Associations between serum homocysteine, holotranscobalamin, folate and cognition in the elderly: a longitudinal study
Babak Hooshmand or Miia Kivipelto, Aging Research Center, Department of Neurobiology, Care Sciences and Society (NVS), Karolinska Institutet, Gävlegatan 16, S 113 30, Stockholm, Sweden. (fax: +46 8 690 59 54; e-mail: firstname.lastname@example.org or email@example.com).
Abstract. Hooshmand B, Solomon A, Kåreholt I, Rusanen M, Hänninen T, Leiviskä J, Winblad B, Laatikainen T, Soininen H & Kivipelto M (Aging Research Center, Karolinska Institutet, Stockholm, Sweden; KI Alzheimer’s Disease Research Center (KI-ADRC), Karolinska Institutet, Stockholm, Sweden; National Institute for Health and Welfare (THL), Helsinki, Finland; University of Eastern Finland, Institute of Clinical Medicine, and University Hospital, Kuopio, Finland). Associations between serum homocysteine, holotranscobalamin, folate and cognition in the elderly: a longitudinal study. J Intern Med 2012; 271: 204–212.
Objectives. To examine the associations between serum homocysteine (tHcy), holotranscobalamin (holoTC, the biologically active fraction of vitamin B12) and folate and cognitive functioning in a longitudinal population-based study of Finnish elderly subjects.
Subjects and design. tHcy, holoTC and folate were measured at baseline in 274 dementia-free subjects aged 65–79 years from the Cardiovascular Risk Factors, Aging and Dementia study. Subjects were re-examined 7 years later, and global cognition, episodic memory, executive functioning, verbal expression and psychomotor speed were assessed.
Results. Higher baseline tHcy levels were associated with poorer performance in global cognition, relative difference: 0.90 [95% confidence interval (CI) 0.81–0.99]; episodic memory: 0.87 (95% CI 0.77–0.99); executive functions: 0.86 (95% CI 0.75–0.98); and verbal expression: 0.89 (95% CI 0.81–0.97) at follow-up. Increased holoTC levels were related to better performance on global cognition: 1.09 (95% CI 1.00–1.19); executive functions: 1.11 (95% CI 1.01–1.21); and psychomotor speed: 1.13 (95% CI 1.01–1.26). After excluding 20 cases of incident dementia, increased tHcy remained associated with poorer performance in episodic memory, execution functions and verbal expression. Higher holoTC levels tended to be related to better performance in executive functions and psychomotor speed, while elevated serum folate concentrations were significantly related to higher scores in global cognition and verbal expression tests.
Conclusions. tHcy, holoTC and folate levels are related to cognitive performance 7 years later even in nondemented elderly subjects. Randomized trials are needed to determine the impact of vitamin B12 and folate supplementation on preventing cognitive decline in the elderly.
Vitamin B12 and folate are essential vitamins that are part of the homocysteine metabolic cycle . Vitamin B12 and folate deficiencies, which are common in the elderly [1–3], can result in increased total homocysteine (tHcy) levels, and this may lead to a variety of disorders including cardiovascular, cerebrovascular and peripheral vascular conditions [1, 2]. In addition, low levels of vitamin B12 or folate or high levels of tHcy have been linked to dementia/Alzheimer’s disease (AD), worse cognitive functioning and structural brain changes in some [4–20] but not all [4–6, 8–10, 12, 14–16, 19–24] longitudinal studies.
In view of the poor predictive values of the standard vitamin B12 assays, it has recently been suggested that holotranscobalamin (holoTC), the biologically active fraction of vitamin B12, may be a more sensitive marker of vitamin B12 status . However, few prospective studies have investigated the association between holoTC and dementia or cognitive performance, and results have been inconsistent [4, 5, 14, 21].
