Abstract
Abnormal hemoglobin levels are a risk factor for Alzheimer’s disease (AD). Although the mechanism underlying these associations is elusive, inadequate micronutrients, particularly folate and vitamin B12, may increase the risk for anemia, cognitive impairment, and AD. In this study, we investigated whether the nutritional status of folate and vitamin B12 is involved in the association between low hemoglobin levels and the risk of AD mortality. Data were obtained from the 1999–2006 National Health and Nutrition Examination Survey (NHANES) and the NHANES (1999–2006) Linked Mortality File. A total of 4,688 participants aged ≥60 years with available baseline data were included in this study. We categorized three groups based on the quartiles of folate and vitamin B12 as follows: Group I (low folate and vitamin B12); Group II (high folate and low vitamin B12 or low folate and high vitamin B12); and Group III (high folate and vitamin B12). Of 4,688 participants, 49 subjects died due to AD. After adjusting for age, sex, ethnicity, education, smoking history, body mass index, the presence of diabetes or hypertension, and dietary intake of iron, significant increases in the AD mortality were observed in Quartile1 for hemoglobin (HR: 8.4, 95% CI: 1.4–50.8), and the overall risk of AD mortality was significantly reduced with increases in the quartile of hemoglobin (p for trend = 0.0200), in subjects with low levels of both folate and vitamin B12 at baseline. This association did not exist in subjects with at least one high level of folate and vitamin B12. Our finding shows the relationship between folate and vitamin B12 levels with respect to the association between hemoglobin levels and AD mortality.
INTRODUCTION
Alzheimer’s disease (AD) is an emerging public health problem. The prevalence of AD has been estimated to be approximately 5 million Americans, and this number is expected to increase by approximately three-fold by 2050 [1]. AD is one of the most expensive chronic diseases for Americans, and the total costs of AD care are expected to increase from $172 billion in 2010 to $1.08 trillion in 2050 [2]. There is currently no cure for AD, although medication is available to temporarily reduce some symptoms. Therefore, understanding the factors that contribute to the development of AD may be of particular importance to prevent or delay the onset of AD.
Anemia or abnormal hemoglobin levels are often associated with the development of AD [3–5]. The risk of AD has been shown to increase by two-fold among older adults with anemia relative to that of adults with those without anemia [3–5]. Although the mechanisms underlying these associations remain elusive, inadequate micronutrients, particularly folate and vitamin B12, may increase the risk of anemia, cognitive impairment, and AD [6–11]. On the basis of previous knowledge, the intake of folate and vitamin B12 may have a role in the relationship between hemoglobin levels and AD mortality, but their potential interrelationship has not been proven. In this study, we hypothesized that the intake of folate and vitamin B12 affects the association between anemia and the risk of AD. We investigated the association between hemoglobin levels and AD mortality in groups of older adults stratified by their levels of red cell folate and vitamin B12 using mortality data from a nationally representative sample of the US.
MATERIALS AND METHODS
Study population
We used data from the 1999–2006 NHANES and the NHANES (1999–2006) Linked Mortality Public File from the United States [12]. The latter was a follow-up study of mortality that matched records from the NHANES with data in the National Death Index (NDI) as of December 31, 2011. The date and cause of death in the NDI were derived from death certificates. The NHANES Linked Mortality files are available in two-year increments (e.g., NHANES 1999–2000, NHANES 2001–2002, NHANES 2003–2004, and NHANES 2005–2006) and only include the mortality follow-up for eligible individuals: those who are 17 years of age and younger are classified as ineligible.
From the 1999–2006 NHANES data and the linked morality file, we excluded 3 participants who had no information matched with NDI records and initially included 7,173 participants aged ≥60 years at the time of the initial survey. We excluded 1,802 older adults for whom the hemoglobin, red cell folate, or serum vitamin B12 levels were unavailable or anemia [hemoglobin < 12 g/dL in female and <13 g/dL in male] was defined. We also excluded 683 participants who were missing data for other variables. The cohort analysis presented in this study was based on 4,688 older adults in the 1999–2006 NHANES. This study’s protocol was approved by the institutional review board of Ajou University Hospital (IRB No. AJIRB-MED-MDB-13-017).
