CI 1.5,
13.2) at age 48 with severe PAD, 2.9 times higher (95% CI 1.39, 5.97) with elevated total cholesterol and mild PAD and 4.14 times higher (95% CI 1.23,
13.98) for obesity with severe PAD. Mobility, vigorous activity and calf pain made the most difference in the relationship between PAD and physical disability. Low prevalence of PAD in the general population (6.1%, 95% CI 4.9,
7.1) makes monitoring for disability impractical. Mean number of physically unhealthy days was higher for mild PAD, so surveillance with the CDC HRQOL-
4 may be helpful in monitoring PAD impact.
iii
DEDICATION
To God be the Glory
iv
ACKNOWLEDGMENTS
Greg, Becca, Keven; Thank you for your steadfast love, prayer and encouragement. Also the large extended family both physically related and through the church, you have all been a continuous source of encouragement and support. Cathy and Chantal Thank you!
Dr. Pamela Salsberry has been a God send to completing this dissertation and learning large data set research, which I will be able to continue using. I thank Dr. Susan Frazier who was my advisor for the major portion of my doctoral study. Thank you also to the current and previous committee members for support and wonderful suggestions.
Thank you to the Society for Vascular Nursing for funding to do this research and the encouragement to keep revising the proposal as needed to complete the research process.
Dr. Lee Cohen and the dissertation support group where would I be without all of you over the years encouraging me in the writing, grieving with me when obstacles were blocking my progress, and celebrating the accomplishment with me.
Last, to my statistical consultants, Amy Lehman and Chris Holloman who helped me understand how to reach my research goals.
v
VITA
June 6, 1955 ...... Born, Pittsfield, IL
1976 ...... Diploma, Mennonite Hospital School of Nursing, Bloomington, IL
1980 ...... B. S. N. Nursing East Tennessee State University
1983 ...... M. S. N. Nursing Vanderbilt University
1976 1977 ...... Staff Nurse University of Missouri Medical Center Columbia, MO.
1977 1981 ...... Staff Nurse Johnson City Medical Center Hospital Johnson City, TN
1981 1983 ...... Staff Nurse II Vanderbilt University Hospital, Nashville, TN
1986 1999 ...... Staff Nurse Pattie A. Clay Hospital Richmond, KY.
1989 1998 ...... Visiting Instructor, Instructor, and Assistant Professor (Tenure: 1995) Eastern Kentucky University Richmond, KY
vi
1999 Present ...... Staff Nurse Mt. Carmel East Hospital Columbus, OH
1999 2000 ...... Graduate Teaching Associate The Ohio State University
2000-2004 ...... Graduate Research Associate The Ohio State University Dr. Debra Moser (2000-2003) Dr. Terry Lennie (2002-2003) Dr. Kathy Stone (2003-2004)
2003 Present ...... Instructor, Basic Life Support and Advanced Cardiac Life Support Mount Carmel Training Center Columbus, OH
2004 2007 ...... Clinical Evaluator (CPNE) Excelsior College Midwest Performance Assessment Center (MPAC) Ohio Columbus and Mansfield, Ohio
PUBLICATIONS
Heo, Seongkum, Doering, Lynn V., Widener, Jeanne, and Moser, Debra K. (2008). Predictors and effect of physical symptom status on health- related quality of life in patients with heart failure. American Journal of Critical Care, 17(2), 124-132.
Widener, J.M. (2007). C-reactive protein measurement in the patient with vascular disease. Journal of Vascular Nursing, 15(3), 51-54.
Heo, Seongkum, Moser, Debra K, Widener, Jeanne. (2007). Gender differences in the effects of physical and emotional symptoms on health- related quality of life in patients with heart failure. European Journal of Cardiovascular Nursing, 6 (2), 146-152.
vii
Frazier, S. K., Stone, K. S., Moser, D., Schlanger, R., Carle, C., Pender, L. and Widener, J. (2006). Hemodynamic changes during discontinuation of mechanical ventilation in medical intensive care unit patients. American Journal of Critical Care, 15(6), 580-593.
Frazier, S. K., Brom, H., Widener, J., Pender, L., Stone, K. S. and Moser, D. K. (2006). Prevalence of myocardial ischemia during mechanical ventilation and weaning and its effects on weaning success. Heart & Lung, 35, 363- 373.
Widener, J., Yang, C., Costello, P. and Allen, K. (1999). Modifications to standard guidelines and changes in blood pressure readings using an automatic blood pressure devise. AAOHN Journal, 47(3), 107 113.
Widener, J. and Reid, C. (1993). Use of inquiry in allied health education. AHN Forum, 9(1), 3-6.
FIELDS OF STUDY
Major Field: Nursing
Cognate: Psycho-Physiology
viii
TABLE OF CONTENTS
Page ABSTRACT...... ii DEDICATION...... iv ACKNOWLEDGMENTS...... v VITA...... vi LIST OF FIGURES...... xiii
CHAPTERS: 1. INTRODUCTION...... 1 Significance of the Problem...... 1 Purpose...... 5 Specific Aims and Hypotheses...... 6 Specific Aim 1...... 6 Specific Aim 2...... 6 Specific Aim 3...... 7 Specific Aim 4...... 7 2. REVIEW OF THE LITERATURE...... 8 Epidemiology of Peripheral Arterial Disease...... 8 Conceptual Model...... 11 Biological Function...... 13 Interaction of local and systemic inflammatory factors. . 17 Peripheral Arterial Disease (PAD) severity...... 20 Health practices, chronic diseases and peripheral 21 arterial disease...... Socioeconomic status and biological function...... 27
ix
Symptoms...... 28 Relation of biological status to symptoms...... 29 Functional Status...... 30 General Health Perception...... 32 Health-Related Quality of Life...... 33 Disability...... 37 Summary...... 38 Specific Aims and Hypotheses...... 39 Specific Aim 1...... 39 Specific Aim 2...... 39 Specific Aim 3...... 39 Specific Aim 4...... 40 3. METHODS...... 41 Study Design...... 41 Description of the Data Source...... 41 Sample...... 42 MEASURES ...... 44 Dependent variable ...... 45 Independent variables ...... 47 Demographics and Socioeconomic status...... 47 Peripheral arterial disease severity ...... 51 Chronic disease...... 52 Biological risks...... 53 Symptoms ...... 55 Functional Status...... 55 General Health Perception ...... 56 Data Analysis ...... 57 Model 1...... 58 Specific Aim 1 ...... 58
x
Order of Model 1 Analysis ...... 58 Model 2...... 60 Specific Aim 2 ...... 60 Order of Model 2 Analysis ...... 61 Model 3...... 62 Specific Aim 3 ...... 62 Order of Model 3 Analysis ...... 63 Model 4...... 65 Specific Aim 4 ...... 65 Order of Model 4 Analysis ...... 66 4. RESULTS ...... 68 Descriptive Statistics ...... 68 Demographic and socioeconomic ...... 69 Symptom, Functional Status, and General Health 71 Perception ...... Biological risk and select chronic diseases...... 73 Health-related quality of life and disability...... 77 Correlations...... 78 Logistic Regression...... 80 Crude estimates odds ratios...... 81 Model 1...... 82 Model 2...... 86 Model 3...... 90 Model 4...... 94 Summary ...... 96 5. DISCUSSION...... 100 Disability and PAD ...... 102 Physical Disability ...... 104 PAD Surveillance...... 106
xi
Conceptual Framework Revisited...... 107 Recommendations for Nursing Practice...... 110 Limitations of the Study...... 112 Recommendation for Future Research...... 113 LIST OF REFERENCES...... 115 APPENDICES...... 130 Appendix A.Listing of NHANES files included in this study by category. Information is copied from the NHANES websites (http://www.cdc.gov/nchs/about/major/nhanes/nhanes01- 02.htm and http://www.cdc.gov/nchs/about/major/nhanes/nhanes2003- 2004/nhanes03_04.htm)...... 130
xii
LIST OF FIGURES
Figure Page
2.1 Conceptual Framework...... 22 3.1 Demographic and Socioeconomic Differences between Eligible Sample and Final Samples...... 42 3.2 Healthy Days Questions 2 through 4 for Self-responders...... 45 3.3 Table of Measures...... 47 4.1 Weighted Demographic and Socioeconomic Prevalence in NHANES 2001-2004 data for Final Sample by Three Ankle- Brachial Index (ABI) levels...... 67 4.2 Weighted Symptom, Functional Status, and General Health Perception Prevalence in NHANES 2001-2004 data for Final Sample by Three Ankle-Brachial Index levels...... 69 4.3 Weighted Biological Risk variables Prevalence in NHANES 2001-2004 data for Final Sample by Three Ankle-Brachial Index levels...... 71 4.4 Weighted Chronic Disease Prevalence in NHANES 2001-2004 data for Final Sample by Three Ankle-Brachial Index levels. . . 73 4.5 Weighted Physical, Mental, and Activity Disability Prevalence In NHANES 2001-2004 data for Final Sample and Three Ankle- Brachial Index levels...... 74 4.6 Correlations: Ankle-Brachial Index with Health-Related Quality of Life and Binary Disability variables...... 75
xiii
4.7 Estimated Percentages and 95% Confidence Interval for Physical, Mental, and Activity Disability Outcomes by Ankle- Brachial Index levels...... 77 4.8 Crude (unadjusted) Ankle-Brachial Index level odds ratios (comparisons) for Physical, Mental, and Activity Disability...... 78 4.9 Step 1: Process for Determining Potential Confounders for Model 1...... 79 4.10 Physical Disability final Model 1 ...... 80 4.11 Application of Model 1 Showing Interaction Odds Ratios and 95% CI for Ankle-Brachial Index and Age at the 25th Quartile, Median, and the 75th Quartile...... 82 4.12 Step 1: Process for Determining Potential Confounders for Model 2...... 84 4.13 Step 2: Process for Determining Potential Confounders for Model 2...... 84 4.14 Physical Disability Model 2 Adjusted Odds Ratios and 95% Confidence Intervals ...... 86 4.15 Step 1: Process for Determining Potential Confounders for Model 3...... 88 4.16 Physical Disability Model 3 Significant Adjusted Odds Ratios and 95% Confidence Intervals...... 91 4.17 Step 1: Process for Determining Potential Confounders for Model 4...... 92 5.1 Peripheral Arterial Disease and Disability Conceptual Framework (Based on Wilson & Cleary, 1995 model for health- related quality of life revised by Farrans, et al., 2005)...... 105 A.1 Appendix A Listing of NHANES files included in this study by category. Information is copied from the NHANES websites (http://www.cdc.gov/nchs/about/major/nhanes/nhanes01-02.htm and http://www.cdc.gov/nchs/about/major/nhanes/nhanes2003- 124 2004/nhanes03_04.htm)......
xiv
CHAPTER 1
INTRODUCTION
Significance of the Problem
The first US nationally representative community survey sponsored by the Peripheral Arterial Disease (PAD) Coalition found 75% of respondents were unaware peripheral arterial diseases existed and the other 25% knew minimally correct information (Hirsch et al., 2007). The problem is so severe that a group of representatives from 17 public health and professional organizations released a mandate to create a public awareness program encouraging adults to reduce risk factors and obtain diagnosis early, when the chances of slowing disease progression are greatest (Hirsch, Gloviczki, Drooz, Lovell, & Creager,
2004). Several public service messages have been aired and educational materials have been made available since 2006 when the Stay in Circulation
Campaign was released jointly by the National Heart, Lung, and Blood Institute
(NHLBI) of the National Institutes of Health and the Peripheral Arterial Disease
Coalition (National Heart Lung and Blood Institute, 2006; Peripheral Arterial
Disease Coalition, 2007).
Atherosclerotic plaque develops in the coronary and peripheral arteries by a process where lipids and other materials accumulate within short sections
1
of the vessel wall. The lipid accumulation progresses to plaque formation that thickens the vessel wall in that area and causes the vessel wall to bulge and narrow the vessel lumen (Stühlinger & Tsao, 2003). Plaque formation has recently been linked to inflammatory changes in the vessel walls. This inflammatory process may be evaluated by measurement of an acute phase protein, C-reactive protein (Danesh et al., 2000; Ross, 1999). Elevated C- reactive protein levels in apparently healthy men have been associated with long-term development of PAD (Majewski et al., 1993; Ridker, Cushman,
Stampfer, Tracy, & Hennekens, 1998), risk of a heart attack or stroke in persons with peripheral arterial disease (Rossi et al., 2002) and risk of any future cardiovascular event in persons without peripheral arterial disease
(Albert, Glynn, & Ridker, 2003; Biasucci et al., 1999; Chew et al., 2001; Koenig et al., 1999; Ridker, Buring, Cook, & Rifai, 2003; Ridker, Rifai, Rose, Buring, &
Cook, 2002; Rifai, Buring, Lee, Manson, & Ridker, 2002).
Peripheral arterial disease is asymptomatic initially, but as pain increases with ambulation (intermittent claudication), physical and social functioning may become more difficult, so disability has been reported by individuals seeking treatment (Klevsgard, Hallberg, Risberg, & Thomsen, 1999). Intermittent claudication progresses gradually to critical limb ischemia, where there is pain in the legs at rest and/or development of ischemic lesions on the feet further limiting use of the leg (Cimminiello, 2002). Amputation has been reported in
1.8% to 2.5% of limbs with intermittent claudication, but up to 25% of limbs with
2
critical limb ischemia (Weitz et al., 1996). Older recipients of amputation frequently become bedfast and institutionalized requiring long-term-care supplemented by Medicare and Medicaid creating a burden for both the individual and society (Aronow, 2004; Pell, Donnan, Fowkes, & Ruckley, 1993;
Shechter, Auslander, Weinmann, & Bass, 2003).
Living with peripheral arterial disease may limit health-related quality of life (Hallin, Bergqvist, Fugl-Meyer, & Holmberg, 2002; Johnstone, 2003;
Klevsgard et al., 1999). Physical and social functioning becomes more difficult and, consequently, most function may be limited as pain increases with ambulation (Klevsgard et al., 1999). Limited physical health-related quality of life contributes to increased depression in some chronic illnesses (Arseven,
Guralnik, O'Brien, Liu, & McDermott, 2001; McDermott et al., 2003b) and increased depression may limit health-related quality of life (Lyness et al.,
1997). Depressive mood or major depression have been reported in about 20% of PAD subjects in the small sample tested (Arseven et al., 2001). Depression has been associated with greater morbidity and mortality in cardiovascular disease populations (McDermott et al., 2003b), so the measurement of both physical and mental aspects of health-related quality of life is important.
The Centers for Disease Control and Prevention (2000) use a brief measure of general health and health-related quality of life (HRQOL) to monitor disability for several chronic diseases at the state and national level, but peripheral arterial disease is not currently monitored. Incorporation of both a
3
measure of peripheral arterial disease severity (Ankle-Brachial Index) and the
CDC disability measure (CDC HRQOL-4) into the National Health and Nutrition
Examination Survey (NHANES) database starting in 2001 provides a means not previously available for studying disability associated with peripheral arterial disease in a community based nationally representative sample who may be unaware of the disease and its risks. The Ankle-Brachial Index (ABI) was completed on willing adults 40 years and over during the NHANES examination, since over 40 years is when PAD changes become measurable with ABI
(http://www.cdc.gov/nchs/data/nhanes/nhanes_01_02/lexab_b_doc.pdf and http://www.cdc.gov/nchs/data/nhanes/nhanes_03_04/lexab_c.pdf).
Current tobacco exposure and uncontrolled diabetes mellitus are the health issues most commonly associated with progression of peripheral arterial disease severity. Current tobacco smoking in subjects over 65 years of age resulted in a six times greater chance of developing severe peripheral arterial disease than either non-smokers or subjects who had quit smoking more than a year before (Hooi et al., 2001). For the same group of patients, the odds ratio of developing PAD with another concurrent disease process was 2.1 for diabetes mellitus, 1.7 for hypertension, and 1.5 for hyperlipidemia (Hooi et al., 2001). The
Diabetes Control and Complications Trial investigators (1993; 1996) found that glycosylated hemoglobin of 7% or less was predictive of fewer diabetic complications, including postponing lower extremity amputation an average of
5.6 years. Since elevated blood glucose, elevated blood pressure, and elevated
4
cholesterol levels can remain undetected and asymptomatic, the associations between these undiagnosed biological risks, peripheral arterial disease and disability need to be investigated.
Purpose
The purpose of this study was to understand the relationship between mild or severe peripheral arterial disease (PAD) and Disability (Health-related
Quality of Life) and determine which factors affect that relationship. Disability is equated with decreased health-related quality of life (Moriarty, Kobau, Zack, &
Zahran, 2005). A Model of the Relationships between Biological Function and
Health-related Quality of Life developed by Wilson and Cleary (1995) with revisions suggested by Farrans and colleagues (2005) provided the foundation for the conceptual framework. Incidence of PAD is known to increase with advancing age, male gender, and African American race, but how adjusting for these factors might change the associations between PAD severity and disability in the general population is unknown. Increased education and income have been associated with more positive lifestyle change related to health risks, but the associations between socioeconomic factors and PAD have not been delineated. Pain, mobility and general health perception have all been associated with PAD severity and health-related quality of life using other instruments; however the associations within the current model are not known.
Other chronic diseases are associated with disability, but it is unknown if adjustment for other chronic diseases will change the associations between
5
PAD severity and disability in the general population. Biological risk factor levels contribute to PAD severity and they may modify the associations between PAD severity and disability. The four specific aims to guide the research are the following:
Specific Aims and Hypotheses
Specific Aim 1. To determine the association between PAD severity and disability in community dwellers 40 years and over using the NHANES 2001-
2004 data while adjusting for other factors.
Hypothesis 1. There will be a significant association between PAD
severity and disability alone and while controlling for:
1. Demographics, and
2. Socioeconomic status.
Specific Aim 2. To determine the potential mediator effects of symptoms, functional status, and general health perceptions on the association between
PAD severity and disability.
Hypothesis 2. There will be a significant mediator effect on the
association between PAD severity and disability from the respondent
perceptions of:
1. symptoms
2. functional status
3. general health.
6
Specific Aim 3. To determine the potential moderator effect of other chronic diseases and biological risks on the association between PAD severity and disability.
Hypothesis 3. There will be a significant moderator effect on the
association between PAD severity and disability from respondent other
diagnosed chronic diseases and biological risks of:
1. High tobacco exposure,
2. Diabetes diagnosis or elevated blood glucose,
3. Hypertension diagnosis or elevated systolic or diastolic blood
pressure,
4. Hyperlipidemia diagnosis or elevated total cholesterol or low
HDL cholesterol, and
5. Adiposity (obesity or large waist measurement).
Specific Aim 4. To determine the effect of integrating Models 1, Model 2, and Model 3 on the association between PAD severity and disability.
Hypothesis 4. There will be a significant effect on the association
between PAD severity and disability when integrating the variables
included in previous models.
7
CHAPTER 2
REVIEW OF THE LITERATURE
The disability from peripheral arterial disease (PAD) is generally thought to be greater as the blood flow decreases with worsening PAD, but the level of disability has not been examined in a large national study. Disability may be a negative aspect of health-related quality of life. Different individuals have different perceptions of the same level of disease; one may perceive disability and another not. Improving Quality and Years of Healthy Life, one of the two overriding goals of Healthy People 2010, has lead to state and national monitoring of disability in several chronic disease populations, however PAD is not one of those diseases (About Healthy People 2010, http://www.cdc.gov/nchs/about/otheract/hpdata2010/abouthp.htm).
Epidemiology of Peripheral Arterial Disease
Several cross-sectional and longitudinal investigations of the epidemiology of peripheral artery disease had been conducted over the past 30
40 years (Bordeaux, Reich, & Hirsch, 2003; Cimminiello, 2002; Collins,
Petersen, Suarez-Almazor, & Ashton, 2003; Hobbs, Wilmink, & Bradbury, 2003;
Hooi et al., 2001; Mya & Aronow, 2003; Westendorp et al., 2000). Differences in disease progression with age, gender, ethnicity and presence of risk factors
8
were shown. Each investigation had a limited focus, and gender and ethnicity were rarely examined in the same investigation. The age range of subjects varied considerably resulting in a variety of symptoms and comorbidities, which makes comparisons difficult. Investigations with a higher percentage of older subjects reported a higher prevalence of both mild and severe disease.
Subclinical hypothyroidism was reported to greatly increase the incidence of severe disease in one very small elderly group (Mya & Aronow, 2003).
Peripheral arterial disease incidence had mainly been studied in subjects who were recruited while seeking treatment from vascular specialists for their symptoms or during screening events in one specific area of the United States.
The prevalence of peripheral arterial disease is highest in older adults with reports of 12 to 27% of those over the age of 60 years diagnosed with the disease; while only 20% of those have symptoms (Cimminiello, 2002; Collins et al., 2003; Hirsch et al., 2004; Ostchega, Paulose-Ram, Dillon, Gu, & Hughes,
2007). The Framingham Study found an annual symptomatic peripheral arterial disease incidence rate of 3.6 for men and 1.8 for women (Kannel & McGee,
1985). Within 5 years after menopause the incidence rate for women may be higher than for men (Hooi et al., 2001), but hormone replacement therapy of longer than a year duration had a protective effect on post-menopausal women helping to keep them symptom free (Westendorp et al., 2000). Collins et al.
(2003) found that the prevalence of peripheral arterial disease was higher in
African American subjects (22.5%) than Caucasian (13.3%) or Hispanic
9
(13.7%) subjects in Texas. A higher incidence of severe PAD (Hobbs et al.,
2003) and amputation (Wrobel, Mayfield, & Reiber, 2001) is also reported in
African Americans.
Data from NHANES 1999-2000 projected that about 5 million PAD cases were present nationally in persons 40 years or older with an ankle-brachial index below 0.90 (Selvin & Erlinger, 2004). Gender prevalence was equivalent in this study with overlapping 95% confidence intervals (CI) for both men (4.5%) and women (4.2%) near the 4.3% national prevalence (Selvin & Erlinger, 2004).
Prevalence did increase with age to 14.5% (95% CI 10.8 to 18.2%) in persons
70 years and older (Selvin & Erlinger, 2004). Non-Hispanic Black prevalence
(7.9%, 95% CI 5.2-10.6%) was higher than Non-Hispanic white (4.4%, 95% CI
2.8-6.0%) and Mexican American (3.0%, 95% CI 1.4-4.6%) as found in previous studies (Selvin & Erlinger, 2004), but the prevalence was not as high in the 40 year and older group as it was in the Texas study (Collins et al., 2003). The prevalence was reported at about 6 million for the NHANES 1999-2002 data
(Resnick & Foster, 2005). One study that included the NHANES 1999-2002 data, also included several other studies with multiple national data collection locations over multiple years (1972 2003) to try to get the best overview of ethnic prevalence of PAD for Asian Americans and American Indians, since those groups are under represented in the NHANES (Allison et al., 2007).
Prevalence increased by age in all groups, but for American Indians both men
(2.6%) and women (3.2%) had the highest prevalence in the 40-49 year old
10
decade and Asian Americans had the lowest prevalence overall. All other race/ethnicities had similar prevalence to the previously mentioned NHANES data (Allison et al., 2007).
Conceptual Model
Wilson and Cleary (1995) developed a conceptual model to explain
Quality of Life related to health factors that depicts the progressive links from the dimensions of Biological Function to Symptoms, Functional Status, General
Health Perception and Quality of Life. Characteristics of the Individual and
Characteristics of the Environment were conceptualized to influence Symptoms,
Functional Status, General Health Perception, and the Overall Quality of Life in different specified ways (Wilson & Cleary, 1995). Farrans and colleagues
(2005) revised the model and the model has become the foundation for the conceptual framework for this study (Figure 1).
The conceptual framework continues to examine the characteristics of the individual as influential on all other aspects of the model, but characteristics of the environment are not included in this study. At the left Biological Function is a continuum from health to life-threatening illness experienced by all humans.
Malfunction at the molecular, cellular, or organ system level can result in poor physical or mental health. Altered biological function leads to symptoms, functional status and general health perception. Symptoms occur as a conscious or unconscious interpretation of the biological malfunctions.
Functional Status is a
11
Characteristics of the Individual
Demographics (age, gender, race/ethnicity, marital status)
Socioeconomic status (education, PIR=income/poverty ratio)
Biological Function
Peripheral arterial disease Symptoms Disability Ankle-brachial index Pain in legs, ------calves and feet Health-related Chronic Disease Quality of Life Diagnosed: Functional Diabetes Status Healthy Days: Hypertension Physical Hyperlipidemia Activity level Mobility Mental Activity Biological risks Nicotine exposure General Health Blood pressures Perception Glycosylated hemoglobin HDL and total cholesterol Adiposity
Figure 2.1 Conceptual Framework
12
physical, social, or psychological ability to be active in response to the biological function (or malfunction). General Health Perception is a global self-perception putting a positive or negative filter on the person s situation. Disability is a perception of limited health-related quality of life. The literature review will be organized using the Biological Function, Symptom, Functional Status, and
Health-related Quality of Life dimension headings while focusing on PAD.
Characteristics of the Individual that relate to a given dimension will be presented in that section of the literature review, since those characteristics have been postulated to influence all the dimensions (Farrans et al., 2005).