The potential impact of high tHcy or low folate or vitamin B12 levels on cognitive function is of importance because they are modifiable risk factors, making them possible candidates for preventative intervention. We have recently shown that tHcy and holoTC levels, measured up to 7 years before diagnosis, were associated with the risk of AD (increased and decreased risk, respectively) in the population-based Cardiovascular Risk Factors, Aging and Dementia (CAIDE) study . The aim of the present study was to investigate the associations between serum levels of tHcy, holoTC and folate and several domains of cognitive functioning 7 years later in this elderly population without folic acid fortification.
Subjects and methods
The design of the CAIDE study has been described previously [4, 26]. Briefly, CAIDE participants were examined in middle age within the framework of the North Karelia project and the FINMONICA (Finnish Multinational Monitoring of Trends and Determinants in Cardiovascular Disease) study in 1972, 1977, 1982 or 1987. Individuals who were still alive, aged 65–79 years and living in the areas of Kuopio and Joensuu in Finland at the end of 1997 were invited to the first re-examination in 1998. A second re-examination of the same cohort was conducted in 2005–2006. Both re-examinations comprised a self-administered questionnaire on sociodemographic characteristics, health-related behaviours and medical history, including cerebrovascular, cardiovascular and renal conditions (such as urinary tract infection, pyelonephritis and kidney stones). Specially trained nurses checked the questionnaires to ensure that they were fully completed. Height, weight and blood pressure were measured. Body mass index (BMI) was calculated as weight (in kilograms) divided by height squared (in metres). Blood pressure was measured from the right arm twice after subjects had been seated for 5 min, and the mean of the two measurements was calculated.
This study included a subsample of 274 dementia-free subjects from the cohort participating in the first re-examination of the CAIDE study in 1998. Subjects were selected based on the availability of serum samples from 1998 for tHcy, holoTC and folate measurements. The mean (standard deviation) duration of follow-up of the CAIDE subsample of participants from the 1998 re-examination (baseline for this study) was 7.4 (0.3) years. There were no significant clinical differences between the CAIDE subsample and the entire dementia-free CAIDE cohort. None of the participants reported using B vitamin or other vitamin supplements. There is no mandatory folic acid fortification in Finland.
The CAIDE study was approved by the local ethics committee (University of Kuopio and Kuopio University Hospital, Kuopio, Finland), and written informed consent was obtained from all participants.
Measurement of cognitive functions
A comprehensive battery of neuropsychological tests to assess several cognitive domains was administered to CAIDE participants. Only identical tests used at both re-examinations were considered for the present study: (i) the Mini-Mental State Examination (MMSE) , regarded as a measure of global cognition; (ii) immediate word recall test, a measure of episodic memory; (iii) the Stroop test, used as a measure of executive functioning; (iv) category fluency test as a measure of verbal expression; and (v) the bimanual Purdue Pegboard test and the letter digit substitution test, with the mean of their normalized scores used as a measure of psychomotor speed .
Venous blood samples were taken at the 1998 re-examination, and the serum samples were stored at or below −20 °C until analysis at the National Institute for Health and Welfare, Helsinki. Serum tHcy was determined by chemiluminescent microparticle immunoassay, and serum folate was determined by chemiluminescent microparticle folate binding protein assay using the Architect i system (Abbott Laboratories, Abbott Park, IL, USA). The interassay coefficients of variation (CVs) for homocysteine were 5.9% and 5.4% at the levels of 6.6 and 11.0 μmol L−1, and the CVs for folate were 13.0% and 11.0% at the levels of 7.5 and 31.0 nmol L−1, respectively. Holotranscobalamin was measured by microparticle enzyme immunoassay using the AxSym System [Active-B12 (Holotranscobalamin); Axis-Schield, Dundee, UK, Abbott Laboratories]. At the levels of 48 and 97 pmol L−1, the interassay CVs were 7.1% and 8.0%, respectively.
Blood leucocyte were analysed to determine the APOE genotype in 1998. A standard phenol–chloroform technique was used to extract DNA; APOE genotypes were analysed by PCR and HhaI digestion . Participants were classified as positive for the APOEε4 allele genotype if they had one or two ε4 alleles.
