|
|
||||||||
Lipids, Lipoproteins, and Cardiovascular Risk Factors |
1 Hadassah-Hebrew University Braun School of Public Health and Community Medicine, Jerusalem, Israel.
aAddress correspondence to this author at: Hadassah-Hebrew University Braun School of Public Health, PO Box 12272, Jerusalem 91120, Israel. Fax 972-2-6449145; e-mail ronitc{at}md.huji.ac.il.
| Abstract |
|---|
|
|
|---|
Methods: A baseline cross-sectional study was conducted between 1985 and 1987, encompassing residents
50 years of age living in a Jewish neighborhood in western Jerusalem. Interviews were followed by examinations and nonfasting blood sampling (available for 1807 participants). Follow-up data to April 1996 on mortality and causes of death were obtained through record linkage with the Israeli Population Registry.
Results: BuChE activity was inversely related to age and was positively associated with serum concentrations of albumin (r = 0.35; P <0.001), cholesterol (r = 0.31; P <0.001), and triglycerides (r = 0.30; P <0.001). Enzyme activity was associated with measures of overweight, obesity, and body fat distribution (e.g., body mass index, r = 0.20; P <0.001). In multivariate analysis, the associations of enzyme activity with serum cholesterol, triglycerides, and albumin persisted strongly. After adjustment by Cox proportional hazards regression for other predictors of mortality in this population, individuals in the lowest quintile of BuChE activity had significantly higher mortality than those in the highest quintile [hazard ratios (95% confidence intervals): all-cause mortality, 1.62 (1.152.30); cardiovascular deaths, 1.79 (1.053.05)]. The association was attenuated by introduction of serum albumin into the models.
Conclusions: This is the first study to report on the association between BuChE and mortality. The relatively strong association of BuChE with serum lipid and albumin concentrations requires elucidation. Our results suggest that low BuChE activity may be a nonspecific risk factor for mortality in the elderly.
| Introduction |
|---|
|
|
|---|
Many studies on BuChE have been reported, as summarized in 2 reviews (2)(3). Most of the studies were conducted in vitro or in animal models. Of the small proportion undertaken in humans, very few were rigorous epidemiologic studies. Interest in BuChE has been growing because of its possible role in Alzheimer disease and the introduction of anticholinesterase treatments for this disorder (2). Alzheimer disease itself was reported to be associated with cardiovascular risk factors such as hypertension and increased serum cholesterol (4). Previous reports suggested that BuChE enzymatic activity was higher in patients who had hypertension, hyperlipidemia, and high body weight and lower in patients who had suffered acute myocardial infarction or undergone treatment with beta blockers (5)(6)(7).
We investigated the association of BuChE activity with risk factors for coronary heart disease and with all-cause and cardiovascular mortality in the setting of a community-based epidemiologic study.
| Materials and Methods |
|---|
|
|
|---|
data collection
Signed informed consent was provided by all participants. The study was authorized by the institutional Helsinki committee. A structured questionnaire was administered by trained interviewers. Data collected included information on sociodemographic characteristics; lifestyle habits such as diet, smoking, alcohol intake, exercise, and sleep; medical history and current drug treatment; self-appraisal of health; and cognitive and psychologic disturbances. Examinations conducted by a nurse immediately after the interview included standardized measurements of blood pressure, anthropometric indices, a 12-lead electrocardiograph, and a nonfasting blood sample (9)(10).
An
10-year follow-up of mortality in the cohort up to April 1996 and the underlying causes of deaths were obtained by record linkage with the Israeli Population Registry.
biochemical measurements
Blood was drawn into plain Vacutainers (Becton Dickinson) and Vacutainers containing EDTA; 90% of samples were taken between 1300 and 1800 in the afternoon. The EDTA tubes were immediately refrigerated for up to 3 h until centrifugation. On the basis of previously reported data (11)(12), we considered this refrigeration to have no effect on BuChE activity measurements. We measured serum BuChE activity and serum concentrations of glucose, cholesterol, triglycerides, total protein, and albumin with a Technicon SMAC (Technicon Instruments Corp.). Serum BuChE activity was determined with a colorimetric method using S-butyryl-thiocholiniodide (Merck Chemicals) as substrate, as described by Knedel and Bottger (13). HDL-cholesterol was measured enzymatically (Laboratories Biotrol) on a Cobas Bio-analyzer (F Hoffman-La-Roche Ltd.). We measured plasma total homocysteine by HPLC with fluorometric detection as described by Araki and Sako (14).
data analysis
Mean BuChE activity was stratified by sex and categorized according to sociodemographic variables, anthropometric measures, lifestyle habits, clinical biochemistry results, and self-reported health conditions. Daily aerobic activity was defined as at least 20 min of leisure time physical activity that induced shortness of breath and sweating.