Variables of interest
For the measurement of hemoglobin, red cell folate, and serum vitamin B12 levels, blood specimens were processed, stored, and shipped to the Division of Environmental Health Laboratory Sciences, National Center for Environmental Health, Centers for Disease Control and Prevention for analysis. Vials were stored under the appropriate frozen (–20°C) conditions until they were shipped to the National Center for Environmental Health for testing. For hemoglobin levels, complete blood count analyses are based on the Beckman Coulter method. Red cell folate and serum vitamin B12 levels are measured using the Bio-Rad Laboratories “Quantaphase II Folate/Vitamin B12” radioassay kit. The assay is performed by combining either serum or a whole blood hemolysate sample with 125I-folate and 57Co-vitamin B12 in a solution containing dithiothreitol (DTT) and cyanide.
The International Classification of Diseases 10th Revision (ICD-10) codes were used for all causes of deaths, and the code for AD was G30 based on the ICD-10 underlying causes of death.
Baseline data were obtained from the interview data of the 1999–2006 NHANES and included the age (60–69, 70–79, and 80–89 y), sex, ethnic background (white, black, Hispanic, or other), education (less than high school, high school graduate, college, or more), and smoking history (yes or no). Body mass index (BMI) was calculated by dividing the measured weight in kilograms by the measured height in meters squared and was categorized into two groups- normal weight (<24.9 kg/m2) and overweight/obesity (≥25 kg/m2). For disease history, hypertension was defined as a systolic blood pressure ≥140 mm Hg and a diastolic blood pressure ≥90 mm Hg, the use of antihypertensive drugs or previous physician-diagnosed hypertension. Diabetes was defined as a fasting plasma glucose level ≥6.99 mmol/L, a nonfasting plasma glucose level ≥11.1 mmol/L, current insulin use or a prior physician diagnosis of diabetes. Data regarding the average daily intake of iron were collected and divided into quartiles (Q1:<8.67 mg, Q2 : 8.67–12.19 mg, Q3 : 12.19–17.34 mg, and Q4:≥17.34 mg).
Statistical analysis
The hemoglobin, red cell folate, and vitamin B12 levels were divided by quartiles. The chi-square test and (t-test were used to compare demographic variables and distributions of hemoglobin, folate, and vitamin B12 between subjects with and without AD mortality. To investigate the interrelationship between folate and vitamin B12, we categorized three groups based on their quartiles: Group I, low for both folate and vitamin B12; Group II, either high folate and low vitamin B12 or low folate and high vitamin B12; and Group III, high for both folate and vitamin B12. Cox proportional hazards regression was used in the multivariate analyses of AD mortality. In each group, the hazard ratios (HRs) and 95% confidence intervals (CIs) of the outcome variable associated with the hemoglobin quartile level were compared with the first quintile (Q1:≤13.5 g/dL). The proportional hazards assumption for the model was confirmed. The model was adjusted by age, sex, ethnicity, education, smoking history, BMI, the presence of diabetes or hypertension, and dietary intake of iron.
The weighted estimates of the population parameters were computed by the National Center for Health Statistics to account for the complex sampling design. All analyses were performed using the PROC SURVEY procedures in SAS 9.2 (SAS Institute, Cary, NC, USA).
RESULTS
Of the 4,688 participants, 49 subjects died due to AD. Table 1 compares the characteristics of the study participants with and without AD mortality. The subjects with AD mortality were significantly more likely to be older and white and have college or more education and normal weight at baseline compared with those without AD mortality. However, no baseline differences between participants who did or did not have AD mortality were detected for sex, cigarette smoking, history of hypertension and diabetes, or dietary intake of iron.