Cognitive test scores and all continuous variables were log-transformed because of skewed results. Continuous variables are presented as geometric means with 95% confidence intervals (CIs) and their range (minimum–maximum), whereas categorical variables are shown as numbers (percentages). Multiple linear regression analyses were performed to investigate the associations between the levels of tHcy, holoTC and folate at the first re-examination (baseline for this study) and cognitive test scores at the second re-examination 7 years later. We analysed each of the primary predictors as continuous variables and within-quartile categories (with the lowest quartile as the reference category: ≤10.4 μmol L−1 for tHcy, ≤57.8 pmol L−1 for holoTC and ≤4.9 nmol L−1 for folate). Models were adjusted for potential confounding or mediating factors known to influence the cognition, including age, sex, education level, duration of follow-up, mean baseline systolic blood pressure (SBP), BMI, history of smoking, stroke, APOEε4 allele and presence of renal conditions (as creatinine values were not available) (model 1) and also cognitive measures in 1998 (model 2). Because folate and holoTC are intrinsically related to tHcy and are often correlated with each other, we examined their associations with the cognitive outcomes when adjusted for each other in a final set of models (model 3). All variables were entered into the models as continuous except for sex, history of stroke, smoking, APOEε4 allele status and presence of renal conditions which were dichotomized. Additional analyses were carried out to investigate the associations between both holoTC and folate and cognitive test scores after excluding subjects with low levels of holoTC (<35 pmol L−1, n = 21) and folate (<5.3 nmol L−1, n = 97) according to the laboratory cut-off values. All analyses were also repeated after excluding 20 subjects who developed dementia at follow-up [diagnosed according to Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) criteria ].
The results are presented as relative difference (RD) with 95% CI, where values >1.0 indicate better performance and those <1.0 indicate worse performance. We analysed the data using stata software version 9 (StataCorp, College Station, TX, USA), and the level of significance was <0.05 in all analyses.
Baseline characteristics of the study population are presented in Table 1. As expected, tHcy was inversely correlated with holoTC (Pearson’s correlation coefficient, r = −0.46; P < 0.001) and folate (r = −0.42; P < 0.001), indicating that elevated tHcy was a marker of low status of both vitamin B12 and folate. In addition, holoTC was positively correlated with folate (r = 0.14; P = 0.02).
Table 1. Characteristics of the study populationa
|Age at baseline, years||70.1 (69.7–70.5)||65.2 to 79.9|
|Duration of follow-up, years||7.4 (7.4–7.5)||6.8 to 8.1|
|Sex, no. of women||169 (61.7)|| |
|Education, years||8.4 (8.1–8.8)||4 to 22|