Adjustment for age was performed with multiple linear regression. Further adjustment for country of birth, grouped into 4 categories (Israel, Europe, North Africa, and Asia), did not materially change the results. Body mass index (BMI) was calculated as the ratio of weight (kg)/height (m2) and was analyzed as a continuous and a categorical variable. Participants were defined as hypertensive if they reported use of antihypertensive drugs or had blood pressure higher than 140 mmHg systolic or 90 mmHg diastolic, according to the classification of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (15). Information on time of blood sampling was available for 1599 participants. Information on time since last meal was available for 1581 participants; we therefore conducted subgroup analyses including these variables. The unadjusted association of BuChE activity with continuous variables was estimated by the Pearson correlation coefficient. The associations of BuChE activity and serum concentrations of lipids, glucose, and albumin and plasma homocysteine are also presented as means of BuChE activity by quintiles of the clinical chemistry tests. Multiple linear regression was used to study independent associations with BuChE activity. Interactions were assessed as deviations from additivity. We used Cox proportional hazards regression with time-to-event as the dependent variable to estimate associations between BuChE activity and all-cause mortality and cause-specific mortality. To obtain adjusted hazard ratios (HRs) for death rates between quintiles of BuChE activity, we used 3 models. The first was adjusted only for age, and the second and third included additional potential confounders. Cause-specific mortality was classified by cardiovascular mortality and mortality attributable to all causes except cardiovascular diseases, referred to as other causes.
| Results |
|---|
|
|
|---|
BMI < 30 kg/m2), and 25% were classified as obese (BMI
30 kg/m2). Only 8% reported engaging in daily aerobic activity.
BUCHE and sociodemographic characteristics
The mean (SD) BuChE activity was 8.98 (2.02) kU/L (range, 2.7018.70 kU/L). Enzyme activities by age, country of birth, and sex are presented in Table 1
. Mean activities were inversely associated with age (P for trend <0.001). Enzyme activities were higher among women than men [9.2 (2.0) vs 8.7 (2.0) kU/L; P <0.001]. The sex difference persisted across age strata (P for age-adjusted difference <0.001) and after we controlled for country of birth, BMI, serum glucose and triglyceride concentrations, plasma homocysteine concentration, and hypertension (data not shown), but it largely disappeared after we controlled for serum cholesterol [mean (SD) BuChE activities among men and women, 8.9 (1.9) and 9.0 (1.9) kU/L, respectively; P = 0.35].
|
Mean (SD) crude BuChE activity was associated with country of birth and was lowest among Asian-born individuals [8.7 (2.0) kU/L] and highest among European/Americanborn individuals [9.2 (2.1) kU/L]. This association with country of birth persisted after we controlled for age, sex, BMI, and serum cholesterol (P for differences between country of birth categories <0.001). We found no association between education level, self-appraisal of health, and BuChE activity (data not shown).
Time since eating was not associated with BuChE activity (data not shown). Neither bivariate nor multivariate analysis indicated that the time of day of blood sampling was associated with BuChE activity. When the model also included serum triglycerides, however, the association with time of blood withdrawal became significant.
BUCHE and cardiovascular risk factors
Anthropometric measures.
Univariate analysis results indicated that BuChE activity was significantly positively correlated with BMI (r = 0.204; P <0.001), hip circumference (r = 0.213; P <0.001), and waist circumference (r = 0.17; P <0.001) but not with waist-to-hip ratio (WHR). The positive association of BuChE activity with BMI persisted after we controlled for sex, age, and country of birth (r = 0.17; P <0.001), as did that with hip and waist circumference [r = 0.15 (P <0.001) and r = 0.21 (P <0.001), respectively]. On adjustment, a positive association with WHR emerged (r = 0.15; P <0.001). ANOVA results for the above associations are shown in Table 2
.
|
Lifestyle.
We found no association between BuChE activity and smoking or with number of cigarettes smoked (data not shown). We also found no association between self-reported aerobic physical activity and BuChE activity (data not shown). Alcohol consumption was associated with BuChE activity only among men. BuChE activity was slightly higher among men who consumed alcoholic beverages once a week or more compared with men who had lower consumption (age- and country-of-originadjusted means, 8.5 vs 8.3, respectively; P = 0.06; Table 2
). A positive association with wine consumption (Table 2
) was noted in men, but there was no association with the frequency of beer or hard liquor consumption (data not shown).
Serum lipid, glucose, and albumin and plasma homocysteine concentrations.
