BALTIC SEA DIETARY PATTERN IN RELATION WITH BONE MINERAL DENSITY IN ELDERLY WOMEN
Homa Sadeghi Master’s thesis
Public Health Nutrition School of Medicine
Institute of Public Health and clinical Nutrition
University of Eastern Finland June 2014
University of Eastern Finland, Institute of Public Health and Clinical Nutrition Public Health Nutrition
Sadeghi, Homa: Baltic Sea Dietary Pattern in relation with Bone Mineral Density in Elderly Women
Master’s thesis, 49 pages, no attachments
Instructors: Arja Erkkilä, Adjunct Professor, PhD and Jaakko Mursu, PhD.
June 2014
Key words: Baltic Sea Diet, Bone Mineral Density (BMD), elderly, Post menopause, osteoporosis
BALTIC SEA DIETARY PATTERN IN RELATION WITH BONE MINERAL DENSITY IN ELDERLY WOMEN
Recent studies on osteoporosis have focused on dietary pattern rather than the intake of individual nutrients. The aim of this observational study was to assess the relationship between bone mineral density (BMD) and Baltic Sea diet characterized by high consumption of Nordic berries and fruits, whole grains, vegetables, fish, and lower consumption of processed meat and alcohol among elderly women.
The data were from interventional “Kuopio Osteoporosis Risk Factor and Fracture
Prevention study (OSTPRE-FPS)”. At the baseline, 554 subjects filled in 3-day food record and questionnaire on life-style, health status and medications. BMD was measured by dual energy X-ray absorptiometry in lumbar spine (L2-L4), femoral neck and total body at the baseline and year 3. The intakes of whole grain cereals, vegetables, fruits, fish, low fat dairy, total fat intake, and polyunsaturated fatty acid to saturated fatty acid ratio, as well as consumption of processed meat and alcohol as negative components were categorized into quartiles. The sum of quartiles was defined as Baltic Sea Diet Score. Associations between dietary score and BMD were analyzed using mixed linear model. We tested energy intake, hormone therapy, calcium and vitamin D supplementation, diseases affecting BMD, mobility, time from menopause, life time smoking, age, body mass index, and intervention group as covariates. In the final model only significant covariates including current
mobility, calcium supplementation, mobility, smoking, BMI, and hormone therapy were retained.
The elderly postmenopausal subjects’ mean age (±SD) was 67.9±1.87, and mean BMI was 28.8±4.71 kg/m2. The mean BMI was in overweight range (BMI≥25g/m2) in all quartiles.
The mean value for lumbar BMD was 1.09g/cm2 and the mean femoral BMD among the four quartiles was 0.87g/cm2. Average total body BMD was 1.07 g/cm2. Almost 90% of the subjects had normal mobility. The difference of root vegetable and alcohol
consumption was not significant among the four BSQs. The highest scores of Baltic Sea diet were associated with highest levels of energy (kcal), total fat (g/d), protein, and carbohydrate (g/d) intakes, calcium (mg/d), vitamin C (mg/d), fiber (g/d), sodium (mg/d);
and the lowest energy percentage from carbohydrate. The subjects with the lowest BSQ gained the greatest percentage of energy intake from carbohydrate.
We found a significant association between Baltic Sea diet Score and lumbar and femoral BMD (p-value=0.046 and 0.027, respectively) so that the highest BMD was in the third quartile. However, the Baltic Sea diet score was not significantly associated with total BMD (p-value=0.089). In conclusion, Baltic Sea diet may be beneficial for BMD in elderly women.
2. Theoretical Background ... 9
2.1 Osteoporosis ... 9
2.2 Bone metabolism ... 10
2.3 Factors related to BMD ... 12
2.3.1 BMI, fat and lean tissue ... 12
2.3.2 Physical activity ... 13
2.3.3 Age and gender ... 13
2.3.4 Genetics ... 13
2.4 Dietary Patterns and bone health ... 14
2.4.1 Mediterranean diet ... 17
2.4.2 Baltic Sea Diet ... 18
2.4.3 Energy and nutrient density of diet... 19
2.5 Particular food groups and nutritional components and bone health ... 20
2.5.1 Fruit and vegetable intake... 20
2.5.2 Phytoestrogens ... 21
2.5.3 Fish and PUFA ... 22
2.5.4 Alcohol consumption ... 22
2.5.5 Alkali foods ... 23
2.6 Nutrient intakes and bone health ... 24
2.6.1 Calcium and vitamin D ... 24
2.6.2 Protein ... 26
3. Aims and Hypothesis ... 28
4. Subjects and methods ... 28
4.1 Questionnaire ... 29
4.2 Dietary assessment ... 29
4.4 Anthropometric measurements ... 30
4.5 Statistical analysis ... 30
5. Results ... 31
6. Discussion ... 37
7. Conclusion ... 41
8. References ... 42
BMC Bone Mineral Content
BMD Bone Mineral Density
BSQ Baltic Sea diet Score Quartile
BTM Bone turnover markers
DASH Dietary Approaches to Stop Hypertension
DXA Dual-energy x-ray
FFQ Food Frequency Questionnaire
HEI Healthy eating index
hPTH Human parathyroid hormone
HT Hormone Therapy
IGF-1 Insulin growth factor-1
MUFA Monounsaturated fatty acid
PTH Parathyroid hormone
PUFA Polyunsaturated fatty acid
SFA Saturated fatty acid
USDA US Department of Agriculture
1. Introduction
Osteoporosis is one of the health challenges in the elderly, particularly among postmenopausal women. In 1995 only in the USA 13.8 billion dollars per year or 38 million per day was spent for treating fractures (Prentice 2004). In 2005, a treatment of total of 20 million fractures cost more than 17 billion dollars, and it is predicted that during the coming 20 years this expenses will increase over 50 % (Langstemo et al. 2011). It has been reported that the overall incidence of hip fracture has doubled during 1992-2002 in Finland (World Health Organization 2011, Dawson-Hughes 2002). During 1996-2008, 7000 hip fractures per year were reported in Finland (Finnish Orthopedics Association, 2011).
Various factors such as physical activity, diet and dietary status, ethnicity, hormonal fluctuations, energy expenditure, and body mass index (BMI), may have an effect on osteoporosis (Prentice 2004). The studies on osteoporosis in relation with diet are branched into two major areas. Most researchers have studied the effect of individual nutrients such as calcium, vitamin D, magnesium, zinc, vitamin A, vitamin K, and vitamin C on
osteoporosis risk. The other studies which are more recent, attempt to investigate the effect of dietary patterns such as Mediterranean diet, Western diet, traditional English diet, etc.
on bone health. The rationale for preference of dietary pattern to individual nutrients is that, there are still many unrevealed interactions among nutrients, and hidden mechanisms affect osteoporosis and its determinants. In other words, studying independent role of a single factor and possible interactions is difficult. In addition, the role of a single factor is most likely small and thus difficult to detect, and studying the effect of whole diet may be more practical for daily life (Bendich 2005). According to my knowledge, no research has been done on the relation of Baltic Sea dietary pattern and bone mineral density (BMD).
Baltic Sea diet is basically characterized with high consumption of Nordic berries and fruits, whole grains, vegetables, fish, and lower consumption of processed meat and alcohol.
Osteoporosis can be assessed in several ways such as; measuring osteocalcin turnover, BMD, etc. or indirectly by following the incidence of hip, wrist, and vertebral fractures (Bouchard 2012).
This study will assess the relationship between Baltic Sea dietary patterns and BMD. Baltic Sea diet is basically characterized by high intake of vegetable, fruit, and whole grain, which are the main characteristics of nutrient-dense diet. In previous studies, nutrient dense or healthy dietary pattern have been found to be related to a decreased risk of fracture (Langstemo et al. 2011). The data are from interventional Kuopio Osteoporosis Risk Factor and Fracture Prevention study “(OSTPRE-FPS)”.
