Increases in Serum Non-High-Density Lipoprotein Cholesterol May Be Beneficial in Some High-Functioning Older Adults: MacArthur Studies of Successful Aging

Arun S. Karlamangla, PhD, MD; Burton H. Singer, PhD; David B. Reuben, MD; Teresa E. Seeman, PhD

J Am Geriatr Soc. 2004;52(4)

Abstract and Introduction

Objectives: To examine the association between changes in serum non-high-density lipoprotein cholesterol (non-HDL-C) over a 2.5-year period and risk of adverse health outcomes in the following 4.5 years in high-functioning older adults.
Design: Prospective cohort, established in 1988, with a follow-up in 1991 and 1995.
Setting: Population-based, community-dwelling men and women.
Participants: A random sample (n = 267) from the MacArthur cohort (N = 1,189). The cohort represented the highest-functioning tertile of 4,030 screened candidates aged 70 to 79.
Measurements: Change in non-HDL-C between 1988 and 1991 was measured as a predictor of health outcomes between 1991 and 1995, including all-cause mortality, and among survivors, incident heart attack or stroke, development of new disability in basic activities of daily living, and decline in performance on the Short Portable Mental Status Questionnaire.
Results: More-positive change in non-HDL-C between 1988 and 1991 was associated with fewer adverse outcomes between 1991 and 1995. In individuals whose total cholesterol at baseline was in the middle two quartiles (195-244 mg/dL), each 10-mg/dL increase in the 1988-to-1991 change in non-HDL-C was associated with an adjusted mortality odds ratio (OR) of 0.67 (95% confidence interval (CI) = 0.51-0.88). In individuals without cardiovascular disease at baseline, the adjusted OR for new physical disability was 0.79 (95% CI = 0.65-0.95) and for cognitive decline was 0.81 (95% CI = 0.67-0.98).
Conclusion: Increases in cholesterol over time have beneficial associations in some older adults. The role of cholesterol changes in the health of older individuals needs further exploration.

Randomized clinical trials in elderly individuals, subgroup analysis of data from broader-based clinical trials, and cohort studies in older adults have demonstrated that treatment of elderly adults with 3-hydroxy 3-methylglutaryl coenzyme A reductase inhibitors (statins) reduces the risk of major vascular events^[1,2,3] and all-cause mortality.^[4,5,6] This benefit of statin therapy is presumed to be the result of the lipid-lowering effect of statins, but many large prospective studies of older adults have found no statistically significant association between elevated levels of total serum cholesterol and all-cause mortality risk,^[7,8,9] and several have found hypercholesterolemia to be protective against mortality in older individuals,^[10] even after adjusting for serum albumin,^[11] excluding individuals with low serum total cholesterol,^[12] and excluding individuals who died within the first few years of cholesterol measurement.^[11,12,13] With respect to the low-density lipoprotein (LDL) fraction of cholesterol, perhaps the main culprit in cardiovascular disease (CVD), Framingham analyses found an inverse association between serum LDL-C and all-cause mortality risk in those who were aged 66 and older at the time of cholesterol determination,^[12] but other large cohort studies of older adults have found no statistically significant association between serum LDL-C and mortality risk, after adjusting for other risk factors.^[13,14]

This contrast between the benefits of statin therapy and the questionable role of hypercholesterolemia as a marker of increased mortality risk in older individuals raises the possibility that the benefits of statin therapy result less from the effects of statins on serum cholesterol levels and more from their antiinflammatory, plaque-stabilizing, and other beneficial effects.^[15,16] In fact, the benefits of statin therapy in clinical trials were independent of the LDL response to therapy,^[1] were greater than that expected from the magnitude of LDL-C reduction,^[17] and occurred faster than predicted from lipid-lowering alone.^[18] Thus, the effects of changes in cholesterol levels in older adults are unclear.

Examining the health effects of recent change in serum cholesterol may resolve this lingering question about the role of lipid lowering in the care of older adults to some extent. The work reported here examines the association between changes in serum cholesterol over an initial 2.5-year period and health outcomes over the subsequent 4.5 years in a cohort of high-functioning elderly men and women.

1 2 3 4

Next Section

Cite this: Increases in Serum Non-High-Density Lipoprotein Cholesterol May Be Beneficial in Some High-Functioning Older Adults: MacArthur Studies of Successful Aging - Medscape - Apr 01, 2004.