|BMI, kg m−2||27.4 (26.9–27.9)||19.3 to 47.2|
|Systolic blood pressure at baseline, mmHg||151.0 (148.5–153.6)||106 to 216|
|Diastolic blood pressure at baseline, mmHg||82.1 (80.9–83.3)||55 to 108|
|APOEε4 allele||92 (33.6)|| |
|MMSE score at baseline||26.7 (26.4–26.9)||18 to 30|
|MMSE score at follow-up||26.5 (26.2–26.8)||10 to 30|
|Immediate word recall at baseline||5.6 (5.4–5.7)||0 to 10|
|Immediate word recall at follow-up||4.8 (4.5–5.0)||0 to 9|
|Executive functioning at baseline||36.8 (34.5–39.2)||−36 to 106|
|Executive functioning at follow-up||46.6 (43.6–49.7)||−22 to 177|
|Verbal expression at baseline||20.5 (19.9–21.2)||7 to 41|
|Verbal expression at follow-up||19.6 (18.8–20.3)||4 to 40|
|Bimanual Purdue Pegboard test at baseline||9.9 (9.7–10.1)||5 to 18|
|Bimanual Purdue Pegboard test at follow-up||7.6 (7.3–7.8)||0 to 12|
|Letter digit substitution test at baseline||20.1 (19.3–20.9)||0 to 47|
|Letter digit substitution test at follow-up||16.6 (15.7–17.4)||0 to 36|
|History of stroke at baseline||19 (6.9)|| |
|Ever smoked at baseline||96 (35.0)|| |
|tHcy, μmol L−1||12.2 (11.9–12.6)||6.4 to 30.8|
|holoTC, pmol L−1||81.9 (76.8–87.3)||6.5 to 365.0|
|Folate, nmol L−1||6.3 (6.0–6.7)||1.8 to 29.2|
tHcy and cognitive functioning
The relationships between tHcy at baseline as a continuous or categorical variable and cognitive performance 7 years later are presented in Table 2. After adjusting for age, sex, education level, follow-up time, APOEε4 status, BMI, SBP, history of stroke, smoking status and presence of renal conditions (model 1), higher concentrations of tHcy were significantly associated with lower performance in global cognition, episodic memory, executive functioning, verbal expression and psychomotor speed. After additional adjustment for baseline cognitive measures (model 2), the association between higher baseline tHcy and psychomotor speed was not significant. After controlling for holoTC and folate (model 3), the relationship between tHcy and MMSE was also no longer significant.
Table 2. Seven-year relative differences (RDs) and 95% confidence intervals (CIs) for the associations between tHcy, holoTC, folate and cognitive functions (n = 274)a
|Global cognition (n = 274)|
| Model 1||(Ref)||1.00 (0.91–1.09)||1.01 (0.91–1.11)||0.93 (0.84–1.02)||0.89 (0.80–1.00)*|
| Model 2||(Ref)||1.00 (0.92–1.09)||1.00 (0.91–1.09)||0.92 (0.84–1.01)†||0.90 (0.81–0.99)*|
| Model 3b||(Ref)||1.01 (0.93–1.10)||1.01 (0.92–1.11)||0.94 (0.85–1.05)||0.92 (0.81–1.03)|
| Model 1||(Ref)||1.03 (0.94–1.12)||1.06 (0.97–1.16)||1.10 (1.00–1.20)†||1.08 (1.00–1.17)†|
| Model 2||(Ref)||1.03 (0.94–1.12)||1.06 (0.98–1.15)||1.09 (1.00–1.19)*||1.07 (0.99–1.15)†|
| Model 3c||(Ref)||1.01 (0.93–1.10)||1.04 (0.96–1.14)||1.07 (0.97–1.17)||1.04 (0.96–1.13)|
| Model 1||(Ref)||1.09 (0.99–1.19)†||1.05 (0.96–1.15)||1.05 (0.95–1.15)||1.02 (0.96–1.08)|
| Model 2||(Ref)||1.08 (0.99–1.17)†||1.08 (0.99–1.17)†||1.04 (0.96–1.14)||1.02 (0.97–1.09)|