Table 2 presents the distribution of hemoglobin, folate, and vitamin B12 according to AD mortality. Participants with AD mortality had significantly lower hemoglobin levels (14.0 g/dL vs. 14.1 g/dL; p-value = 0.0089) and were differentially distributed across the quartile of hemoglobin (p-value = 0.0435) compared with those without AD mortality. However, there was no significant difference between the two cohorts for the mean (SD) red cell folate and serum vitamin B12 and number of subjects in the quartile of folate and vitamin B12.
We categorized three groups based on the quartiles of folate and vitamin B12 - Group I (low folate and vitamin B12), Group II (high folate and low vitamin B12 or low folate and high vitamin B12), and Group III (high folate and vitamin B12) and compared their hemoglobin means (Table 3). There were no significant differences in mean hemoglobin levels by the quartiles among Group I-III.
Table 4 shows the HRs (95% CI) for AD mortality according to hemoglobin in the three categories stratified for the levels of folate or vitamin B12 at baseline. Mortality events for participants in Group I (low levels for both folate and vitamin B12) were significantly increased with the quartile of hemoglobin (p for trend = 0.0318), whereas no trend was observed among those in both Group II (either high folate and low vitamin B12 or low folate and high vitamin B12) and Group III (high of both folate and vitamin B12). Additionally, for Group I significant increases in the AD mortality risk were observed in Q1 for hemoglobin (HR: 8.4, 95% CI: 1.4–50.8), after adjusting for age, sex, ethnicity, education, smoking history, BMI, the presence of diabetes or hypertension, and dietary intake of iron. The overall risk of AD mortality was significantly reduced with increases in the quartile of hemoglobin (p for trend = 0.0200; Fig. 1). By contrast, for participants in Group II or Group III, there was no significant association between hemoglobin and AD mortality risk.
DISCUSSION
The present study investigated the associations among red cell folate, vitamin B12, hemoglobin, and AD mortality in older US adults. We found an interrelationship between folate and vitamin B12 levels with respect to the association between hemoglobin levels and AD mortality. Specifically, participants with low levels of both folate and vitamin B12 (Group I) at baseline had significantly increased risk of AD mortality with increases in the quartiles of hemoglobin levels; however, those who had at least one high level of the two variables at baseline (i.e., high folate/ low vitamin B12, low folate/high vitamin B12, or high folate/ high vitamin B12) did not show a significant linear trend between hemoglobin and AD mortality. This finding suggests that the nutritional intake of folate and vitamin B12 levels plays a role in developing AD or subsequent AD mortality associated with low hemoglobin levels.
There have been no other reported studies on the associations among folate, vitamin B12, hemoglobin, and AD mortality. However, much concern has been expressed regarding folate and vitamin B12 in terms of cognitive impairment and dementia [8, 13–15]. Subjects with low folate or vitamin B12 levels in the blood or because of dietary intake had an increased risk of cognitive impairment and incidence of AD [10, 13–15]. A randomized controlled trial of subjects with mild cognitive impairment showed that the mean plasma total homocysteine, a strong and independent risk factor for developing AD [16], was 30% lower in those with B vitamin treatment than in those with the placebo [17]. Overall, these findings are heterogeneous but suggest adverse effects of low folate and vitamin B12 on the development of AD by promoting oxidative stress and hyperhomocysteinemia [18, 19]. Our results also support the hypothesis that low intake of folate and vitamin B12 contributes to AD risk.
