Chapter 5. Discussion
5.2 Homocysteine and cerebral white matter lesions (WML)
In the present study, we found that Hcy was significantly associated with WML.
After adjustment for known cardiovascular risk factors, including age, sex, hypertension, diabetes mellitus and smoking, Hcy and age remained as independent risk factors for WML. These findings suggest that Hcy had clinical utility in terms of identifying patients at increased risk of WML.
The strong association observed in our study between Hcy levels and risk of WML is concordant with previous studies that Hcy levels were higher in patients with silent brain infarction 67, 126. Silent brain infarcts and WML are thought to have a common vascular origin. Both involve the cerebromicorvascular system. Cerebral small vessel disease is a major cause of vascular cognitive impairment and dementia 127. Baseline WML and lacunes predicted both WML progression and new lacunes 128. Postmortem studies have shown that WML correspond to heterogeneous pathological substrates with varying degrees of demyelination, arteriosclerosis, and gliosis representing
“incomplete” infarction, but also tissue degeneration 129, 130. Silent brain lacunar infarction, although “silent” in clinical manifestation, is a “complete” infarction.
Lacunes are small cavities located in the white matter or subcortical gray matter. They have been regarded as small ischemic infarcts, but several pathogenetic mechanisms may exist 131. The MR images of lacunes are different from that of WML. Infarcts were defined as focal hyperintensities on T2 weighted images and had to have corresponding prominent hypointensities on T1 weighted images in order to distinguish them from cerebral WML. WML were considered present if visible as hyperintense on proton density and T2 weighted images, without prominent hypointensity on T1 weighted scans 132. Previous studies performed in Japan and Korea focused on the association
between Hcy and silent brain infarcts without investigating the association with WML 67,
126. To the best of our knowledge, this is the first study to demonstrate an association between Hcy and risk of WML among Asian population. The present study
demonstrated the effect of Hcy on the incomplete infarction of WML. The findings obtained in this study were consistent with the results from the Rotterdam study 52. In the Rotterdam Scan Study, periventricular WML were rated semiquantitatively (grade range 0-9) and subcortical WML were measured by total volume 52. Elevated serum Hcy was associated with severe grade of WML and this study was performed in elderly (aged 60-90 years). In our study, the severity of WML of case group was varying from mild to severe grade. And our participants were younger than Rotterdam’s study, with mean age 50.9 years. Our study confirmed the association of Hcy with the risk of WML even in young asymptomatic adults.
The mechanisms through which elevated Hcy levels might cause vascular damage to the brain are unclear. Elevated Hcy may promote atherosclerosis by damaging the vascular wall 1 or by its direct toxic effect on nerve cells 133. The hypotheses of atherothrombosis by Hcy could not explain the effect of Hcy on the risk of WML.
Results of present study were consistent with Fassbender’s finding that Hcy was associated with small vessel disease 62. Further histopathological studies are needed to examine the association with microvascular disease.
In conclusion, we found a relation between serum Hcy and cerebral WML.
However, this is a cross-sectional study and results should be confirmed by prospective longitudinal studies.
5.3 Homocysteine and clinical vascular events
Our data suggest serum total Hcy concentration was significantly associated with the risk of CHD and death among ethnic Chinese adults but not significantly related with stroke. Furthermore, we determined the optimal cutpoints of Hcy by selecting the values of highest sensitivity and specificity. The use of cutpoints of Hcy added appreciably to the overall prediction of CHD and death risk. These findings highlight the strength of using new definition of hyperhomocysteinemia for the prediction of future
cardiovascular risk and death in asymptomatic persons.
The association between Hcy and stroke was inconclusive. Two large prospective studies demonstrated significant linear trend for an increased relative risk of incident stroke with each quartile of Hcy levels compared with lowest quartile among elderly people 134, 135. Nonetheless, in the Physicians’ Health Study, baseline serum Hcy for 109 participants who developed ischemic stroke during 5 years of follow-up were compared with those of 427 controls matched for age, smoking status, and length of follow-up 136. There was no significant difference in mean Hcy at baseline. The results of some prospective studies are also conflicting 2. It remains to be determined whether the association is causal or whether the association is confounded due to factors associated with high Hcy, such as vitamin deficiency, diabetes, age, male sex, hypertension, smoking, social class and preexisting vascular disease 137, 138. Besides, the underlying pathogenetic mechanism in stroke is complex. Effects of Hcy on different subtype of stroke such as embolic, large artery, small artery and other etiologies may be different
139, 140.