| Model 3d||(Ref)||1.07 (0.98–1.16)||1.07 (0.98–1.16)||1.00 (0.92–1.10)||1.00 (0.94–1.07)|
|Episodic memory (n = 274)|
| Model 1||(Ref)||0.94 (0.84–1.05)||0.94 (0.84–1.06)||0.85 (0.75–0.95)*||0.82 (0.72–0.93)*|
| Model 2||(Ref)||0.95 (0.85–1.05)||0.93 (0.84–1.04)||0.89 (0.80–1.00)*||0.87 (0.77–0.99)*|
| Model 3b||(Ref)||0.93 (0.83–1.03)||0.91 (0.81–1.02)||0.86 (0.76–0.98)*||0.83 (0.71–0.96)*|
| Model 1||(Ref)||1.09 (0.97–1.22)||1.02 (0.91–1.14)||1.10 (0.98–1.23)||1.04 (0.94–1.15)|
| Model 2||(Ref)||1.11 (1.00–1.23)†||1.02 (0.92–1.13)||1.05 (0.94–1.17)||1.01 (0.92–1.11)|
| Model 3c||(Ref)||1.08 (0.98–1.20)||0.99 (0.89–1.10)||1.01 (0.90–1.13)||0.96 (0.87–1.06)|
| Model 1||(Ref)||0.96 (0.85–1.07)||0.99 (0.89–1.11)||1.01 (0.90–1.13)||1.02 (0.94–1.10)|
| Model 2||(Ref)||0.94 (0.84–1.04)||0.99 (0.89–1.09)||0.97 (0.87–1.08)||0.99 (0.92–1.07)|
| Model 3d||(Ref)||0.92 (0.83–1.03)||0.96 (0.87–1.07)||0.92 (0.83–1.04)||0.96 (0.88–1.03)|
|Executive function (n = 260)|
| Model 1||(Ref)||0.90 (0.80–1.01)†||0.84 (0.74–0.95)*||0.86 (0.77–0.97)*||0.81 (0.70–0.93)*|
| Model 2||(Ref)||0.87 (0.79–0.97)*||0.86 (0.77–0.96)*||0.89 (0.80–1.00)†||0.86 (0.75–0.98)*|
| Model 3b||(Ref)||0.89 (0.79–0.99)*||0.87 (0.77–0.98)*||0.92 (0.81–1.05)||0.88 (0.75–1.03)|
| Model 1||(Ref)||1.08 (0.96–1.21)||1.09 (0.97–1.22)||1.10 (0.98–1.24)||1.13 (1.02–1.25)*|
| Model 2||(Ref)||1.02 (0.92–1.14)||1.06 (0.95–1.18)||1.08 (0.96–1.20)||1.11 (1.01–1.21)*|
| Model 3c||(Ref)||1.00 (0.90–1.12)||1.03 (0.92–1.15)||1.04 (0.92–1.16)||1.07 (0.96–1.18)|
| Model 1||(Ref)||1.02 (0.91–1.14)||1.10 (0.98–1.23)||1.06 (0.94–1.17)||1.02 (0.94–1.10)|
| Model 2||(Ref)||1.00 (0.90–1.12)||1.04 (0.94–1.16)||1.06 (0.95–1.19)||1.01 (0.94–1.09)|
| Model 3d||(Ref)||0.99 (0.89–1.11)||1.03 (0.92–1.14)||1.02 (0.91–1.14)||0.98 (0.91–1.06)|
|Verbal expression (n = 273)|
| Model 1||(Ref)||0.94 (0.86–1.04)||0.99 (0.90–1.10)||0.89 (0.80–0.98)*||0.86 (0.78–0.97)*|
| Model 2||(Ref)||0.92 (0.85–1.00)*||0.97 (0.89–1.06)||0.89 (0.82–0.97)*||0.89 (0.81–0.97)*|
| Model 3b||(Ref)||0.93 (0.86–1.01)†||0.98 (0.90–1.07)||0.91 (0.83–1.00)*||0.90 (0.81–1.01)†|
| Model 1||(Ref)||1.04 (0.95–1.15)||1.08 (0.98–1.19)||1.10 (1.00–1.22)†||1.06 (0.98–1.16)|
| Model 2||(Ref)||1.04 (0.96–1.12)||1.04 (0.96–1.13)||1.08 (0.99–1.17)†||1.05 (0.98–1.13)|
| Model 3c||(Ref)||1.03 (0.95–1.11)||1.02 (0.94–1.11)||1.04 (0.96–1.14)||1.02 (0.94–1.10)|
| Model 1||(Ref)||1.09 (0.99–1.20)†||1.03 (0.94–1.14)||1.08 (0.98–1.20)||1.02 (0.95–1.09)|
| Model 2||(Ref)||1.04 (0.97–1.13)||1.04 (0.96–1.12)||1.06 (0.98–1.15)||1.04 (0.98–1.09)|
| Model 3d||(Ref)||1.04 (0.96–1.12)||1.02 (0.95–1.11)||1.03 (0.94–1.12)||1.01 (0.95–1.07)|
|Psychomotor Speed (n = 260)|
| Model 1||(Ref)||0.87 (0.78–0.99)*||0.86 (0.75–0.99)*||0.80 (0.70–0.92)*||0.72 (0.62–0.84)*|
| Model 2||(Ref)||0.94 (0.85–1.05)||0.94 (0.84–1.05)||0.95 (0.85–1.07)||0.91 (0.80–1.04)|
| Model 3b||(Ref)||0.95 (0.85–1.06)||0.95 (0.84–1.07)||0.97 (0.85–1.10)||0.92 (0.78–1.08)|