Biological Function
Atherosclerosis is the term used to identify a process that occurs in arterial vessels, initially as a benign accumulation of lipid rich, fatty streaks, but can progress to the pathologic lesions found in peripheral, coronary and cerebral artery disease (Ross, 1993, 2004). The Response-to-Injury hypothesis
(Ross, 1993, 1999, 2004) indicates that atherosclerosis begins as an inflammation of the endothelial and smooth muscle cells of the artery wall in response to injury from mechanical (excessive hemodynamic pressure within the vessel and sheering forces along the wall), irritant (homocysteine or toxins), and/or infectious (viruses and bacteria) causes within the vessel. Damaged endothelial cells release interleukin-6, interleukin -1 , and other pro- inflammatory reactants into the bloodstream that attract monocytes to the area
(Greaves & Channon, 2002). Blake and Ridker (2002) found that damaged
13
endothelial cells release adhesion molecules and leukocytes adhere and roll along the arterial wall until the monocytes and T-cells are attracted inside the vessel wall covered by these endothelial cells. Monocytes levels in the white blood cell differential counts were found to be significantly higher (p = 0.002) in person with peripheral arterial disease than those without the disease in the
NHANES 1999-2002 data (Nasir, Guallar, Navas-Acien, Criqui, & Lima, 2005).
Monocytes that engulf foreign proteins are called macrophages (Ross,
2004). The macrophages initially adhere to the vessel wall, but also invade the intimal layer of the vessel wall by squeezing between endothelial cells. Low- density lipoproteins are oxidized by exposure to free radicals released by the macrophages along the vessel wall, and this oxidized low-density lipoprotein increases endothelial inflammation. When macrophages phagocytize oxidized low-density lipoproteins they develop a foamy appearance, so are called foam cells. Fatty streaks develop along the vessel wall as T-cells and foam cells accumulate in the vessel s intimal layer (Ross, 2004). Strong and colleagues
(1999) found fatty streaks in patchy areas in up to 40% of 15-19 year olds, but almost continuously in 50% of the aortic and 10% of the coronary walls of 30-35 year olds examined on autopsy from death due to an unrelated cause.
Research supports the concept that fatty streaks do not inhibit blood flow, but can regress or progress to a fibrous lesion (Stühlinger & Tsao, 2003).
The Response-to-Injury hypothesis postulates that fibrous lesions develop through a complex interaction of several processes (Ross, 1993, 1999,
14
2004). Macrophages exposed to oxidized LDL recruit smooth muscle cell into the intimal layer by the release of several cytokines including tumor necrosis factor alpha and interleukin 1 that may also escape into the blood stream
(Greaves & Channon, 2002). Smooth muscle cells in the intimal and medial layers of the arterial wall are stimulated to proliferate by platelet derived growth factor, macrophages and endothelial cells. This increases smooth muscle cell numbers and stimulates their migration to the intimal layer. Low-density lipoproteins that have diffused into the intimal layer can bind to the cell wall of smooth muscle cells and are absorbed in large amounts leading to a foam cell appearance for those smooth muscle cells, in addition to the macrophage derived foam cells. Some foam cells rupture dumping their contents into a necrotic lipid mass in the center of the plaque. Multiple layers of the lipid and smooth muscle can develop in the early stages (Ross, 1993, 1999, 2004).
The fibrous cover of the plaque, when intact, is smooth like the lumen wall, so even as the plaque narrows the lumen circumference, the flow is not hindered enough for the person to become symptomatic (Stühlinger & Tsao,
2003). Smooth muscle cells are instrumental in the development of the matrixes and collagen of the fibrous cap layer that form on the lumen aspect over the lesion (Ross, 1993, 1999, 2004). The growth factors that activate smooth muscle proliferation can also inhibit the matrix building of the fibrous cap, but antagonistic growth factors influence stabilization of the matrix. A strong collagen matrix over the lesion keeps the plaque from rupturing (Ross, 1999,
15
2004). Inflammatory macrophages and T-cells have been found in high concentrations adjacent to plaque rupture in weakened fibrous caps but not in other non-ruptured plaques, so the growth factors breaking down the matrix are thought to be released by these cells, which contributes to the inflammation theory of plaque progression (Boyle, 1997).
The fibrous plaque can become calcified with mineral deposits that stiffen the artery wall making it less able to dilate to meet flow needs of the distal tissues, but the individuals may still not become symptomatic (Stühlinger
& Tsao, 2003). Turbulence at artery bifurcations or other areas of restricted flow makes those areas highly susceptible to plaque formation. Kiechl and Willeit
(1999a; 1999b) examined carotid plaque development prospectively over a 5 year period in 40-84 year old adults and found that plaque growth and vascular remodeling occurs to maintain the lumen patency. Plaque growth increased more rapidly in the lesions with greater than 40% occlusion and at that occlusion, the artery was unable to remodel to maintain the original lumen size
(Kiechl & Willeit, 1999a, 1999b).
When endothelial cells are unable to completely cover either early or fibrous lesions, due to lesion size or hemodynamic forces, spaces develop between these cells and platelets are exposed to oxidized LDL in the intimal layer of the artery wall. This can trigger clot formation either in the lumen or within the plaque (Ross, 1993, 1999, 2004). Complement and other clotting cascade activating factors in the intimal layer of the vessel wall can trigger clot
16
formation inside the plaque. The presence of clotting factors, macrophages, T- cells, smooth muscle cells, and other deposits form a complex lesion.
Calcification from mineral deposits within the lesion, in addition to a liquid lipid necrotic center from ruptured foam cells, creates a less stable lesion (Ross,
1993, 1999, 2004). Dirksen and colleagues (1998) found a greater concentration of macrophages in the upstream area and a greater concentration of smooth muscle cells in the downstream areas of plaques collected at autopsy, indicating that the sheer force upstream may be responsible for more damage to the proximal area of the plaque leaving the downstream area of the plague more stable. Plaque rupture activates the clotting cascade and thrombosis can block part or the entire artery lumen
(Ross, 1993, 2004).
Interaction of local and systemic inflammatory factors. Localized inflammation in the vessel wall and systemic inflammation both play a role in the atherosclerotic process (Greaves & Channon, 2002). Damaged endothelial cells and macrophages in the atherosclerotic lesions that become plaques release interleukin-6, interleukin-1 , and other reactants into the bloodstream that can enhance the inflammatory response both in the local plaque and systemically (Greaves & Channon, 2002). Interleukin-6 stimulates hepatic cells to produce and release the acute phase protein, C-reactive protein (CRP), into the systemic circulation. CRP can be produced by vessel endothelial cells and released in or near the inflammatory lesions as well (Weinhold & Rüther, 1997).
17
C-reactive protein is one acute phase protein, which can rise as much as a ten-thousand fold in response to acute infection, trauma, tissue necrosis, or malignant neoplasia (Pepys, 2003; Pepys & Hirschfield, 2003). The acute phase response will peak in 48 hours and then drop, since C-reactive protein has a 19 hour half-life and levels are dependent on hepatic production. Chronic infection or inflammation will cause a steady, sustained elevation of C-reactive protein, although the exact mechanism of action is not fully understood.
Genetics has some role in the level where an individual s steady state remains, but 90% of all individuals will have a steady level of less than 3 milligrams per liter (mg/L) and 99% will be less than 10 mg/L. Since very few individuals would normally have a CRP level between 3 and 10, special screening of these individuals in necessary to determine their relative risk of future cardiac events
(Pepys, 2003; Pepys & Hirschfield, 2003).
C-reactive protein (CRP) initiates the complement cascade near inflammatory sites by binding with Complement 1q (Weinhold & Rüther, 1997), which then binds with the plaque in the artery wall to enhance the immune response (Yasojima, Schwab, McGeer, & McGeer, 2001). CRP has stimulated expression of adhesion molecules in endothelial cells in laboratory experiments.
This response may encourage enlargement of a smaller lesion (Pasceri,
Willerson, & Yeh, 2000). CRP encourages foam cell development by acting as a trigger for LDL uptake by macrophages in atherosclerotic plaques (Zwaka,
Hombach, & Torzewski, 2001). When levels of CRP in the plasma are elevated,
18
CRP binds to foam cells and some smooth muscle cells within the atherosclerotic lesion, which marks those cells for apoptotic rupture. This process may be the reason why early Type III lesions contain liquid deposits of lipid (Kobayashi et al., 2003). CRP has been found as a component of human plaque removed during surgery (Kobayashi et al., 2003; Torzewski et al., 2000) implicating CRP in all phases of the plaque development process.
Prospective studies have attempted to identify a useful serum marker of future cardiovascular events. Scientists have examined cytokines (interleukin-6 and interleukin-1 ) and the acute-phase protein C-reactive protein (CRP) in systemic circulation. CRP has been found to be the most stable and reliable serum marker of future myocardial infarction, cerebral infarction, and peripheral arterial disease (Albert et al., 2003; Biasucci et al., 1999; Chew et al., 2001;
Koenig et al., 1999; Majewski et al., 1993; Ridker et al., 2003; Ridker et al.,
1998; Ridker et al., 2002; Rifai et al., 2002; Rossi et al., 2002). The cytokines
(interleukin-6 and interleukin-1 ), though frequently elevated and positively correlated with CRP levels, have more variable normal and circadian levels within the same individual, so have not proven to be a reliable predictor of cardiovascular events (Weinhold & Rüther, 1997). Since CRP has become the serum marker of choice, several studies have used the NHANES data to examine national trends in the relationship between inflammation (CRP) and
PAD (Wildman, Muntner, Chen, Sutton-Tyrrell, & He, 2005), CRP and developing peripheral arterial disease earlier than age 60 (Lane, Vittinghoff,
19
Lane, Hiramoto, & Messina, 2006), and CRP with PAD and risk of future cardiovascular events (Menke, Muntner, Wildman, Dreisbach, & Raggi, 2006;
Vu et al., 2005).
The abdominal aorta branches into the inguinal arteries, which further branch into the femoral arteries (Graham & Ford, 1994; Stühlinger & Tsao,
2003). The femoral, popliteal, peroneal, tibial, and pedal arteries and their branches provide blood supply to the thigh, calf and foot. The arteries in the legs are long and bifurcate multiple times providing likely sites for plaque development. Tight clothing, sitting and crossing the legs may further contribute to turbulent blood flow in the legs. Diffuse, symmetrical lesions in the femoral, popliteal, and tibial arteries occur in about half of the people with peripheral arterial disease. Tibial artery branches are involved one fourth of the time when lesions are only present in one vessel segment. The location, type and size of the lesions plays a role in the pathological change that lead to signs and symptoms of peripheral arterial disease (Graham & Ford, 1994; Stühlinger &
Tsao, 2003).
Peripheral Arterial Disease (PAD) severity. The pathological changes that occur with PAD develop slowly and progressively over time. The first change is a reduction in arterial pressure distal to the lesion (Graham & Ford,
1994; Stühlinger & Tsao, 2003). During healthy, laminar blood flow, Poiseuille s law predicts an increase in driving pressure with decreasing radius of the vessel, usually due to vasoconstriction of all vessels in an area. Atherosclerotic
20
lesions gradually narrow the artery lumen to a critical stenosis at which decreased flow will reduce arterial pressure distal to the lesion. A critical stenosis is present when the diameter is reduced by half or the lumen decreases to only one-fourth the original area. One long lesion in an artery may produce a critical stenosis, but usually there are multiple small lesions that additively decrease the flow pressure enough to cause a significant drop in ankle arterial pressure (Graham & Ford, 1994; Stühlinger & Tsao, 2003).
Peripheral Arterial disease (PAD) severity can be approximated by measuring the Ankle-Brachial Index (ABI), the doppled systolic pressure in an ankle divided by the doppled brachial systolic pressures to calculate the index on that side of the body (Treat-Jacobson & Walsh, 2003). An ABI of greater than 0.9 is considered healthy or without disease, 0.7 to 0.9 is thought to be mild disease and is usually asymptomatic, values below 0.7 are severe disease and present with intermittent ischemic pain (claudication) and critical limb ischemia (Hirsch et al., 2006; Treat-Jacobson & Walsh, 2003). Once PAD becomes severe, it is more likely to progress and be associated with disability.
Health practices, chronic diseases and peripheral arterial disease.
Several health practices and chronic diseases contribute to the development and progression of peripheral arterial disease. For the same group of patients, the odds ratio of developing PAD with another concurrent disease process was
2.1 for diabetes mellitus, 1.7 for hypertension, and 1.5 for hyperlipidemia (Hooi et al., 2001). Vu and colleagues (2005) found the odds ratio of PAD presence
21
with elevated C-reactive protein was 3.9 when metabolic syndrome was also present, 4.8 when metabolic syndrome and diabetes were present and 8.6 when only diabetes was present in addition to PAD and elevated CRP.
Maintaining tight control in treatment of diabetes, hypertension, hyperlipidemia and metabolic syndrome limits the advancement of peripheral arterial disease.
The increased prevalence of peripheral arterial disease in African Americans correlates with the increased incidence of hypertension and diabetes mellitus within this racial group (Campinha-Bacote, 1998), two morbidities related to the development of peripheral arterial disease. Compared to persons without peripheral arterial disease, adverse outcomes include 2-fold increase in likelihood of heart failure (Anand, Ventura, & Mehra, 2007), a 3 to 5-fold increase in likelihood of a myocardial infarction or stroke, high likelihood of severe functional limitations with amputation of one or both legs, and 2 to 3-fold increase in mortality (Hirsch et al., 2004).
Current tobacco smoking in subjects over 65 years of age resulted in a 6 times greater chance of developing severe peripheral arterial disease than either non-smokers or subjects who had quit smoking more than a year before
(Hooi et al., 2001). Tobacco use in any form is associated with the development of atherosclerosis by elevation of serum nicotine levels (Grundy et al., 1998;
Ockene & Miller, 1997). Nicotine has a half-life of about 2 hours and approximately 80% of the nicotine is metabolized by the liver into cotinine, which has a half-life of 20 hours (Wall et al., 2007). Smoking additionally
22
exposes individuals to carbon monoxide and thiocyanate (Bordeaux et al.,
2003; Grundy et al., 1998; Ockene & Miller, 1997). These toxins circulate in the blood and initiate inflammation of the arterial endothelium, which contributes to the development and progression of atherosclerosis (Bordeaux et al., 2003;
Frangos, Gahtan, & Sumpio, 1999). Carbon monoxide increases coagulation of the blood and decreases oxygen carrying capacity when bound to red blood cells, which both contribute to hypoxia and ischemia at the peripheral tissue level. Nicotine activates nicotinic receptors, which may induce alterations in hemodynamic pressures, alter sheering forces on the vessel wall, vasoconstrict the vessels and decrease blood flow to the periphery (Bordeaux et al., 2003).
Most studies ask about current, past or never smoking, but do not measure a biological variable, like cotinine, which is available in the NHANES data base as a serum cotinine level (http://www.cdc.gov/nchs/nhanes.htm).
Diabetic individuals develop endothelial changes in the large and small vessels of the body. High blood glucose levels may encourage hypertrophy of insulin independent cells, like the endothelium lining the arteries, which can narrow the vessel lumen over time. Since the vessels are smaller in the microcirculation at the pre-capillary and capillary level, hypertrophy of the endothelial lining can block these vessels worse than in the larger vessels.
Calcification of the vessel walls may make the arteries difficult to compress and pressure measures in the ankles inaccurate, so ankle-brachial index may not always be possible (Currie, Wilson, Baird, & Lamont, 1995). Intensive control of
23
glucose levels is necessary to keep the glycosylated hemoglobin levels in diabetic individuals below a recommended 7%, which has been shown to slow advancement of vascular complications (Diabetes Control and Complications
Trial, 1993; Marso & Hiatt, 2006). Muntner and colleagues (2005b) recommended close monitoring for the development of PAD or other cardiovascular disease in any person with a glycosylated hemoglobin higher than 5.3, even those who are not diagnosed with diabetes mellitus based on
NHANES 1999-2002 ABI values. Most NHANES data studies which examined ankle-brachial index have included self-reported diabetes but have not included the glycosylated hemoglobin values in their reports (Lane et al., 2006; Menke et al., 2006; Nasir et al., 2005; Resnick & Foster, 2005; Selvin & Erlinger, 2004;
Vu et al., 2005; Wildman et al., 2005), therefore there is a need to examine this biological risk factor and the relationship with PAD.
The Hemodynamic stress of hypertension causes (a) changes in the endothelial cell s ability to approximate in the arterial lining, (b) the stiffening of the large vessels when elastin is replaced by collagen and (c) smooth muscle cell hypertrophy in the medial layer of the vessels (Kiechl & Willeit, 1999b;
Ross, 1999). Uncontrolled hypertension, systolic or diastolic blood pressure consistently above 140 mm Hg or 90 mm Hg respectively, is a major risk factor of all cardiovascular disease through the promotion of atherosclerosis (Burns,
Gough, & Bradbury, 2003; Grundy et al., 1998; Treat-Jacobson & Walsh, 2003;
Wilson & Grundy, 2003). Several PAD studies have included the diagnosis of
24
hypertension in their sample descriptions, but actual systolic and diastolic blood pressure levels were only reported when renal disease or risk factors of PAD were studied (O'Hare et al., 2003; Vu et al., 2005; Wildman et al., 2005).
The National Cholesterol Education Program recommends maintaining the following serum lipid levels in individuals with known cardiovascular disease like peripheral arterial disease: total cholesterol level less than 200 mg/dl, high- density lipoproteins higher than 60 mg/dl, low-density lipoprotein levels lower than 100 mg/dl, non-high-density lipoprotein levels less than 130 mg/dl and triglycerides (very-low-density lipoproteins) level less than 200 mg/dl
(McKenney, 2003; National Heart Lung and Blood Institute, 2002). A majority of studies of the PAD population that include cholesterol measures are drug studies aimed at lowering bad cholesterol and raising good HDL or studies of awareness of need for monitoring and treatment in PAD populations (Fabsitz et al., 1999; Hirsch & Gotto, 2002; McDermott et al., 2002; McDermott et al.,
2003a; McDermott et al., 2003d; Mohler, Hiatt, & Creager, 2003; Selvin &
Erlinger, 2004).
Metabolic syndrome X has been identified as a contributing factor in the development of cardiovascular disease, diabetes mellitus and hypertension, among other chronic diseases (Grundy, Hansen, Smith, Cleeman, & Kahn,
2004). Obesity and insulin resistance are dominant characteristics of individuals with the metabolic syndrome (Grundy et al., 2004; St-Onge & Heymsfield). The prevalence of metabolic syndrome in the general public is 24%, but 50% in
25
individuals with a body mass index higher than 25 are diagnosed with more than two other characteristics of metabolic syndrome. Older individuals,
Hispanic and South Asian ethnic groups and individuals with excessive visceral fat are most likely to be diagnosed. Other factors include fasting serum glucose
100 mg/dl or higher, fasting serum triglycerides 150 mg/dl or higher, high- density lipoprotein cholesterol less than 40 mg/dl in men or less than 50 mg/dl in women, waist circumference 40 inches or larger in men and 35 inches or greater in women; and arterial blood pressure higher than 135/85 mg Hg
(Grundy et al., 2004; National Heart Lung and Blood Institute, 2002; St-Onge &
Heymsfield). The metabolic syndrome recommendations are the basis for most of the biological risks examined in this study, since those factors also contribute to peripheral arterial disease risk and progression.
Patients with higher knowledge levels about peripheral arterial disease and their care are more likely to decide to modify their lifestyle (Treat-Jacobson et al., 2002). Decisions by individuals with PAD to modify lifestyle met with resistance from bodily addictions to tobacco, lack of measurable results (Gibson
& Kenrick, 1998), lack of payment by Medicare for exercise training and physical inability to comply consistently (Buck & Ciccone, 2004; Feinglass,
Morasch, & McCarthy, 2000). The more formal education an individual had received, the more likely the person was to incorporate healthy practices into the lifestyle and report better health overall independent of income level in the
General Social Survey 1972-2000 and the Community Tracking Study 1996-
26
1997 (Schnittker, 2004). Level of education and income may have an impact on the disability perceptions of individuals with PAD, but these variables have not been included in the NHANES investigations that examined ABI (Lane et al., 2006; Menke et al., 2006; Nasir et al., 2005; O'Hare et al., 2003; Resnick &
Foster, 2005; Selvin & Erlinger, 2004; Vu et al., 2005; Wildman et al., 2005).
Socioeconomic status and biological function. An inverse relationship between socioeconomic status and health is considered well established by many researchers (Adler & Newman, 2002; Marmot, 1999, 2003;
Oakes & Rossi, 2003). Measurement of socioeconomic status has varied among investigations, but currently, income in relation to the poverty level and education level are recommended over occupation or work status as indicators of a person s ability to meet current needs, especially in a retired or disabled population (Adler & Newman, 2002; Andresen & Miller, 2005; Marmot, 2003;
Oakes & Rossi, 2003; Schnittker, 2004). Smoking, hypertension, hyperlipidemia and diabetes mellitus; all risk factors for development and progression of PAD are highly associated with poverty and lower education (Harwood, 2006;
Kanjilal et al., 2006). Income is a difficult variable to study, especially in the retired population where the social security check amount may be the answer given even when pensions are also received. PIR will be used with caution in the data analysis. The relationships among peripheral arterial disease, poverty and education have not been specifically studied.
27
Symptoms
Pain is the most common initial symptom for an individual with PAD. Pain occurs in the form of intermittent claudication or ischemic foot pain. Gibson and
Kenrick (1998) found that preoperatively, patients indicated that the pain was so severe it was difficult to describe to caregivers, but the pain became the trigger to get treatment. Intermittent claudication is a cramping, aching, heaviness with fatigue or numbness during walking or other activity. With intermittent claudication, the symptoms are relieved within 5 minutes of cessation of the exercise (Halperin, 2002). Since the femoral arteries provide circulation mainly to the thigh, intermittent claudication in the thigh is an indication of possible femoral blockages, and likewise, intermittent claudication in the calf indicates likely popliteal blockages (Treat-Jacobson & Walsh, 2003). Resnick and Foster
(2005) reported symptoms for respondents in the 1999-2002 NHANES with an
ABI less than 0.90 as 40.3% had pain in their legs with walking, 13.6% had numbness, 16.5% had tingling in their feet and 22.5% had lost some feeling in their feet. Respondents in the 1999-2002 NHANES reported pain with walking equally for those respondents with or without diabetes, but foot symptoms were more closely associated with a Diabetes diagnosis ("Mobility limitation among persons aged > or =40 years with and without diagnosed diabetes and lower extremity disease--United States, 1999-2002.," 2005). Previous NHANES studies reported PAD symptoms for one group instead of the mild (ABI 0.7-0.9) and severe (ABI less than 0.7) PAD levels examined in this investigation.
28
As the PAD becomes more severe, ischemic foot pain may be present constantly and intensify at night when the feet are elevated, since in that position gravity is not assisting the flow of blood to the feet, so ischemia is more likely to occur (Halperin, 2002). Diabetic individuals may have peripheral neuropathy and not have feeling in their feet. The loss of feeling may contribute to diabetic individuals not having pain even when critical limb ischemia with gangrenous or ulcerative lesions is present.
Relation of biological status to symptoms. Arterial pressure can decrease in the distal vessels long before ischemia is noticed (Graham & Ford,
1994). Ischemia to the distal tissues initially will occur with exercise, when the need for increased velocity of blood flow cannot be met and the lack of flow leads to anaerobic metabolism in the distal tissues. The most distal areas, usually the calf muscle and foot are the first to show signs and symptoms of decreased circulation. In the lower extremities, fatigue and ischemic pain called claudication occur. If the pain is relieved by stopping an activity, it is called intermittent claudication (Graham & Ford, 1994). As the number and size of arterial lesions increase, the blood flow, even during rest, may become marginal and conserved for the muscles, so the skin and hair follicles lose their circulation. Hair losses on the feet and lower limbs are accompanied by pale, cool, dry skin that can become shiny and transparent. The changes can be subtle, but progressive. Ischemic lesions or ulcers are dry, with even edges where the circulation has been inadequate to maintain the tissue, so it has dried
29
and become necrotic or fallen away. Eventually, beginning with the most distal and progressing toward the trunk, nutrient loss will lead to increasing fatigue, claudication, and atrophy of the tissues in the ischemic areas (Graham & Ford,
1994).
Functional Status
Person with PAD will complain about inability to walk very far, but the usual level of physical activity influences an individual s perception of current disability (Ekers, 1986). Women have been found to have poorer physical functioning than men with the same pathology (Collins, Suarez-Almazor, Bush,
& Petersen, 2006; Gardner, 2002; McDermott et al., 2003c). Disease progression (drop in ABI, walking speed and distance) is slower in persons with mild PAD who walk at least 3 times a week than those who exercise less
(McDermott et al., 2006).
Gardner and Killewich (2001) found that subjects perceived that they walked better and farther 3 to 4 months after peripheral bypass operations, but there were not significant differences post-operative for speed of a 20 meter walk at clinic, distance walked in 6 minutes at clinic, pedometer readings over 2 days at home, or accelerometer measures of physical activity over 2 days at home. Moderate activity within the last 30 days was reported for NHANES
1999-2002 respondents by 61% with ankle brachial index of 0.90 to 1.40, but only 34.5% when the ABI was less than 0.90 (Resnick & Foster, 2005).