In contrast with the negative association between Hcy and stroke, our findings demonstrates clearly the association of Hcy with the risk of incident CHD and death,
which was consistent with results of studies performed in healthy women 141,
postmenopausal women 142 and in high risk participants 143-146. Compatible with some previous prospective cohort studies 142-147, top quartile of Hcy independently increases 2-3 times the risk of future CHD as compared with the lowest level of Hcy and our results showed the graded relationship between serum Hcy concentrations and risk of coronary events. Our data also showed positive association of high Hcy with not only cardiovascular death but also non-cardiovascular mortality. Previous study revealed that hyperhomocysteinemia can predict total and cardiovascular mortality in high risk subjects 145, and this study, furthermore, demonstrated the predictive ability of Hcy in community based asymptomatic population without preexisting vascular disease.
We are interested in defining hyperhomocysteinemia for the use of Hcy screening in primary prevention. However, there is no general agreement on the optimal value of Hcy that confers the vascular risk. Clarke and colleagues first establish the diagnostic criterion for hyperhomocysteinemia by comparing post methionine loading Hcy levels between those with cystathionine β-synthase deficiency and 27 age- and sex-matched normal subjects. A level of 24.0 μmol per liter or more of Hcy was 92% sensitive and 100% specificity in distinguishing the two groups 148. The Rotterdam Study results among the elderly showed the risk of CHD increased (OR, 3.0; 95%CI, 1.5-6.1) in particular above 85th percentile (≧18.6 μmol/L). In another prospective study of 21,520 men aged 35 to 64, the risk of ischemic heart disease in highest quartile (15.17 μmol/L) of Hcy level was 3.7 (95% CI, 1.8-4.7) as compared with the lowest quartile 149.
According to a statement by the American Heart Association 150, a serum level of 15 μmol/L, corresponding to the 90th percentile of the distribution, was chosen as the cutoff point for an elevated level. This Hcy value was significantly correlated to CHD among patients with type 2 diabetes mellitus 146. A prospective study among postmenopausal
women found the adjusted RRs of myocardial infarction or stroke in Hcy level ≧ 50th (11.19 μmol/L), 75th (13.26 μmol/L) , 90th (17.29 μmol/L), and 95th (20.70 μmol/L) were 1.9 (95%CI, 1.0-3.4), 2.2 (1.2-4.0), 1.9 (0.9-4.3) and 4.6 (1.7-12.3) respectively 142. Another long-term cohort of women population in Sweden found the fifth quintile of Hcy (≧14.18 μmol/L) had the relative risk as 1.86 ( 95% CI, 1.06-3.26) for myocardial infarction events 141. Our data showed the participants with Hcy level of 9.47 μmol/L or more had highest specificity and sensitivity rate in association with the risk of future CHD. Furthermore, we observed people with high Hcy level had higher total and cardiovascular mortality. We determined that the Hcy value of 11.84 μmol/L may be an optimal cutpoint with better specificity and sensitivity rate for prediction of all-cause of death. As compared with previous studies 141-143, 145, our data showed mild or moderate elevation, not extremely high, of Hcy is significantly related to the risk of CHD and death in the future.
Optimal cutpoints of Hcy were selected across possible values with the range of test results. The value of maximal Youden’s index was determined to be the optimal value. However, this data-driven selection of optimal cutpoints is prone to bias, meaning that it can systematically lead to overestimation of the sensitivity and specificity 151. The amount of bias in sensitivity and specificity predominantly depended on the sample size.
Small sample size (<200) may lead to overly optimistic estimate of sensitivity and specificity in studies. Nonetheless, such bias can be lowered in study with big sample size as in this data 151.
The c statistic is the most commonly used method of determining model
discrimination. In this study, the c statistics were calculated to compare the additional value of Hcy in predicting the risk of vascular events and death in the model fully adjusted for all the confounders. The c statistics increased with Hcy adding in the model
of CHD and death, but not of stroke. Two new measures, net reclassification
improvement (NRI) and integrated discrimination improvement (IDI), were used on the comparison of predicted values from two models. Tests for the IDI were based on t-tests comparing differences in predictions between two models. The majority of individuals in this cohort may be at low risk, and changes in their predicted probabilities are small and lead to a low overall IDI. Nonetheless, the differences in predictive probabilities between the model of conventional risk factors without Hcy and that with Hcy were statistically significant in both the models for CHD and death respectively.
Our study had several potential limitations. First, our study lacked information on some determinants of total Hcy levels such as dietary patterns, food fortification and vitamin supplements. The use of a single Hcy measurement to classify persons may also have underestimated the strength of any association because of regression dilution 152. Second, stroke was classified into ischemic, hemorrhagic and mixed type. Without definite image documentation, misclassification on the stroke type may develop. In addition, the subtypes of ischemic stroke, such as large vessel atherosclerotic infarcts, lacunar infarcts, embolic infarcts, and ischemic stroke of unknown or other origin, were not specified and Hcy may have different impacts on different subtype of stroke 139, 140.