| Model 1||(Ref)||1.21 (1.06–1.39)*||1.19 (1.04–1.36)*||1.28 (1.12–1.47)*||1.18 (1.05–1.03)*|
| Model 2||(Ref)||1.11 (0.99–1.23)†||1.04 (0.93–1.15)||1.13 (1.01–1.26)*||1.08 (0.98–1.18)|
| Model 3c||(Ref)||1.10 (0.99–1.23)†||1.03 (0.92–1.14)||1.11 (0.98–1.24)†||1.05 (0.95–1.17)|
| Model 1||(Ref)||1.07 (0.93–1.22)||1.03 (0.90–1.13)||1.06 (0.92–1.22)||1.03 (0.94–1.14)|
| Model 2||(Ref)||1.04 (0.93–1.15)||1.02 (0.91–1.13)||0.97 (0.87–1.08)||0.99 (0.92–1.06)|
| Model 3d||(Ref)||1.03 (0.92–1.14)||1.00 (0.90–1.12)||0.93 (0.83–1.05)||0.96 (0.89–1.04)|
HoloTC, folate and cognitive functioning
Higher baseline holoTC was significantly associated with better performance in global cognition, executive functioning and psychomotor speed and approached significance in relation to verbal expression (P < 0.1) 7 years later, after taking into account baseline study covariates (Table 2, model 2). In model 2, the association between holoTC as a continuous variable and global cognition or verbal expression became stronger when excluding subjects with holoTC below the cut-off value of 35 pmol L−1: RD changed from 1.07 to 1.11 (95% CI 1.01–1.21) for global cognition and from 1.05 to 1.08 (95% CI 1.00–1.18) for verbal expression. Including tHcy and folate into the models (model 3) resulted in associations that were no longer significant, although the size and direction of the associations did not change substantially. No significant relationships between folate and any of the cognitive domains were detected, even after restricting the analyses to subjects with serum folate levels higher than the cut-off value of 5.3 nmol L−1.
tHcy, holoTC, folate and cognition in nondemented elderly subjects
Analyses were repeated excluding 20 individuals with incident dementia at follow-up. These subjects were older [73.0 (4.1) vs. 70.5 (3.5) years], had a lower BMI [25.7 (3.9) vs. 28.0 (4.0) kg m−2], lower systolic [143.1 (22.1) vs. 153.3 (21.5) mmHg] and diastolic [76.7 (9.7) vs. 82.8 (10.0) mmHg] blood pressures and higher serum tHcy [14.7 (5.5) vs. 12.6 (3.1) μmol L−1] and lower holoTC [65.7 (27.9) vs. 93.3 (51.6) pmol L−1] levels compared with nondemented participants. In addition, they had lower scores on episodic memory [4.9 (1.4) vs. 5.6 (1.5)], executive functioning [52.6 (18.4) vs. 37.6 (18.0)] and psychomotor speed [−0.46 (1.03) vs. 0.06 (0.80)] at baseline. After controlling for all study covariates (model 2), increasing tHcy concentrations were significantly associated with worse performance in episodic memory, executive functioning and verbal expression, but not global cognition 7 years later: RDs for tHcy as a continuous variable were 0.87 (95% CI 0.76–0.99) for episodic memory, 0.88 (95% CI 0.78–0.99) for executive functioning and 0.91 (95% CI 0.83–1.00) for verbal expression. The relationship between baseline tHcy and episodic memory remained unchanged even after including all three biomarkers into the model (model 3).