Mean (SD) concentrations of serum cholesterol in study participants were 5.5 (1.1) mmol/L among men and 6.3 (1.2) mmol/L among women (P <0.001 for sex differences). BuChE activity was significantly positively correlated with serum concentrations of total cholesterol [r = 0.31 (P <0.001) in univariate analysis; r = 0.29 (P <0.001) when controlled for age, sex, and country of birth]. Mean BuChE activity by quintiles of cholesterol concentrations after controlling for age and origin is presented in Table 3
. A similar correlation was found between serum triglyceride concentrations and BuChE activity [r = 0.30 (P <0.001) in both univariate and multivariate analysis]. Mean BuChE activity by quintiles of serum triglycerides, after controlling for age and origin, is presented in Table 3
.
|
We found no association between serum HDL-cholesterol and BuChE activity, and the correlation with the cholesterol/HDL-cholesterol ratio was similar to the correlation with total cholesterol alone (data not shown). We found a weak inverse association between BuChE activity and plasma homocysteine concentrations [r = 0.11 (P <0.001) on univariate analysis]. After we controlled for age and country of origin, the inverse association between mean BuChE activity and quintiles of homocysteine did not reach statistical significance.
In women, but not in men, we found a positive correlation between blood glucose concentrations and BuChE activity. After we excluded data from diabetic patients, this association disappeared (r = 0.03; P = 0.37). Because serum glucose is associated with time since last meal, we analyzed the association of BuChE with serum glucose after restricting the data to samples obtained from participants who were fasting for 2 h or more (n = 1231), but the resulting correlation coefficients between BuChE activity and glucose were largely unchanged (data not shown).
We found a positive association between BuChE activity and serum albumin (r = 0.35; P <0.001) that persisted strongly after we controlled for age and country of birth (Table 3
).
History of hypertension, diabetes, coronary heart disease, and cerebrovascular accidents.
History of hypertension or diabetes was associated with increased BuChE activity. Hypertensive individuals had higher BuChE activities than did normotensive individuals [means adjusted for age, country of birth, and BMI: 8.8 vs 8.5 kU/L, respectively, among men (P = 0.041); and 9.2 vs 8.9, respectively, among women (P = 0.002)]. This association persisted after the introduction of cholesterol into the model (data not shown). We repeated the analysis excluding data from 404 treated hypertensive patients, and the age- and BMI-adjusted mean BuChE activities were unchanged (data not shown).
BuChE activity was significantly higher in female patients with reported diabetes than in women without reported diabetes [9.5 vs 9.0 kU/L, respectively (P = 0.001), after controlling for age, country of birth, and BMI]. This association was also unaltered when serum cholesterol concentrations were introduced into the model (data not shown).
We found no association between BuChE activity and a self-reported history of angina, myocardial infarction, or cerebrovascular accident (data not shown).
Multivariate analysis.
The results of the multivariate regression are presented in Table 4
. The variables included in this model were sex; age; country of birth; BMI; hypertension; serum concentrations of cholesterol, glucose, triglycerides, and albumin; plasma homocysteine; and the hour of blood sampling. Age, country of birth; BMI; serum concentrations of cholesterol, triglycerides, and albumin; and time of blood sampling had highly significant independent associations with BuChE. The explained variance (R2) of this model reached 0.26.
|
BuChE activity and all-cause and cardiovascular mortality.
During the 9 to 11 years of follow-up, 409 (22.6%) participants died, of whom 193 were men and 216 women. The median follow-up was 9.4 years for both men and women. All-cause mortality was significantly associated with BuChE quintiles, with the lowest survival rates among those in the lowest quintile of BuChE activity (P <0.001; data not shown). After adjustment for age; sex; country of birth; serum concentrations of glucose, cholesterol, triglycerides, and creatinine; plasma homocysteine; and hypertension, individuals in the lowest quintile of BuChE activity had significantly higher mortality than those in the highest quintile [HR (95% confidence interval), 1.62 (1.152.30); P = 0.006, Cox proportional hazards models; Table 5
]. This result held true when data from patients with either diabetes or cardiovascular morbidity were excluded from the analysis (data not shown). In assessing the association of BuChE activity with cause-specific mortality (i.e., cardiovascular diseases and other causes), we observed no overall significant effects. However, individuals in the lowest quintile of BuChE activity had a higher risk of cardiovascular death [HR = 1.79 (1.053.05); P = 0.032] and a nonsignificantly increased risk for death from other causes [HR = 1.50 (0.912.47); P = 0.11]. The introduction of serum albumin into the models reduced the significance of the associations between BuChE activity and mortality, but the trends of higher mortality among those in the lowest quintile of BuChE activity persisted (Table 5
).
|
| Discussion |
|---|
|
|
|---|
The results of this study indicate an inverse association between BuChE activity and age among adults older than 50 years, and no independent association with sex.