Separately, whether hemoglobin is a risk factor for AD is interesting. Hemoglobin is the major iron-containing protein of erythrocytes. Studies have reported that hemoglobin is specifically expressed in neurons of the cortex [3, 20], which suggests that hemoglobin has a role in intraneuronal oxygen homeostasis [21]. Decreased hemoglobin levels have been observed in neurons of AD brains [21]. A recent study showed low hemoglobin to be an accurate diagnostic biomarker panel for AD diagnosis [22]. Longitudinal and cross-sectional studies have provided evidence of a significant association between anemia or low hemoglobin levels and an increased risk of AD [3–5]. Beard et al. (1997) evaluated the association between AD and anemia using both case-control and cohort methodologies [3]. In the case-control study, people with anemia had an approximately two-fold (OR: 1.88, 95% CI: 1.17–3.03) increased risk for occurrence of AD, whereas no overall increased risk of AD was observed in the cohort study. A recent cross-sectional study reported a strong association between anemia and AD (OR: 2.43, 95% CI: 1.30–4.54) after adjusting for age and APOE-ɛ4 [4]. Moreover, AD emerged as a stronger risk factor (OR: 3.41, 95% CI: 1.68-6.92) for anemia than did aging 5 years (OR: 1.95, 95% CI: 1.71–2.21) [4]. In a prospective study, older adults with anemia had a 60% (95% CI 1.02–2.52) higher risk of developing AD over 3.3 years than did those with clinically normal hemoglobin [5]. The authors suggested that a nonlinear relationship between baseline hemoglobin levels and AD risk existed that showed an increased risk of AD for each unit of hemoglobin lower and higher than 13.7 g/dL [5]. Consistent with these findings, we found that subjects with AD mortality had lower hemoglobin levels than did those without AD mortality and showed that AD mortality increased with decreasing quartiles of hemoglobin (Table 2).
Notably, the association between hemoglobin and AD mortality risk was robust, in subjects with low levels of both folate (<172.1 nmol/L) and vitamin B12 (<259.1 pmol/L) at baseline (Table 3). This association did not exist in subjects with either a high level of folate or vitamin B12 or both. The present data suggest that folate and vitamin B12 status are involved in the hemoglobin-AD mortality link. In this context, two studies by Morris and Selhub appear to slightly support our hypothesis. Using data from the 1999–2002 NHANES, they found an interaction effect of vitamin B12 and folate status on anemia (p = 0.03) or cognitive impairment (p < 0.001). The risks of anemia and cognitive impairment significantly increased in older Americans with low vitamin B12 levels (<148 pmol/L) and high serum folate levels (>59 nmol/L) [8, 23]. Unfortunately, direct comparison of these studies with ours is impossible, because of the small number of subjects with AD mortality (n = 49) and different methods (i.e., time periods, serum or red cell folate, and concentration of low or high folate and vitamin B12). The possible explanation for the interaction between folate and vitamin B12 status on the risk of anemia, AD incidence, and hemoglobin-AD mortality link is abstruse. Moreover, we cannot rule out the possibility that AD risk attributed to low intake of folate and vitamin B12 actually may have been due to other nutritional factors. For example, poor diet results in low folate intake and also in low carotenoids intake. Recent studies have suggested that lutein and zeaxanthin play a role in cognitive function and mortality risk for AD [24,25, 24,25]. However, our findings suggest that it is important to assess whether the nutritional status of folate and vitamin B12 modulates the risk for AD and subsequent mortality in connection with low hemoglobin (or anemia), which improves with increased levels of folate and vitamin B12.
To our knowledge, this is the first study to show the potential role of folate and vitamin B12 in the association between low hemoglobin and the risk of AD mortality. We analyzed the data from the large-scale high-power NHANES study and included potential covariates to establish the independence of association between them. However, the limitations of our study should be considered. First, the data included a single measurement of red cell folate and serum vitamin B12 levels obtained at baseline, and there was the long interval between the measurement and follow-up period. Second, participant characteristics were available only from the initial survey, and we did not consider changes in health behaviors or risk factors from baseline. Because of the observational nature of this investigation, we cannot rule out residual confounding effects due to unmeasured confounders. The study is also not free from bias because several variables were dependent on self-reported data.
In conclusion, the study results showed that low levels of folate and vitamin B12 were associated with a higher risk of AD mortality with decreasing hemoglobin levels. Our findings suggest that adequate intake of folate and vitamin B12 is a preventive strategy for AD mortality, particularly for those with a high risk of developing anemia.
Footnotes
ACKNOWLEDGMENTS
We thank members of the National Center for Health Statistics (NCHS) of the Centers for Disease Control (CDC) and Prevention and the participants who enrolled in the study.
This study was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (Grant No. 2015R1A1A3A04000923, 2015R1D1A1A01059048).