In nondemented subjects, elevated holoTC concentrations were significantly related to better performance in executive functioning [holoTC as a continuous variable: 1.10 (1.00–1.21)] and psychomotor speed [highest holoTC quartile: 1.22 (1.08–1.39)], but not global cognition or verbal expression. These estimates changed slightly after controlling for other study covariates (P < 0.10). The associations between holoTC as a continuous variable and global cognition or executive functioning became slightly stronger after restricting the analyses to those subjects with serum holoTC levels higher than the cut-off value of 35 pmol L−1: RD changed from 1.05 to 1.08 (95% CI 0.99–1.17) for global cognition and from 1.09 to 1.11 (95% CI 1.01–1.24) for executive functioning. Additional adjustment for tHcy and folate attenuated the results, although the direction of associations remained unchanged (data not shown).
Compared with the first quartile, the second and third quartiles of baseline folate were significantly associated with higher MMSE scores: RD was 1.12 (95% CI 1.03–1.21) for the second quartile and 1.09 (95% CI 1.01–1.18) for the third quartile; whereas the second and fourth quartiles were related to better performance in verbal expression: RD was 1.08 (95% CI 1.01–1.16) for the second quartile and 1.08 (95% CI 1.01–1.17) for the fourth quartile, after controlling for baseline study covariates. The association between the second folate quartile and MMSE remained significant even when tHcy and holoTC were incorporated into the models simultaneously: 1.11 (1.03–1.20). No relationship between folate and cognitive performance was found after excluding subjects with low folate levels.
In this prospective population-based study of elderly subjects, higher tHcy concentrations were related to worse cognitive performance 7 years later, irrespective of several potential confounders, including common vascular risk factors. In addition, higher holoTC levels had a protective role for several cognitive domains. The protective effect of holoTC was present over the whole normal range of holoTC concentrations.
These results are consistent with the findings from some previous longitudinal studies that reported a relationship between higher tHcy and worse cognitive functioning [7–10, 12–14, 17]. By contrast, no associations between tHcy and cognition were found in the Rotterdam study (follow-up 2.7 years) , the Leiden 85-Plus Study (follow-up 4 years) , the Mac Arthur Studies of Successful aging (follow-up 7 years)  or the Chicago Health and Aging Project (follow-up 6 years) . Differences in follow-up periods, cognitive measurement methods and other characteristics of the study population [e.g. relatively high tHcy levels ], or implementing the study after mandatory folic acid fortification , could explain some of the discrepancies between the studies.
Compared with tHcy, there have been few previous studies of the association between holoTC and cognitive performance in prospective settings. In agreement with the present results, lower levels of holoTC were associated with greater cognitive decline over 10 years in the Oxford Healthy Aging project . However, there was no relationship between holoTC and cognition in the Rotterdam scan study .
Vitamin B12/holoTC and folate are required for the remethylation of tHcy to methionine and the subsequent formation of S-adenosyl methionine (SAM). Biologically plausible mechanisms that may explain the association between tHcy and cognition include the impact of tHcy on cerebrovascular pathology, direct neurotoxic effects of tHcy or the possibility that elevated tHcy is only a marker of low levels of vitamin B12 and folate, which may themselves be related to cognition [2, 31]. Alternatively, the effects of vitamin B12/holoTC and folate may be mediated through SAM, which is the primary methyl donor in many biochemical reactions involved in normal brain functions, including the production of cell membrane phospholipids, myelin, monoaminergic neurotransmitters and nucleic acid [21, 32]. SAM deficiency may be related to white matter damage and brain atrophy , factors associated with cognitive decline and dementia .