Inversely, 1/3 of respondents with ABI greater than 0.90 and 60% with ABI less
30
than 0.90 reported physical inactivity during the NHANES 1999-2002 cycles
(Wildman et al., 2005). Mobility limitation was defined as difficulty walking a quarter mile, walking up 10 steps without resting, or walking one room to another on the same level from the NHANES physical function questionnaire in one report ("Mobility limitation among persons aged > or =40 years with and without diagnosed diabetes and lower extremity disease--United States, 1999-
2002.," 2005). A higher percentage of adults with diabetes and lower extremity disease (meaning PAD or peripheral neuropathy) had greater limitation (39%) than those with only diabetes (23%), only lower extremity disease (25%) or with no disease (14%) ("Mobility limitation among persons aged > or =40 years with and without diagnosed diabetes and lower extremity disease--United States,
1999-2002.," 2005). Mobility issues for persons with PAD separate from peripheral neuropathy have not been studied using the NHANES data.
CRP levels for adults over 60 years of age were found to correlate with difficulty in lower extremity mobility, leisure and social activities, instrumental activities of daily living, and general physical activity domains derived from some NHANES 1999-2002 physical function questions (Kuo, Bean, Yen, &
Leveille, 2006). Cognitive function interacted with gait speed to determine disability in older adults (Kuo, Leveille, Yu, & Milberg, 2007). PAD was not examined in these studies, but the correlations found between PAD and CRP encourages study of the relationships between PAD and functional status.
31
General Health Perception
General health perception is a global self-perception of a person s overall outlook on their health status (Wilson & Cleary, 1995). Health status is reported to reflect morbidity more closely than health-related quality of life (Verbrugge,
Merrill, & Liu, 1999). The general health perception may reflect the person s attitude toward their health status more than their actual physical or mental status, since proxy measures do not correlate well. Proxy ratings on a five-point scale (from excellent to poor) by family members and health care workers in long-term care settings correlated poorly with the actual rating reported by the individual who typically self-rated the level higher (Andresen, Vahle, & Lollar,
2001b). Pell (1995) found a similar result with ratings on the global item of the
SF-36; subjects with PAD rated their general health higher than their surgeons
(kappa = 0.4) following an initial visit to determine interventions.
The Behavior Risk Factor Surveillance System (BRFSS), implemented at the state and territory level in the US has included general health perception as part of the CDC HRQOL-4 since 1993 (Zahran et al., 2005). The five levels are dichotomized as excellent-very good-good or fair-poor in much of the reporting
(Centers for Disease Control and Prevention, 2000). In a BRFSS report covering 1993 to 2001, poor and fair health combined were chosen by women
(15.5%), adults over age 75 (33.2%), Hispanics (21.9%), non-Hispanic blacks
(19.8%) widowed (31.3%), separated (23.1%), divorced (18.5%), less than high
School educated (34.6%), and lowest income (31.5%) groups more than by
32
other groups (Zahran et al., 2005). When the BRFSS 2001 data was compared to NHANES 2001-2002 data, the same gender, education and income trends were found (Zahran et al., 2005). Higher percentages of fair-poor ratings were also found for those persons with chronic medical conditions, especially those with multiple conditions: 13.6% for one condition to 75.4% with 5 chronic medical conditions (Zahran et al., 2005).
Health-Related Quality of Life
Health-related quality of life (HRQOL) is considered an important outcome measure (Liles, Kallen, Petersen, & Bush, 2006), especially in the past
20 years when a goal of Healthy People 2000 and 2010 has been improving quality of life (Healthy People 2000, http://www.cdc.gov/nchs/about/otheract/hp2000/hp2k.htm; About Healthy
People 2010, http://www.healthypeople.gov/About/goals.htm).
The SF-36 (Ware, Kosinski, & Dewey, 2000; Ware & Sherbourne, 1992) has been the HRQOL tool used most frequently in the United States with PAD patients, while in Europe the Nottingham Health Profile (Hunt, McEwen, &
McKenna, 1986; Hunt & McEwen, 1980) has been used more frequently.
Qualitative research has also been conducted to develop a more appropriate disease specific HRQOL tool (Treat-Jacobson et al., 2002). Several studies have examined HRQOL by comparing different levels of PAD (Hallin et al.,
2002; Klevsgard et al., 1999; Long et al., 2004; Pell, 1995) or interventions for
PAD (Currie et al., 1995; Thorsen, McKenna, Tennant, & Holstein, 2002;
33
Tretinyak, Lee, Kuskowski, Caldwell, & Santilli, 2001; Wann-Hansson, Hallberg,
Risberg, Lundell, & Klevsgard, 2005).
Pell (1995) reported on the measurement of health-related quality of life in 201 individuals with intermittent claudication before appointments with vascular surgeons to consider intervention. HRQOL showed a significant reduction from normal on six of the eight SF-36 sub-scales: physical role, physical functioning and bodily pain at p < 0.001; social functioning at p < 0.01; vitality and general health at p < 0.05 (Pell, 1995).
Hallin and colleagues (2002) conducted a study on 40 individuals with
PAD: divided equally between diagnosed intermittent claudication and diagnosed critical limb ischemia to determine the effects of different levels of disease and intervention on HRQOL. Physical functioning, physical role, and bodily pain were significantly lower than published age matched normal values in both groups, but the physical role and bodily pain were more greatly reduced for those with critical limb ischemia. The SF-36 was found to measure the individual s capabilities, but not their satisfaction with those capabilities (Hallin et al., 2002).
Long and colleagues (2004) compared HRQOL in individuals (n=108) found to have non-lifestyle limiting peripheral arterial disease (n=50), lifestyle limiting claudication (n=19) & limb-threatening ischemia (n=39) during an initial vascular clinic evaluation. The physical summary score average was significantly reduced in the 58 individuals with life-style limiting disease
34
compared to the 50 individuals with non-lifestyle limiting disease (24.3 + 1.02 &
30.7 + 1.18 respectively, p < 0.001), but the mental summary scores were similar. Ankle-brachial index correlated poorly with physical summary scores
(r=0.25) and mental summary score (r=0.01). The three traditional measures used by surgeons to determine quality of life: ankle-brachial index scores, reported symptoms and presence of comorbidities, were used in multiple regression analysis to determine the percentage of the SF-36 summary scores they predicted. Only 10% of the SF-36 mental summary score and 19% of the
SF-36 physical summary score were explained by ankle-brachial index, reported symptoms, and comorbidities combined, thus subject reported HRQOL was found important to measure (Long et al., 2004).
Currie and colleagues (1995) compared health-related quality of life
(HRQOL) among 186 individuals before and three months after revascularization surgery (n=34), angioplasty (n=74), or prescription of an unsupervised structured home exercise program (n=78). The interventions were not randomly assigned. Pre-intervention scores on physical function, physical role, social function and pain were all significantly lower for the revascularization group, but similar between the angioplasty and exercise groups. Post intervention, the angioplasty and revascularization groups showed significant improvement in all sub-scales, but the exercise group showed significant change (improvement) only in the pain sub-scale (Currie et al.,
1995).
35
Tretinyak and colleagues (2001) measured health-related quality of life
(HRQOL) using the SF-36 in 46 male US veterans with critical limb ischemia
(ABI < 0.4 in non-diabetic and < 0.6 in diabetic individuals) pre and post- operatively for revascularization surgery. Compared to normal scores for males over age 65, the veterans (average age 67, range 51-83 years) pre-operative scores were all significantly lower (p < 0.025). Post-operatively, only the physical function sub-scale showed a significant change, from 27.9 pre- operative to 33.7 post-operative (Wilcoxon test, p < 0.05).
Klevsgard and colleagues (1999) found that the Nottingham Health
Profile measured significantly reduced health-related quality of life in individuals with intermittent claudication (n = 93) or critical limb ischemia (n = 75) compared to disease free controls (n = 102). Kruskal-Wallace analysis indicated significant differences (p = 0.001) between the three groups for all dimensions, except social isolation (p = 0.004). Mann-Whitney U-tests indicated significantly worse HRQOL existed for individuals with critical limb ischemia than with intermittent claudication in the dimensions of pain, sleep and physical mobility.
Social isolation was only a problem for those with critical limb ischemia with the worst quality of life (the upper quartile) who would have been the least likely to be able to leave their homes without assistance (Klevsgard et al., 1999).
Thorsen and colleagues (2002) prospectively measured HRQOL at pre- operative baseline, 3 and 12 months in 48 individuals with critical limb ischemia
(25 with diabetes mellitus) scheduled for revascularization surgery. Pain and
36
sleep dimension values improved significantly (p < 0.05) from baseline to 3 months, but there was little additional improvement after the initial post- operative healing occurred (Thorsen et al., 2002). Interestingly, physical mobility and energy did not show a significant change typically reported on similar aspects of the SF-36.
Wann-Hansson and colleagues (2005) found that the improvements in pain reported on the Nottingham Health profile continued for the entire 4 years of the study in persons initially diagnosed with intermittent claudication or critical limb ischemia following revascularization surgery. Physical mobility had a slower decline for persons with initial intermittent claudication, but a rapid decline between 1 and 4 years following surgery for the critical limb ischemia group. Health-related quality of life was improved for longer and to a greater extent in persons with less initial disease (Wann-Hansson et al., 2005).
HRQOL has not been examined in PAD patients in a large national study, but the general HRQOL items in the CDC HRQOL-4 may not have provided the specificity the smaller studies were seeking. The CDC HRQOL-4 has been used for monitoring for disability in other chronic diseases though, therefore a study examining HRQOL in PAD patients using the NHANES data is needed.
Disability
Disability is defined by the CDC as an inability to work, do usual activities and contribute to society (Centers for Disease Control and Prevention, 2000).
37
The CDC HRQOL-4 are used to monitor disability with chronic diseases
(arthritis, diabetes, heart disease and several others, but not PAD) in the population by creating dual level items using the Physical-Mental summed score and Activity score for 30 days divided so 0-13 days is not disability and
14-30 days is disability (Centers for Disease Control and Prevention, 2000).
Disability is not always felt by the individual as it looks from the outside, so a person s perception of their disability can be a valuable measure and contribute to the knowledge we have about PAD persons reaction to the disease. A lack of disability is the goal of treatment allowing the person to have healthy days.
Summary
The progression from symptoms to functional status to general health perception in the model on which the conceptual framework is based was not completely supported by the literature, but the current conceptual framework was supported. The literature supports a possible mediator effect from the symptoms, function status, and general health perception grouping on the association between PAD (biological function) and disability (HRQOL). Some characteristics of the individual, chronic diseases and biological risks may have a moderator effect on the relationship between PAD and disability.
38
Specific Aims and Hypotheses
Specific Aim 1. To determine the association between PAD severity and disability in community dwellers 40 years and over using the NHANES 2001-
2004 data while adjusting for other factors.
Hypothesis 1. There will be a significant association between PAD
severity and disability alone and while controlling for:
1. Demographics, and
2. Socioeconomic status.
Specific Aim 2. To determine the potential mediator effects of symptoms, functional status, and general health perceptions on the association between
PAD severity and disability.
Hypothesis 2. There will be a significant mediator effect on the
association between PAD severity and disability from the respondent
perceptions of:
1. symptoms
2. functional status
3. general health.
Specific Aim 3. To determine the potential moderator effect of other chronic diseases and biological risks on the association between PAD severity and disability.
39
Hypothesis 3. There will be a significant moderator effect on the
association between PAD severity and disability from respondent other
diagnosed chronic diseases and biological risks of:
1. High tobacco exposure,
2. Diabetes diagnosis or elevated blood glucose,
3. Hypertension diagnosis or elevated systolic or diastolic blood
pressure,
4. Hyperlipidemia diagnosis or elevated total cholesterol or low
HDL cholesterol, and
5. Adiposity (obesity or large waist measurement).
Specific Aim 4. To determine the effect of integrating Models 1, Model 2, and Model 3 on the association between PAD severity and disability.
Hypothesis 4. There will be a significant effect on the association
between PAD severity and disability when integrating the variables
included in previous models.
40
CHAPTER 3
METHODS
Study Design
A descriptive cross-sectional design is used for this doctoral dissertation. The NHANES data was chosen, since it is a nationally representative data source for answering the specific aims of this study.
Description of the Data Source
The NHANES has been conducted periodically since the 1960 s but became continuous in 1999 with data released in 2-year cycles
(http://www.cdc.gov/nchs/nhanes.htm). The NHANES website provides access to the cycle specific survey questionnaires and operation manuals, examination and laboratory components, brochures, and consent documents
(http://www.cdc.gov/nchs/nhanes.htm). Approval for these questionnaires, age appropriate examinations, and laboratory testing was given to NHANES by the
CDC/ATSDR institutional review board of the Centers for Disease Control and
Prevention.
The NHANES uses a complex sampling design (multistage, probability, cluster and stratified design) for continuous collection and release of two-year
41
cross-sectional data sets. Respondents from chosen counties in the United
States participated in a detailed in-home face-to-face interview and at least one eligible person per household participated in a health examination specific for their age in a Mobile Examination Center (MEC). Trained interviewers administer the in-home questionnaires using detailed guidelines for item inclusion and response coding. The MEC provides a consistent and controlled environment for computer-assisted questionnaires about sensitive topics, physical examination and collection of laboratory samples. Data from this large
Nationally representative study is stripped of specific responder identifiers and is released free online after processing. Sampling weights and variance estimates are provided in the cycle specific demographics file to allow unbiased national estimation (http://www.cdc.gov/nchs/nhanes.htm).
Sample
The eligible sample of N= 6135 for this study were non-pregnant persons
40 years or older who completed both the interview and examination sections for NHANES 2001-2004 themselves (not a proxy responder). Inclusion criteria for the analysis sample included a) useable Ankle-Brachial Index from the mobile examination center data and b) completion of selected interview questions from health status and demographic questionnaires. Eligible sample persons (SP) with missing data on any of the inclusion criteria were summarily deleted from the final sample, since the main focus of the study could not be tested with these variables missing. A final sample of 4559 SP met the
42
Eligible Final Sample T-Test Variables Sample (N = 4559*) P values (N = 6135*) Age Mean 56.55 years 55.86 years < 0.0001 (+ 12.534) (+ 11.981) Age categories 40-60 66.4% 68.4% < 0.0001 61-85 33.6% 31.6% < 0.0001 Gender Male 47.6% 49.7% < 0.0001 Female 52.4% 50.3% < 0.0001 Race/Ethnicity Non-Hispanic white 78.3% 82.1% < 0.0001 Non-Hispanic black 10.1% 9.5% < 0.0001 Hispanic 8.6% 8.3% < 0.0001 (Other) (3.0%) Education Less than HS 18.9% 16.6% < 0.0001 HS or GED 26.2% 26.4% < 0.0001 More than HS 54.8% 57.0% < 0.0001 * N are the unweighted values. Mean Years and Percentages were determined and tested (T-tests) using weighted data.
Figure 3.1 Demographic and Socioeconomic Differences between Eligible
and Final Samples
inclusion criteria. A total of 1576 SP had to be removed from the eligible sample. Ankle-brachial index was not available or usable for 1034 SP because of unattainable blood pressures, sample person refusal, physical limitations
(casts, sores, or size of legs), time constraints limiting participation, or values that were invalid (higher than 1.5). Health status information was missing for 26 additional SP. Age, gender, and race/ethnicity were available for all eligible SP, but to ensure the statistical analysis would have adequate cell counts to provide
43
useful information about Non-Hispanic White, Non-Hispanic Black, and
Hispanic (Mexican American and other Hispanic) races, 101 other race SP were removed. Education (7 SP), poverty income ratio (306 SP), and marital status (2 SP) were missing, finally leaving 4559 SP for Specific Aim 1 with complete data. The N=1576 sample persons removed from the final sample were significantly older, more women, more racially diverse and with less education.
Some additional missing data occurred with variables for specific aim 2
(4 SP) and specific aim 3 (as low as 4390 SP), but the subjects are only eliminated from the analyses that include those variables. Completion of a) information from questionnaires on blood pressure, diabetes, medical conditions, physical activity, and physical function; b) mobile examination center data for body mass index, systolic and diastolic blood pressure readings
(reported to the sample person); and c) laboratory data for cotinine, glycosylated hemoglobin, total cholesterol and high-density lipoproteins will all be included in analyses of specific aims 2 or 3.
MEASURES
Data for the NHANES were collected in a variety of ways and locations.
Questionnaires were completed in face-to-face home interviews or computer assisted personal interviews (CAPI) within the mobile examination center
(MEC). The at home interview segment included questionnaires about demographic information, hypertension, hypercholesterolemia, diabetes,
44
medical conditions, physical activity, and physical functioning. Health status questions were administered using computer assisted personal interview
(CAPI) in the mobile examination center. Sociodemographic information included age, gender, race/ethnicity, education, and income information. A poverty income ratio (PIR) was a CDC calculation based on the household income divided by poverty threshold for that year. The physical examination occurred in a mobile examination center (MEC) and included measurements of height, weight, waist circumference, manual blood pressure, and ankle-brachial index. Venipuncture occurred in the MEC with serum isolated and frozen for later analysis at a variety of laboratory sites. Details about each variable and the reliability and validity of measurement will be addressed later in this chapter.
The variables are listed in the Table of Measures but greater detail is given in the text.
Dependent variable
The Dependent variable in this study is disability related to PAD.
Disability is a restriction or limit on the ability to work and maintain usual activities of life that limits quality of life. Disability is operationally defined using items 2-4 from the 4-item CDC Measuring Healthy Days questionnaire (Centers for Disease Control and Prevention, 2000). The Healthy Days scale was developed by a working unit from the CDC as a way to measure HRQOL in population surveys (Centers for Disease Control and Prevention, 2000). The
MOS SF-36 (Ware & Sherbourne, 1992) and the Quality of Well-Being scale
45
(Kaplan & Bush, 1982) were considered too long for use in public health surveillance for phone interviews, so the 3-item parsimonious Healthy Days approach was adopted (Hennessy, Moriarty, Zack, Scherr, & Brackbill, 1994;
Moriarty, Zack, & Kobau, 2003; Verbrugge et al., 1999). The three questions are listed in Figure 3.2.
2. Thinking about your physical health, which includes physical illness and injury, for how many days during the past 30 days was your physical health not good?
3. Now thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good?
4. During the past 30 days, for about how many days did poor physical or mental health keep you from doing your usual activities, such as self-care, work, school or recreation?
Figure 3.2 Healthy Days Questions 2 through 4 for Self-responders
Items 2, 3 and 4 each have a range of 0-30 days when health was not good or activity was limited. Items 2 and 3 are tallied together and a score of any days when physical or mental health was not good up to a total of 30 days is calculated. Mean number of unhealthy days for persons in the general public were reported as ranging from 5.1 at age 40 to 6.7 days at age 75 or greater
46
(Centers for Disease Control and Prevention, 2000). Disability for several chronic diseases is monitored at the state and territory level by the Behavioral
Risk Factor Surveillance System (BRFSS) using the CDC HRQOL-4 in population surveys (Centers for Disease Control and Prevention, 2000; Moriarty et al., 2003). No disability is recorded for 0 to 13 days, but 14 to 30 days is recorded as disability for questions 2 through 4 or the combined question 2 and
3 (Centers for Disease Control and Prevention, 2000). The same operational definition of disability for Physical, Mental, Combined physical and mental and
Activity will be used in this study: 0-13 days = no disability and 14-30 days = disability.
Reliability and validity have been well established during the extensive use of this scale in state and national health monitoring through the BRFSS
(Andresen, Catlin, & Wyrwich, 2001a; Moriarty & Zack, 1999). The items were first included in the NHANES in 2000, but data reporting started with the 2001-
2002 and 2003-2004 cycles (http://www.cdc.gov/nchs/nhanes.htm).
Independent variables
Demographics and Socioeconomic status. Demographics are age, gender, race/ethnicity and marital status of the respondents as summarized in the Table of Measures. Age is a continuous variable from 40 to 85 years. The ages of respondents over 85 were truncated in the available NHANES data to
47
Concepts Operational NHANES Coding Definitions File Sources (in analysis) Dependent Variables Disability Healthy Days Questionnaires - 0-30 continuous as unhealthy days Health Status and - Physical (P) (HSQ) Disability - Mental (M) categories of: - Combined 0= 0-13 days (P+M) 1=14-30 days - Activity Independent variables Peripheral Peripheral Arterial Examination - Into 3 levels of Arterial Disease severity LEXABPI arterial blockage: Disease Severe= <0.70 Ankle-Brachial Index Mild= 0.70-0.90 (indicates arterial Negligible= blockage) 0.90-1.5 Demographics Age Questionnaires Continuous 40-85 Demographics Groupings: (DEMO) 0 = 40-60 1 = 61-85 Race/Ethnicity 1= Non-Hispanic White 2= Non-Hispanic Black 3= Hispanic and Mexican American Gender 0=MALE 1=FEMALE Marital Status 1=married /partner 2= never married 3= widowed /divorced /separated
Figure 3.3 Table of Measures Continued
48
Figure 3.3 (continued)
Concepts Operational NHANES Coding Definitions File Sources (in analysis) Socioeconomic Education Questionnaires 1=< high school Status Demographics 2= HS graduate (DEMO) or GED 3= more than HS Poverty Income Ratio 0-4.99 = actual ratio 5 = 5 or more
Pain Symptoms Lesions or ulcers, Questionnaires 0= no pain or tingling in Diabetes (DIQ) 1= Yes to pain/ feet, and pain in legs tingling in feet or or calves with walking legs/calves pain Physical Usual daily physical Questionnaires 1 = sitting activity activity level Physical Function 2 = light (PFQ) 3 = moderate 4 = heavy Vigorous 10 minutes of 0 = yes activity vigorous activity in 1 = not done past 30 days 2 = Unable to do Mobility Difficulty walking a Questionnaires 0= no difficulty quarter mile, up 10 Physical Function 1= difficulty or steps without stopping (PFQ) unable to do any or walking from one of the 3 items room to another on the same level. General Health Perception Questionnaires - 1= negative (poor Global rating of overall health as Health Status or fair rating) excellent to poor. (HSQ) 0= positive (good, very good or excellent) Diabetes Questionnaires - Yes= yes to any Told you have Diabetes (DIQ) of the questions. Diabetes diagnosis? Take insulin or pill? No= no to all
Continued
49
Figure 3.3 (continued)
Concepts Operational NHANES Coding Definitions File Sources (in analysis) Chronic Hypertension or Questionnaires - Yes= yes to disease Hypercholesterolemia Blood Pressure either question (diagnosed) Told diagnosis? (BPQ) No= both No Take medication? Cardiac and Vascular Questionnaires Yes= yes to any Diseases Medical Condition of the questions. Congestive failure (MCQ) Coronary disease No= no to all Angina/ pectoris Heart attack Stroke Cotinine Laboratory 0= low exposure 06 - Cotinine (<15 ng/ml) 1= high exposure (> 15.0 ng/ml) Glycosylated Laboratory 0= 7% or less Hemoglobin 10 - (L10) 1= >7% Total Cholesterol Laboratory - 0= < 240 mg/dl 13 - Cholesterol 1= > 240 mg/dl HDL Cholesterol 0= > 40 + male (mg/dl) > 50 + female (HDL = high density 1= <40 + male lipoproteins) <50 + female Body Mass Index Examination - 0= < 30 Kg/m(squared) (BMX) 1= >30 (regression) (descriptive statistics) 0= < 25 1= 25 - 29.9 2= > 30 Waist circumference 0=<102cm (male) <88cm (female) 1= >102cm(male) >88 cm (female) Systolic Blood Examination - 0= <140 mmHg Pressure (BPX) 1= >140 mmHg Diastolic Blood 0= <90 mmHg Pressure 1= > 90 mmHg
50
85 years, since subject identification was likely at older ages. Gender is male and female. Race/ethnicity is at three levels, since the percentage of respondents drops greatly after the first three categories: non-Hispanic White, non-Hispanic Black, and Hispanic. Marital status six levels can be seen in
Appendix A, but these were combined to three levels for data analysis: a) married or partner, b) never married, and c) widow, divorced, or separated.
Socioeconomic status includes education and poverty income ratio. Education is operationally defined as 3 levels; a) less than high school diploma ( (HS/GED), and c) some college or vocational school beyond high school (>HS). The poverty income ratio is a CDC calculation of the actual household reported income divided by the poverty level for that year which yields a continuous number (to two decimal places) up through 5 times the poverty level. The number of cases drops significantly at a PIR of 5; so 5 or greater were truncated by the CDC for confidentiality purposes. Peripheral arterial disease severity, a decrease in blood flow to the lower extremities, is the main independent variable in this study. The operational definition will be an ankle-brachial index measure below 0.7 on either leg for severe arterial blockage (severe PAD when symptoms are most likely to occur), 0.70 to 0.90 for mild arterial blockage (mild PAD usually asymptomatic), and higher than 0.90 for negligible arterial blockage when no PAD is indicated (Hirsch et al., 2006; Treat-Jacobson & Walsh, 2003). ABI 51 higher than 1.50 are invalid and those cases have been dropped because of this exclusion criteria. Ankle-brachial index (ABI) is an indirect comparison of central to lower limb perfusion that indicates the degree of PAD in that limb with 95% sensitivity and specificity of 97.3% at rest and 100% after exercise (Treat- Jacobson & Walsh, 2003; Yao, Hobbs, & Irvine, 1969). Reproducibility of the ABI is high when the procedure is done by the same examiner (r = 0.967 [95% CI 0.957, 0.975]) or different examiners (r = 0.908 [95% CI 0.88, 0.93]). ABI measurement for the NHANES was done in the mobile examination center (MEC) by a group of trained vascular examiners. Ankle-brachial index was measured in the MEC by placing a blood pressure cuff on the right (or left) upper arm and both calves for the measurement of systolic blood pressure. The systolic pressure was measured in the brachial and posterior tibial pulse locations twice for each site for SP younger than 60 and the mean was obtained. If only one pressure was obtained at a site, that pressure was used in all calculations. The computer calculated the index by dividing the mean posterior tibial systolic by the mean brachial systolic for each leg resulting in a right ABI and a left ABI. The minimum ABI for each SP was used in the analysis, since that would indicate the leg with the greatest arterial blockage (most severe PAD). Chronic disease diagnosis of diabetes mellitus, hypertension, and/or hyperlipidemia was noted. Questions about diagnosis of health conditions and prescription medication usage were self-reported by respondents during the in 52 home interview. Since individuals may not always recognize a diagnosis by the words used in a survey, the self-reported use of prescription medications appropriate to each diagnosis was considered evidence of a diagnosed co- morbidity (See Appendix A for specific NHANES items and coding). For analysis purposes, if a yes answer to diagnosed diabetes mellitus, insulin, or oral diabetic medication usage was given, that respondent was scored as diabetic. If hypertension, hyperlipidemia or medications for either of these disorders were reported, the respondent was scored as positive for hypertension or hyperlipidemia. For descriptive purposes, if congestive heart failure, coronary artery disease, angina (or angina pectoris) or heart attack diagnosis were reported, the respondent was recorded as having cardiovascular disease and stroke was reported alone. Biological risks indicate level of tobacco exposure and/or disease risk. Each biological risk was a continuous measure in the NHANES data, but each was categorized at the risk cut points for analysis. Tobacco exposure was operationally defined as serum cotinine level in nanograms per milliliter (ng/ml). The higher the cotinine level the greater the risk even from second hand smoke (Wortley, Caraballo, Pederson, & Pechacek, 2002). The Society for Research on Nicotine and Tobacco guidelines for high cotinine exposure were used and are defined as greater than 15 ng/ml, which was consistent with smoking or tobacco use (Society for Research on Nicotine and Tobacco, 2002). Diabetes risk was operationally defined as serum glycosylated hemoglobin level greater 53 than 7%, since that was the cut-off found to prolong the development of PAD complications (Diabetes Control and Complications Trial, 1996; Marso & Hiatt, 2006). Hpertension risk was operationally defined as an average systolic blood pressure (reported) higher than 140 millimeters of mercury (mmHg) or an average diastolic blood pressure (reported) higher than 90 mmHg from recommendations of the seventh report of the Joint Commission on High Blood Pressure prevention and treatment (Chobanian et al., 2003). Recommendations for lower blood pressures are 135/85 mmHg for metabolic syndrome prevention (Grundy et al., 2004) and 130/80 mmHg for high risk of coronary artery disease (Rosendorff et al., 2007), but those values only applied to a limited part of the sample. Hypercholesterolemia risk was operationally defined as serum total cholesterol higher than 240 milligram per deciliter (mg/dl) or serum high-density lipoprotein cholesterol (HDL-C) below 40 mg/dl for males and below 50 mg/dl for females (National Heart Lung and Blood Institute, 2002). Lower total cholesterol and higher HDL-C levels are recommended for persons at high risk of cardiovascular disease, but the higher (lower) cut points were used in this investigation, since most of the SPs were not at cardiac risk. Low density lipoproteins and triglycerides were only available on NHANES SPs who had been fasting when the sample was taken, so these variables with greatly reduced sample sizes were not used in this investigation. Adiposity was 54 operationally defined as body mass index (BMI) greater than 30 kilograms per meter squared (obesity) and/or waist measure greater than 102 centimeters (cm) for men or 88 cm for women (National Heart Lung and Blood Institute, 2002). Symptoms are any reported discomfort (Wilson & Cleary, 1995). Symptoms were operationally defined as SPs self-reported yes response to any of 4-items selected from the Diabetes questionnaire; pain in the legs or calves with walking, numbness and tingling in the feet or ulcer on a foot (see Appendix A for specific wording). Functional Status is the ability to complete required or desired physical, social, or psychological tasks (Wilson & Cleary, 1995). Typical physical activity was noted, since the usual level of activity may determine an individuals perception of related disability (Ekers, 1986). Three items from the physical activity questionnaire were included. Usual level of daily activity ranging from sitting all day to very strenuous activity was reported in descriptive statistics. The other 2 items asked about 10 minutes of moderate or vigorous activity within the past 30 days (details in Appendix A). Respondents used CAPI to complete the items during the home interview. Only vigorous activity was entered into analysis, since collinearity was likely between these measures. Since PAD involves the lower extremities, mobility was chosen as the indicator of functional status. The operational definition of mobility for this study ws the self-reported difficulty to 1) walk a quarter mile, 2) walk up 10 steps 55 without resting, and/or 3) walk from room to room on the same level. The 3-item mobility indicators from the Physical Function CAPI questionnaire were administered in the home interview and answered by the respondent. Possible responses ranged from no difficulty to do not do this activity (see Appendix A, PFQ061 for details). Limited functional status was indicated by an answer of some difficulty, much difficulty or unable to do any of the three mobility items (see Table of Measures). If the respondent answered yes to a screening question about requiring special equipment to walk, the first two items were skipped in the question sequence. Respondents that answered yes to the screening question were coded as having limited functional status for this analysis. General Health Perception was the subjective interpretation of input from all other sources in the model. The Center for Disease Control and Prevention (CDC) global measure of general health perception was chosen from the Medical Outcome Survey Short Form 36-item (Ware & Sherbourne, 1992) when the health-related quality of life 4-item core measure (CDC HRQOL-4) was originally developed for use in surveillance of public health over time (Hennessy et al., 1994; Moriarty et al., 2003). The question reads Would you say your health in general is excellent, very good, good, fair, or poor? (Centers for Disease Control and Prevention, 2000, p. 10). BRFSS surveillance dichotomizes this variable as excellent-very good-good = good health 56 perception and fair-poor = poor health perception (Zahran et al., 2005), which was used in this investigation also. Data Analysis All statistical analysis was completed using SAS version 9.1 (SAS Institute Inc., Cary, NC, USA). WTMEC2YR is the NHANES weighting variable (see Appendix A). The 2001-2004 four year estimate uses one-half of the weighting for each of the 2-year cyles, so weight was adjusted by the equation: WT4yrMEC = WTMEC2YR x 0.5 (Analytic and Reporting Guidelines, http://www.cdc.gov/nchs/about/major/nhanes/nhanes2003- 2004/analytical_guidelines.htm). All statistical analysis incorporated the survey techniques for complex sampling designs (weighting, stratum, and cluster variables). Descriptive statistics were used to present the general characteristics of the respondents. Means and standard deviations or frequencies appropriate for the level of the variables were presented to characterize the final sample and three levels of PAD: no disease indicated, mild and severe PAD (Figure 4.1 to 4.5). Chi square analysis of weighted frequency proportions were used to compare the PAD groups for each variable and reported in the tables. The three PAD level means were compared when appropriate with post-hoc comparisons. Correlations were run of all ordinal, interval, and continuous variables with all others to look for potential collinearity. Linearity was measured for continuous variables used in the logistic regression. 57 Model 1 Specific Aim 1. To determine the association between PAD severity and disability, as measured by health-related quality of life in community dwellers 40 years and over using the NHANES 2001-2004 data while adjusting for other factors. Hypothesis 1. There will be a significant association between PAD severity and disability alone and while controlling for: 1. Demographics, and 2. Socioeconomic status. Order of Model 1 Analysis: 1. Logistic regression was run using the dual disability form of the HRQOL variables (physical, mental and activity) as the dependent variable by the three-level categorical ABI to determine the overall association between HRQOL disability and PAD. The highest values for ABI were used as the reference (ABI > 0.9 = no disease indicated), so the severe and mild PAD CRUDE odds ratios were examined for effect. If no effect was found for one or more of the disability variables for both levels of the ABI that disability variable was eliminated from further analysis. 2. Logistic regression was run with each demographic (age, gender, race/ethnicity and marital status), and each socioeconomic (SES: education and poverty income ratio) variable added one at a time with the 58 disability and ABI variables from the CRUDE model (unadjusted for other variable influence). 3. The percentage change from the ABI CRUDE odds ratio was calculated for each variable separately. Variables that cause a 10% or greater change in at least one of the ABI CRUDE odds ratios were considered for retention as confounders (spurious or mediators) or effect moderators (Hosmer & Lemeshow, 2000). Variables with the strongest theoretical support for inclusion were considered first. The variable which caused the greatest change overall had high consideration for retention as a confounder with ABI initially. Each of the other variables was entered in a logistic regression one at a time with ABI and the confounder variable as independent variables. 4. The percentage change from the ABI CRUDE odds ratios and the percentage change from the ABI ADJUSTED odds ratios (due to the addition of the first confounder) was calculated and examined for effect. Variables that caused a 10% or greater change in at least one of the ABI ADJUSTED odds ratios were considered for retention in the model as additional confounders. 5. Variables were added to the model in this way until there were no longer any variables that caused a 10% or greater change in at least one of the ABI ADJUSTED odds ratios. 6. The confounder variables were then tested as effect moderators. An interaction variable was created between ABI and each confounder. The 59 interaction variables were entered separately into the established model to test their interaction effect using the Wald Chi square Type 3 analysis of effect. A conservative p-value of 0.01 or less was considered significant for interaction effect and that interaction variable was included in the model as an effect moderator. Added variables were either a confounder or an effect moderator. 7. Model 1 retained a disability variable, ABI levels, confounders and effect moderators with separate models for Physical, Mental and Activity Disability as appropriate. The Model 1 confounders were included in the integrated Model 4 development. Model 2 Specific Aim 2. To determine the potential mediator effects of symptoms, functional status, and general health perceptions on the association between PAD severity and disability. Hypothesis 2. There will be a significant mediator effect on the association between PAD severity and disability from the respondent perceptions of: 1. symptoms 2. functional status 3. general health. 60 Order of Model 2 Analysis: 1. Logistic regression was run with each symptom: leg pain, calf pain, or foot pain with walking; functional status: mobility difficulty, physical activity level, and vigorous activity (ability to do at least 10 minutes in the past 30 days); and general health status (dichotomous) variable added one at a time with the disability and ABI variables from the CRUDE model (unadjusted for other variable influence). 2. The percentage change from the ABI CRUDE odds ratio was calculated for each variable separately. Variables that caused a 10% or greater change in at least one of the ABI CRUDE odds ratios were considered for retention as confounders. Variables with the strongest theoretical support for inclusion were considered first. The variable which caused the greatest change overall had highest consideration for retention as a confounder (mediator or spurious) with ABI initially. Each of the other variables was entered in a logistic regression one at a time with ABI and the confounder variable. 3. The percentage change from the ABI CRUDE odds ratios and the percentage change from the ABI ADJUSTED odds ratios (due to the addition of the first confounder) were calculated and examined for effect. Variables that caused a 10% or greater change in at least one of the ABI ADJUSTED odds ratios were considered for retention in the model as confounders. 61 4. Variables were added to the model in this way until there are no longer any variables that cause a 10% or greater change in at least one of the ABI ADJUSTED odds ratios. 5. The variables to be retained were then tested as effect moderators. An interaction variable was created between ABI and each confounder. The interaction variables were entered separately into the established model to test their interaction effect using the Wald Chi square Type 3 analysis of effect. A conservative p-value of 0.01 or less was considered significant for interaction effect and that interaction variable was included in the model as an effect moderator if the model remained stable. If the model developed extremely wide confidence intervals for the odds ratios, the variable was not trusted and the interaction variable was not used in the model. 6. Model 2 retained a disability variable, ABI levels, confounders and effect moderators with separate models for Physical, Mental and Activity Disability as appropriate. The Model 2 confounders were included in the integrated Model 4 development. Model 3 Specific Aim 3. To determine the potential moderator effect of other chronic diseases and biological risks on the association between PAD severity and disability. 62 Hypothesis 3. There will be a significant moderator effect on the association between PAD severity and disability from respondents biological risks and diagnosed chronic diseases of: i) High nicotine exposure, ii) Diabetes diagnosis or elevated blood glucose, iii) Hypertension diagnosis or elevated systolic or diastolic blood pressure, iv) Hyperlipidemia diagnosis or elevated total cholesterol or low HDL cholesterol, and v) Adiposity (obesity or large waist measurement). Order of Model 3 Analysis: 1. Biological risk variables were dichotomized at cut points suggested from the literature for testing odds ratios of elevated levels with the suggested normal as the indicator. Cotinine was cut at 15 ng/ml to indicate elevated nicotine exposure, glycosylated hemoglobin was cut at 7% to indicate consistently elevated glucose (or one very extreme recent episode), SBP was cut at 140 mmHg, DBP was cut at 90 mmHg, total cholesterol was cut at 240 mg/dl, HDL was cut at 40 mg/dl for males and 50 mg/dl for females (with the high level as the reference), body mass index was cut at 30 KG/m(sq) to indicate obesity, and waist measure was cut at 88 cm for females and 102 cm for males to indicate abdominal adiposity. 63 2. Logistic regression was run with each biological risk (cotinine, glycosylated hemoglobin, systolic and diastolic blood pressure, total cholesterol, HDL, obesity, and waist measurement) and selected chronic disease (diabetes, hypertension, and hyperlipidemia) variable added one at a time with the disability and ABI variables from the CRUDE model (unadjusted for other variable influence). 3. The percentage change from the ABI CRUDE odds ratio was calculated for each variable separately. Variables that caused a 10% or greater change in at least one of the ABI CRUDE odds ratios were considered for retention as confounders. Variables with the strongest theoretical support for inclusion were considered first. The variable which caused the greatest change overall had highest consideration for retention as a confounder (mediator or spurious) with ABI initially. Each of the other variables was entered in a logistic regression one at a time with ABI and the confounder variable. 4. The percentage change from the ABI CRUDE odds ratios and the percentage change from the ABI ADJUSTED odds ratios (due to the addition of the first confounder) were calculated and examined for effect. Variables that cause a 10% or greater change in at least one of the ABI ADJUSTED odds ratios were considered for retention in the model as confounders. 64 5. Variables were added to the model in this way until there were no longer any variables that caused a 10% or greater change in at least one of the ABI ADJUSTED odds ratios. 6. The confounder variables were then tested for effect moderation. An interaction variable was created between ABI and each confounder. The interaction variables were entered separately into the established model to test their interaction effect using the Wald Chi square Type 3 analysis of effect. A conservative p-value of 0.01 or less was considered significant for interaction effect and that interaction variable was included in the model as an effect moderator if the model remained stable. If the model developed extremely wide confidence intervals for the odds ratios, the variable was not trusted and the interaction variable was not used in the model. 7. Model 3 will retain a disability variable, ABI levels, confounders and effect moderators with separate models for Physical, Mental and Activity Disability as appropriate. The Model 3 confounders were included in the integrated Model 4 development. Model 4 Specific Aim 4. To determine the effect of integrating Models 1, Model 2, and Model 3 on the association between PAD severity and disability. Hypothesis 4. There will be a significant effect on the association between PAD severity and disability when integrating the confounders from previous models. 65 Order of Model 4 Analysis: 1. Variables that were confounders in Model 1, Model 2, or Model 3 were used to develop an integrated Model 4. The adjusted odds ratio information from the Logistic regressions for Step 1 in previous models was collated for these variables. No interaction variables were included, since the model development process restarted for this group of variables and identifying confounders is the first step. 2. The calculated percentage change from the ABI CRUDE odds ratio was compared and variables that caused a 10% or greater change in at least one of the ABI CRUDE odds ratios were considered for retention as confounders. The variable which caused the greatest change overall had highest consideration for retention as a confounder in Model 4 initially. Each of the other variables were entered one at a time with ABI and the first confounder variable. 3. The percentage change from the ABI CRUDE odds ratios and the percentage change from the ABI ADJUSTED odds ratios (due to the addition of the first confounder) were calculated and examined for effect. Variables that caused a 10% or greater change in at least one of the ABI ADJUSTED odds ratios were considered for retention in Model 4 as confounders. 66 4. Variables were added to the model in this way until there were no longer any variables that cause a 10% or greater change in at least one of the ABI ADJUSTED odds ratios. 5. The confounders were tested for effect moderation. An interaction variable was created between ABI and each confounder. The interaction variables were entered separately into the established Model 4 to test their interaction effect using the Wald Chi square Type 3 analysis of effect. A conservative p-value of 0.01 or less was considered significant for interaction effect and that interaction variable was included in the model as an effect moderator if the model remained stable. 6. Model 4 retained a disability variable, ABI levels, confounders and effect moderators with separate models for Physical, Mental and Activity Disability as appropriate. 