5.4 Homocysteine-lowering vitamin therapy on dementia
In this randomized, double-blind, placebo controlled, 26-week trial of subjects with mild to moderate AD, the daily use of over-the-counter supplements containing folic acid, vitamin B12, and B6 caused plasma homocysteine concentrations to decrease by 3.02 μmole per liter in the vitamin group than in placebo group. However, there were no significant differences in the change of score of cognition and daily living function between these two groups, even after stratification by age and gender. There was no
significant cognitive improvement in each group and no significant association between the change of cognition score and the change of Hcy concentration in either treatment group.
As for the Hcy-lowering therapy, different dosage and combinations have been studied and recommended in several reports. B6, B12, and folic acid supplementation has been successfully used in selected settings to reduce moderate hyperhomocysteinemia in patients with normal or abnormal B12 concentration 153-157. Among the three vitamins, however, only folic acid seems to have a universal potential for reducing fasting plasma levels of Hcy 158. Daily supplementation with both 0.5-5 mg folic acid and about 0.5 mg vitamin B12 should be efficient 159. Additional combination with 5 to 10 mg of vitamin B6 has been reported to be safe and tolerable in reducing Hcy concentration159. In viewing the entire over-the-counter vitamin supplements in Taiwan, the combined use of Methycobal capsule (Eisai Co.), containing methyl based vitamin B12 0.5 mg, and a nutritional regimen registered for pregnant women named Pramet®FA (Abbott
Laboratories Services Corp.), containing folic acid 1 mg and pyridoxine HCl 5mg, may meet the requirement of Hcy-lowering vitamin regimens. These regimens were also accessible and affordable over-the-counter daily supplements.
Although observational studies suggest an inverse association between Hcy concentration and cognitive performance 112-114, a Cochrane review of randomized controlled trials of folic acid with or without vitamin B12160, as well as another review for vitamin B6 supplement 161, concluded that there was no benefit for healthy but cognitively impaired people. The participants in those previous studies had cognitive impairment but were not all diagnosed to have Alzheimer’s dementia. Besides, some studies did not mention the subjects’ serum level of B6, folic acid, or vitamin B12, which could influence the results of vitamin therapy. Moreover, the intervention periods were
no more than 12 weeks. A two-year, controlled trial of homocysteine lowering therapy showed no improvement for healthy older people 162. B12 or folic acid deficiency can be reversible factors in cognition performance, but most of the aforementioned studies did not clarify the vitamin status in the included subjects 153-155. A randomized controlled trial with homocysteine-lowering therapy in patients with mild to moderate AD and with normal B12 and folic acid status has never been reported before in reviewed
literatures. This study was specifically designed to address the diagnosis of Alzheimer’s dementia as an inclusion criterion. We not only assessed the cognition performance but also the daily living function, as well as the homocysteine concentration change after 6 months of intervention.
Our study showed significant effects on lowering of the homocysteine
concentration by these over-the-counter multi-vitamin supplements as compared to the placebo group. Though there were evidences that plasma Hcy is associated with cognition function and Alzheimer’s disease, the Hcy-lowering therapy had no
significant clinical beneficial effects on cognition and daily living function in these AD patients. These results were similar to other clinical trials on healthy elderly or patients with vascular disease 163, 164. In terms of homocysteine association with atherosclerotic vascular disease 32, 165, 166, several randomized controlled trials also showed that the Hcy-lowering therapy had no consistent effects on the incidence of vascular events
167-172. The possible explanations for the lack of statistic significance of the
Hcy-lowering therapy for cognition function in this study are as follows: 1. although there was no significant effect in our study, there still showed trend of benefit favoring vitamin treatment (Fig IV-2). Our sample size may be too small for a significant statistic power. 2. The treatment duration may be too short to have significant effects. It might take a longer duration to improve the cognition function by Hcy-lowering therapy. 3.
The pathogeneses of Alzheimer’s dementia are complicated and are related to amyloid deposition in addition to vascular factor 173. Controlling one risk factor such as Hcy may not be sufficient to obtain the clinical benefit. 4. Our study participants were in the mild or moderate stage of dementia. The clinical effects of vitamin therapy may be little or not obvious on such irreversible and significant damage of brain. Further study with selection of target participants with mild cognition impairment and very early stage of dementia may be indicated for further exploring the efficacy of Hcy-lowering therapy on Alzheimer’s dementia.