2. Theoretical Background 2.1 Osteoporosis
Osteoporosis is the major cause of disability in the world and in the USA 10 million people are osteoporotic and 18 million have low bone mass (Zalloua et al. 2007). Osteoporotic fractures are the main cause of morbidity and disability in elderly. Hip fractures can lead to early death and 20% higher mortality, which results in many billions of dollars cost every year (Prentice 2004, Siris et al. 2001).
The prevalence of low BMD was approximately 50 to 80% among women over 50 years old in a study done in the USA between 1997 and 1999 (Prentice 2004, Siris et al. 2001).
According to World Health Organization (WHO), low BMD is defined as -2.5<T score< -1 and is also named osteopenia; and the BMD lower than mean peak bone mass of young healthy adults with 2.5 or higher standard deviation; and T-score lower than -2.5 is defined as osteoporosis. In addition, T-score is calculated from manufacturers’ healthy, white young adult reference databases via the standard formula of:
T score=BMD of participant - mean BMD of reference population/SD of reference population mean BMD (Prentice 2004, Siris et al. 2001).
The prevalence of fracture in older women is 1.5 million per year. The mechanism behind the higher incidence of fracture and osteoporosis among postmenopausal women is related to decreasing of estrogen concentration (McTiernan et al.2009). Osteoporosis is becoming a challenge in Asian countries because of the increase in mean age. Whites have higher BMD than Asians and lower BMD than blacks (Kumar et al. 2010). In Finland, the prevalence of osteoporosis among 70-79 year old women is 38.5% (World Health Organization 2011).
BMD is a reliable marker for osteoporosis. Although, BMD is one of the most concordant factors of fracture, it is better to choose fracture as the outcome rather than BMD in research studies, because fracture risk is the main outcome for osteoporosis; whereas, BMD is one of the factors in relation with fracture risk (Langstemo et al. 2011). Age, gender, and menopause status are three main predictors of BMD (Zalloua et al. 2007).
BMD is strongly related to geographic region, life style, and ethnicity. For instance, the incidence of hip fracture is low among African-Americans. BMD and BMC have positive relationship with energy expenditure, lean body mass, calcium intake, bone area, body fat, and skeletal age (Ilich et al. 2003).
2.2 Bone metabolism
Bone is both an organ and a tissue, 80% of which is made up from cortical tissue consisting of osteons or Haversian systems, and the remaining 20% is trabecular or cancellous bone tissue (Mahan et al. 2012). Bone modeling during growth and bone remodeling after growth is done by both trabecular and cortical tissue (Ross et al. 2013).
Bone modeling is defined as the growth of skeleton to the final maturity height, and bone remodeling is the continuous resorption and reform of bone by osteoclasts and osteoblasts, respectively (Mahan et al. 2012). In normal young adults, there is a balance between resorption and formation phases. As every tissue in human body, bone is a complex of different nutrients forming the bone cells and bone matrix or osteoid (Mahan et al. 2012).
Osteoblasts are bone cells forming the bone structure, and osteoclasts are responsible for bone resorption and breakdown (Mahan et al. 2012). Matrix is formed from collagen,
proteoglycans, growth factors, osteocalcins and hydroxyapatit. Collagen is a triple helix of amino acids mainly glycine and prolin and the main content of hydroxyapatit is calcium and phosphorus (Mahan et al. 2012).
Calcium homeostasis aims to stabilize the serum calcium concentration. In the case of inadequate dietary calcium, homeostasis is dependent on drawing on minerals from the bone to reach the calcium of serum to 10 mg/dL (Ross et al. 2013). Calcium drawing from bone is done via two sources including mobilizable calcium ions of bone fluid or
osteocalcin resorption from the bone tissue. The regulating of calcium is mainly done by parathyroid hormone (PTH) and 1, 25 dihydroxy vitamin D3. PTH adjusts blood’s calcium affecting skeleton, kidney, and indirectly gut. Vitamin D improves calcium digestion via active or passive transportation of vitamin D across intestinal cells. The role of vitamin D is noticeable in insufficient intakes of calcium (Bendich 2005) via formation of bone matrix proteins and suppression of bone degradation. Nutrients other than calcium and phosphorus play important role in bone formation and activity including vitamins C, K and minerals including copper, manganese and zinc (Mahan et al. 2012).
Studies have shown that the deficiency of calcium in elderly results in increase of PTH levels and secretion dynamics. The intake of 2400 mg calcium can regulate the function of PTH in elderly (Bendich 2005). Human parathyroid hormone (hPTH) (1-34) leads to significant increase in lumbar spine BMD and slight increase in femoral neck and total hip BMD (Cranney et al. 2006). The effect of gonad hormones on bone mass is significant and studies have shown that bone mass in the absence of menopausal hormones is 12-15 % lower compared with the presence of estrogen. The effect of estrogen lack on bone loss is not long lasting. In the case of adequate nutritional status, the bone loss only lasts for few years, and then flattens. The effect of estrogen on calcium retention is mediated not only by skeletal mechanisms, but also by non-skeletal factors including intestinal calcium absorption and renal calcium preservation. In other words, in the absence of gonad hormones, the intake of calcium should be increased (Prentice 2004). According to Lumachi and coworkers (2013), BMD is highly affected by serum concentration of
Insulin-like Growth Factor 1 (IGF-1) (R=0.64, P=0.0016). According to this study, elderly
postmenopausal women represented increased BMD via the effect of IGF-1 hormone and PTH.
Sodium and calcium have similar transport process in proximal tubules of kidney, so that every 2300 mg of sodium excretion leads to 20 to 60 mg of calcium out. It is worth noting that the calciuric effect of sodium (Na) is only concerned to sodium chloride (table salt) and sodium bicarbonate (Bendich 2005). However, according to more recent studies, higher sodium intake of 3000 mg/day in elderly women does not have detrimental impact on bone density in the case of sufficient intake of calcium and vitamin D (Ilich et al. 2010).
Besides, the studies have shown that sodium intake of Asian and African population is as high as that of Western populations (Prentice 2004).
2.3 Factors related to BMD
2.3.1 BMI, fat and lean tissue
Body weight and body lean mass are challenging variables in relation to BMC; because although higher leanness (the ratio of lean tissue to fat tissue) is a determinant of higher bone mass, in postmenopausal women this equivalence is not true, and fatness is associated with higher bone mass. In other words, in the youth higher leanness is beneficial because of osteogenic effect of muscles (Parsons et al. 1996), whereas, in the elderly, adipose tissue compensates the lack of osteogens after menopause and produces sexual hormones
(Baumgartner et al. 1996).
The average rate of BMD loss has been reported to be 1% per year in women and 0.8 % per year in men 67-79 years old (Kaptoge et al. 2003). Weight gain and high force expiratory volume per second (FEV1) were protective against BMD loss (Kaptoge et al.
2003). The positive relationship between BMD and BMI makes an idea that higher energy intake and excess intake of energy leads to higher BMD (Langsetmo, 2010); On the
contrary, if weight gain is the result of higher energy intake from Western diet, it may have
negative impact on bone health (McNaughton et al. 2011). Weight loss through calorie reduction has been found to reduce BMD (MCTiernan et al. 2009).
2.3.2 Physical activity
High physical activity increases the likelihood of better bone parameters (Bendich 2005).
Studies have shown that weight bearing exercises either in children or adults (4-20 years old) have positive effect on BMD (Ilich et al. 2003). Physical inactivity is one of the most threatening behavioral risk factors for osteoporosis (World Health Organization, 2011).