BuChE activity was positively correlated with measures of obesity, i.e., BMI, waist and hip circumference, and the WHR, which reflects body shape. These associations are consistent with the findings of studies in the United Kingdom, Denmark, and Canada (7)(16)(17).
Our findings indicate a positive association between BuChE activity and serum cholesterol and triglyceride concentrations, and support earlier studies (5)(6)(18)(19). Reports of an inverse association with HDL-cholesterol in bivariate analysis (6)(17) that disappeared on multivariate analysis (17) were not confirmed in our study. Rustemeijer et al. (19) showed that BuChE activity was reduced after treatment with benzafibrate but did not change significantly after treatment with statins. Proposed explanations for the association of BuChE with serum lipids include effects of lipids on BuChE and effects of BuChE on lipids. High serum lipid concentrations may induce stereoscopic alteration in the enzymatic configuration that modifies BuChE activity or altered expression of the enzyme-encoding gene that determines BuChE concentration and activity (18). Recently, it has been suggested that hypercholesterolemia modulates the efficacy of synthetic cholinesterase inhibitors used for treatment of patients with Alzheimer disease (20). Similarly, the activity of natural cholinesterase inhibitors may be modulated by serum lipids. On the other hand, BuChE may play a role in lipid metabolism (21), whether directly(19) or through a synergistic action with cholesterol esterase, as has been demonstrated in saliva (22). Finally, this association might be confounded by unrecognized factors.
In women only, we found a positive association of enzyme activity with diabetes and serum glucose concentrations. Hashim et al. (23) reported that type 2 diabetes was associated with the presence of the K variant of BuChE, a finding contradicted in a large-scale study by Johansen et al. (24), who observed no difference in the prevalence of the K variant between patients with diabetes and persons with normal glucose metabolism. Because the K variant has been associated with an
30% decrease in enzyme activity (3), the lack of an association with diabetes among men and a higher activity in women with diabetes than in glucose-tolerant women in our study do not support the inference of a higher prevalence of the K variant among patients with diabetes. A limitation of our study is the lack of analysis of variant alleles of the enzyme.
In this study, we used nonfasting samples. However, time elapsed since last meal had no meaningful effect on the results. We found only a weak association of BuChE activity with time of blood sampling. When both time of blood sampling and time since last meal were introduced into a multivariate model, the association with time of blood sampling disappeared. Consequently, we are unable to unravel whether the time of blood sampling reflects circadian characteristics or duration of fasting. Furthermore, it may reflect patient characteristics other than the many controlled for in our analysis.
Serum albumin concentrations were strongly correlated with BuChE activity. The suggestion by Masson (25) that the monomeric form of BuChE is conjugated to plasma albumin by the C2 compound might explain an association. Another explanation is that because albumin and BuChE are synthesized in the liver, their correlation is determined by their production site, except for cases of genetic variants in which BuChE synthesis is diminished (26).
To the best of our knowledge, no other studies have investigated the association of BuChE activity with mortality. The prospective association with mortality suggests a higher risk for individuals with very low BuChE activity. The association with both cardiovascular and noncardiovascular mortality suggests that BuChE activity is perhaps a nonspecific risk factor. The attenuation of this association by the introduction into the models of serum albumin, a nonspecific risk marker by itself, supports this assumption. Furthermore, although BuChE activity was positively associated with cardiovascular risk factors such as serum lipids and BMI, it was inversely related to cardiovascular mortality, suggesting that the latter association is not mediated through the BuChE association with cardiovascular risk factors.
The possible role of low BuChE activity as a risk factor for mortality does not necessarily point to its value as a risk marker, except perhaps in conjunction with other risk markers. A value in the lowest quintile had a sensitivity of only 17% as a predictor of death; i.e., 83% of decedents had not had low BuChE activity. This sensitivity was similar to that of the lowest quintile of albumin in this study and was considerably lower than the sensitivities reported for various predictors in this population, e.g., 89% for a doctors appraisal of ill-health and 72% for a battery of nonspecific indicators (27).
In conclusion, the current study confirms and extends the unexplained association of BuChE activity and cardiovascular risk factors, namely serum lipids and BMI. The mechanism underlying these associations needs to be further investigated. Low BuChE activity seems to be a risk marker for death in middle aged and elderly people residing in the community.
| Acknowledgments |
|---|
| Footnotes |
|---|
| References |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |