European Journal of Clinical Nutrition (2001) 55, 97±106
ß 2001 Nature Publishing Group All rights reserved 0954±3007/01 $15.00
www.nature.com/ejcn
The realization of a project aimed at reducing the plasmatic lipid
level in a large Italian population improves the mean calcium
daily intake: the Brisighella Study
A Gaddi1*, AFG Cicero1, FO Odo Wani1, A Dormi1, V Pasquarelli2 and S D'Addato1
1
Atherosclerosis and Metabolic Diseases Study Centre `G. Descovich', Department of Internal Medicine and Applied Biotechnology,
University of Bologna, Italy; and 2Novartis Ð Sandoz Prodotti Farmaceutici SpA, Milan, Italy
Objectives: Evaluation of the impact of a coronary heart disease prevention program on calcium, magnesium,
phosphorus and vitamin D dietary intake in respect of recommended daily allowances in a large Italian rural
population.
Design: Retrospective analysis of the Brisighella Study dietary data. The Brisighella Study started in 1972 as a
longitudinal study on atherosclerosis risk factors.
Setting: Brisighella, a rural North Italian village.
Subjects: The Brisighella population's dietary habits were monitored from 1980 every 4 h through a dietary record
sheet. 1350 constantly tested subjects were subdivided according to NHI Consensus Conference on Calcium RDA.
Intervention: In 1986, the studied subjects were invited to reduce their consumption of animal fats and cholesterol
through a Nutrition Educational Program (NEP).
Results: Before NEP, calcium intake was low in each sex and age category: 20 ± 40% of the populatioin had a
daily intake < 550 mg. In 1988, among the 1350 subjects who constantly completed the questionnaire (M ˆ 651,
F ˆ 699), the mean calcium intake signi®cantly rose in all age categories: M ˆ 1003 (25 ± 65 y) and 877 ( > 65)
mg=24 h (P < 0.001 vs 1984); F ˆ 923 (25 ± 50), 860 (51 ± 65) and 767 ( > 65) mg=24 h (P < 0.05). In 1992, 3 y
after the NEP conclusion, calcium intake dropped in each sex and age category. The NEP in¯uenced vitamin D,
phosphorus and magnesium intakes less.
Conclusions: A collective NEP aimed at lowering saturated fats and cholesterol intakes, improves the calcium
intake; in order to maintain their ef®cacy on nutritional habit changes, these programs must become an ongoing
item.
Descriptors: calcium; magnesium; vitamin D; epidemiology; nutritional programs; coronary heart disease;
prevention project
European Journal of Clinical Nutrition (2001) 55, 97±106
Introduction
Bone is the mineral storage tank of the human body: it
contains 99% of all body calcium in hydroxyapatite crystals,
85% of body phosphorus and 65% of body magnesium. The
*Correspondence: A Gaddi, Dean of the Bologna Medicine Faculty,
Dipartimento di Medicina Clinica e Biotecnologie applicate `D.
Campanacci', Policlinico S. Orsola, via Massarenti 9, 40138 Bologna, Italy.
E-mail [email protected]
Guarantor: A Gaddi.
Contributors: Prof Gaddi is responsible for the Brisighella Heart Study and
ideator, guarantor and coordinator of the present research; Dr D'Addato
represents the original Brisighella Heart Study medical staff and supervised
the research; Dr AFG Cicero is responsible for the ®nal version of the
manuscript and for the study data processing with Dr Ada Dormi; Dr Odo
Wani and Dr Pasquarelli co-operated at various levels in the paper
elaboration.