Physical activity is positively related to BMD gain and common practiced exercises such as climbing the flight stairway are protective against BMD loss (Kaptoge et al. 2003).
2.3.3 Age and gender
The trend of bone mass is increasing to the peak bone mass after the adolescence height, and afterward, the density of bone remains stable; whereas, during aging the higher activity of osteoclasts comparing osteoblasts, leads to decrease in BMD (Ross et al. 2013).
Moreover, the onset of menopause attenuates the BMD two to six times more than before, which makes the female gender as a risk factor for BMD loss. Decrease in the levels of gonadal hormones is one of the outcomes of aging leading to bone loss (Ross et al. 2013).
Sex steroid hormones have significant effects on bone tissue maintenance. Waiving the probable adverse effects of sex hormone therapy (HT), it can maintain BMD and prevent bone loss (Imai, 2013).
2.3.4 Genetics
According to the estimations, about 80% of the variance in peak bone mass is related to hereditability (Ross et al. 2013). Bone loss with aging is associated to genetics with a weaker association as compared to peak BMD. Some of the genes affecting BMD are
genes coding for the low-density lipoprotein receptor related protein 5 (LRP5), estrogen receptor alpha (ESR1), and osteoprotegerin (OPG). The genetic determinant of vitamin D receptors (which affect gastrointestinal absorption of calcium) has been associated with BMD. The polymorphism in LRP5 results in decreased spinal BMD in the case of inadequate calcium intake (less than 680 mg/day) (Ross et al. 2013).
2.4 Dietary Patterns and bone health
Dietary pattern as a prominent determinant of healthy life style has been the matter of research in medical prospective and retrospective studies attempting to investigate their probable associations with osteoporosis (Kontogianni et al. 2009). For more than 30 years ago the majority of nutritional health studies concentrated on single food compounds or nutrients. However, this kind of study has some limitations. Human diet includes mixture of foods and nutrients. Studies on single nutrients do not consider the interactions and inter-correlations among nutrients. Since the effect of a single nutrient factor is small, it is difficult to investigate the small effects of separate nutrients. Moreover, results from dietary pattern studies are more practical and useful to make recommendations for populations (Newby et al. 2004). Conclusively, the study of dietary patterns can lead to finding more probable effective factors on BMD (Tucker et al. 2002).
Food preferences are dependent on various determinants such as health, lifestyle,
economic, cultural, environmental, and ecological factors. To assess the diet quality in a systematic and practical frame, dietary patterns are organized. Dietary pattern is a general format of main food groups with particular nutritional characteristics presenting a special population, geographical area, and ecological possibilities. Moreover, the patterns may be based on special dietary index guidelines. Two methods of dietary pattern have been used in studies, including a priori and posteriori. A priori method is defined so that predefined criteria are used to evaluate dietary quality. There are many criteria for evaluation of dietary intake according to a priori method such as dietary diversity score, simple food group based score, nutritional risk scores, and Mediterranean diet score. Mediterranean diet score is whole dietary pattern based on reduced biomarkers of CVD, hypercholesterolemia,
diabetes, and hypertension and based on traditional Mediterranean consumption pattern.
Eating patterns may be defined as posteriori, independent of authors’ definition for healthy pattern. In posteriori method the intake of various components is assessed and the main characteristics of dietary intake are assessed among the study population via statistical methods such as factor analysis. The most common tools of nutritional epidemiology are factor and cluster analysis. Factor analysis is based on inter-correlation among dietary items; whereas, in cluster analysis patterns are initiated according to personal intake differences (Newby et al. 2004, Whittle et al. 2012).
Healthy eating index is established by the United States Department of Agriculture (USDA) to evaluate the diet quality according to the population requirements and also to monitor dietary changes of the US population. Healthy dietary index is characterized by the frequency of particular food groups and also amount of different food groups and nutrients (Kennedy et al. 1995). HEI score is based on 12 components (including total fruits, total vegetables, total grains, dark green and orange vegetables, whole grains, whole fruit and legumes which are given 0-5 points and meat, milk, beans and oils which are given 0-10 points). Each component has a minimum and maximum rate per 1000 kcal, so that the highest point is achieved by recommended intake of components, and zero score is assigned in the case of not eating from the special food group. HEI score is not
representative of only one type of diet; there may be several varieties of dietary pattern with the same HEI score. The majority of US population does not have a healthy diet (high intake of fruits, vegetables, and dairy products), so there are few people in higher ends of score. In one of the studies the effect of HEI on bone turnover markers (BTM) was
assessed in 827 postmenopausal women aged ≥45 year. There was no association between BTM and HEI 2005 score in that study (Hamidi et al. 2011, Kennedy et al. 1995).
Healthy dietary patterns may have represented positive effects on BMD. Healthy dietary patterns are diets with high intake of fruits and vegetables and rice or pasta, white meat, oily fish and dairy products. On the contrary, snack food patterns with positive scores for confectionary, crisps or nuts and sauces; and processed food patterns with positive scores for ready-made foods; including cakes and desserts; and negative scores for fats or oils and breads have negative impact on BMD. The negative effect of processed foods is hypoth-
esized to be due to low protein content of this type of diet (Heaney et al 1996, Hardcastle et al. 2011). There is no consensus about the role of protein in bone health; however, some studies have stated that low protein diets have negative effects on bone health (Heaney et al. 1996, Hardcastle et al. 2011). The potential factors of healthy dietary pattern probably are because of higher intake of alkaline foods such as vegetables and fruits, more dairy products which contain higher calcium and the PUFA content of oily fishes (Heaney et al 1996, Hardcastle et al. 2011).
The effect of Japanese dietary pattern on BMD has been assessed in two studies. In Japanese premenopausal farmwomen, healthy dietary pattern, which was defined by high intake of dark green and yellow vegetables, fruit, processed fish, mushrooms, fish and shellfish, was positively correlated with BMD (Monma et al. 2010). In the other study population, no significant relationship between traditional Japanese diet and fall related fractures was found (Monma et al. 2010). The traditional Japanese diet is characterized by high amount of rice and soy bean products, seafood, fish and vegetables and low intake of red meat and dairy products (Monma et al. 2010).
Western pattern characterized with high intake of fat and oil, meat and processed meat was in non-significant negative relationship with BMD of premenopausal Japanese farmwomen (p=0.08) (Okubo et al. 2006). Western dietary patterns with high consumption of fats, processed meats and energy dense foods are suggested to increase C-reactive protein and interleukin-6 (McNaughton et al. 2011). Inflammatory factors are shown to have negative effects on BMD and positive relationship with bone loss (McNaughton et al. 2011).
Traditional English diet of 20th century including high intake of fried fish, fried potato, legumes (baked beans), red and processed meats, savory pies and cruciferous vegetables (cauliflower and cabbage), had negative relationship with BMD in the Twins UK study (Fairweather-Tait et al. 2011). In the Framingham study higher intake of fruit and
vegetable was related to higher BMD (Tucker 2002), whereas in the aforementioned Twins UK study, no such association was found (Fairweather-Tait et al. 2011). Probably, the difference is hypothesized to originate from the energy intake from fruit and vegetable in
these two studies, so that in the Framingham study 29.7% of energy was derived from fruit and vegetable, but in the Twins study the percentage was 17.1%.
Diets withlow fat content, increased fruit and vegetable and grain intake had some positive aspects such as decrease in bone resorption and remodeling. The 8 year follow up of 215 women aged 50-79 year represented that dietary intervention of increased fruit and vegetables, low fat content, and grains reduced the risk of multiple falls (p< 0.01) and hip fracture (p=0.21) in the subjects (McTiernan et al. 2009). Moreover, DASH diet (Dietary approaches to Stop Hypertension) is defined as diet low in salt, high in vegetables and fruits, and low fat dairy products, low in cholesterol, saturated fatty acid, and total fat, and high in potassium, magnesium, dietary fiber, calcium, and protein. DASH diet was
reported to decrease serum osteocalcin by 8-11 % and carboxy-terminal telopeptide of type I collagen (CTX) by 16-18% (Bullo et al. 2009) and decrease calcium excretion
(Kontogianni et al. 2009).
2.4.1 Mediterranean diet
Although many studies using healthy eating indices have been performed, some particular dietary habits are studied more than other indices. For instance, Mediterranean diet has been the area of various research projects (Kontogianni et al. 2009).
Mediterranean diet is the major food pattern of Mediterranean countries and includes high content of MUFA, fiber, calcium, magnesium, phosphate, folate, and β-carotene; and is characterized with higher intake of vegetables, legumes, fruits and nuts, fish and seafood, non-refined cereals, olive oil, low intake of red meat and full fat dairy products; and moderate intake of poultry and alcoholic beverages (McNaughton et al 2011, Kontogianni et al. 2009, Trichopoulou 2009).According to Kontogianni and coworkers (2009), there has not been a significant association between Mediterranean diet and BMC in 220 Greek women of 48±12 year old (p=0.048). However, a diet with low intake of red meat, high
intake of fish and olive oil was found to be related to high BMD and total body BMC (Kontogianni et al. 2009).
The effect of three interventional dietary patterns during 12 months in 238 elderly men and women aged 60-80 years were studied at University of Rovira, Spain (Bullo et al. 2009).
The three dietary patterns included Mediterranean diet supplemented with virgin olive oil, Mediterranean diet supplemented with mixed nuts, and low fat diet (as the control group).
Supplementation with mixed nuts increased dietary potential renal acid load, net
endogenous acid production (NEAP) and also PTH levels; however, it did not have any significant effect on bone metabolism. Traditional Mediterranean diet with virgin olive oil was found to be positively associated with BMD but intake of nuts may have detrimental effect, and higher intake of fruit led to increase in BMC (Weiss et al. 2005). In general, the studies on the influence of Mediterranean diet are controversial, and more studies are required for stronger conclusions.
2.4.2 Baltic Sea Diet
Baltic Sea (Nordic) dietary pattern is the common eating pattern of Nordic countries and mainly characterized by foods such as boiled potatoes, berries, whole grain wheat and rye bread, breakfast cereals, and fermented milk as healthy food groups (Åkesson et al. 2013).
The main criterion for an original Baltic Sea diet includes growth in Nordic nature, frequency of daily intake, and healthy effects (Olsen et al. 2011).
Explaining about Baltic Sea dietary pattern, two concepts of modern and traditional may be defined. Modern Baltic Sea diet is characterized by various unhealthy features such as high intake of sugar, margarine, high fat dairy products, red meat and low intake of fruits and vegetables (Slimani et al. 2002). Whereas, traditional Baltic Sea diet is specialized for Nordic ecology and contains healthy food compounds such as fish and shellfish provided by the coastline, cabbages, whole grains, rye bread, oatmeal, apples, pears, and berries, and
root vegetables (Olsen et al. 2011). The Baltic Sea diet is similar with healthy diets such as DASH and the Mediterranean diets (Olsen et al. 2011).
The healthy aspects of Baltic Sea diet are meeting the requirements for a healthy diet according to the criterion for European, American and international dietary guidelines (Olsen et al. 2011). Healthy Nordic food index significantly decreases the risk of mortality in both female and male. The effect of rye bread has been more prominent than other Nordic dietary elements in men (Olsen et al. 2011). However, no study was done on the association of Baltic Sea dietary pattern with bone health.
In January 2011, the SYSDIET study, initiated the Baltic Sea food pyramid. The aim of this project was to recommend healthier options for eating patterns in Nordic countries (Denmark, Finland, Iceland, Norway and Sweden) (Kanerva et al. 2012). The diet is rich in local foods of Nordic countries such as apples and berries, roots and cabbages, rye, oats and barley, low fat dairy products, rapeseed oil, and fish (salmon and Baltic herring); and low in red and processed meat and alcohol (Kanerva et al. 2012). Baltic Sea food pyramid includes vegetables and fruits in the basement. The main vegetable and fruit category consists of cabbages, mushrooms, root vegetables, berries, and apples. The main sources of carbohydrate are rye breads, whole cereals, and potatoes in the upper section of pyramid.
Dairy products and fish are placed in the third section accompanied by sources of proteins including chicken, meatballs, eggs, cheese, etc. The top of the pyramid is ascribed to sweets, chocolates, sausages and other ready foods. Accordingly, a diet following
mentioned nutritional recommendations sustains normal waist circumference and body fat percentage (Kanerva et al. 2013), (Ruokakolmio, lautasmali 2013).
2.4.3 Energy and nutrient density of diet
Energy-dense diet is defined as a diet with high concentration of energy in every
measurement unit of included foods. In other words, energy dense dietary pattern is a diet consisting of foods with high concentration of energy such as soft drinks, potato chips,
French fries, processed meats (hamburger, hot dog, lunch meat, smoked meat, sausage, and bacon) and sweet desserts such as chocolate, doughnuts, and ice cream (Langsetmo et al.
2010). Energy dense diet is more popular among the youth than elderly (Tucker et al.
2002). High rate of candies and sweets had detrimental effect on proximal right femur BMD of 907 adults aged 69-93 year old in Framingham Study. Comparing BMD of
subjects with this dietary pattern with other patterns containing bigger proportions of meat, dairy, and bread; meat and sweets; sweet baked products; alcohol; vegetables, fruits, and cereals; the lowest BMD was allocated to candy group which had the smallest nutrient density (Tucker et al. 2002). However, nutrient dense diet (prudent or healthy diet) is characterized by high intake of vegetable, fruit, and whole grain, and has been studied to have positive effect on BMD in postmenopausal, premenopausal, older and younger men and women and children (Tucker et al. 2002).
2.5 Particular food groups and nutritional components and bone health
2.5.1 Fruit and vegetable intake
The effect of vegetable and fruit on bone health has been studied in various projects. The majority of studies represented the positive effect of vegetables (Monma et al. 2010).
Vegetables with light green leaves such as lettuce, and cabbages decreased the risk of fall related fractures in 1178 elderly people in Japan in a 4 year prospective study (Monma et al. 2010). Moreover, higher intake of fruit (>250 g of fruit per week) was significantly associated with higher total BMD in 1324 men and 1479 women (Zalloua et al. 2007).
Fixed anions (chloride) lead to higher obligatory urinary calcium loss compared to metabolized anions such as acetate or bicarbonate. This can be an explanation for higher anti-fractural effect of vegetables and fruits (Bendich 2005).
Diets high in fruit and vegetable contain more vitamin C, magnesium, potassium and vitamin K. The positive effect of fruit, vegetable, cereal group is more obvious and visible in men than in women (Tucker et al. 2002). According to Framingham study, high intake of fruit, vegetable, milk and cereals is effective on bone health in comparison with people who have a diet high in soda, pizza, salty snacks, meat, bread and potatoes (Kontogianni et
al. 2009). The same study concluded that high intake of vegetable increases the hazard ratio for cumulative fall related fractures in 291 postmenopausal farmwomen (p=0.048) (Okubo et al. 2006).
Intervention with low fat diet and increased fruit and vegetable and grain intake (with the target daily intake ≥5 servings of fruit and vegetables, ≤20% energy from fat, ≥ 6 servings of grains ) was related to decreased multiple falls and hip BMD, however, no significant relationship was shown for risk of osteoporotic fractures. This diet decreased plasma estrogen (McTiernan et al. 2009, Bullo et al. 2009). Most of the studies represent positive effect of fruits and vegetables on bone health. The positive effects of vegetables and fruits on bone health are supposed to be related to antioxidant and anti-inflammatory compounds such as vitamin C and β-carotene, and negative potential for renal load, alkaline nature of most of fruits and vegetables or other mechanisms. However, more studies are still required to prove the theory.
2.5.2 Phytoestrogens
Phytoestrogens, also called isoflavones, are botanical compounds with the structure like estradiol and estrogen-like biological effects. Isoflavones bind to estrogen receptors in animals and human beings (Clarkson, 2011). The main sources of isoflavones are soybeans and soy products, red clover, kudzu, and the American groundnut; among which, soy is the most common food used by humans (Clarkson, 2011). The main soy isoflavones are genistein, diadzein (with equal rate in whole soybean), and glycitein (in less amount in whole soybean) (Clarkson, 2011). The soy germ from which some soy supplements are produced, consists higher rates of isoflavones including diadzein about four times greater than genistein, and high content of glycitein (Clarkson, 2011).The studies in animal models have proven the effect of estrogens on BMD; however, more studies on human models are required(Prentice 2004). Accordingly, it is hypothesized that phytoestrogens such as soy- derived isoflavons, genistein, daidzein and lignans derived from cereals are effective on bonehealth(Prentice 2004).Ipriflavone is a synthetic isoflavone from soy products used in dietary supplements. Various studies have shown ipriflavone to be favorably affecting
bone metabolism which can increase BMD(Seaman 2004). The standard recommendation for ipriflavone is 600 mg/day (Seaman 2004). According to the review of 60 clinical studies in Italy, Japan, and Hungary, the long-term intake of ipriflavon is safe and may increase bone density and prevent fracture in elderly osteoporotic patients; however, another study has suggested that ipriflavon leads to intense lymphocytopenia(Seaman 2004). This intervention study lasted for 4 years with 234 women in the ipriflavone intervention group and 240 subjects in placebo group. In 29 women, a reduction in lymphocyte concentration (from 33% -27%) was observed. However, this reduction was still in the normal range of 12%-50%(Seaman 2004).
2.5.3 Fishand PUFA
High sea food consumption is related to higher hip and total BMD in premenopausal and post-menopausal women (p<0.001) who consumed more than 250 g sea food per week (Zallaoua et al. 2007). Sea food derived PUFA’s can increase the fluidity of intestine’s membrane and increase the absorption of calcium by enhancing the action of vitamin D3 (Zallaoua et al. 2007). In the study of OSTPRE-FPS in Kuopio the results represented positive effects of PUFA on lumbar spine and total BMD in 554 elderly women (Jarvinen et al. 2012). Moreover, the ratio of dietary linoleic acid to alpha-linolenic acid represented significant inverse association with hip BMD in 642 men and 564 women not using hormone therapy, and 326 women using hormone therapy since 1988 to 1992 (Weiss et al.
2005). According to a review by Lau and coworkers (2013), essential and long chain n-3 PUFA’s have probably positive effects on bone health.
2.5.4 Alcohol consumption
There have been various controversial results about the effect of alcohol on BMD and fracture risk. In the cross sectional study by New and colleagues (2000), the higher alcohol intake was related to higher BMD in 62 women aged 45-55 year old, and even after energy adjustment alcohol consumption was positively related to forearm trabecular BMD.
Although there is some idea that moderate alcohol is effective on postmenopausal period,
the mechanisms are unclear. The positive effect of alcohol is hypothesized to be because of production of androstenedione in adrenal glands and the gonads; and increase in adrenal androgen or estrogen concentrations (New et al. 2000, Tucker et al. 2002). Furthermore, in 1232 postmenopausal twin pairs from the UK co-twin control study, use of 1 unit of wine per day, but not beer or spirits was related to increase in spine BMD (Fairweather-Tait et al. 2011). High amounts of alcohol have been shown to be harmful, whereas, moderate servings are suggested to have some protective effect in osteoporosis (Prentice 2004).
Alcohol consumption has a weak inverse association with bone fracture risk in
postmenopausal women (Baron et al. 2001). In the study of Whittle and coworkers (2012) the social diet which is loaded highly for alcohol was associated with higher BMD;
however, this relationship lost its significance after adjustment for energy intake.
Conclusively, moderate intake of alcohol (particularly wine) may be useful for bone maintenance; however, to clarify either positive or negative effect of alcohol on bone, more studies are required (Tucker et al. 2002), and various alcoholic drinks may have different effects on bone health.
2.5.5 Alkali foods
Serum bone specific alkaline phosphatase is one of the markers of bone formation. There is a theory by Wachman and Berstein (1968) that bone consists of labile bases that can be mobilized to balance blood acidity. Alkaline foods are defined as foods increasing alkaline ash in the body and are high in magnesium and potassium. Main alkali food groups are vegetables and fruits. Potassium intake can balance acid-base homeostasis by diminishing urinary calcium excretion (New et al. 2000). Although white meat and cheese are included in healthy dietary pattern, they can have a negative effect on bone health by their potential renal acid load. In healthy pattern high intake of vegetables and fruits compensates this effect by buffering high acid load of diet (Heaney 1996, Aucott et al. 2011). Potassium and magnesium have positive effect on BMD because they increase the alkalinity of blood and reduce renal load as well as acid and endogenous acid production. The positive effect of fruit and vegetables is also related to this mechanism (Okubo et al. 2006). Western diet increases acidity and urinary calcium and C-teleopeptide excretion, whereas diet high in
vegetable and fruit or diets supplemented with potassium reduce bone resorption (McNaughton et al. 2011, Kontogianni et al. 2009).
2.6 Nutrient intakes and bone health
2.6.1 Calcium and vitamin D
Calcium intake can be considered as a barometer for bone health (Seaman 2004). Calcium requirement is measured according to skeletal maturity and turnover rate mixed with calcium absorption and excretion. However, the optimal recommendation for calcium is still debated (Bendich 2005). Many cross sectional and retrospective studies have represented the positive correlation of calcium intake with BMC and BMD (Alissa et al.
2011, Quesada-Gomez et al. 2013). Calcium is a threshold element; In lower than a particular intake, the availability of calcium for bone tissues is decreased, and amounts higher than the threshold have no positive effect on bone function and mass (Bendich 2005). Moreover, the relationship between calcium retention and the dietary intake of calcium is different during various life stages. The retention of calcium is positive during growth, and is zero in maturity. However, elderly people have negative retention, which increases 0.3 to 5 years after menopause, and eventually, for the next 10 to 15 years it becomes increasingly negative (Bendich 2005). Several factors interact with calcium absorption. Caffeine and fiber have minor effects on decreasing the absorption in comparison with protein and sodium which impact the urinary excretion of calcium and phosphorus have minor to nonexistent effects (Alissa et al. 2011). The type of fiber is significant; so that green, leafy vegetables, have no effects and fibers in wheat bran reduce the absorption of co-ingested calcium (Bendich 2005). Phytate and oxalate have no effect on co-ingested calcium absorption (Bendich 2005). Bioavailability of calcium in fruits and vegetables is half of that in dairy products (Bendich 2005). To illustrate the effect of different types of fibers, the bioavailability of calcium in beans is 50%, whereas spinach and rhubarbs have 0% bioavailability (Bendich 2005). Foods containing more than 100 mg calcium per serving, include dairy products (except cottage cheese), greens of mustard family (collards, kale, mustard), calcium-enriched tofu, sardines, a few nuts (hazelnut and almonds). Meats and fishes are all low in calcium except shellfish eaten with bones (Bendich 2005).
There is no study representing the effect of only calcium supplementation even in high amounts such as 3100 mg/day in menopausal loss. Moreover, calcium intake and
absorption efficacy decreases with increasing age. For those over 50 years old, the average recommendation for calcium is 800-1000 mg per day (Ross et.al 2011, Bendich 2005);
whereas, the median daily intake is 500-600 mg in the USA (Bendich 2005). Reviewing various studies, it is rational to mention that in countries with average calcium intake close to the recommended intakes, no significant relationships are observed between calcium supplementation and BMD, so that increasing the intake of calcium does not lead to lower risk of hip fracture. However, in countries with low average calcium intake, higher intake of calcium decreases the rate of hip fractures. Daily 1000 mg calcium intake reduces the risk of fracture to 24% (Prentice 2004). Generally, calcium supplementation is
recommended only among elderly (over 50-60 years) with a low calcium intake (lower than 400 mg per day), people who live with Westernized life style, and in patients with osteoporosis, whereas for children and young adults there is inadequate evidence to support supplementation (Prentice 2004).
Moreover, calcium supplementation has been considered non-effective on BMD of trabecular region which loses the highest rate of calcium during five years of menopause (Ross et al. 2011). However, in older women calcium supplementation has shown to increase BMD by about 1-3 % (Ross et al. 2011). Most studies have assessed calcium supplementation by milk intake, whereas the effect of milk differs from calcium as an individual mineral element. The new recommended intake of calcium is 1100 mg for teenagers (14-18 year old) and 1000 mg for 14-18 years old pregnant or lactating, and 800- 1000 mg for elderly in the US (Ross et al. 2011).
Studies about the effect of calcium and vitamin D supplementation represent considerable decrease in the incidence of non-vertebral fracture. The hypothesis that geographical differences affect BMD is not true; since low bone mineral status in old age appear to be universal phenomena and the studies suggest that there are no basic differences in calcium biology in developed and developing countries (Prentice 2004). The general conclusion of
all studies has estimated the recommended intake of at least 1000 mg calcium for elderly.
Clinical trial studies have shown that the effect of calcium and vitamin D supplementation has stronger effect on the risk of fracture than on bone mass (Bendich 2005, Ross et al.
2011).
Vitamin D is one of the vital components for bone health, deficiency of which leads to rickets in childhood and osteomalasia in adulthood. It can be obtained from dietary sources such as milk, fish and vitamin D enriched margarines or produced by the action of sunlight and sterols in skin (7-dehydrocholesterol). Dietary vitamin D intake is positively correlated with higher intake of cholesterol, all fatty acids, total fat, and fiber in elderly
postmenopausal women in Saudi Arabia (Alissa et al. 2011). Adequate intake is dependent even in the youth on dietary sources of vitamin D. In elderly, the endogenous production of vitamin D declines and old people get more dependent on dietary intake of vitamin D. In this period of life concentration of plasma 25- hydroxyl vitamin D is a proxy for vitamin D status of bone. The general recommendation of vitamin D is 15μg/d for elderly people (51- 70 year old) (Ross et.al 2011). Although it is hypothesized that hyperparathyroidism enhances the requirement for vitamin D during elderly, supporting data for this hypothesis is not sufficient yet (Alissa et al. 2011, Prentice 2004).
2.6.2 Protein
Protein plays a significant role in establishment of bone matrix as collagens (Kontogianni et al. 2009). Bone calcium resorption, quicker hip fracture recovery, and blunting elderly related bone loss, are some roles of protein on bone structure that offset the corrupting effect of bone calcium resorption in the case of urinary calcium loss (Bendich 2005). A recent study suggested that either vegetable or animal protein is in relation with higher BMD (Darling et al. 2009). Since dietary protein increases urinary calcium loss, more than 95 g protein per day in comparison to 68 g protein per day was related to increased risk of forearm fracture in women (Feskanich et al. 1996). Similar results were found for animal proteins, whereas for vegetable proteins there was no such association (Feskanich et al.
1996). In contrary, Dawson-Hughes and coworkers (2002) found animal proteins as
augmenter on bone calcium (Dawson-Hughes et al. 2002, Darling et al. 2009). In two randomized trials, protein was found to improve bone metabolism after hip fracture (Seaman 2004). The explanation may be the increase of the insulin-like growth factor which is an osteotrophic substance (Seaman 2004). Low intake of protein in elderly has been associated with the increasing risk of fracture (Prentice 2004), energy and protein intakes are positively related to higher lumbar BMD in northern Indians (Kumar et al.
2010).
The issue of animal protein and its effect on bone health and calcium resorption can be explained by two mechanisms. Animal proteins and other high sulfur sources elevate the urinary calcium excretion. In other words, sulfur containing amino acids increase blood acidity. On the other hand, high urinary calcium excretion due to animal proteins is compensated by meat phosphorus. Accordingly, the studies around animal protein are entirely controversial, because in some studies high intake of meat and dairy products have been associated with higher hip fracture risk; whereas, high intake of meat at young age decreases the risk of forearm fracture in postmenopausal women (Xu et al. 2006, Kaptoge et al. 2003), and low intakes of meat have been associated with increased hazard ratio for fall related fractures (Monma et al. 2010). In the Framingham study, low intake of animal and total protein was related to higher rate of bone loss; whereas, in the cohort study of Osteoporotic Fracture (SOF), higher ratio of animal to vegetable protein and not total protein intake was associated with hip bone fracture, and both vegetable and animal proteins seemed to be effective in protecting BMD loss (Hannan et al. 2000, Kaptoge et al.
2003). Legumes and grains are main sources of protein in vegan diet and contain same amount of sulfur as meat; whereas, meat has a better ratio of n-6/n-3 fatty acids than grains and do not contain gliadine or phytate (Kaptoge et al. 2003).
3. Aims and Hypothesis
This study aims to assess the relationship between Baltic Sea diet score and BMD in elderly women who live in Kuopio, Finland during three years of follow-up. Our hypothesis was that the Baltic Sea diet with berries, cabbages, apples, pears, root vegetables, oats, rye and fish), can positively affect BMD.
4. Subjects and methods
Current study is using the data from “Kuopio Osteoporosis Risk Factor and Fracture Prevention study (OSTPRE-FPS)” in Kuopio, Finland during 2003-2007. Among 13100 peri- and postmenopausal women born in 1932-1941, totally 5407 women were sent an invitation mail to participate the study. The including criteria for participants was minimum age of 65 at the end of November 2002, residency in Kuopio province in the beginning of study, and not being included in the OSTPRE study sample in which BMD measurements were conducted. The percentage of population who responded to the mails whether they are interested to participate in the intervention was 63.5%. These 3432 women volunteered to take part the OSTPRE-FPS study. A sub-sample of 750 women was randomly selected for this project. In the baseline; 606 women were included in baseline measurement which took place from February 2003 to May 2004, and follow up
measurements were carried out between January 2006 and May 2007. From 593 women who completed the study, 544 women had BMD measurement in femoral neck and 480 of them had lumbar spine. Explain the intervention groups. The study was approved in October 2001 by the ethical committee of Kuopio University Hospital. The study was registered in Clinicaltrials.gov by the identification NCT00592917. The subjects were involved voluntarily.
4.1 Questionnaire
Questionnaires were mailed to the women’s homes and returned at study visit.
Questionnaire assessed alcohol consumption, smoking, use of dietary supplements, diseases, time of menopause, use of medicines such as hormone therapies, and current mobility. Mobility was assessed with a question about current ability to move, with the options of fully mobile; able to move, but not running; not able to walk more than 1 km;
not able to walk more than 100 m; able to move only indoors; and immobile. The proxies for restricted mobility were: not able to walk more than 100 m, only able to move indoors, and immobile. In the assessment of alcohol use, subjects were asked to quantify their intake of beer/cider bottles (1 bottle equals 330 ml), wine glasses (one glass consists of 120 ml), spirits or strong alcohol portions (40 ml for each shot) during the last 4 weeks.
Diseases which may affect BMD were included in the questionnaire. These diseases included hyperthyroidism, disease of parathyroid gland, chronic liver disease, chronic intestinal disease, celiac disease, ventricle operation, chronic nephropathy arthritis, osteoporosis, and lactose intolerance. Furthermore, medications influencing BMD were included and they included loop-diuretics, insulin, antiepileptic, glucocorticoids, and cancer chemotherapy.
4.2 Dietary assessment
At the baseline, among all of the participants who had BMD measurement data, 554 had valid dietary intake data. At the baseline, the subjects filled in 3-day food record at home and returned it during the research visit. The instructions how to fill the forms were sent beforehand to the subjects and they were asked to record their food intake for three
consecutive days, two of which were week days and one weekend day. Type of fat used by subjects was asked separately for type of fat on bread, cooking and baking. In the case of ambiguity, subjects were contacted via phone by a nutritionist. Calculation of food and nutrient intake was done with the software Nutrica program (version 2.5, Finnish Social Insurance Institute, Turku, Finland).
To create a Baltic Sea diet score, the intakes of whole grain cereals, vegetable, fruit, processed meat, fish, low fat dairy products, total fat intake, PUFA/SFA ratio from food record data, and alcohol consumption from the questionnaire were categorized into quartiles. The method has been published earlier by Kanerva et al. 2012. For fish, fruit, vegetable, low fat dairy, whole grain cereals, PUFA/SFA ratio and total fat intake the highest score was allocated to the highest rate of consumption, whereas, for processed meat the lowest intake got the highest dietary score. Moreover, for alcohol consumption, 1 portion per week or less got the highest score (score 3) and the higher portion amount got the 0 score (the lowest score). The quartile categories were summed up. The total score was categorized into quartiles and defined as Baltic Sea diet Score Quartile (BSQ).
4.3 Bone mineral density measurement
Measurement of BMD was performed by dual energy X-ray absorptiometry (DXA; Lunar DPX, Madison, WI, USA) for the lumbar spine (L2-L4), femoral neck and total body.
Measurements were done at the baseline and after 3 years. The quality and technical monitoring was done every day. Measurement errors were excluded from the analysis.
4.4 Anthropometric measurements
Weight and height measurement was done at the baseline using a calibrated scale (Philips, type HF 351700) and wall meter by a trained person. BMI (kg/m2) calculation was done by dividing the weight (kilograms) by the square of height (in meter).
4.5 Statistical analysis
The analysis of the data was done using the IBM SPSS statistics 19. Descriptive analysis of baseline characteristics and food and nutrient intake was done by one-way Anova and respective non-parametric test (Kruskall-Wallis and Pearson’s Chi-Square test). The BMD
and BSQ relationship were analyzed by mixed linear model. Among the BMD indicators, lumbar BMD, femoral BMD, and total body BMD at the baseline and at year 3 were included in the mixed model. Covariates in the mixed model included energy intake, HT, calcium and vitamin D supplementation, diseases affecting BMD, restricted mobility, time from menopause, life time smoking, age, BMI, and intervention group. We studied the results of two models including the full model and final model. In the full model we included all of the potential confounders. Studying the p-value for each covariate, we step- wisely omitted those which were not significantly confounding the dependent variable (lumbar, femoral, and total body BMD). The final model included restricted mobility, calcium supplementation, smoking, BMI, and HT as covariates.
5. Results
The elderly postmenopausal women’s mean age (±SD) was 67.9±1.87, and mean BMI was 28.8±4.71 kg/m2. Table 1 represents baseline characteristics of the subjects across the quartiles of BSQ. The mean BMI was in overweight range (BMI≥25g/m2) in all quartiles.
At the baseline 34% of subjects (198 subjects) had used HT on average for 11 years, and they had passed menopause about 18 years before. The mean value for lumbar BMD was 1.09g/cm2 and the mean femoral BMD among the four quartiles was 0.87g/cm2. Average total body BMD was 1.07 g/cm2. Baseline measurement of BMD was not significantly different among BSQ groups (Table1). In each quartile at least one fifth of the subjects were calcium supplement users and almost the same was observed about the percentage of people who had been taking vitamin D supplements. Furthermore, the highest proportion of vitamin D supplementation was 27% in the third quartile; however, it did not differ from the other groups. Almost 90% of the subjects had normal mobility. The percentage of women who had never smoked was about 80% in all quartiles. Diseases affecting BMD were significantly different among the BSQ groups, so that the highest percentage (45%) of the participants who had disease was in the lowest BSQ, and the largest percentage of the subjects with no disease (70%) were in the fourth BSQ.
Table 1 Baseline Characteristics*
BSQ1 (n=161) BSQ2 (n=125) BSQ3 (n=133) BSQ4 (n=135) p-value**
Age, year 67.7±1.8 68.0±1.8 67.6±1.8 68.0±1.8 0.426 Height, cm 158.1±5.2 158.3 ±5.5 159.0 ±5.2 159.1±5.0 0.300 Weight, kg 73.5 ±13.2 72.7±12.2 71.7±9.8 71.8±12.7 0.782 BMI(kg/m2) 29.4±5.2 29.01±4.6 28.3±3.9 28.3±4.8 0.675 Length of HT use
(y), N=198
10.6±7.5 11.1±6.9 11.385±7.4 10.7±6.9 0.224
Time from menopause, y
18.79±5.47 18.66±4.69 17.86±4.97 18.47±5.04 0.001
Lumbar BMD (g/cm2)
1.073±0.183 1.087±0.193 1.112±0.188 1.115±0.178 0.472
Femoral BMD(g/cm2)
0.855±0.01 0.867±0.12 0.888±0.011 0.867±0.011 0.454
Total body BMD(g/cm2)
1.070±0.097 1.077±0.094 1.086±0.086 1.072±0.095 0.618
Current mobility
Normal% 90% 93% 95% 91% 0.412
Limited% 10% 7% 5% 9%
Disease affecting bone health (N %)
No 55% 68% 61% 70% 0.04
Yes 45% 32% 39% 30%
Vitamin D
supplementation (N
%) No 81% 75% 73% 75% 0.338
Yes 19% 25% 27% 25%
Ca supplement use, yes %
21% 24% 30% 29% 0.286
Smoking
Never 83% 83% 84% 80% 0.093
Previous smoker 8% 15% 9% 15%
Current smoker 9% 2% 7% 5%
*Mean± SD, **Test used: one way ANOVA (for height, lumbar BMD, total BMD) and Kruskall-Wallis;(for weight, femoral BMD, age, time from menopause, BMI) and Pearson’s Chi-Square (for calcium
supplementation, vitamin D supplementation, smoking, current mobility, length of HT, disease affecting bone health). HT: hormone therapy; BMD: bone mineral density, BSQ: Baltic Sea score quartile, N: number.
Table 2 describes the average intake of main food groups according to the BSQ groups.
The highest amount of fruits and berries, vegetables, fish, and root vegetables was used in the highest BSQ; whereas, subjects in the second BSQ had the highest consumption of whole grain consumption and low fat dairy products. The difference of root vegetables and alcohol consumption was not significant among the four BSQs.
Table 2 Food consumption at baseline Food intake(g/d) BSQ 1
(n=161)
BSQ 2 (n=125)
BSQ 3 (n=133)
BSQ 4 (n=135)
p-value*
Fruit and berries 105±120 112±116 158±146 195±164 <0.001
Vegetables 52±23 107±13 160±17 253±54 <0.001
Root vegetable 12±15 26±23 32±27 52±46 0.598
Whole grain 112±61 117±55 113±50 114±48 <0.001
Processed meat 17±24 12±19 18±28 15±26 <0.001
Low fat dairy products
820±645 830±751 782±641 826±638 <0.001
Alcohol (portion per week)
0.95±0.12 0.72±0.1 0.80±0.13 0.82±0.12 0.447
Fish 33±37 43±45 43±42 45±45 <0.001
* Test used: respective Non-parametric test (Kruskall-Wallis), BSQ: Baltic Sea score quartile.
Table 3 summarizes the nutrient intake across the 4 quartiles. PUFA/SFA ratio, intake of fiber, potassium, magnesium, and phosphorus were higher in the fourth BSQ. The highest scores of Baltic Sea diet were associated with highest levels of energy (kcal/d), total fat (g/d), protein (gram and percentage of energy intake from protein), and carbohydrate (g/d);
and the lowest energy percentage from carbohydrate. The percentage of energy intake from total fat was the highest in the second BSQ. The highest quartile of BSQ contains the greatest intake of calcium (mg/d), vitamin C (mg/d), fiber (g/d), sodium (mg/d); however, vitamin D (µg/d) was the highest in the second quartile of BSQ.
Table 3 Energy and nutrient intakes at baseline in quartiles of Baltic Sea Diet Score Quartiles of Baltic Diet Score
n=554
Nutrient BSQ =1
(n=161)
BSQ =2 (n=125)
BSQ =3 (n=133)
BSQ =4 (n=135)
p-value*
Energy (kcal/d) 1466±364 1599±386 1569±329 1639±389 <0.001 Total fat(%energy
intake)
30.7±5.5 31.4±4.8 30.9±5.4 31.0±6.4 0.076
Fat(g/d) 49.4±15.5 55.9±18.4 54.0±17.3 57.2±20.4 <0.001 PUFA/SFA ratio 0.40±0.16 0.43±0.16 0.46±0.15 0.49±0.23 <0.001 Protein(g/d) 61.0±17.1 69.2±18.2 69.0±16.1 73.7±18.4 <0.001 Protein(% En) 16.7±0.2 17.9±0.3 17.9±0.2 18.1±0.2 0.454 Carbohydrate(g/d) 186.5±51.5 196.0±48.4 194.6±42.2 197.4±51.0 <0.001 Carbohydrate(%
energy)
50.7±0.4 49.09±0.4 48.9±0.4 46.9±0.5 <0.001
Ca(mg/d) 930±363 1008±373 999±355 1104±366 <0.001
Vitamin D(µg/d) 6.89±4.13 8.11±5.02 7.61±4.50 7.94±5.64 <0.001 Vitamin C(mg/d) 72.4±49.5 89.1±52.2 103.2±56.0 132.5±66.7 0.058 Fiber(g/d) 19.59±6.82 21.53±6.30 22.80±5.78 24.98±6.67 <0.001 Potassium(mg/d) 2871±709 3220±798 3312±652 3794±810 <0.001 Na(mg/d) 2532±703 2877±816 2853±673 3001±726 <0.001 Mg(mg/d)
Phosphorus(mg/d)
300±79 1325±385
330±79 1424±385
337±66 1447±348
360±75 1536±376
<0.001
<0.001 PUFA: polyunsaturated fatty acid, SFA: saturated fatty acid, Mg: magnesium, Na: Sodium, Ca: calcium, BSQ: Baltic Sea score quartile. * Test used: one way ANOVA(for Mg, potassium, energy, fiber) and Non- parametric ; Kruskall-Wallis (for protein, Total fat(%energy intake), Fat(g/d), PUFA/SFA ratio, Protein(g/d), Carbohydrate(g/d), Carbohydrate(% energy), calcium (mg/d), vitamin D, Vitamin C, sodium, phosphorus).
Results of the mixed model analysis are shown in table 4 where we analyzed the association of Baltic Sea diet score with three bone sites in lumbar, femoral and total BMD. For the assessment of the results and association between BMD and BSQ, we made three models. In model 1 the analysis was done without any adjustment. In this model BSQ was in non-significant relationship with lumbar, femoral, and total BMD. Model 2 assessed the relationship between BMD and BSQ with full adjustment. In this model the adjustment included all potential covariates including life time smoking, intervention group, vitamin D and calcium supplementation, disease or medication reducing BMD, current mobility, age, and time from menopause, BMI, HT, and energy intake. In the fully adjusted models we did not found significant association between BSQ and femoral, lumbar, and total BMD.
Final model is adjusted for selected covariates. After making the fully adjusted model, we excluded those covariates with p-value bigger than 0.05, and we adjusted the final model only for life time smoking, intervention group, vitamin D supplementation, disease or medication reducing BMD, and HT. In the final model (minimally adjusted model),
femoral BMD was significantly in positive relation with BSQ (p=0.027), and lumbar BMD was positively related to BSQ with p=0.046. However, the association of total BMD and BSQ was not significant (p=0.089). In the three models, the increasing trend of BMD was not linear with the increase of BSQ. BMD increased from 1st quartile group to the third, but decreased in the fourth BSQ group.
Table 4 Bone mineral density across the quartiles of Baltic diet score Quartiles of Baltic diet score
Femoral BMD, Mean (95% CI) P-value
(1)
BSQ=1
0.847(0.806,0.887)
BSQ =2
0.849(0.806,0,891)
BSQ=3
0.880(0.841,0.919
BSQ=4
0.832(0.792,0.872) 0.060 (2) 0.852(0.832,0.871) 0.861(0.839,0.883) 0.882(0.860,0,903) 0.856(0.835,0.877) 0.195 (3) 0.849(0.810,0.888) 0.849(0.808,0.891) 0.885 (0.846,0.923 0.833(0.794,0.872) 0.027 Lumbar BMD, Mean (95% CI)
(1) 1.086(1.049,1.111) 1.107(1.058,1,130) 1.141(1.088,1.157) 1.129(1.088,1.157) 0.203 (2) 1.080(1.049,1.111) 1.094(1.058,1,130) 1.123(1.088,1.157) 1.123(1.088,1.157) 0.183 (3) 1.060(1.00,1.110) 1.090(1.040,1.150) 1.130(1.070,1.180) 1.110(1.050,1.160) 0.046 Total BMD, Mean(95%CI)
(1) 1.068(1.00,1.110) 1.085(1.040,1.150) 1.096(1.070,1.180) 1.067(1.050,1.160) 0.139 (2) 1.076(1.060,1.092) 1.090(1.071,1.108) 1.093(1.075,1.110) 1.076(1.059,1.094) 0.392 (3) 1.038(1.010,1.066) 1.056(1.026,1.086) 1.067(1.038,1.096) 1.041(1.011,1.070) 0.089 (1) The full model: It has all potential covariates without any adjustment. No covariate was excluded
from the model.
(2) Fully adjusted: The model is adjusted for all potential covariates including life time smoking, intervention group, vitamin D and Calcium supplementation disease or medication reducing BMD, current mobility, age, and time from menopause, BMI, hormone replacement therapy, energy intake.
(3) Minimally adjusted (Final model): The model is adjusted for life time smoking, intervention group, vitamin D supplementation disease or medication reducing BMD, and hormone replacement therapy.
BMD: Bone Mineral Density, BMI: Body Mass Index, BSQ: Baltic Sea score quartile