In the present study, both tHcy and holoTC were related to executive functioning, a domain that may be particularly sensitive to vascular disease . In addition, tHcy and holoTC were associated with performance in the category fluency test, which partly relies on executive functions. The association between holoTC and executive functioning or category fluency (verbal expression) was attenuated when adjusting for tHcy, suggesting that the effects of holoTC on cognition may be partly explained by the vascular effects of tHcy. However, as tHcy was also related to episodic memory after controlling for several vascular risk factors, the effects on cognition may also be partly independent of the vascular pathway.
The associations between tHcy or holoTC and MMSE scores were weakened when study participants with incident dementia were excluded from analyses. It is possible that the effects of tHcy and holoTC on global cognition may take longer to develop, and become manifest closer to dementia onset. Of interest, higher folate concentrations became significantly related to the measures of global cognition and verbal expression in nondemented subjects. A possible explanation for this could be that the protective effects of folate are no longer manifest when the disease processes related to dementia become too advanced. In agreement with our findings, an earlier prospective study showed an association between low folate or vitamin B12 levels and an increased risk of AD, particularly amongst subjects with baseline MMSE scores >26 . However, the results from studies on the association between folate and cognitive performance have been notoriously inconsistent [8–10, 12–15, 23, 24], and such findings need to be further investigated.
The main strengths of the present study are the population-based design, follow-up time of at least 7 years, use of multiple tests to measure cognitive functioning, available data for a large number of potential confounders and evaluation of tHcy, holoTC and folate simultaneously in relation to cognitive performance. Compared with vitamin B12, levels of holoTC may represent a more sensitive assay of vitamin B12 status . The long follow-up period, the comprehensive evaluation and diagnostic protocol at each examination, recruitment of dementia-free subjects at baseline and adjustment of analyses for baseline cognitive measures make the findings less prone to the influence of reverse causality.
The main limitations of our study are the small sample size and availability of serum measurements at only one time-point, which may underestimate the relationship with cognition [1, 34]. tHcy was measured in serum, and although the stability of tHcy in serum samples that have been stored for a long time has been reported previously , folate levels decline during storage . This may have influenced our results and, as a consequence, associations may be weaker in our study than those found in populations in which folate has been measured in fresh samples.
Because creatinine values were not available, self-reported history of renal conditions was considered in the analyses. Selective survival may also have contributed to an underestimation of the relationship between tHcy and cognition, given that elevated tHcy has previously been linked to increased mortality .
In conclusion, our results suggest that tHcy, holoTC and folate might be independent predictors of cognitive performance 7 years later, even in nondemented elderly individuals. Because of the observational study design, we must caution against a causal interpretation of the findings. Further studies to confirm and refine the observed associations are warranted. Although there may be a limited window of opportunity for folate intervention (before disease processes become too advanced), intervention with tHcy and vitamin B12 may be effective at any time-point. Adequately timed and powered randomized controlled trials are needed to determine the potential efficiency, as well as the type of intervention required (e.g. dietary recommendations and combinations of B vitamin supplements), for better cognitive performance later in life.
Conflict of interest statement
No conflict of interest was declared.
The authors thank all the CAIDE participants and members of the CAIDE study group for their cooperation and data collection and management. The study was supported by the Karolinska Institutet (Sweden), the Swedish Research Council for Medical Research (Vetenskapsrådet), EU FP7 project LipiDiDiet 211696, EVO grant 5772720 (Finland), Academy of Finland grants 120676 and 117458, Strategic Research Program in Epidemiology (SFO) at the Karolinska Institutet L129395, Loo och Hans Ostermans stiftelse (Sweden), Stiftelsen Ragnhild och Einar Lundströms Minne Lindhés Foundation (Sweden), Stohnes Stiftelse Foundation (Sweden), Gamla Tjänarinnor Foundation (Sweden), Alzheimerfonden (Sweden), Alzheimer’s Association/Senator Mark Hatfield AwardAgreement No HAT-10-173121 (USA), Demensfondens Forskningsstipendier (Sweden) and Stiftelsen Dementia (Sweden). The funding sources did not have any role in the design and conduct of the study, the collection, management, analysis and interpretation of data or preparation of the manuscript.