67 CHAPTER 4 RESULTS Descriptive Statistics Study data from NHANES 2001-2004 are described in Figures 4.1 to 4.5. The tables contain columns of information across with Final sample and 3 levels of PAD: no PAD indicated (ABI 0.91-1.50), mild PAD (ABI 0.70-0.90) and severe PAD (ABI < 0.70). The demographic (age, gender, race/ethnicity and marital status) and socioeconomic (education and poverty income ratio) characteristics from Specific Aim 1 are listed in Figure 4.1. Figure 4.2 contains variables introduced in Specific Aim 2: symptoms (3 types of lower extremity pain), functional status (2 physical activity items and a physical mobility item), and general health perception (dual). The biological risk variables are in Figure 4.3: serum cotinine, glycosylated hemoglobin, systolic and diastolic blood pressure, HDL, total cholesterol, body mass index and waist measurement. Prevalence of diabetes, hypertension, hyperlipidemia, cardiac disease and stroke are presented in Figure 4.4. Figure 4.5 presents the health-related quality of life variable means for Physical, Mental, and Activity unhealthy days in 30 days. Wald chi square test adjusted F value, degrees of freedom (df, 68 numerator, denominator), and p-value were used to statistically compare the 3 levels of PAD by each variable level using a survey design adjustment. Demographic and socioeconomic (Figure 4.1). The weighted age in the final sample contained more cases age 40 to 60 (68.4%, 95% CI 66.4%,70.4%) than age 61-85 (31.6%, 95% CI 29.6%, 33.6%). The percentages reversed in the mild PAD group (ABI 0.7-0.9) and progressed for severe PAD (ABI < 0.70). Women had a higher prevalence of mild PAD (62.7%), and men had a higher prevalence with severe PAD (53.6%). Non- Hispanic blacks had a progressively higher prevalence of PAD as the ABI dropped (9.2%, 12.7%, and 17.7%). More than half of the cases were married or partnered in all columns, but the prevalence of widowed, divorced or separated weighted cases almost doubled from 22.6% to 42.2% in the severe PAD group. Almost twice as many (39.6%) had not completed high school in the severe PAD group. The Wald Chi square adjusted F-value (numerator and denominator degrees of freedom (df), and p-value) was used to statistically compare the groups using a survey design adjustment. The test results indicated significant differences among: age categories (F 46.96, df 2,29, p <0.0001), gender (F 3.81,df 2,29, p=0.03), race/ethnicity (F 4.40, df 4,27, p = 0.0075), marital status 69 Final Ankle-Brachial Index Sample 0.91-1.50 0.70-0.90 0.28-0.69 F-value (No PAD) (mild) (severe) (df) N=4559 N=4147 N=283 N=129 p-value Weighted % 100% 93.9% 4.6% 1.5% Age categories 46.96 40-60 68.4% 70.8% 35.7% 17.0% (2,29) 61-85 31.6% 29.2% 64.3% 83.0% <0.0001 Gender 3.81 Male 49.7% 50.2% 37.5% 53.6% (2,29) Female 50.3% 49.8% 62.5% 46.4% 0.03 Race/Ethnicity Non-Hispanic white 82.2% 82.3% 82.2% 76.4% 4.40 Non-Hispanic black 9.5% 9.2% 12.5% 17.6% (4,27) Hispanic 8.3% 8.5% 5.3% 6.0% 0.0075 Marital Status Married/ partner 70.9% 71.6% 60.5% 54.4% 4.09 Never married 5.7% 5.8% 4.7% 3.4% (4,27) Widowed /Divorced 23.4% 22.6% 34.8% 42.2% 0.01 /Separated Education Less than HS 16.6% 16.0% 20.5% 39.6% 8.54 HS or GED 26.4% 26.3% 26.7% 29.6% (4,27) More than HS 57.0% 57.7% 52.8% 30.8% 0.0001 PIR (rounded) 0.00-0.99 (lowest) 8.8% 8.5% 12.3% 16.4% 1.00-1.99 17.4% 16.7% 26.4% 34.1% 11.86 2.00-2.99 15.5% 15.0% 23.1% 20.9% (10,21) 3.00-3.99 15.0% 15.3% 9.0% 12.1% <0.0001 4.00-4.99 13.4% 13.8% 7.8% 8.2% 5.00 or more 29.9% 30.7% 21.4% 8.3% Means and 95% confidence intervals Age 55.9 55.1 65.3 72.0 (55.2, 56.4) (54.6, 55.7) (63.5, 67.2) (67.6, 76.5) PIR 3.3 3.3 2.7 2.4 (poverty income ratio) (3.2, 3.4) (3.2, 3.5) (2.4, 3.1) (2.0, 2.7) Figure 4.1 Weighted Demographic and Socioeconomic Prevalence in NHANES 2001-2004 data for Final Sample by Three Ankle-Brachial Index levels 70 (F 4.09, df 4,27, p=0.01), education (F 8.54, df 4,27, p=0.0001) and categorized PIR levels (F 11.86, df 10,21, p <0.0001) across the 3 PAD levels. Those with mild PAD tend to be a decade older, female, non-Hispanic, widowed, divorced, separated, slightly lower educated and have less income than those with no PAD. Those with severe PAD tended to be older, male, widowed, divorced, separated, non-Hispanic black, less educated and have much lower income than those with no PAD. Symptoms, Functional Status, and General Health Perception (Figure 4.2). Pain in the legs with walking was present in one-fourth of the overall sample, but the prevalence rose to one-third of the cases in the mild PAD (0.70-0.90 ABI) and half of the cases in the severe PAD group (ABI < 0.70). Calf pain with walking was less common, but doubled to one-fourth of the cases in the mild PAD (0.70-0.90 ABI) and 42.8% (95% CI 28.9%, 56.8%) of the cases in the severe PAD group (ABI < 0.70). Wald chi square adjusted F values were significant for leg pain (8.5653, df 2, 29, p = 0.0012) and calf pain (13.7, df 2, 29, p < 0.0001). Differences in foot pain were not significant. Weighted cases showed less usual physical activity and less tolerance of vigorous activity as the ABI decreased. Sitting was a usual daily activity for one- fourth of the overall final sample, but increased to one-third of the cases in the 0.70-0.90 ABI and half of the cases in the lowest ABI group (< 0.70). Moderate or heavy activity is usual for only one-fifth of the overall sample, but dropped to 6% (95% CI 1.8%, 11.2%) in the lowest ABI group. Vigorous activity for 10 71 Final Ankle-Brachial Index Sample 0.91-1.50 0.70-0.90 0.28-0.69 F-value (No PAD) (mild) (severe) (df) N=4559 N=4147 N=283 N=129 p-value Weighted % 100% 93.9% 4.6% 1.5% Symptoms (df 2,29) Leg pain with F 8.5653 walking 24.8% 23.9% 34.5% 51.3% p=0.0012 Calf pain with F 13.7 walking 12.9% 11.9% 23.7% 42.8% p<0.0001 Pain or tingling in NS foot 11.2% 10.9% 17.1% 16.5% (p = 0.12) Functional Status Physical Activity Sitting 25.2% 24.4% 33.4% 50.8% 15.53 Light 52.8% 53.0% 52.3% 43.1% (2, 29) Moderate 16.8% 17.2% 13.0% 5.2% <0.0001 Heavy 5.2% 5.4% 1.3% 0.9% (usual daily level) Vigorous Activity >10 minutes in 28.1% 29.3% 12.8% 4.2% 15.92 past 30 days (4, 27) No 67.3% 66.6% 78.8% 72.7% <0.0001 unable to do 4.6% 4.1% 8.4% 23.1% Mobility difficulty 41.65 20.6% 18.8% 42.9% 69.0% (2,29) <0.0001 General Health Perception 8.4 Good health 81.9% 82.6% 73.9% 58.2% (2,29) Poor health 18.1% 17.4% 26.1% 41.8% 0.0013 Figure 4.2 Weighted Symptom, Functional Status, and General Health Perception Prevalence in NHANES 2001-2004 data for Final Sample by Three Ankle-Brachial Index levels 72 minutes in the past 30 days was not done in a majority of all cases, but the prevalence of doing vigorous activity and inability to do vigorous activity reversed between the highest and lowest ABI groups. Wald chi square adjusted F values were significant for usual physical activity (15.53, df 2, 29, p < 0.0001) and vigorous activity (15.92, df 4, 27, p < 0.0001). Mobility difficulty rose significantly as the ABI decreased (Wald adjusted F value 41.65, df 2,29, p < 0.0001). Good general health perception dropped from 82.6% (95% CI 80.4%, 84.9%) to 58.2% (95% CI 42.2%, 74.2%) as the ABI decreased (Wald adjusted F value 8.4, df 2,29, p = 0.0013). The change in prevalence of symptoms, activity, mobility difficulty and health perception began dramatically at the 0.70- 0.90 ABI level where the mild PAD is typically asymptomatic. Biological Risk and Select Chronic Diseases. Most biological risk variables have some missing data, so the N for each variable is listed in each cell at the top left side. Published cut points for risk are used as the basis for the categorical variables in Figure 4.3. Serum cotinine mean levels are well above the 15 nanogram per milliter (ng/ml) cut point that indicates heavy tobacco use for all levels of ABI despite about 75% of individuals in the groups measuring below the 15 ng/ml cut point. The overall mean cotinine was 60.4 ng/ml (95% CI 52.8, 68.0 ng/ml). Mild PAD mean cotinine was 63.6 ng/ml (95% CI 43.5, 83.6 ng/ml) and 77.7 ng/ml (95% CI 46.9, 108.5 ng/ml) for severe PAD (a much wider CI due to smaller group size). Glycosylated hemoglobin means were all below the 7% cut point, so the 17% to 27% (95% CI 9.0%, 26.3% and 17.2%, 73 Final Ankle-Brachial Index Sample 0.91-1.50 0.70-0.90 0.28-0.69 (No PAD) (mild) (severe) p-values Total N N=4559 N=4147 N=283 N=129 Weighted % 100% 93.9% 4.6% 1.5% serum cotinine (N) (4390) (4003) (264) (123) NS Negligible or mild 75.2% 75.2% 75.4% 70.3% (0.5688) heavy (>15.0ng/ml) 24.8% 24.8% 24.6% 29.7% glycosylated (N) (4454) (4057) (272) (125) hemoglobin NS < 7% 94.9% 95.1% 93.1% 90.7% (0.2192) > 7% 5.1% 4.9% 6.9% 9.3% systolic BP (N) (4447) (4047) (275) (125) < 140 mmHg 78.4% 79.6% 61.3% 53.7% < 0.0001 > 140 mmHg 21.6% 20.4% 38.7% 46.3% diastolic BP (N) (4447) (4047) (275) (125) NS < 90 mmHg 92.6% 92.4% 94.9% 95.5% (0.1747) > 90 mmHg 7.4% 7.6% 5.1% 4.5% total cholesterol (N) (4401 (4012) (266) (123) NS < 240 mg/dl 79.6% 79.8% 75.8% 81.0% (0.5969) > 240 mg/dl 20.4% 20.2% 24.2% 19.0% HDL Combined (N) (4400) (4011) (266) (123) Above cut point* 70.9% 71.5% 63.8% 56.1% 0.027 Below cut point 29.1% 28.5% 36.1% 43.9% body mass index(N) (4482) (4088) (271) (123) < 25 Kg/m2 28.5% 28.7% 26.3% 23.7% NS Overweight (> 25) 38.4% 38.6% 33.6% 39.3% (0.2120) Obese (> 30) 33.1% 32.7% 40.1% 37.0% Waist (4595) (4095) (275) (125) measurement < 0.0001 Above cut point* 58.4% 57.6% 69.5% 75.3% Below cut point 41.6% 42.4% 30.5% 24.7% *HDL cut points: >40 mg/dl for men, >50 mg/dl for women; waist cut points: <102 cm for men, < 88 cm for women; NS not significant Figure 4.3 Weighted Biological Risk variables Prevalence in NHANES 2001-2004 data for Final Sample by Three Ankle-Brachial Index levels 74 36.5%, respectively) of cases with both PAD and diabetes (Wald adjusted F value 5.55, df 2, 29, p = 0.0091) in this study were within fairly good control of their disease. One in 5 weighted cases overall had an elevated SBP reported, but a dramatic increase to 38.7% (95% CI 32.6%, 44.6%) of the mild PAD group and 46.3% (95% CI 34.6%, 57.9%) of the severe PAD group had an elevated systolic pressure (Wald adjusted F value 23.69, df 2, 29, p < 0.0001). Hypertension had been diagnosed in 60.3% (95% CI 53.0%, 67.6%) and 71.4% (95% CI 64.4%, 78.3%) of the mild and severe PAD cases (Wald adjusted F value 23.35, df 2, 29, p < 0.0001). Very few cases had an elevated diastolic BP. Elevated total cholesterol was present in about one fifth of the cases at any ABI level, but the difference was not significant. Hyperlipidemia had been diagnosed in 36.8% (95% CI 34.4%, 39.2%) of cases without PAD, but almost half of the PAD cases (Wald adjusted F value 5.39, df 2, 29, p = 0.01). The mean good high density lipoprotein (HDL) for both males and females was well above their respective cut points, but the prevalence of HDL below the male or female cut points 28.5% (95% CI 26.2%, 30.8%) rises to 36.2% (95% CI 28.8%, 43.6%) and 43.9% (95% CI 33.2%, 54.5%) within both the 0.70-0.90 and < 0.70 respective ranges (Wald adjusted F value 4.08, df 2, 29, p = 0.027). Disease progression is the most likely to be slowed by change in lifestyle with mild PAD (0.7-0.9 ABI). Body mass index means were well within the overweight designation (BMI 25 to 30) for all ABI levels and one third of the 75 weighted sample were obese when the cut point of 30 kilograms per meter squared (Kg/m2) was applied, but differences between ABI levels were not significant. Waist measurement was larger than the cut point recommendations for over half of the overall sample (58.4%, 95% CI 56.2%, 60.6%). Waist measurements were significantly larger in 3/5ths of the 0.70-0.90 ABI cases and 3/4ths of the < 0.70 ABI cases (Wald adjusted F value 15.47, df 2, 29, p < 0.0001). Cardiac disease and stroke diagnoses are common among the cases in the lower ABI ranges. One fourth of the ABI 0.70-0.90 cases and 44.4% (95% CI 30.8%, 58.0%) of the < 0.70 cases had been diagnosed with congestive Final Ankle-Brachial Index Sample 0.91-1.50 0.70-0.90 0.28-0.69 (No PAD) (mild) (severe) p-values N=4559 N=4147 N=283 N=129 Weighted % 100% 93.9% 4.6% 1.5% Chronic 9.5 8.8 17.6 26.8 0.0091 Diabetes Mellitus Chronic 37.3 35.7 60.3 71.4 < 0.0001 Hypertension Chronic 37.5 36.8 48.5 49.8 0.01 Hyperlipidemia Cardiovascular Disease 10.7 9.5 25.6 44.4 0.0002 (CHF, CAD, MI, and/or angina) Stroke Diagnosis 2.9 2.5 7.5 13.9 0.0383 Figure 4.4 Weighted Chronic Disease Prevalence in NHANES 2001-2004 data for Final Sample by Three Ankle-Brachial Index levels 76 heart failure (CHF), coronary artery disease (CAD), myocardial infarction (MI), and/or angina (Wald adjusted F value 11.65, df 2, 29, p = 0.0002). Stroke had previously occurred in 7.5% (95% CI 4.1%, 10.9%) of the ABI 0.70-0.90 cases and 13.9% (95% CI 6.0%, 21.8%) of the < 0.70 cases (Wald adjusted F value 2.95, df 4, 27, p = 0.0383). Health-Related Quality of Life and Disability. The CDC health-related quality of life mean number of unhealthy days within 30 days is presented in Figure 4.5. The final sample mean number of unhealthy days for each HRQOL variable was similar to the population means for this age group (Centers for Disease Control and Prevention, 2000). The confidence intervals for mean CDC Final Ankle-Brachial Index means* Sample 0.91-1.50 0.70-0.90 0.28-0.69 (No PAD) (mild) (severe) N=4559 N=4147 N=283 N=129 Weighted % 100% 93.9% 4.6% 1.5% Mean number of unhealthy days (95% CI) Physical 4.2 4.1 6.1 4.9 5.6 (3.7, 4.6) (3.6, 4.5) (4.6, 7.5) (1.7, 8.2) Mental combined 3.7 3.7 3.2 2.0 (5.3-6.7) (3.3, 4.0) (3.3, 4.0) (2.3, 4.0) (0.5, 3.4) Activity 2.1 1.8 1.8 2.0 2.0 (1.7-3.0) (1.5, 2.1) (1.4, 2.1) (1.0, 3.0) (0.1, 3.8) * mean at 55-64 years with range for 45 to >75 years. Figure 4.5 Weighted Physical, Mental, and Activity Disability Prevalence in NHANES 2001-2004 data for Final Sample by Three Ankle-Brachial Index levels 77 number of physical unhealthy days do not overlap between the mild and no PAD groups, so there is a significant increase in physical unhealthy days for mild PAD. The number of unhealthy days for severe PAD has a wider 95% CI since the sample is small, so no difference is discernible between mild and severe PAD. The binary form of each HRQOL variable uses a cut-point at 14-30 unhealthy days as the definition of disability (Centers for Disease Control and Prevention, 2000), so the mean for the individual HRQOL question does not reflect disability prevalence. Disability prevalence will be addressed with the logistic regression. Correlations. Weighted correlations were done between every dichotomous, ordinal and scale variable with every other to examine the data for high correlations and potential multicollinearity. Correlations between ABI and the continuous or binary (disability) HRQOL variables (physical, mental and activity) were quite low but mostly negative as would be expected, since severe PAD has a lower ABI and higher values mean more unhealthy days for HRQOL interval or binary values (Figure 4.5). Correlations among the continuous physical, mental and activity variables were similar to the correlations among the binary HRQOL variables. The correlations among the HRQOL variables were fairy high, but since they were not in any analysis at the same time, multicollinearity was not an issue. 78 ABI Physical Mental Activity continuous continuous continuous continuous Physical -0.042 1 continuous Mental 0.013 0.288 1 continuous Activity -0.020 0.517 0.329 1 continuous ABI Physical Mental Activity continuous Binary Binary Binary Physical -0.037 1 Binary Mental 0.016 0.283 1 Binary Activity -0.010 0.473 0.307 1 Binary ABI Ankle-Brachial Index (lower = worse blockage), Binary Disability form of HRQOL variables, Figure 4.6 Correlations: Ankle-Brachial Index with Health-Related Quality of Life and Binary Disability variables Logistic regression will be used with the dichotomous disability form of the HRQOL variables (physical, mental, and activity). Logistic regression has been chosen over regression using the interval form of the HRQOL variables for two reasons. First, line graphs of mean values showed that the relationship between PAD (ABI) and HRQOL is not linear (see means in Figure 4.5). Second, the HRQOL data is extremely skewed low, so a non-parametric analysis is more appropriate for examining disability with a cut point of 14-30 days (Hosmer & Lemeshow, 2000). 79 Logistic Regression Specific Aim 1. To determine the association between PAD severity and disability, as measured by health-related quality of life in community dwellers 40 years and over using the NHANES 2001-2004 data while adjusting for other factors. Hypothesis 1. There will be a significant association between PAD severity and disability alone and while controlling for: 1. Demographics, and 2. Socioeconomic status. Ankle-Brachial Index Dependent Level variables > 0.9 0.7 0.9 < 0.7 (No PAD) (mild PAD) (severe PAD) 87.3% 80.7% 84.1% No Physical (86.0%, 88.8%) (76.2%, 85.3%) (72.0%, 96.1%) Disability 12.6% 19.2% 15.9% Yes (11.2%, 14.0%) (14.7%, 23.8%) (3.9%, 28.0%) 89.2% 90.8% 95.8% No Mental (87.7%, 90.7%) (89.2%, 94.3%) (91.2%, 100%) Disability 10.8% 9.2% 4.2% Yes (9.3%, 12.3%) (5.7%, 12.8%) (0%, 8.8%) 94.6% 93.3% 94.6% No Activity (93.2%, 96.0%) (89.5%, 97.2%) (88.7%, 100%) Disability 5.4% 6.7% 5.4% Yes (4.0%, 6.8%) (2.8%, 10.5%) (0%, 11.3%) Figure 4.7 Estimated Percentages and 95% Confidence Interval for Physical, Mental, and Activity Disability Outcomes by Ankle-Brachial Index levels 80 Crude estimated odds ratios. Estimated odds ratios for physical, mental, and activity disability by the 3-category ABI presented in Figure 4.8 were calculated using prevalence data from Figure 4.7. Rows in Figure 4.8 indicate which ABI levels are being compared for the estimated odds ratios. The PAD equivalent of the ABI comparisons was mild vs. no PAD, severe vs. no PAD, and severe vs. mild PAD. The yes disability line (Figure 4.7) was used for the lower ABI range as the numerator and the highest ABI range always used as the denominator in the ratio calculated by SAS for survey data with confidence ABI Overall Row p- Outcome Est. OR* 95% CI comparisons p-value values 0.7 0.9 vs. > 0.9 1.652 (1.264, 2.157) 0.0002 Physical < 0.7 vs. > 0.9 1.314 (0.538, 3.212) 0.0002 0.5488 Disability < 0.7 vs. 0.7 - 0.9 0.796 (0.289, 2.192) 0.6587 0.7 0.9 vs. > 0.9 0.839 (0.550, 1.279) 0.4146 Mental < 0.7 vs. > 0.9 0.361 (0.115, 1.129) 0.1286 0.0799 Disability < 0.7 vs. 0.7 - 0.9 0.430 (0.122, 1.512) 0.1891 0.7 0.9 vs. > 0.9 1.259 (0.633, 2.503) 0.5115 Activity < 0.7 vs. > 0.9 1.002 (0.316, 3.183) 0.7626 0.9972 Disability < 0.7 vs. 0.7 - 0.9 0.796 (0.281, 2.253) 0.6673 *Estimated odds ratio (numerator vs. denominator) Figure 4.8 Crude (unadjusted) Ankle-Brachial Index level odds ratios (comparisons) for Physical, Mental, and Activity Disability 81 intervals and p-values. Physical disability has the only significant odds ratio (see bold in Figure 4.8), so the mental and activity disability variables were not tested further. The comparison between the mild versus no PAD equivalent was the only significant crude estimated odds ratio (bold in Figure 4.8), but both mild and severe PAD versus no PAD were tested as adjustments were added. Model 1. A manual forward step-wise process was used to test each sociodemographic variable as a potential confounder. Initially, separate logistic regressions were run for the categorical ABI with each of the six variables in Figure 4.9 and the results examined for change in the ABI odds ratio. The percentage change from the crude odds ratio to the adjusted odds ratio was calculated and the variables with greater than a 10% change in the odds ratio % Change from Adjusted OR* Crude OR Potential crude confounder 0.7 0.9 < 0.7 0.7 0.9 < 0.7 0.7 0.9 < 0.7 Gender 1.601 1.326 1.652 1.314 -3.1% 0.9% Age 1.427 1.031 1.652 1.314 -13.6% -21.5% Marital 1.561 1.197 1.652 1.314 -5.5% -8.9% Status Education 1.585 1.010 1.652 1.314 -4.1% -23.1% PIR 1.365 0.956 1.652 1.314 -17.4% -27.2% Race/ 1.660 1.304 1.652 1.314 0.5% -0.8% Ethnicity * ABI > 0.9 is the reference level in these models; models have ABI + potential confounder as predictors Figure 4.9 Step 1: Process for Determining Potential Confounders for Model 1 82 were considered potential confounders. The generous 10% change was used, since the crude odds ratio was small (Hosmer & Lemeshow, 2000). Age was the first confounder entered into the logistic regression with ABI, since age made a greater than 10% change in both ABI categories. Education and poverty income ratio (PIR) also were considered, but PIR was not chosen initially, since income can be highly correlated with age and poorly reported when a large portion of the sample is retired, so more likely spurious. Logistic regressions were then run with gender, marital status, education, PIR and Race/Ethnicity added individually to the ABI and age model. In step 2, a change in adjusted odds ratio was calculated using the ABI odds ratios reported when ABI, age and an additional potential confounder were entered divided by the adjusted odds ratio for ABI with age in Figure 4.9. A new Crude odds ratio was also calculated for information only, since the change in adjusted odds ratio was used to decide if additional confounders needed to be added to the logistic regression model. Adding confounders to the logistic regression and running each new model with all the remaining variables separately was followed until age, education, and PIR had all been added to the regression model 1, and gender, marital status and race/ethnicity made less than a 10% change in the adjusted estimated odds ratio. A fractional polynomials procedure was used to test for linearity between PIR and physical disability, since PIR is a continuous variable. Transformations of PIR did not improve the model fit more than the linear form. Age, education and PIR were all accepted as confounders 83 of the relationship between ABI and physical disability, so each was then tested for effect moderation by entering an interaction with ABI term into the model. Separate logistic regressions were run for physical disability with ABI categories, age, education, PIR and one of the interaction variables (ABI by age, ABI by each education level, or ABI by PIR). The type 3 Analysis of effects Wald chi square was examined for a conservative p-value less than 0.01. Age was found to be an effect moderator, since the interaction term was significant (p-value < 0.0001). The final Model 1 (Figure 4.10) equation is: Physical Disability level = -1.602 -0.303(ABI: 0.7-0.9) + 5.462(ABI: <0.7) + 0.007(Age) + 0.497(< HS education) + 0.151(HS/GED) 0.294(PIR) + 0.008(Age x ABI: 0.7-0.9) 0.083(Age x ABI: <0.7). Parameter Estimate SE p-value Intercept -1.602 0.326 --- ABI: 0.7 0.9 (Mild PAD) -0.303 1.078 0.7784 ABI: < 0.7 (Severe PAD) 5.462 1.296 <.0001 Age 0.007 0.005 0.1232 Education: < HS 0.497 0.200 0.0131 Education: HS graduate or 0.151 0.153 0.3235 GED PIR -0.294 0.049 <.0001 Age × ABI: 0.7 0.9 (mild) 0.008 0.016 0.6082 Age × ABI: < 0.7 (severe) -0.083 0.018 <.0001 Figure 4.10 Physical Disability final Model 1 84 Comparison Crude OR 95% CI p-value ABI: 1.31 (0.54, 3.21) 0.5488 < 0.7 vs. > 0.9 ABI: 1.65 (1.26, 2.16) 0.0002 0.7-0.9 vs. > 0.9 ABI: 0.80 (0.29, 2.19) 0.6587 < 0.7 vs. 0.7-0.9 Age Comparison OR* 95% CI p-value ABI: 4.42 (1.48, 13.18) 0.0077 < 0.7 vs. > 0.9 th ABI: 48 (25 1.10 (0.58, 2.08) 0.7695 percentile) 0.7-0.9 vs. > 0.9 ABI: 4.02 (1.25, 12.93) 0.0198 < 0.7 vs. 0.7-0.9 ABI: 1.64 (0.67, 3.96) 0.2762 < 0.7 vs. > 0.9 th ABI: 60 (50 1.22 (0.86, 1.72) 0.2688 percentile) 0.7-0.9 vs. > 0.9 ABI: 1.35 (0.51, 3.58) 0.5527 < 0.7 vs. 0.7-0.9 ABI: 0.66 (0.28, 1.54) 0.3339 < 0.7 vs. > 0.9 th ABI: 71 (75 1.33 (0.95, 1.86) 0.0921 percentile) 0.7-0.9 vs. > 0.9 ABI: 0.49 (0.19, 1.29) 0.1495 < 0.7 vs. 0.7-0.9 * Odds ratio at a specific age, controlling for education and PIR. Figure 4.11 Application of Model 1 Showing Interaction Odds Ratios and 95% CI for Ankle-Brachial Index and Age at the 25th Quartile, Median, and the 75th Quartile 85 The age by ABI interaction effect on the relationship between ABI and Physical Disability is illustrated using age 48 years (25th percentile), 60 years (median), and 71 years (75th percentile). Figure 4.11 contains the results of the comparisons with the crude estimated odds ratio for comparison. It is interesting to note that the 0.7-0.9 vs. >0.9 (mild versus no PAD) had the highest and only significant crude estimated odds ratio, but adjustments for education level and poverty index ratio made all those age comparisons not significant. No comparisons for age 60 or 71 were significant. Age 48 comparisons for < 0.7 ABI with > 0.9 ABI and 0.7-0.9 ABI (severe versus no or mild PAD) were both significant. The 95% confidence intervals for Age 48 comparisons were wide, but remained stable. Model 2 Specific Aim 2. To determine the potential mediator effects of symptoms, functional status, and general health perceptions on the association between PAD severity and disability. Hypothesis 2. There will be a significant mediator effect on the association between PAD severity and disability from the respondent perceptions of: 1. symptoms 2. functional status 3. general health. 86 A manual forward step-wise process was used to test each symptom, functional status and general health perception variable separately as a potential confounder. Figure 4.12 contains the adjusted odds ratios for the seven variables that were run with the categorical ABI. The percentage change from the crude odds ratio to the adjusted odds ratio was calculated and the variables with greater than a 10% change in the odds ratio were considered potential confounders. Most of the variables had a greater than 10% change in at least one of the ABI levels. Mobility was the first confounder entered into the logistic regression with the ABI categories, since mobility caused more than a 20% change in both of the ABI categories. % Change Potential Adjusted OR* Crude OR from crude Confounders 0.7 0.9 < 0.7 0.7 0.9 < 0.7 0.7 0.9 < 0.7 Leg Pain 1.419 0.851 1.652 1.314 -14.1 -35.2 Calf Pain 1.352 0.762 1.652 1.314 -18.2 -42.0 Foot Pain 1.502 1.192 1.652 1.314 -9.1 -9.3 Mobility 1.009 0.514 1.652 1.314 -38.9 -60.9 Physical Activity 1.563 1.101 1.652 1.314 -5.4 -16.2 Vigorous 1.354 0.74 1.652 1.314 -18.0 -43.7 GHP 1.376 0.698 1.652 1.314 -16.7 -46.9 * ABI > 0.9 is the reference level in these models; models have ABI + potential confounder as predictors Figure 4.12 Step 1: Process for Determining Potential Confounders for Model 2 87 % Change from % Change Potential Adjusted OR* previous adj. OR from crude Confounders 0.7 0.9 < 0.7 0.7 0.9 < 0.7 0.7 0.9 < 0.7 Leg Pain 1.019 0.472 1.0 -8.2 -38.3 -64.1 Calf Pain 0.972 0.431 -3.7 -16.1 -41.2 -67.2 Foot Pain 1.009 0.533 0.0 3.7 -38.9 -59.4 Physical Activity 0.995 0.475 -1.4 -7.6 -39.8 -63.9 Vigorous 0.958 0.437 -5.1 -15.0 -42.0 -66.7 GHP 1.051 0.447 4.2 -13.0 -36.4 -66.0 * ABI > 0.9 is the reference level in these models; models have ABI + mobility + potential confounder as predictors Figure 4.13 Step 2: Process for Determining Potential Confounders for Model 2 The change in adjusted odds ratio was examined to determine the next potential confounder to be included in Model 2. Figure 4.13 shows the change in percentages from the mobility adjusted odds ratio for Physical Activity. Calf pain and vigorous activity were the variables with the highest percentage of change in the adjusted odds ratio. Calf pain was chosen as the next variable to be added to the model as a confounder, since pain in the calf with walking is a frequent symptom of PAD. Step 3, vigorous was the only variable with a greater than 10% change in the mobility plus calf pain adjusted odds ratio (-13.2% for < 0.7 ABI), so it was also added to the model. No further variables made a significant change in the adjusted odds ratio. Mobility, calf pain and vigorous activity were tested for effect moderation with ABI in separate logistic regressions using interaction variables. Mobility 88 and calf pain remained confounders. A significant interaction between vigorous activity and ABI (type 3 Analysis of Effects, Wald chi square test, p-value <0.0001) implies that vigorous activity is an effect modifier of the relationship between ABI and physical disability. The model became very unstable when the interaction variable was entered due to the small cell counts for severe PAD, so the interaction variable has been left out of the current Model 2. Pain symptoms (leg or foot pain), usual physical activity level and general health perception were not found to be significant confounders or effect modifiers. Crude ABI Comparison 95% CI p-value OR ABI: < 0.7 vs. > 0.9 1.31 (0.54, 3.21) 0.5488 ABI: 0.7-0.9 vs. > 0.9 1.65 (1.26, 2.16) 0.0002 ABI: < 0.7 vs. 0.7-0.9 0.80 (0.29, 2.19) 0.6587 Model 2 ABI OR* 95% CI p-value and confounders ABI: < 0.7 vs. > 0.9 2.674 (1.076, 6.667 ) 0.0342 ABI: 0.7-0.9 vs. > 0.9 0.924 (0.675, 1.264) 0.6194 Mobility: 4.026 (2.991, 5.418) <0.0001 difficulty vs. none Calf Pain: 2.314 (1.788, 2.994) <0.0001 with walking vs. none Vigorous: 2.168 (1.595, 2.946) <0.0001 Not done vs. yes Cannot do vs. yes 3.995 (2.600, 6.138) <0.0001 *OR- adjusted odds ratio, highlights for significant OR. Figure 4.14 Physical Disability Model 2 Adjusted Odds Ratios and 95% Confidence Intervals 89 Model 2 equation: Physical Disability = -3.1642 - 0.0796 (ABI: 0.7-0.9) 0.9846(ABI: <0.7) + 1.3927(Mobility) + 0.839(Calf Pain) + 0.7738 (Vigorous activity not done) + 1.385(Vigorous activity: cannot do) means more when applied with the odds ratios listed in Figure 4.14. Persons age 40 and older with severe PAD were found to be 2.7 (95% CI 1.1, 6.7) times as likely to have physical disability as those without PAD after adjusting for mobility difficulty, calf pain with walking and the vigorous activity for 10 minutes in the past 30 days. Odds of mobility difficulty (4.0, 95% CI 3.0, 5.4), calf pain with walking (2.3, 95% CI 1.8, 3.0) and inability to do vigorous activity (4.0, 95% CI 2.6, 6.1) were also very significant after adjusting for all the other variables in Model 2. Model 3 Specific Aim 3. To determine the potential moderator effect of other chronic diseases and biological risks on the association between PAD severity and disability. Hypothesis 3. There will be a significant moderator effect on the association between PAD severity and disability from respondent other diagnosed chronic diseases and biological risks of: 1. High tobacco exposure, 2. Diabetes diagnosis or elevated blood glucose, 3. Hypertension diagnosis or elevated systolic or diastolic blood pressure, 90 4. Hyperlipidemia diagnosis or elevated total cholesterol or low HDL cholesterol, and 5. Adiposity (obesity or large waist measurement). A manual forward step-wise process was used to test each biological risk and chronic disease variable separately as a potential confounder. Figure 4. contains the adjusted odds ratios for the eleven variables that were run with the categorical ABI. The percentage change from the crude odds ratio to the adjusted odds ratio was calculated and the variables with greater than a 10% change in the odds ratio were considered potential confounders. Hypertension was the first confounder entered into the logistic regression with ABI, since hypertension was the only variable to cause more than a 10% change for both mild and severe PAD (-13% and -18.2% respectively). The change in adjusted odds ratio was examined to determine the next potential confounder to be included in Model 3. Total cholesterol was the only variable with a 20% change for either ABI odds ratios from the previous model with hypertension. Total cholesterol made the largest percentage (-21%) change in either adjusted hypertension odds ratio, so at Step 2, total cholesterol was entered into model 3. Obesity caused a higher than 10% change in both the adjusted odds ratios (13.6% for mild and 12.4% for severe PAD). No further potential confounders were identified. 