Our study had several limitations. The duration was relatively short and thus, longer cognitive benefits from Hcy-lowering in AD patients cannot be excluded. Some of the participants also suffered from symptoms of gastric upset, nausea, and vomiting, which were often related to AChEIs 174. Some subjects who dropped out were possibly due to these adverse effects, causing a smaller-than-desired final sample size. Further large randomized controlled trials with longer periods are needed.
Chapter 6. Summary, Conclusions and Future direction
6.1 Summary of the findings in these serial studies Study 1:
In the first cross-sectional, observational study, total of 542 asymptomatic adults (mean age 55 years old ) were enrolled. Carotid duplex studies were performed in all the participants to measure the flow velocities, resistance or volume in the common, internal and external carotid arteries (CCA, ECA and ICA) and the vertebral arteries (VA). Subjects with higher serum Hcy had slower velocities on the CCA, ICA and VA.
After adjustment of vascular risk factors of age, sex, smoking habit, body mass index and serum lipid profile, the effect of Hcy on the hemodynamic status of the extracranial cerebral arteries.
Study 2:
We conducted a case control study in a total of 258 asymptomatic adults (mean age 51 years old) which were classified into case group with microangiopathy related cerebral white matter change noted on the MRI and control group with normal brain MRI study.
The mean serum Hcy levels was 9.1 µmol/L in the 65 participants of case group and 8.9 µmol/L in control group (p=0.83). After fully adjustment of the confounding variables such as age, sex, blood pressure and other vascular factors, Hcy was an independent risk factor for cerebral white matter change. Every 1 µmol/L increase of Hcy, the relative risk for having cerebral white matter lesions was 1.15 (95% CI 1.01-1.31, P=0.03).
Study 3:
We conducted a community-based prospective cohort study of 2009 participants (56%
women; age range, 38-91 years), who were free from stroke, coronary heart disease (CHD) and cancer at baseline in 1994, and were followed up to 2007 (median 11.95
years). We documented 114 cases of stroke, 95 cases of CHD and 380 deaths. Cox proportional hazard model was used to examine the association between Hcy and the incidence of stroke, CHD, and all-cause death. The receiver-operating characteristic curve was preformed to identify the best Youden’s index for determining the cutpoint of Hcy in risk prediction. Homocysteine levels remained significantly associated with the cardiovascular events and death in fully adjusted models. Participants with Hcy more than 9.47 µmol/L (sensitivity 81.1%, specificity 54.3%) had a 2.3-fold risk for cardiovascular events (95% CI, 1.24-4.18, P=0.008), and participants with Hcy more than 11.84 µmol/L (sensitivity 49.7%, specificity 84.0%) had a 2.4 fold risk for death (95% CI, 1.76-3.32, P<0.0001).
Study 4:
A total of 89 patients (44 women and 45 men; mean age: 75 ± 7.3 years) were
randomized to receive multi-vitamin or placebo for 26 weeks. Homocysteine-lowering therapy was a daily supplement containing folic acid 1 mg, pyridoxine 5 mg and
mecobalamin (B12) 500mcg. Plasma levels of homocysteine and tests of cognition were conducted at baseline and after 26 weeks of treatment. In vitamin group, the
concentration of vitamin B12 and folic acid were significantly elevated and the mean plasma homocysteine concentration was 3.02 μmole per liter lower than in the placebo group (p=0.008) after 26 weeks’ treatment. Overall, there were no significant
differences between the vitamin and placebo groups in the scores of cognition and activities of daily living. The proportion of adverse events was not significant different in both groups (p=0.1).
6.2 Strength and limitations Study 1.
Strength:
Previous reports showed high serum Hcy levels were associated with the slow flow phenomenon in non-stenotic coronary arteries. But there was no study exploring the relationship between Hcy and hemodynamics status of the brain. To the best of our knowledge, our study was the first to investigate the effect of Hcy on the flow velocity of brain vessels. Our results demonstrated that the Hcy induced slow flow phenomenon in coronary artery does not exist for the cerebral arteries after fully adjustment for known cardiovascular risk factors. Besides, our sample size were much bigger than previous studies on the coronary flow and this made for greater statistical power.
Limitations:
The study limitations are as following. First, the arteries we sampled were extracranial carotid and vertebral arteries which may not completely indicate the hemodynamic status of the intracranial cerebral blood flow. The results of our study might not be extrapolated to the whole brain circulation. Second, as compared to the
The study limitations are as following. First, the arteries we sampled were extracranial carotid and vertebral arteries which may not completely indicate the hemodynamic status of the intracranial cerebral blood flow. The results of our study might not be extrapolated to the whole brain circulation. Second, as compared to the