Abstract and Introduction
Methods
Results
Discussion
References

Table 1. Descriptive Statistics for Baseline Variables in the MacArthur Cohort and the Study Sample
Variable	MacArthur Cohort^* n = 1,189	Study Sample n = 267
Variable	Median (Q1, Q3)
Age	74 (72, 76)	74 (72, 76)
Smoking, pack-years	0 (0, 37)	0 (0, 28)
Body weight, kg	72 (63, 79)	73 (63, 82)
Systolic blood pressure, mmHg	136 (125, 148)	136 (125, 148)
Diastolic blood pressure, mmHg	76 (70, 83)	76 (70, 81)
Serum albumin, g/L	41 (39, 43)	41 (39, 43)
Glycosylated hemoglobin, %	6.4 (5.9, 7.1)	6.2 (5.7, 6.9)
Serum HDL cholesterol, mg/dL	47 (37, 54)	47 (37, 58)
Serum total cholesterol, mg/dL	218 (191, 245)	218 (191, 247)

*Participants who had measurements at baseline (1988).
Participants who had measurements at baseline and measurements of total and high-density lipoprotein (HDL) cholesterol in 1991.
Two-sided P <.05 for t test of difference between means of (transformed) variables in the study sample and the rest of the MacArthur cohort.
Q1 = first quartile cutoff point (i.e., the 25th percentile); Q3 = third quartile cutoff point (i.e., the 75th percentile).

Table 2. Demographics of the MacArthur Cohort and the Study Sample
Variable	MacArthur Cohort^* n = 1,189	Study Sample n = 267
Variable	%
Female	55.4	58.4
White	81.1	85.7
Prevalent cardiovascular disease^§	23.9	25.5
Use of statins for lipid lowering	1.0	0.4
Use of other lipid lowering agents	1.3	1.5

*Participants who had measurements at baseline (1988).
Participants who had measurements at baseline and measurements of total and high-density lipoprotein (HDL) cholesterol in 1991.
Two-sided P <.05 for chi-square test of difference between study sample and rest of MacArthur cohort.
^§Previous diagnosis of diabetes mellitus, heart attack, or stroke.

Table 3. Descriptive Statistics for Change Variables in the MacArthur Cohort Versus the Study Sample
Variable	MacArthur Cohort^*	Study Sample
	n = 1,189	n = 267
	Median (Q1, Q3)
Change, 1988-1991
Body weight, kg	0.0 (2.7, 1.4)	0.0 (2.7, 1.8)
Systolic blood pressure, mmHg	1.0 (13,9)	1.0 (14, 9)
Diastolic blood pressure, mmHg	4.0 (10,3)	4.0 (10, 1)
Serum albumin, g/L	0.0 (2, 2)	0.0 (2, 2)
Glycosylated hemoglobin, %		0.4 (0.2, 0.8)
Serum HDL cholesterol, mg/dL		2.0 (6, 2)
Serum total cholesterol, mg/dL		9.0 (22, 5)

^*Everyone with measurements at baseline (1988).
People with measurements of total and high-density lipoprotein (HDL) cholesterol in 1988 and 1991.
Change refers to the 1991 value minus the 1988 value.
Q1=first quartile cutoff point (i.e., the 25th percentile); Q3=third quartile cutoff point (i.e., the 75th percentile).

Table 4. Outcomes in the MacArthur Cohort Versus the Study Sample
Event	MacArthur Cohort^*	Study Sample
	n = 1,189	n = 267
	%
1988 to 1991
Myocardial infarction or stroke	6.2	4.5
Increase in ADL disability	4.6	2.6
Decrease in SPMSQ score to ≤6	3.8	3.4
1991 to 1995
Mortality	18.1	14.2
Myocardial infarction or stroke	11.2	12.8
Increase in ADL disability	16.7	17.0
Decrease in SPMSQ score to ≤6	12.6	13.9

*Everyone with measurements at baseline (1988).
People with measurements of total and high-density lipoprotein cholesterol in 1988 and 1991.
ADL = activity of daily living; SPMSQ = Short Portable Mental Status Questionnaire.

Table 5. Unadjusted and Adjusted Odds Ratios for Each 10 mg/dL Increase in the 1988-to-1991 Change in Non-High-Density Lipoprotein Cholesterol for 1991 to 1995 Events
Model	Mortality	Heart Attack or Stroke	New Activity of Daily Living Disability	Cognitive Decline
Model	Odds Ratio (95% Confidence Interval)
Unadjusted	0.92 (0.80-1.06)	1.07 (0.91-1.27)	0.85 (0.74-0.97)	0.83 (0.72-0.97)
Adjusted^*	0.84 (0.62-1.00)	1.16 (0.95-1.40)	0.84 (0.71-0.99)	0.86 (0.73-1.02)

^*Adjusted for sex, ethnicity, baseline level of total cholesterol, prevalent cardiovascular disease status at baseline, baseline weight (continuous), and 1988-to-1991 weight loss (continuous and squared term).
Confidence intervals do not cross 1.