91 % Change from Potential Adjusted OR* Crude OR crude Confounder 0.7 0.9 < 0.7 0.7 0.9 < 0.7 0.7 0.9 < 0.7 Cotinine 1.578 1.103 1.652 1.314 -4.5 -16.1 Glycosylated Hemoglobin 1.581 1.061 1.652 1.314 -4.3 -19.3 Systolic BP 1.648 1.307 1.652 1.314 -0.2 -0.5 Diastolic BP 1.719 1.389 1.652 1.314 4.1 5.7 Total Cholesterol 1.554 1.128 1.652 1.314 -5.9 -14.2 HDL ( good ) 1.537 1.087 1.652 1.314 -7.0 -17.3 Obese 1.548 1.434 1.652 1.314 -6.3 9.1 Waist measurement 1.601 1.255 1.652 1.314 -3.1 -4.5 Diagnosed: Diabetes 1.564 1.173 1.652 1.314 -5.3 -10.7 Hypertension 1.438 1.075 1.652 1.314 -13.0 -18.2 Hyperlipidemia 1.58 1.249 1.652 1.314 -4.4 -4.9 * ABI > 0.9 is the reference level in these models; models have ABI + potential confounder as predictors Figure 4.15 Step 1: Process for Determining Potential Confounders for Model 3 Hypertension, total cholesterol and obesity were tested for effect moderation by entering interaction terms into the model one at a time with the confounders previously selected. The type 3 Analysis of Effects, Wald Chi square test showed total cholesterol (p-value = 0.0058) and obesity (p-value = 0.0079) were significant individually and borderline significant when both interaction terms were entered together (p-values = 0.0121 and 0.0118, respectfully). Both interaction terms were guardedly left in model 3. 92 Crude Comparison 95% CI p-value OR ABI: < 0.7 vs. > 0.9 1.31 (0.54, 3.21) 0.5488 ABI: 0.7-0.9 vs. > 0.9 1.65 (1.26, 2.16) 0.0002 ABI: < 0.7 vs. 0.7-0.9 0.80 (0.29, 2.19) 0.6587 Comparison OR* 95% CI p-value Total ABI: > 0.9 vs. < 0.7 13.56 (1.31, 140.65) 0.0288 Cholesterol ABI: 0.7-0.9 vs. > 0.9 2.88 (1.39, 5.97) 0.0044 >240 mg/dl ABI: 0.7-0.9 vs. < 0.7 39.06 (3.11, 492.61) 0.0045 Total cholesterol < 240 did not have a significant effect on ABI comparisons. ABI: < 0.7 had a small enough N to make the comparison unsafe. Obesity ABI: < 0.7 vs. 0.7-0.9 4.14 (1.23, 13.98) 0.0219 (BMI > 30) Body Mass Index < 30 Kg/m2 and other obesity did not have a significant effect on ABI comparisons Hypertension Diagnosed vs. not 1.84 (1.54 , 2.20) < 0.0001 * Odds ratio at an elevated total cholesterol or obesity, controlling for Hypertension. Figure 4.16 Physical Disability Model 3 Significant Adjusted Odds Ratios and 95% Confidence Intervals Figure 4.16 contains application of the final Model 3 equation: Physical Disability = -2.3410 + 0. 3989(ABI: 0.7-0.9) - 0.4663(ABI: < 0.7) + 0.6096 (Hypertension) + 0.1689(Total Cholesterol category) + 0.2477(Obese) + 0.6593(Total Cholesterol x ABI: 0.7-0.9) -2.1420(Total Cholesterol x ABI: < 0.7) + 0.8998(Obese x ABI: 0.7-0.9) + 1.3870(Obese x ABI: < 0.7). 93 The adjusted odds ratios displayed in Figure 4.16 showed that the interaction between mild PAD (ABI 0.7-0.9) and elevated total cholesterol lends an individual more likely for physical disability than those with severe or no PAD. The comparisons for severe PAD (ABI < 0.7) were reversed to show how extreme the range became and the inability to use that part of the equation to safely predict the likelihood of physical disability. Obese persons with severe PAD (ABI < 0.7) were 4 times more likely to report physical disability than persons with mild PAD. Other body mass index and ABI comparisons did not make a significant difference in likelihood of physical disability. Model 4 Specific Aim 4. To determine the effect of integrating Models 1, Model 2, and Model 3 on the association between PAD severity and disability. Hypothesis 4. There will be a significant effect on the association between PAD severity and disability when integrating the variables included in previous models. Model 4 used a manual forward step-wise approach to test variables that had been confounders in Model 1 (age, education level, and poverty income ratio), Model 2 (mobility, calf pain, and vigorous) or Model 3 (hypertension, total cholesterol, and obesity) against each other. The previous model selections decreased the number of variables to be tested for entry into Model 4. The adjusted odds ratio obtained in Step 1 of each original Model was assembled into one table and mobility was clearly the greatest confounder. Each other 94 % Change from Potential Adjusted OR* Crude OR crude Confounder 0.7 0.9 < 0.7 0.7 0.9 < 0.7 0.7 0.9 < 0.7 AGE 1.427 1.031 1.652 1.314 -13.62% -21.54% Education 1.585 1.01 1.652 1.314 -4.06% -23.14% PIR 1.365 0.956 1.652 1.314 -17.37% -27.25% Mobility 1.009 0.514 1.652 1.314 -38.92% -60.88% Calf Pain 1.352 0.762 1.652 1.314 -18.16% -42.01% Vigorous 1.354 0.74 1.652 1.314 -18.04% -43.68% Hypertension 1.438 1.075 1.652 1.314 -12.95% -18.19% Total Cholesterol 1.554 1.128 1.652 1.314 -5.93% -14.16% Obese 1.548 1.434 1.652 1.314 -6.30% 9.13% * ABI > 0.9 is the reference level in these models; models have ABI + potential confounder as predictors Figure 4.17 Step 1: Process for Determining Potential Confounders for Model 4 potential confounder was entered into the logistic regression with ABI and mobility. Reference categories were specified for each variable in model 4 logistic regressions, so the ABI odds ratios were slightly different than the ones in model 2 for calf pain and vigorous activity. The percentage of change from the ABI and mobility adjusted odds ratio was calculated. Vigorous activity and calf pain each had a greater than 20% change in the adjusted odds ratio for severe PAD (-22.6% and -24.1%, respectively) and a greater than 10% change for mild PAD (-12.6% and -10.5%, respectively). No other potential confounder 95 had a 20% change in either ABI. Vigorous activity was chosen as the second confounder to enter Model 4. Again, each other potential confounder was added to the logistic regression individually with ABI, mobility and vigorous activity by physical disability and the change in the adjusted odds ratio was calculated. Calf pain (- 16.3%) and age (-11.1%) were the only potential confounders with a significant change in either percentage, so calf pain was entered into Model 4. No additional confounders were identified, so Model 4 ended with the same confounders present in Model 2. Vigorous activity again had a significant interaction with ABI, but since the interaction was so unstable in Model 2, it was not added to Model 4 either. Summary The purpose of this study was to understand the relationship between mild or severe peripheral arterial disease (PAD) and Disability (Health-related Quality of Life) and determine which factors affect that relationship. Specific Aim 1 was to determine the association between PAD severity and disability, as measured by health-related quality of life in community dwellers 40 years and over using the NHANES 2001-2004 data while adjusting for other factors. Hypothesis 1 stated that there will be a significant association between PAD severity and disability alone and while controlling for demographics and socioeconomic status. 96 The crude (unadjusted) model estimated odds ratios for physical disability by mild PAD (ABI: 0.7-0.9) and severe PAD (ABI: < 0.7) were quite low, but the 0.7-0.9 estimated odds ratio was significant, so there was a relationship between mild PAD and physical disability when no adjustment was made. Mental disability and activity disability as measured by the CDC HRQOL questions were not related to PAD severity. Age was an important confounder and effect moderator with ABI in Model 1 and provided some evidence of an interaction between younger age and greater physical disability perception with severe PAD. This means that severe PAD at a younger age may be related to a sense of physical disability perception when a person is not expecting to have physical problems. The median age of 60 and upper quartile age 71 did not have the same reaction to lower ABI and PAD progression. Lower education and lower poverty income ratio were both closely related to perception of physical disability. Hypothesis 1 was partially met by the significant crude odds ratio between mild PAD and physical disability and the effect that the interaction between younger age and the severe PAD had to the likelihood of physical disability. Specific Aim 2 was to determine the potential mediator effects of symptoms, functional status, and general health perceptions on the association between PAD severity and disability. Hypothesis 2 stated that there would be a significant mediator effect on the association between PAD severity and disability from the respondent perceptions of symptoms, functional status and 97 general health. Mobility difficulty and calf pain were both found to be confounders and vigorous activity level was both a confounder and effect moderator. Mobility was the main confounder, since the percentage change from the crude estimated odds ratio was so strongly negative bringing the adjusted odds ratios back toward 1.0. Calf pain and vigorous activity also confounded the relationship between ABI and physical disability, but the effect was not as strong as for mobility. Confounders can be a mediator that occurs between an independent and dependent variable set or spurious meaning it has an effect on the dependent variable, but not as an intermediary for another independent variable (Hosmer & Lemeshow, 2000). Hypothesis 2 was found to be marginally true for functional status and symptoms, since mobility, calf pain and vigorous activity were confounders with theoretical support as mediators. General health perceptions were not related. Specific Aim 3 was to determine the potential moderator effect of other chronic diseases and biological risks on the association between PAD severity and disability. Hypothesis 3 stated that there was a significant moderator effect on the association between PAD severity and disability from respondents other diagnosed chronic diseases and biological risks of high cotinine, diabetes diagnosis or elevated blood glucose, hypertension diagnosis or elevated systolic or diastolic blood pressure, hyperlipidemia diagnosis or elevated total cholesterol or low HDL cholesterol, and adiposity (obesity or large waist measurement). 98 Hypertension was a confounder in Model 3 while total cholesterol and obesity were effect moderators, so the hypothesis was partially supported for specific aim 3 also. Hypertension is likely a mediator type of confounder Specific Aim 4 was to determine the effect of integrating Models 1, Model 2, and Model 3 on the association between PAD severity and disability. Hypothesis 4stated that there was a significant effect on the association between PAD severity and disability when integrating the variables included in previous models. Model 4 showed that the mediator effect of the confounder mobility was stronger than any other variable in this study. Strong support was present in Model 4 for symptoms (specifically calf pain) and functional status (specifically mobility and vigorous activity) placement at mediators in the conceptual framework, since all were confounders. Support for general health perception to remain in a potential mediator area of the conceptual framework was not present. Age, total cholesterol and obesity did not show a moderator effect in Model 4, but they may have their moderator effects more on PAD development and progression than on disability resulting from PAD. Poverty income ratio and education level trend loosely in the same direction as ABI; higher income and higher education level are more associated with higher ABI which is little or no PAD. Likewise, severe PAD and its lower ABI are more associated with lower income and less education, but there is not a direct causal relationship, so these confounders are seen as spurious. 99 CHAPTER 5 DISCUSSION The purpose of this study was to understand the relationship between mild or severe peripheral arterial disease (PAD) and Disability (Health-related Quality of Life) and determine which factors affect that relationship. The first national phone survey about PAD discovered that 75% of those responding were not aware PAD was a disease and the informed 25% knew little or erroneous information (Hirsch et al., 2007). The Stay in Circulation campaign started in September 2007 to increase awareness about PAD in the community as an initiative of the P. A. D. Coalition with the National Heart, Lung and Blood Institute (http://www.nhlbi.nih.gov/health/public/heart/pad/index.html). Approximately 9 million Americans meet the diagnostic criteria for PAD, but many of them do not know they have the disease or that the disease even exists. The National Health and Nutrition Examination Survey (NHANES) 2001- 2004 data base was used as the source of the community dwelling individuals, since the NHANES included both the ankle-brachial index (ABI, a measure of PAD severity) and the CDC HRQOL-4 questions used in the Behavior Risk 100 Factor Surveillance System (BRFSS) for several chronic diseases including diabetes, hypertension and heart failure (Zahran et al., 2005). The ABI has been part of the NHANES since 1999, and over 20 published studies have examined a variety of topics in relation to PAD. Nine studies have reported the prevalence of PAD alone or with other diseases (Anand et al., 2007; Andresen et al., 2001b; Gregg et al., 2004; Lane et al., 2006; "Lower extremity disease among persons aged > or =40 years with and without diabetes--United States, 1999-2002.," 2005; Nelson, Reiber, Kohler, & Boyko, 2007; O'Hare, Glidden, Fox, & Hsu, 2004; Ostchega et al., 2007; Selvin & Erlinger, 2004) . Six additional studies have reported PAD prevalence with CRP or monocytes as markers of inflammation (Menke et al., 2006; Nasir et al., 2005; Resnick & Foster, 2005; Shankar, Li, Nieto, Klein, & Klein, 2007; Vu et al., 2005; Wildman et al., 2005). Eight additional studies have examined heavy metals or indicators of liver or renal function with prevalence of PAD (Guallar et al., 2006; Kuo, Lin, & Yu, 2007; Muntner, Menke, DeSalvo, Rabito, & Batuman, 2005a; Navas- Acien et al., 2004; Navas-Acien et al., 2005; Perlstein, Pande, Beckman, & Creager, 2008; Shankar, Klein, Nietoc, & Klein, 2008; Shankar, Li, Klein, Nietoc, & Klein, 2007). Health-related quality of life and PAD association has been studied in small, disease specific studies using a variety of HRQOL tools (Aquarius, Denollet, de Vries, & Hamming, 2007; Hiatt, 1997; Kalbaugh et al., 2004; Klevsgard, Hallberg, Risberg, & Thomsen, 2000; Menard et al., 2004; Pell et al., 1993; Treat-Jacobson et al., 2002). No previously published study has 101 addressed PAD with disability using the CDC HRQOL-4 in a population based study. There were 4559 cases with complete data that were included in this study. Weighted complete cases with PAD (ABI 0.90 or less) were 6.1% (95% CI 4.9, 7.1) for adults age 40-85 and 13.2% (95% CI 9.9, 16.5) for age 60-85 weighted cases. The purported prevalence of PAD in the Stay in Circulation campaign information is about 5% of persons over age 50 (http://www.nhlbi.nih.gov/health/public/heart/pad/index.html) which agrees with the overall findings of this study. PAD occurs between age 40 and 50, but less often, so the campaign is aimed at a slightly older group than was available through the NHANES data. The CDC HRQOL-4 is a group of 4 questions that ask about general health, number of physical or mental unhealthy days in the past 30, and number of days in past 30 when usual activity was limited by physical or mental health. The definition of disability used in this study was 14-30 unhealthy days of the last 30 days for physical, mental or activity, which matches the CDC definition used in chronic disease surveillance. The small relative prevalence of PAD and the even smaller prevalence of severe PAD made the disability definition difficult to test accurately even in this large data set. Disability and PAD Depression has been reported by patients with advanced PAD, but the prevalence of cases with both PAD and perceived mental disability in the 102 population as a whole was low enough that mental disability was not reported as often by individuals with PAD as individuals without PAD. Mental disability was not a valuable measure of PAD related disability. Oddly, those with severe PAD had the lowest prevalence (4.2%, 95% CI 0%, 8.8%) of reported mental disability (see Figure 4.7), but it is unclear whether they really had fewer mental unhealthy days or all unhealthy days were attributed to their physical health. Physical and mental unhealthy days are typically reported as a combined number of days in 30 for the purpose of chronic disease monitoring, but question 4 about activity limitation due to the physical/mental unhealthy days was actually the item used most for monitoring disability with chronic diseases (Centers for Disease Control and Prevention, 2000). Individuals may have unhealthy physical or mental days, but the real disability occurs when the unhealthy days interfere with the ability to do usual activities. Many individuals keep on with activities even if it is a physical or mental unhealthy day or over time have adjusted their activity level down to the point they do not notice a difference in activity. Activity disability as measured by the CDC HRQOL-4 is not a good measure of PAD related disability. Additional questions added to some surveys in the 1995-1997 BRFSS (but not available in the NHANES) indicated that persons that needed assistance with routine or personal care were the ones who had a mean number of activity limited days over 14 in 30 days (Centers for Disease Control and Prevention, 2000). Only a smaller portion of the 1.5% in this study with 103 severe PAD (ABI < 0.7) would be likely to require assistance with care due to their PAD, so it is reasonable to have little association between ABI and Activity disability. The small overall prevalence of severe PAD does not warrant chronic disability monitoring. Physical Disability Mild PAD had a small but significant predictive ability in the crude (unadjusted) estimated odds of physical disability. The adjusted odds ratios in all the models tended to negate that significant predictive ability. Adjusted odds were used to remove the effect of a confounder (either a mediator or a spurious variable). In this case, inability to do vigorous activity, mobility difficulty and calf pain with walking are known problems (likely mediators) for persons with severe PAD, since that is the reason most persons agree to treatment, especially surgery (Gibson & Kenrick, 1998). A mediator effect from these confounders is likely, since persons with worse levels of PAD had lower physical component summary scores (PCS) using the SF-36, (Hallin et al., 2002; Pell et al., 1993; Society for Vascular Nursing, 2006) a HRQOL measure used in the development of the CDC HRQOL-4 (Hennessy et al., 1994; Moriarty et al., 2003). Only 1.5% of the weighted sample had severe PAD (ABI below 0.70), the level at which symptoms of pain and limited function surface, so that is the group of cases most likely to have been seen by a physician for their PAD. Limited circulation is least expected at a younger age and could make the 104 greatest impact on the perception of physical disability, so finding a 4.0-4.4 adjusted odds ratio (95% CI 1.3, 13.2) of physical disability at age 48 with severe PAD compared to no PAD or mild disease was not surprising. Lower levels of total cholesterol in the severe PAD group may have indicated successful treatment of a concurrent chronic disease. When the two PAD groups were checked for SBP at a 130 mmHg cut point (recommended for persons with high cardiovascular risk), 60% of each the mild and severe PAD groups had a SBP higher than the recommended 130 mmHg. A single BP measure would not be used to start treatment, but the high prevalence of SBP above the cut point indicated poor BP control, since 60-70% of those with PAD reported a hypertension diagnosis and treatment. The significant crude odds ratio present for mild PAD cases was lost when adjustments for pain and mobility occurred, which indicated that this group of individuals is having some pain or mobility difficulty before they are distressed enough to seek treatment. The mild PAD group also tested high in total cholesterol and other biological risks that are typically asymptomatic, so few individuals might even be aware of a problem when both biological risks and PAD are present. The Stay in Circulation campaign is aimed at this group of individuals who can decrease their disease progression with lifestyle changes. The NHANES data could be used to monitor overall biological risk levels, but physical disability would not be a good monitoring measure of effectiveness of the campaign. 105 PAD Surveillance The CDC HRQOL-4 has been shown to lack usefulness in disability measurement for PAD, so the BHFSS disability surveillance would not be appropriate to use for this population. The prevalence of PAD is on the rise, but the small percentage of the population (1.5%) with severe PAD makes the problem seem less critical. The CDC HRQOL-14 that includes an additional 5 questions about activity limitation and 5 questions for number of days of pain, anxiety, depression, vitality and/or sleeplessness might be helpful in making the disability measure more specific for PAD, since the current 4-items included in the NHANES do not measure activity disability with PAD very well (Centers for Disease Control and Prevention, 2000). Monitoring the mean number of unhealthy Physical or Activity limited days may mean more in this small segment of the adult population over age 40. The BRFSS 1993-1997 reported mean number of unhealthy days (combined physical/mental) in 30 days ranged from 5.3 days (for 45-55 year olds) to 6.7 days (for >= 75 year olds) and the mean activity limitation was 1.7 to 3.0 days for the same age groups, so activity limitation due to unhealthy days is reported about half as often as the unhealthy days (Centers for Disease Control and Prevention, 2000). The mean number of Activity limitation days in this study was 2.0 for severe PAD (95% CI 0.11, 3.8 days) and 1.7 overall (95% CI 1.5, 2.1 days). The mean number of physical unhealthy days was highest in the mild PAD group (6.1 days, 95% CI 4.6, 7.5), but 4.2 days overall (95% CI 3.7, 4.6), 106 so monitoring the mean number of physical unhealthy days may be appropriate in this population. The addition of PAD awareness monitoring to the cardiovascular section of the chronic disease indicator site at the CDC Data & Statistics website (http://apps.nccd.cdc.gov/cdi/) would be appropriate. Monitoring the awareness of PAD as a disease would be a first step. The goal of the Stay in Circulation campaign would be to avoid PAD progression to the point of causing disability. Monitoring mean number of physical unhealthy days and activity limited days through the NHANES for persons who received ABI testing may provide a glimpse of the effect of PAD nationally. An important procedure to remember with the disability surveillance is to be sure the subjects are answering for themselves, since proxy responders, both relatives and health care workers have reported poorer general health and more unhealthy days than the subjects themselves in previous studies (Andresen et al., 2001a; Pell, 1995). Conceptual Framework Revisited Age, education and PIR were more influential characteristics of the individual than gender, race/ethnicity or marital status. Age, education and PIR were bolded to indicate their importance. In the conceptual framework, gender, race/ethnicity and marital status have been left in without bolding to indicate that their monitoring needs to continue, but their effect may be limited. Further testing is needed before they are eliminated from the conceptual framework. 107 Biological risks are more likely to be predictive of PAD than the disability that follows; therefore the biological risks placement within Biological Function in the model is appropriate. The elevated glucose in the blood is more influential than diagnosis of diabetes alone, but further research is needed to verify this fact. Hyperlipidemia presence was important to PAD and may moderate the relationship between PAD and HRQOL. Hypertension and disease appropriate cut points for blood pressure confound the relationship between PAD and HRQOL. The Symptoms category was reduced to pain variables and Functional status was reduced to mobility difficulty and vigorous activity. General Health Perception was not predictive of the relationship between PAD and disability. The general health question is asked with the CDC HRQOL-4, so it makes the most sense to move it to the HRQOL box. Disability was the focus of this study and thus was placed at the top of the outcome in the original conceptual framework. Disability is one aspect of HRQOL, but in this study disability has not been found as useful as physical HRQOL for monitoring PAD. Figure 5.1 illustrates the modified form of the conceptual framework based on this study: the Peripheral Arterial Disease and Health-Related Quality of Life Conceptual Framework. 108 Characteristics of the Individual Demographics (Age, Gender, Race/Ethnicity, Marital status) Socioeconomic status (Education, PIR=income/poverty ratio) Biological Function Health-Related Peripheral Arterial Symptoms Quality of Life Disease (PAD) Pain in legs Disability Ankle-brachial index Pain in calves ------vs. Pain in feet Chronic Disease Healthy Days: Hypertension Physical or Activity Diabetes Functional Hyperlipidemia Status Biological risks General Health Nicotine exposure Mobility Perception Glycosylated hemoglobin Vigorous activity Systolic/diastolic BP Total and HDL cholesterol Adiposity (obesity, waist) Figure 5.1. Peripheral Arterial Disease and Health-Related Quality of Life Conceptual Framework (Based on Wilson & Cleary, 1995 model for Health-Related Quality of Life revised by Farrans, et al., 2005.) 109 Recommendations for Nursing Practice In this study only 6.1% of adult age 40 or over had an ABI less than 0.9, indicating few community dwellers have PAD. How do we locate the persons with mild disease when lifestyle changes may limit the progression of the PAD and prevent the development of disability? The Stay in Circulation campaign is encouraging adults over 50 to be screened for PAD. People self select themselves to be screened usually because they are symptomatic. Several questionnaires are used to monitor PAD patients after treatment, but more simple questionnaires in general physicians offices, in clinics and at health fairs may screen better for the 6-7% of the population that have ABI of 0.7-0.9, who are typically asymptomatic. Items from the NHANES that correlated with current recommendations for PAD screening (Hirsch et al., 2006) included: middle aged to older diabetic with hypertension and hyperlipidemia that had discomfort in the legs. Other items from the NHANES that correlated with mild PAD included: 1) more than 5 physically unhealthy days in the past 30 days, 2) inability to do 10 minutes of vigorous activity, 3) difficulty walking 2-3 blocks or walking up 10 steps without stopping, and 4) female. These are simple questions that can be asked quickly to determine who will benefit most from the time and cost of screening for PAD. The ankle-brachial index is reliable and fairly easy to perform when done by a nurse on a routine basis, but the ABI requires time to complete, training to be 110 accurate, and practice to be consistent in the measurement. Nurse practitioners need to be mindful of the type of patient to target for this testing. The portrait of a typical person with mild PAD based on the results of this study would be a non-Hispanic (black or white) woman around age 60 who was widowed, divorced, or separated, had a high school or lower education and had an income two to three times the poverty level. She sometimes had pain in her legs or calves with walking, sat or did light activity most days (office worker), choose not to or could not do vigorous activity (even 10 minutes in 30 days), could not walk 2-3 blocks or up 10 steps without stopping (at least briefly), and felt in poor health about a fourth of the time. She was treated for hypertension, but her blood pressure was higher than 130/80 mmHg. Her cotinine level was elevated from tobacco exposure (likely smoker). She was obese with a waist larger than 35 inches (88 cm), was borderline diabetic with glycosylated hemoglobin less than 7%, had chronic hypercholesterolemia with elevated total cholesterol, and she reported about 6 physically unhealthy days a month. The portrait of a typical person with severe PAD based on the results of this study would be a non-Hispanic black man in his early to mid 70 s who was widowed, divorced, or separated, had a high school or lower education and had an income twice the poverty level. He had pain in his legs or calves with walking, sat or did light activity most days, could not do vigorous activity (even 10 minutes in 30 days), could not walk 2-3 blocks or up 10 steps without stopping (at least briefly), and felt in poor health about half of the time, but did 111 not report more than 4-5 physically unhealthy days in the past month. He was treated for hypertension, but his systolic blood pressure was higher than 130 mmHg and the diastolic BP was close to 80 mmHg. He was overweight with a waist larger than 40 inches (102 cm), was diabetic with glycosylated hemoglobin less than 7%, had chronic hypercholesterolemia, but took medication to lower total cholesterol and increase HDL, and had a history of cardiovascular disease. Limitations of the Study The number of cases with PAD and disability were limited even in this large national data base, so the number of variables that could be introduced at one time was limited. This study was a secondary analysis consequently available variables had missing data and several desired variables were not available. Sample bias existed due to missing data in the eligible sample in a pattern similar to the characteristics of persons with PAD (older, more racially diverse, more females, less educated and lower income). There may be reverse causality between income and disability: disability may have limited the ability to earn income as much as the lower income contributed to the disability. Income figures may not be accurate for all the older adults in the study if only social security was reported and not pensions or other forms of income, so some of the older age with lower income may be spurious. Variables related to cardiovascular risk including triglyceride and low density lipoprotein levels, hormone replacement therapy history for females, exercise history, vascular 112 disease diagnosis or vascular intervention history, and drug therapies were totally unavailable or had excessive amounts of missing data. Recommendation for Future Research A prospective study using the conceptual framework and incorporating the variables missing from the secondary analysis is recommended. Repeating the regressions with additional years of NHANES data as it becomes available may provide a better understanding of PAD on a national level and direction for interventions. Additional NHANES data cycles release may allow the cell size to enlarge enough to keep the Models with interaction variables stable, so future testing is recommended. Chi square statistics dividing the severe PAD group may become possible when more years of data are available. Interventions that would be recommended from this study need to be tested for sensitivity and specificity. The CDC HRQOL-4, physical activity and mobility questions from NHANES are public domain, so adding 6-7 brief questions to assessment sheets and testing the usefulness in increasing identification of persons with PAD in office practice or screens at health fairs is recommended. Persons with low ABI would be encouraged to increase activity by joining a walking group and participating at least 3 times a week, since research has reported stabilization or improvement in ABI with that amount of activity (McDermott et al., 2006). These persons would also be encouraged to lessen other risk factors: monitor and treat elevated blood pressure, monitor and treat elevated blood sugar levels, decrease or stop exposure to tobacco 113 products, monitor and treat abnormal blood lipid levels, and reduce weight and waist measurements to nationally recommended levels. Peripheral arterial disease is an insidious threat to health-related quality of life that can lead to disability for those with severe disease. The goal is to prevent the disease for becoming severe enough to cause disability in the lower extremities or through myocardial infarction or stroke. Currently the prevalence of disability related to peripheral arterial disease is small, so early identification of PAD when disability can be avoided is a priority shared with the Stay in Circulation campaign. 114 LIST OF REFERENCES Adler, N. E., & Newman, K. (2002). Socioeconomic disparities in health: pathways and policies. Health Affairs, 21(2), 60-76. Albert, M., Glynn, R., & Ridker, P. (2003). Plasma concentration of C-reactive protein and the calculated Framingham Coronary Heart Disease Risk Score. Circulation, 108(2), 161-165. Allison, M. A., Ho, E., Denenberg, J. O., Langer, R. D., Newman, A. B., Fabsitz, R. R., et al. (2007). Ethnic-Specific Prevalence of Peripheral Arterial Disease in the United States. American Journal of Preventive Medicine, 32(4), 328-333. Anand, R. G., Ventura, H. O., & Mehra, M. R. (2007). Is heart failure more prevalent in patients with peripheral arterial disease? A meta-analysis. Congestive Heart Failure, 13(6), 319-322. Andresen, E. M., Catlin, T., & Wyrwich, K. (2001a). Retest reliability and validity of a surveillance measure of health-related quality of life. Quality of Life Research, 10(3), 199. Andresen, E. M., & Miller, D. K. (2005). The future (history) of socioeconomic measurement and implications for improving health outcomes among African Americans. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 60(10), 1345-1350. Andresen, E. M., Vahle, V., & Lollar, D. (2001b). Proxy reliability: Health-related quality of life (HRQOL) measures for people with disability. Quality of Life Research, 10, 609-619. Aquarius, A. E., Denollet, J., de Vries, J., & Hamming, J. F. (2007). Poor health- related quality of life in patients with peripheral arterial disease: Type D personality and severity of peripheral arterial disease as independent predictors. Journal of Vascular Surgery, 46(3), 507-512. 115 Aronow, W. (2004). Management of peripheral arterial disease of the lower extremities in elderly patients. The Journals of Gerontology Series A, 59(2), 172-177. Arseven, A., Guralnik, J. M., O'Brien, E., Liu, K., & McDermott, M. M. (2001). Peripheral arterial disease and depressed mood in older men and women. Vascular Medicine, 6(4), 229-234. Biasucci, L. M., Liuzzo, G., Grillo, R. L., Caligiuri, G., Rebuzzi, A. G., Buffon, A., et al. (1999). Elevated levels of C-reactive protein at discharge in patients with unstable angina predict recurrent instability. Circulation, 99(7), 855-860. Blake, G., & Ridker, P. (2002). Inflammatory bio-markers and cardiovascular risk prediction. Journal International Medicine, 252(4), 283-294. Bordeaux, L., Reich, L., & Hirsch, A. (2003). Epidemiology and natural history of peripheral arterial disease. In J. Coffman & R. Eberhardt (Eds.), Contemporary cardiology: Peripheral arterial disease: Diagnosis and treatment. (pp. 21-34). Totowa, NJ: Humana Press Inc. Boyle, J. (1997). Association of coronary plaque rupture and atherosclerotic inflammation. Journal of Pathology, 181(1), 93-99. Buck, M., & Ciccone, C. (2004). Does evidence exist on whether specific interventions can improve adherence to a home exercise program in a patient with intermittent claudication? Physical Therapy, 84(5), 465-473. Burns, P., Gough, S., & Bradbury, A. W. (2003). Management of peripheral arterial disease in primary care. BMJ (Clinical research ed), 326(7389), 584-588. Campinha-Bacote, J. (1998). African-Americans. In L. Purnell & B. Paulanka (Eds.), Transcultural Health Care: A Culturally Competent Approach (pp. 53-73). Philadelphia, PA: F.A.Davis Company. Centers for Disease Control and Prevention. (2000). Measuring Healthy Days (November ed.). Atalnta, Georgia: CDC. Chew, D. P., Bhatt, D. L., Robbins, M. A., Penn, M. S., Schneider, J. P., Lauer, M. S., et al. (2001). Incremental prognostic value of elevated baseline C- reactive protein among established markers of risk in percutaneous coronary intervention. Circulation, 104(9), 992-997. 116 Chobanian, A. V., Bakris, G. L., Black, H. R., Cushman, W. C., Green, L. A., Izzo, J. L., Jr, et al. (2003). Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension, 42(6), 1206-1252. Cimminiello, C. (2002). PAD. Epidemiology and pathophysiology. Thrombosis Research, 106(6), V295-301. Collins, T. C., Petersen, N. J., Suarez-Almazor, M., & Ashton, C. M. (2003). The prevalence of peripheral arterial disease in a racially diverse population. Archives of Internal Medicine, 163(12), 1469-1474. Collins, T. C., Suarez-Almazor, M., Bush, R. L., & Petersen, N. J. (2006). Gender and Peripheral Arterial Disease. Journal of the American Board of Family Medicine, 19(2), 132- 140. Currie, I. C., Wilson, Y. G., Baird, R. N., & Lamont, P. M. (1995). Treatment of intermittent claudication: the impact on quality of life. European Journal of Vascular and Endovascular Surgery, 10(3), 356-361. Danesh, J., Whincup, P., Walker, M., Lennon, L., Thomson, A., Appleby, P., et al. (2000). Low grade inflammation and coronary heart disease: prospective study and updated meta-analyses. BMJ (Clinical research ed), 321(7255), 199-204. Diabetes Control and Complications Trial. (1993). The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. The Diabetes Control and Complications Trial Research Group. The New England Journal of Medicine, 329(14), 977-986. Diabetes Control and Complications Trial. (1996). Lifetime benefits and costs of intensive therapy as practiced in the diabetes control and complications trial. The Diabetes Control and Complications Trial Research Group. JAMA, 276(17), 1409-1415. Dirksen, M. T., van der Wal, A. C., van den Berg, F. M., van der Loos, C. M., & Becker, A. E. (1998). Distribution of inflammatory cells in atherosclerotic plaques relates to the direction of flow. Circulation, 98(19), 2000-2003. Ekers, M. A. (1986). Psychosocial considerations in peripheral vascular disease. Cause or effect? The Nursing Clinics of North America, 21(2), 255-263. 117 Fabsitz, R. R., Sidawy, A. N., Go, O., Lee, E. T., Welty, T. K., Devereux, R. B., et al. (1999). Prevalence of peripheral arterial disease and associated risk factors in American Indians: the Strong Heart Study. American Journal of Epidemiology, 149(4), 330-338. Farrans, C. E., Zerwic, J. J., Wilber, J. E., & Larson, J. L. (2005). Conceptual model of health-related quality of life. Journal of Nursing Scholarship, 37(4), 336-342. Feinglass, J., Morasch, M., & McCarthy, W. J. (2000). Measures of success and health-related quality of life in lower-extremity vascular surgery. Annual Review of Medicine, 51, 101-113. Frangos, S. G., Gahtan, V., & Sumpio, B. (1999). Localization of atherosclerosis: role of hemodynamics. Archives of Surgery, 134(10), 1142-1149. Gardner, A., & Killewich, L. (2001). Lack of functional benefits following infrainguinal bypass in peripheral arterial occlusive disease patients. Vascular Medicine, 6(1), 9-14. Gardner, A. W. (2002). Sex differences in claudication pain in subjects with peripheral arterial disease. Medicine and Science in Sports and Exercise, 34(11), 1695-1698. Gibson, J., & Kenrick, M. (1998). Pain and powerlessness: the experience of living with peripheral vascular disease. Journal of Advanced Nursing, 27(4), 737-745. Graham, L., & Ford, M. (1994). Arterial disease. In V. Fahey (Ed.), Vascular Nursing (pp. 3-20). Philadelphia: W. B. Saunders Company. Greaves, D. R., & Channon, K. M. (2002). Inflammation and immune responses in atherosclerosis. Trends in Immunology, 23(11), 535-541. Gregg, E. W., Sorlie, P., Paulose-Ram, R., Gu, Q., Eberhardt, M. S., Wolz, M., et al. (2004). Prevalence of lower-extremity disease in the U.S. adult population >=40 years of age with and without diabetes: 1999-2000 national health and nutrition examination survey. Diabetes Care, 27(7), 1591-1597. Grundy, S. M., Balady, G., Criqui, M., Fletcher, G., Greenland, P., Hiratzka, L., et al. (1998). Primary prevention of coronary heart disease: guidance from Framingham: a statement for healthcare professionals from the 118 AHA Task Force on Risk Reduction. American Heart Association. Circulation, 97(18), 1876-1887. Grundy, S. M., Hansen, B., Smith, S. C. J., Cleeman, J. I., & Kahn, R. A. (2004). Clinical management of metabolic syndrome: report of the American Heart Association/National Heart, Lung, and Blood Institute/American Diabetes Association conference on scientific issues related to management. Circulation, 109(4), 551-556. Guallar, E., Silbergeld, E. K., Navas-Acien, A., Malhotra, S., Astor, B. C., Sharrett, A. R., et al. (2006). Confounding of the relation between homocysteine and peripheral arterial disease by lead, cadmium, and renal function. American Journal of Epidemiology, 163(8), 700-708. Hallin, A., Bergqvist, D., Fugl-Meyer, K., & Holmberg, L. (2002). Areas of concern, quality of life and life satisfaction in patients with peripheral vascular disease. European Journal of Vascular and Endovascular Surgery, 24(3), 255-263. Halperin, J. L. (2002). Evaluation of patients with peripheral vascular disease. Thrombosis Research, 106(6), V303-311. Harwood, G. A. (2006). Nicotine dependence and socioeconomic status in hard core smokers. Unpublished Doctoral Dissertation, The Ohio State University, Columbus. Hennessy, C. H., Moriarty, D. G., Zack, M. M., Scherr, P. A., & Brackbill, R. (1994). Measuring health-related quality of life for public health surveillance. Public Health Reports, 109, 665-672. Hiatt, W. (1997). Quality of life assessment in peripheral arterial disease. Atherosclerosis, 131, S35-S36. Hirsch, A., & Gotto, A. J. (2002). Undertreatment of dyslipidemia in peripheral arterial disease and other high-risk populations: an opportunity for cardiovascular disease reduction. Vascular Medicine, 7(4), 323-331. Hirsch, A. T., Gloviczki, P., Drooz, A., Lovell, M., & Creager, M. A. (2004). Mandate for creation of a national peripheral arterial disease public awareness program: an opportunity to improve cardiovascular health. Journal of Vascular Surgery, 39(2), 474-481. Hirsch, A. T., Haskal, Z. J., Hertzer, N. R., Bakal, C. W., Creager, M. A., Halperin, J. L., et al. (2006). ACC/AHA 2005 Practice Guidelines for the 119 Management of Patients With Peripheral Arterial Disease (Lower Extremity, Renal, Mesenteric, and Abdominal Aortic): A Collaborative Report from the American Association for Vascular Surgery/Society for Vascular Surgery,* Society for Cardiovascular Angiography and Interventions, Society for Vascular Medicine and Biology, Society of Interventional Radiology, and the ACC/AHA Task Force on Practice Guidelines (Writing Committee to Develop Guidelines for the Management of Patients With Peripheral Arterial Disease): Endorsed by the American Association of Cardiovascular and Pulmonary Rehabilitation; National Heart, Lung, and Blood Institute; Society for Vascular Nursing; TransAtlantic Inter-Society Consensus; and Vascular Disease Foundation. Circulation, 113(11), e463-465. Hirsch, A. T., Murphy, T. P., Lovell, M. B., Twillman, G., Treat-Jacobson, D., Harwood, E. M., et al. (2007). Gaps in public knowledge of peripheral arterial disease. The first national PAD public awareness survey. Circulation, 116(18), 2086-2094. Hobbs, S. D., Wilmink, A. B. M., & Bradbury, A. W. (2003). Ethnicity and peripheral arterial disease. European Journal of Vascular and Endovascular Surgery, 25(6), 505-512. Hooi, J., Kester, A., Stoffers, H., Overdijk, M., van Ree, J., & Knottnerus, J. (2001). Incidence of and risk factors for asymptomatic peripheral arterial occlusive disease: a longitudinal study. American Journal of Epidemiology, 153(7), 666-672. Hosmer, D. W., & Lemeshow, S. (2000). Applied logistic regression (2nd ed. ed.). New York Wiley. Hunt, S., McEwen, J., & McKenna, S. (1986). Measuring Health Status (1 ed.). London: Crim Helm. Hunt, S. M., & McEwen, J. (1980). The development of a subjective health indicator Sociology of Health and Illness, 2(3), 231-246. Johnstone, C. C. (2003). Living with peripheral vascular disease: a review of the literature. Professional Nurse, 18(8), 446-449. Kalbaugh, C., Taylor, S., Cull, D., Blackhurst, D., Gray, B., Langan, E. r., et al. (2004). Invasive treatment of chronic limb ischemia according to the Lower Extremity Grading System (LEGS) score: a 6-month report. Journal of Vascular Surgery, 39(6), 1268-1276. 120 Kanjilal, S., Gregg, E. W., Cheng, Y. J., Zhang, P., Nelson, D. E., Mensah, G., et al. (2006). Socioeconomic status and trends in disparities in 4 major risk factors for cardiovascular disease among US adults, 1971-2002. Archives of Internal Medicine, 166(21), 2348-2355. Kannel, W., & McGee, D. (1985). Update on some epidemiologic features of intermittent claudication: The Framingham study. Journal of the American Geriatrics Society, 33, 13-18. Kaplan, R. M., & Bush, J. W. (1982). Health-related quality of life measurement for evaluation research and policy analysis. Health Psychology, 1, 61. Kiechl, S., & Willeit, J. (1999a). The natural course of atherosclerosis. Part I: incidence and progression. Arteriosclerosis, Thrombosis, and Vascular Biology, 19(6), 1484-1490. Kiechl, S., & Willeit, J. (1999b). The natural course of atherosclerosis. Part II: vascular remodeling. Bruneck Study Group. Arteriosclerosis, Thrombosis, and Vascular Biology, 19(6), 1491-1498. Klevsgard, R., Hallberg, I. R., Risberg, B., & Thomsen, M. B. (1999). Quality of life associated with varying degrees of chronic lower limb ischaemia: comparison with a healthy sample. European Journal of Vascular and Endovascular Surgery, 17(4), 319-325. Klevsgard, R., Hallberg, I. R., Risberg, B., & Thomsen, M. B. (2000). The effects of successful intervention on quality of life in patients with varying degrees of lower-limb ischaemia. European Journal of Vascular and Endovascular Surgery, 19(3), 238-245. Kobayashi, S., Inoue, N., Ohashi, Y., Terashima, M., Matsui, K., Mori, T., et al. (2003). Interaction of oxidative stress and inflammatory response in coronary plaque instability: important role of C-reactive protein. Arteriosclerosis Thrombosis and Vascular Biology, 23(8), 1398-1404. Koenig, W., Sund, M., Frohlich, M., Fischer, H. G., Lowel, H., Doring, A., et al. (1999). C-Reactive protein, a sensitive marker of inflammation, predicts future risk of coronary heart disease in initially healthy middle-aged men: results from the MONICA (Monitoring Trends and Determinants in Cardiovascular Disease) Augsburg Cohort Study, 1984 to 1992. Circulation, 99(2), 237-242. 121 Kuo, H.-K., Bean, J. F., Yen, C.-J., & Leveille, S. G. (2006). Linking C-reactive protein to late-life disability in the National Health and Nutrition Examination Survey (NHANES) 1999-2002. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 61(4), 380-387. Kuo, H.-K., Leveille, S. G., Yu, Y.-H., & Milberg, W. P. (2007). Cognitive function, habitual gait speed, and late-life disability in the National Health and Nutrition Examination Survey (NHANES) 1999-2002. Gerontology, 53(2), 102-110. Kuo, H.-K., Lin, L.-Y., & Yu, Y.-H. (2007). Microalbuminuria is a negative correlate for cognitive function in older adults with peripheral arterial disease: Results from the U.S. National Health and Nutrition Examination Survey 1999 2002. Journal of Internal Medicine, 262(5), 562-570. Lane, J. S., Vittinghoff, E., Lane, K. T., Hiramoto, J. S., & Messina, L. M. (2006). Risk factors for premature peripheral vascular disease: results for the National Health and Nutritional Survey, 1999-2002. Journal of Vascular Surgery, 44(2), 319. Liles, D. R., Kallen, M. A., Petersen, L. A., & Bush, R. L. (2006). Quality of Life and Peripheral Arterial Disease. Journal of Surgical Research, 136, 294- 301. Long, J., Modrall, J., Parker, B., Swann, A., Welborn, M. r., & Anthony, T. (2004). Correlation between ankle-brachial index, symptoms, and health- related quality of life in patients with peripheral vascular disease. Journal of Vascular Surgery, 39(4), 723-727. Lower extremity disease among persons aged > or =40 years with and without diabetes--United States, 1999-2002. (2005). MMWR. Morbidity And Mortality Weekly Report, 54(45), 1158-1160. Lyness, J., Noel, T., Cox, C., King, D., Conwell, Y., & Caine, E. (1997). Screening for depression in elderly primary care patients: a comparison of the center for epidemiologic studies-depression scale and the geriatric depression scale. Archives of Internal Medicine, 157(4), 449-454. Majewski, W., Zielinski, A., Laciak, M., Staniszewski, R., Gorny, A., Zapalski, S., et al. (1993). C-reactive protein and alpha-1-acid glycoprotein in monitoring of patients with chronic arterial occlusion of the lower limbs. European Journal of Vascular Surgery, 7(6), 628-632. 122 Marmot, M. G. (1999). Epidemiology of socioeconomic status and health: are determinants within countries the same as between countries? Annals of the New York Academy of Sciences, 896, 16-29. Marmot, M. G. (2003). Understanding social inequalities in health. Perspectives in Biology and Medicine, 46(3 Suppl), S9-23. Marso, S. P., & Hiatt, W. R. (2006). Peripheral Arterial Disease in Patients With Diabetes. Journal of the American College of Cardiology, 47(5), 921 929. McDermott, M., Hahn, E., Greenland, P., Cella, D., Ockene, J., Brogan, D., et al. (2002). Atherosclerotic risk factor reduction in peripheral arterial diseasea: results of a national physician survey. Journal of General Internal Medicine, 17(12), 895-904. McDermott, M., Mandapat, A., Moates, A., Albay, M., Chiou, E., Celic, L., et al. (2003a). Knowledge and attitudes regarding cardiovascular disease risk and prevention in patients with coronary or peripheral arterial disease. Archives of Internal Medicine, 163(18), 2157-2162. McDermott, M. M., Greenland, P., Guralnik, J. M., Liu, K., Criqui, M. H., Pearce, W. H., et al. (2003b). Depressive symptoms and lower extremity functioning in men and women with peripheral arterial disease. Journal of general internal medicine, 18(6), 461-467. McDermott, M. M., Greenland, P., Liu, K., Criqui, M. H., Guralnik, J. M., Celic, L., et al. (2003c). Sex differences in peripheral arterial disease: leg symptoms and physical functioning. Journal of the American Geriatrics Society, 51(2), 222-228. McDermott, M. M., Guralnik, J. M., Greenland, P., Pearce, W. H., Criqui, M. H., Liu, K., et al. (2003d). Statin use and leg functioning in patients with and without lower-extremity peripheral arterial disease. Circulation, 107(5), 757-761. McDermott, M. M., Liu, K., Ferrucci, L., Criqui, M. H., Greenland, P., Guralnik, J. M., et al. (2006). Physical Performance in Peripheral Arterial Disease: A Slower Rate of Decline in Patients Who Walk More. Annals of Internal Medicine, 144(1), 10-20. McKenney, J. (2003). Update on the National Cholesterol Education Program Adult Treatment Panel III guidelines: getting to goal. Pharmacotherapy, 23(9 Pt 2), 26S-33S. 123 Menard, J., Smith, H., Riebe, D., Braun, C., Blissmer, B., & Patterson, R. (2004). Long-term results of peripheral arterial disease rehabilitation. Journal of Vascular Surgery, 39(6), 1186-1192. Menke, A., Muntner, P., Wildman, R. P., Dreisbach, A. W., & Raggi, P. (2006). Relation of borderline peripheral arterial disease to cardiovascular disease risk. The American Journal of Cardiology, 98(9), 1226-1230. Mobility limitation among persons aged > or =40 years with and without diagnosed diabetes and lower extremity disease--United States, 1999- 2002. (2005). MMWR. Morbidity And Mortality Weekly Report, 54(46), 1183-1186. Mohler, E. R. r., Hiatt, W. R., & Creager, M. A. (2003). Cholesterol reduction with atorvastatin improves walking distance in patients with peripheral arterial disease. Circulation, 108(12), 1481-1486. Moriarty, D., & Zack, M. (1999). Validation of the Centers for Disease Control and Prevention's Healthy Days measures. Quality of Life Research, 8(7), 617. Moriarty, D., Zack, M., & Kobau, R. (2003). The Centers for Disease Control and Prevention's Healthy Days Measures - Population tracking of perceived physical and mental health over time. Health and Quality of Life Outcomes, 1(1), 37. Moriarty, D. G., Kobau, R., Zack, M. M., & Zahran, H. S. (2005). Tracking Healthy Days -- a window on the health of older adults. Preventing Chronic Disease, 2(3), A16. Muntner, P., Menke, A., DeSalvo, K. B., Rabito, F. A., & Batuman, V. (2005a). Continued decline in blood lead levels among adults in the United States. Archives of Internal Medicine, 165(18), 2155-2161. Muntner, P., Wildman, R. P., Reynolds, K., Desalvo, K. B., Chen, J., & Fonseca, V. (2005b). Relationship between HbA1c level and peripheral arterial disease. Diabetes Care, 28(8), 1981 1987. Mya, M. M., & Aronow, W. S. (2003). Increased prevalence of peripheral arterial disease in older men and women with subclinical hypothyroidism. The Journals of Gerontology Series A Biological Sciences and Medical Sciences, 58(1), 68-69. 124 Nasir, K., Guallar, E., Navas-Acien, A., Criqui, M. H., & Lima, J. A. C. (2005). Relationship of monocyte count and peripheral arterial disease: results from the National Health and Nutrition Examination Survey 1999-2002. Arteriosclerosis, Thrombosis, and Vascular Biology, 25(9), 1966-1971. National Heart Lung and Blood Institute. (2002). Third report of the National Cholesterol Education Program (NCEP) Expert Panel on the Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel [ATP] III). (NIH Publication No. 02-5215). Bethesda, MD: National Heart, Lung and Blood Institute, National Institute of Health, US. National Heart Lung and Blood Institute. (2006). NHLBI Launches National Campaign To Educate About Peripheral Arterial Disease Clogged Arteries in the Legs. Retrieved. from http://www.nih.gov/news/pr/sep2006/nhlbi-13.htm. Navas-Acien, A., Selvin, E., Sharrett, A., Calderon-Aranda, E., Silbergeld, E., & Guallar, E. (2004). Lead, cadmium, smoking, and increased risk of peripheral arterial disease. Circulation, 109(25), 3196-3201. Navas-Acien, A., Silbergeld, E., Sharrett, A., Calderon-Aranda, E., Selvin, E., & Guallar, E. (2005). Metals in Urine and Peripheral Arterial Disease. Environmental Health Perspectives, 113(2), 164-169. Nelson, K. M., Reiber, G., Kohler, T., & Boyko, E. J. (2007). Peripheral arterial disease in a multiethnic national sample: the role of conventional risk factors and allostatic load. Ethnicity & Disease, 17(4), 669-675. O'Hare, A. M., Feinglass, J., Sidawy, A. N., Bacchetti, P., Rodriguez, R. A., Daley, J., et al. (2003). Impact of renal insufficiency on short-term morbidity and mortality after lower extremity revascularization: data from the Department of Veterans Affairs' National Surgical Quality Improvement Program. Journal of the American Society of Nephrology, 14(5), 1287-1295. O'Hare, A. M., Glidden, D. V., Fox, C. S., & Hsu, C.-Y. (2004). High prevalence of peripheral arterial disease in persons with renal insufficiency: results from the National Health and Nutrition Examination Survey 1999-2000. Circulation, 109(3), 320-323. Oakes, J. M., & Rossi, P. H. (2003). The measurement of SES in health research: current practice and steps toward a new approach. Social Science & Medicine, 56(4), 769-784. 125 Ockene, I., & Miller, N. (1997). Cigarette smoking, cardiovascular disease, and stroke: a statement for healthcare professionals from the American Heart Association. American Heart Association Task Force on Risk Reduction. Circulation, 96(9), 3243-3247. Ostchega, Y., Paulose-Ram, R., Dillon, C. F., Gu, Q., & Hughes, J. P. (2007). Prevalence of peripheral arterial disease and risk factors in persons aged 60 and older: data from the National Health and Nutrition Examination Survey 1999-2004. Journal of the American Geriatrics Society, 55(4), 583-589. Pasceri, V., Willerson, J. T., & Yeh, E. T. (2000). Direct proinflammatory effect of C-reactive protein on human endothelial cells. Circulation, 102(18), 2165-2168. Pell, J. P. (1995). Impact of intermittent claudication on quality of life. The Scottish Vascular Audit Group. European Journal of Vascular and Endovascular Surgery, 9(4), 469-472. Pell, J. P., Donnan, P. T., Fowkes, F. G., & Ruckley, C. V. (1993). Quality of life following lower limb amputation for peripheral arterial disease. European Journal of Vascular Surgery, 7(4), 448-451. Pepys, M. (2003). Amyloid, familial Meditertanean fever, and acute phase response. In D. A. Warrell (Ed.), Oxford Textbook of Medicine (4th ed., Vol. 2, pp. 150-156). New York: Oxford University Press. Pepys, M., & Hirschfield, G. (2003). C-reactive protein: a critical update. The Journal of Clinical Investigation, 111(12), 1805-1812. Peripheral Arterial Disease Coalition. (2007). Stay in Circulation Campaign [Electronic Version]. PAD Coalition E-News, , 09 October 2007, 1. Retrieved 12-5-2007 from http://www.padcoalition.org/wp/awareness- campaign/ and home is http://www.padcoalition.org/wp/. Perlstein, T. S., Pande, R. L., Beckman, J. A., & Creager, M. A. ( 2008). Serum total bilirubin level and prevalent lower-extremity peripheral arterial disease national health and nutrition examination survey (NHANES) 1999 to 2004. Arteriosclerosis, Thrombosis, and Vascular Biology, 28(1), 166-172. 126 Resnick, H. E., & Foster, G. L. (2005). Prevalence of elevated ankle-brachial index in the United States 1999 to 2002. The American Journal of Medicine, 118(6), 676-679. Ridker, P., Buring, J., Cook, N., & Rifai, N. (2003). C-reactive protein, the metabolic syndrome, and risk of incident cardiovascular events: an 8- year follow-up of 14 719 initially healthy American women. Circulation, 107(3), 391-397. Ridker, P., Cushman, M., Stampfer, M., Tracy, R., & Hennekens, C. (1998). Plasma concentration of C-reactive protein and risk of developing peripheral vascular disease. Circulation, 97, 425-428. Ridker, P., Rifai, N., Rose, L., Buring, J., & Cook, N. (2002). Comparison of C- reactive protein and low-density lipoprotein cholesterol levels in the prediction of first cardiovascular events. The New England Journal of Medicine, 347(20), 1557-1565. Rifai, N., Buring, J. E., Lee, I.-M., Manson, J. E., & Ridker, P. M. (2002). Is C- reactive protein specific for vascular disease in women? Annals of Internal Medicine, 136(7), 529-533. Rosendorff, C., Black, H. R., Cannon, C. P., Gersh, B. J., Gore, J., Izzo Jr, J. L., et al. (2007). Treatment of hypertension in the prevention and management of ischemic heart disease: A scientific statement from the American heart association council for high blood pressure research and the councils on clinical cardiology and epidemiology and prevention. Circulation, 115, 2761-2788. Ross, R. (1993). The pathogenesis of atherosclerosis: a persective for the 1990's. Nature, 362, 801-809. Ross, R. (1999). Atherosclerosis--an inflammatory disease. The New England Journal of Medicine, 340(2), 115-126. Ross, R. (2004). Pathophysiology of atherosclerosis. In R. I. Hobson, S. Wilson & F. Weith (Eds.), Vascular surgery: Principles and practice (3rd, revised and expanded ed., pp. 15-29). New York, NY: Marcel Dekker, Inc. Rossi, E., Biasucci, L. M., Citterio, F., Pelliccioni, S., Monaco, C., Ginnetti, F., et al. (2002). Risk of myocardial infarction and angina in patients with severe peripheral vascular disease: predictive role of C-reactive protein. Circulation, 105(7), 800-803. 127 Schnittker, J. (2004). Education and the changing shape of the income gradient in health. Journal of Health and Social Behavior, 45(3), 286-305. Selvin, E., & Erlinger, T. P. (2004). Prevalence of and risk factors for peripheral arterial disease in the United States: results from the National Health and Nutrition Examination Survey, 1999-2000. Circulation, 110(6), 738-743. Shankar, A., Klein, B. E. K., Nietoc, F. J., & Klein, R. (2008). Association between serum uric acid level and peripheral arterial disease Atherosclerosis 196(2), 749-755. Shankar, A., Li, J., Klein, B. E. K., Nietoc, F. J., & Klein, R. (2007). Serum gamma-glutamyltransferase level and peripheral arterial disease. Atherosclerosis Shankar, A., Li, J., Nieto, F. J., Klein, B. E. K., & Klein, R. (2007). Association between C-reactive protein level and peripheral arterial disease among US adults without cardiovascular disease, diabetes, or hypertension. American Heart Journal, 154(3), 495-501. Shechter, M., Auslander, G., Weinmann, E. E., & Bass, A. (2003). Quality of life and social support following distal arterial bypass in elderly patients. The Israel Medical Association Journal, 5(5), 322-325. Society for Research on Nicotine and Tobacco. (2002). Biochemical verification of tobacco use and cessation. Nicotine & Tobacco Research 4(2), 149- 159. Society for Vascular Nursing, S. (2006). Peripheral arterial disease education and ABI training for vascular nursing [booklet and CD powerpoint presentation]. Salem, MA: Society for Vascular Nursing. St-Onge, M., & Heymsfield, S. (May 25, 2004). Reducing CVD Risk Through Appropriate Weight Management. Retrieved June 2, 2004, from http://www.medscape.com/viewprogram/3118_pnt Strong, J. P., Malcom, G. T., McMahan, C. A., Tracy, R. E., Newman, W. P., 3rd, Herderick, E. E., et al. (1999). Prevalence and extent of atherosclerosis in adolescents and young adults: implications for prevention from the Pathobiological Determinants of Atherosclerosis in Youth Study. JAMA, 281(8), 727-735. Stühlinger, M., & Tsao, P. (2003). Etiology and pathogenesis of atherosclerosis. In J. Coffman & R. Eberhardt (Eds.), Contemporary cardiology: 128 Peripheral arterial disease: Diagnosis and treatment. (pp. 1-19). Totowa, NJ: Humana Press Inc. Thorsen, H., McKenna, S., Tennant, A., & Holstein, P. (2002). Nottingham health profile scores predict the outcome and support aggressive revascularisation for critical ischaemia. European Journal of Vascular and Endovascular Surgery, 23(6), 495-499. Torzewski, M., Rist, C., Mortensen, R. F., Zwaka, T. P., Bienek, M., Waltenberger, J., et al. (2000). C-reactive protein in the arterial intima: role of C-reactive protein receptor-dependent monocyte recruitment in atherogenesis. Arteriosclerosis, Thrombosis, and Vascular Biology, 20(9), 2094-2099. Treat-Jacobson, D., Halverson, S., Ratchford, A., Regensteiner, J., Lindquist, R., & Hirsch, A. (2002). A patient-derived perspective of health-related quality of life with peripheral arterial disease. Journal of Nursing Scholarship, 34(1), 55-60. Treat-Jacobson, D., & Walsh, M. E. (2003). Treating patients with peripheral arterial disease and claudication. Journal of Vascular Nursing, 21(1), 5- 14; quiz 15-16. Tretinyak, A., Lee, E., Kuskowski, M., Caldwell, M., & Santilli, S. (2001). Revascularization and quality of life for patients with limb-threatening ischemia. Annals of Vascular Surgery, 15(1), 84-88. Verbrugge, L. M., Merrill, S. S., & Liu, X. (1999). Measuring disability with parsimony. Disability and Rehabilitation, 21(5-6), 295 306. Vu, J. D., Vu, J. B., Pio, J. R., Malik, S., Franklin, S. S., Chen, R. S., et al. (2005). Impact of C-reactive protein on the likelihood of peripheral arterial disease in United States adults with the metabolic syndrome, diabetes mellitus, and preexisting cardiovascular disease. The American Journal Of Cardiology, 96(5), 655-658. Wall, T., Schoedel, K., Ring, H., Luczak, S., Katsuyoshi, D., & Tyndale, R. (2007). Differences in pharmacogenetics of nicotine and alcohol metabolism: Review and recommendations for future research. Nicotine & Tobacco Research, 9(Suppl 3), 459-474. Wann-Hansson, C., Hallberg, I. R., Risberg, B., Lundell, A., & Klevsgard, R. (2005). Health-related quality of life after revascularization for peripheral 129 arterial occlusive disease: long-term follow-up. Journal of Advanced Nursing 51(3), 227-235. Ware, J., Kosinski, & Dewey. (2000). How to score version 2 of SF-36 Health Survey. Lincoln, RI: QualityMetric Incorporated. Ware, J., & Sherbourne, C. (1992). The MOS 36-item short-form health survey (SF-36): I. conceptual framework and item selection. Medical Care, 30(6), 473-483. Weinhold, B., & Rüther, U. (1997). Interleukin-6-dependent and -independent regulation of the human c-reactive protein gene. The Biochemical Journal, 327, 425-429. Weitz, J., Byrne, J., Clagett, G., Farkouh, M., Porter, J., Sackett, D., et al. (1996). Diagnosis and treatment of chronic arterial insufficiency of the lower extremities: a critical review. Circulation, 94(11), 3026-3049. Westendorp, I., in't Veld, B., Grobbee, D., Pols, H., Meijer, W., Hofman, A., et al. (2000). Hormone replacement therapy and peripheral arterial disease: the Rotterdam study. Archives of Internal Medicine, 160(16), 2498-2502. Wildman, R., Muntner, P., Chen, J., Sutton-Tyrrell, K., & He, J. (2005). Relation of inflammation to peripheral arterial disease in the National Health and Nutrition Examination Survey, 1999-2002. American Journal of Cardiology, 96(11), 1579-1583. Wilson, I. B., & Cleary, P. D. (1995). Linking clinical variables with health- related quality of life; A conceptual model of patient outcomes. JAMA, 273(1), 59-65. Wilson, P. W. F., & Grundy, S. M. (2003). The metabolic syndrome: a practical guide to origins and treatment: Part II. Circulation, 108(13), 1537-1540. Wortley, P., Caraballo, R., Pederson, L., & Pechacek, T. (2002). Exposure to secondhand smoke in the workplace: serum cotinine by occupation. Journal of Occupatioinal and Environmental Medicine, 44(6), 503-509. Wrobel, J., Mayfield, J., & Reiber, G. (2001). Geographic variation of lower- extremity major amputation in individuals with and without diabetes in the medicare population. Diabetes Care, 24(5), 860-864. 130 Yao, S. T., Hobbs, J. T., & Irvine, W. T. (1969). Ankle systolic pressure measurements in arterial disease affecing the lower extremities. British Journal of Surgery, 56(9), 676-679. Yasojima, K., Schwab, C., McGeer, E., & McGeer, P. (2001). Generation of C- reactive protein and compliment componebts in atherosclerotic plaques. American Journal of Pathology, 158(3), 1039-1051. Zahran, H. S., Kobau, R., Moriarty, D. G., Zack, M. M., Holt, J., Donehoo, R., et al. (2005). Health-related quality of life surveillance--United States, 1993- 2002. Morbidity and Mortality Weekly Report. Surveillance Summaries / CDC, 54(4), 1-35. Zwaka, T. P., Hombach, V., & Torzewski, J. (2001). C-Reactive protein- mediated low density lipoprotein uptake by macrophages: implications for atherosclerosis. Circulation, 103, 1194-1197. 131 APPENDIX A Listing of NHANES files included in this study by category. Information is copied from the NHANES websites (http://www.cdc.gov/nchs/about/major/nhanes/nhanes01-02.htm and http://www.cdc.gov/nchs/about/major/nhanes/nhanes2003- 2004/nhanes03_04.htm) 132 NHANES File Items Item descriptions Code used for item in Names included NHANES Demographics and Questionnaires. The respondent completed most questionnaires, including demographic information, during a face-to-face interview in their home. Some questionnaires were completed using computer assisted personal interview (CAPI) within the home or mobile examination center (MEC) for items about sensitive material (noted with *). DEMO_B SEQN Respondent sequence number Sequential numbering DEMO_C Found in every file, but only of respondents listed here (in this table) . through all cycles SDDSRVYR Data release cycle number 2=2001-2002 3=2003-2004 RIDSTATR Interview/Examination status 1=interview only 2=interview and examination RIAGENDR Gender - adjudicated 1=male 2=female RIDAGEYR Age at screening adjudicated- 0-84= years rounded recode 85=> 85 RIDRETH2 Race/Ethnicity linked to NHANES 3 1=non-Hispanic white recoded 2= non-Hispanic black 3=Mexican American 4=other race 5=other Hispanic DMDEDUC Education recode 1=less than HS 2=HS diploma/GED (HS=High School) 3=more than HS 7=refused 9=don t know INDFMPIR CPS family PIR 0-4.99 = actual value (poverty income ratio) 5= > 5.00 DMDMARTL Marital status 1=married 2=widowed 3=divorced 4=separated 5=never married 6=living with partner 77=refused 99=don t know RIDEXPRG Pregnancy status at time of MEC 1=yes, pregnant 2=no 3=cannot ascertain Figure A.1 Appendix A Listing of NHANES files included in this study by category. Information is copied from the NHANES websites (http://www.cdc.gov/nchs/about/major/nhanes/nhanes01-02.htm and http://www.cdc.gov/nchs/about/major/nhanes/nhanes2003- 2004/nhanes03_04.htm) Continued 133 Figure A.1 (continued) NHANES File Items Item descriptions Code used for item in Names included NHANES DEMO_B SIAPROXY Proxy used in sample person (SIA), 1=yes DEMO_C FIAPROXY family (FIA), or MEC (MIA) 2=no (continued) MIAPROXY interviews? SIAINTRP Interpreter used in sample person FIAINTRP (SIA), family (FIA) or MEC (MIA) MIAINTRP interviews? WTMEC2YR Full sample 2 year MEC exam 0 to weights 211850.66405 SDMVPSU Pseudo-PSU and Pseudo-Stratum PSU = 1 to 2 SDMVSTRA Masked variance units for variance Stratum = 14 to 43 estimation BPQ_B BPQ020 Told you have hypertension? 1=yes BPQ_C BPQ050A Now taking BP prescription? 2=no BPQ080 Told you have high cholesterol? 7=refused BPQ100D Now taking cholesterol-lowering 9=don t know prescription? DIQ_B DIQ010 Told you have diabetes now? 1=yes DIQ_C DIQ050 Taking insulin now? 2=no DIQ070 Taking oral diabetic medication? 7=refused 9=don t know DIQ090 Ulcer on foot? 1=yes DIQ120 Pain or tingling in past 3 months? 2=no DIQ140 Pain in legs with walking? 7=refused DIQ150 Pain in calves when walking? 9=don t know DIQ130 Pain or tingling where? 1=hands 2=feet 3=both HSQ_B * HSD010 I have some general questions 1=excellent HSQ_C * about your health. 2=very good Would you say your health in 3=good general is excellent, very good, 4=fair good, fair, or poor? 5=poor 7=refused 9=don t know HSQ470 Thinking about your physical health, 0 30 DAYS which includes physical illness and 77=refused injury, for how many days during the 99=don t know past 30 days was your physical health not good? HSQ480 Now thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good? Continued 134 Figure A.1 (continued) NHANES File Items Item descriptions Code used for item in Names included NHANES HSQ_B * HSQ490 During the past 30 days, for about 0 30 DAYS HSQ_C * how many days did poor physical or 77=refused (continued) mental health keep you from doing 99=don t know your usual activities, such as self- care, work, school or recreation? MCQ_B Ever been told you have: 1=yes MCQ_C MCQ160B Congestive heart failure 2=no MCQ160C Coronary heart disease 7=refused MCQ160D Angina/angina pectoris 9=don t know MCQ160E Heart attack MCQ160F Had a Stroke PAQ_B * PAQ180 Please tell me which of these four 1= you sit during the PAQ_C * sentences best describes your day and do not walk usual daily activities? about very much [Daily activities may include your 2=you stand or walk work, housework if you are a about a lot during the homemaker, going to and attending day, but do not have classes if you are a student, and to carry or lift things what you normally do throughout a very often typical day if you are a retiree or 3=you lift light load or unemployed.] have to climb stairs or hills often. 4=you do heavy work or carry/ heavy loads. 7=refused, 9=do not know PAD200 Over the past 30 days, did you do 1=yes any vigorous activities for at least 2=no 10 minutes that caused heavy 3=unable to do sweating, or large increases in activity breathing or heart rate? Some 7=refused examples are lap swimming, 9=don t know running, aerobics or fast bicycling. PAD320 [Over the past 30 days], did you do 1=yes moderate activities for at least 10 2=no minutes that cause only light 3=unable to do sweating or a slight to moderate activity increase in breathing or heart rate? 7=refused Some examples are brisk walking, 9=don t know bicycling for pleasure, golf, and dancing. Continued 135 Figure A.1 (continued) NHANES File Items Item descriptions Code used for item in Names included NHANES PFQ_B * PFQ048 (B) Does a physical, mental or 1=yes PFQ_C * PFQ049 (C) emotional problem now keep you 2=no from working at a job or business? 7=refused PFQ054 (B) Because of a health problem, do 9=don t know PFQ055 (C) you have difficulty walking without using any special equipment? PFQ090 Do you now have any health problem that requires you to use special equipment, such as a cane, a wheelchair, a special bed, or a special telephone? PFQ_B The next questions ask about difficulties you may have doing certain activities PFQ_C because of a health problem. By health problem we mean any long-term (continued) physical, mental, and/or emotional problem or illness (not including pregnancy). (B) PFQ060 = By yourself and not using any 1 = No difficulty (C) PFQ061 special equipment, how much 2 = Some difficulty difficulty do you have 3 = Much difficulty PFQ061B (b) Walking for a quarter mile 4 = Unable to do [about 2 to 3 blocks]? 5= do not do this PFQ061C (c) Walking up 10 steps without activity resting? 7 = Refused PFQ061H (h) Walking from one room to 9 = Don't know another on the same level? Examination and Laboratory. Examination and venipuncture were scheduled in the mobile examination center following the in home interview( at a later date). BMX_B BMXWT Weight in Kilograms 2.4 - 209.1 Kg BMX_C BMXHT Height in centimeters standing 79 204.4 cm BMXBMI Body Mass Index (Wt in Kg / Ht in 7.9 65.41 Kg/m2 meters squared) BMXWAIST Waist circumference 32 170.7 cm BMIWT Reasons not reported 1=could not obtain 3=clothing 4=medical appliance BMIHT Reasons not reported 1=could not obtain 3=not straight BMIWAIST Reason not reported 1=could not obtain BPX_B BPXSAR Systolic Blood Pressure average 70 266 BPX_C reported to respondent. BPXDAR Diastolic Blood Pressure average 0 - 128 reported to the respondent Continued 136 Figure A.1 (continued) NHANES File Items Item descriptions Code used for item in NHANES Names included LEXAB_B LEDSCCT2 Reasons 0=none LEXAB_C respondent may 1=safety exclusion not have been 2=refusal tested 3=no time 4=physical limitation 5=communication problem 6=equipment failure 7=SP ill/emergency 14=interrupted 22=pain or discomfort 23=poor cuff fit 43=weight limitation on equipment 56=came late/left early 58=unable to obtain all BPs 72=error 84=SP with child 101=bandage, stocking or obstruction 103=doctor s request 104=ankle edema 999=other (lesion, rash, or amputation) NHANES File Items Item descriptions Code used for item in Names included NHANES LEXAB_B LEXLABPI Left ankle-brachial index value 0.29 2.16 LEXAB_C LEXRABPI Right ankle-brachial index value 0.23 2.07 (continued) LEXANK Ankle(s) tested 1=right 2=left 3=both 8=could not obtain LEXANKLC Reasons for not using both ankles 1=edema, lesions. bandages 2=rash 3=amputation 4=other LO6_B LBXCOT Serum Cotinine 0.011 1639 ng/ml L06cot_C L10_B LBXGH Glycosylated Hemoglobin 3.3 18.8 % L10_C L13_B LBXTC Direct Total Cholesterol 68-727 mg/dl L13_C LBDHDL (B) Direct High Density Lipoprotein 16-160 mg/dl or Cholesterol LBXHDD (C) 137