Table 6. Adjusted Odds Ratios Within Strata for Each 10 mg/dL Increase in the 1988-to-1991 Change in Non-High-Density Lipoprotein Cholesterol from Models Including Effect Modification
Stratum	Mortality	Heart Attack or Stroke	New ADL Disability	Cognitive Decline
Stratum	Point Estimate of Odds Ratio (95% Confidence Interval)
Baseline total cholesterol stratum, mg/dL	P = .027	P = .40	P = .62	P = .41
<191	0.90 (0.61-1.32)
191-244	0.67 (0.51-0.88)^*
≥245	1.42 (0.89-2.27)
Prevalent CVD status	P = .65	P = .38	P = .147	P = .150
No CVD at baseline			0.79 (0.65-0.95)^*	0.81 (0.67-0.98)^*
With CVD at baseline			1.05 (0.75-1.48)	1.13 (0.76-1.71)

Note: Adjusted for sex, ethnicity, baseline level of total cholesterol, prevalent cardiovascular disease status at baseline, baseline weight (continuous), 1988-to-1991 weight loss (continuous and squared term), and an interaction (product) between the stratification variable (categories of baseline total cholesterol or prevalent cardiovascular disease (CVD) status) and the 1988-to-1991 change in non-high-density lipoprotein cholesterol.
Odds ratios within strata are shown only when interaction P-values are <.2.
*Confidence intervals do not cross 1.

Table 6. Adjusted Odds Ratios Within Strata for Each 10 mg/dL Increase in the 1988-to-1991 Change in Non-High-Density Lipoprotein Cholesterol from Models Including Effect Modification
Stratum	Mortality	Heart Attack or Stroke	New ADL Disability	Cognitive Decline
Stratum	Point Estimate of Odds Ratio (95% Confidence Interval)
Baseline total cholesterol stratum, mg/dL	P = .027	P = .40	P = .62	P = .41
<191	0.90 (0.61-1.32)
191-244	0.67 (0.51-0.88)^*
≥245	1.42 (0.89-2.27)
Prevalent CVD status	P = .65	P = .38	P = .147	P = .150
No CVD at baseline			0.79 (0.65-0.95)^*	0.81 (0.67-0.98)^*
With CVD at baseline			1.05 (0.75-1.48)	1.13 (0.76-1.71)

Table 5. Unadjusted and Adjusted Odds Ratios for Each 10 mg/dL Increase in the 1988-to-1991 Change in Non-High-Density Lipoprotein Cholesterol for 1991 to 1995 Events
Model	Mortality	Heart Attack or Stroke	New Activity of Daily Living Disability	Cognitive Decline
Model	Odds Ratio (95% Confidence Interval)
Unadjusted	0.92 (0.80-1.06)	1.07 (0.91-1.27)	0.85 (0.74-0.97)	0.83 (0.72-0.97)
Adjusted^*	0.84 (0.62-1.00)	1.16 (0.95-1.40)	0.84 (0.71-0.99)	0.86 (0.73-1.02)

Table 6. Adjusted Odds Ratios Within Strata for Each 10 mg/dL Increase in the 1988-to-1991 Change in Non-High-Density Lipoprotein Cholesterol from Models Including Effect Modification
Stratum	Mortality	Heart Attack or Stroke	New ADL Disability	Cognitive Decline
Stratum	Point Estimate of Odds Ratio (95% Confidence Interval)
Baseline total cholesterol stratum, mg/dL	P = .027	P = .40	P = .62	P = .41
<191	0.90 (0.61-1.32)
191-244	0.67 (0.51-0.88)^*
≥245	1.42 (0.89-2.27)
Prevalent CVD status	P = .65	P = .38	P = .147	P = .150
No CVD at baseline			0.79 (0.65-0.95)^*	0.81 (0.67-0.98)^*
With CVD at baseline			1.05 (0.75-1.48)	1.13 (0.76-1.71)

processing....

Increases in Serum Non-High-Density Lipoprotein Cholesterol May Be Beneficial in Some High-Functioning Older Adults: MacArthur Studies of Successful Aging

Abstract and Introduction

Tables

Tables

Tables

Tables

Tables

Tables

Tables

Tables

Tables

References

Authors and Disclosures

Authors and Disclosures

Email This

Feedback