Received 8 June 2000; revised 2 October 2000;
accepted 9 October 2000
plasma concentrations of these ions are regulated by: (a)
absorption from the gut; (b) deposition on and re-absorption
from bone; and (c) by urinary excretion. The main cause for
calcium, magnesium and phosphorus absorption is their
intake through food (Broadus, 1993). In 1994, a NIH
consensus development panel on optimal calcium intake
(Conference NioHC, 1994) emphasized the role of dietary
calcium intake in populations, also because the possible
mechanisms, which compensate dietary calcium de®ciencies, do not always allow a correct bone metabolism homeostatis. Furthermore, since the absorption of calcium by the
gut partially takes place by means of calcitriol-regulated
active saturable mechanism (Ghijsen & Van Os, 1982;
Christakos, 1989; Gross & Kumar, 1990), a reduction in
vitamin D intake and=or synthesis is detrimental for bone,
particularly in subjects with low calcium intake. The
recommended daily allowances (RDA) of calcium differ
according to the age as bone metabolism changes over the
Reducing lipid level improves calcium intake
A Gaddi et al
98
course of a lifetime (Conference NioHC, 1994). However,
550 mg=day is the dietary calcium intake limit under which
the calcium balance become negative, while an adequate
calcium intake increases bone mineral density during
skeletal growth and prevents bone loss and osteoporotic
fractures in the elderly. Some studies have shown that daily
calcium intake is low in most Western and Eastern countries
(Pennington & Young, 1991; Horwath et al, 1991; Haines et
al, 1994), although, especially for the Mediterranean basin
area, there are not suf®cient epidemiological studies on
calcium intake among populations. Magnesium improves
bone and calcium metabolism too (Tucker & Hannan,
1999). The elderly and postmenopausal women are often
at high risk for cardiovascular disease so that these subjects
are submitted to a low-fat diet. The risk of a diet poor in
dairy fat products could be a dietary calcium intake reduction. In this paper, we describe the trends of calcium,
phosphorus, magnesium and vitamin D dietary intakes in a
large Italian rural population, which was subdivided by sex
and age according to the Consensus Development Panel of
the NHI (Conference NioHC, 1994) in an 18-year-plus
period of observation to evaluate if the nutrition educational
programs (NEP) aimed at reducing CHD risk factors
modi®ed these intakes.
Methods
Study design
The Brisighella Study is an epidemiological study on
chronic diseases of social impact in the Northern Italian
rural town of Brisighella, promoted in 1972 by Professor G
Descovich (Descovish et al, 1974); in 1986 it became part of
the WHO European Risk Factors Co-ordinated Analysis
(ERICA Project, 1988), of the Multicenter Italian Study of
Cholelithiasis (Attili et al, 1995) and, in 1990, of the Risk
Factors and Life Expectancy Project (Menotti et al, 1994).
The study structure includes different phases and subprojects (Menotti et al, 1994; Descovich et al, 1990a,b;
Gaddi et al, 1996). It was observatioinal from 1972 to 1986
and became an intervention study in 1986 (phase still in
progress). Throughout both study phases, all causes of
mortality and morbidity, as well as the incidence of CHD
risk factors, have been recorded. The study design included
an updating of the database on fatal and non-fatal new
Table 1
Nutrition Educational Program (NEP). From 1986 to
1989, a Nutritional Information Centre was opened in
Brisighella and four dieticians provided study participants
with nutritioinal advice aimed at reducing their animal fats
and cholesterol daily intake. The centre was open 3 days per
week and the access was free (no charge nor appointment
necessary). The main characteristics of the people involved
in the NEP are summarized in Table 1. All Brisighella
families were informed about NEP by a mailing containing
brochures on nutritional guidelines and by posters. As well
as the literature (Descovich, 1990), people were generically
invited to: (a) substitute whole fat milk and yoghurt with
low fat products; (b) substitute saturated fat-rich meat with
white meat and ®sh, and limit meat consumption to less than
four times=week; (c) prefer fresh cheese and limit its
consumption to less than twice a week, considering cheese
as a substitute for meat; (d) limit egg consumption (as
cooked eggs or as part of noodles and pastry); (e) prefer
vegetable consumption; (f) limit sugar intake and fruit
consumption to less than 300 g=day in case of hypertriglyceridaemia or hyperglycaemia; (g) prefer oil and margarine
to butter and animal fats. Since 1986, approximately 2200
citizens have come to the Nutritional Centre; at least 0.7
persons for every Brisighella family participated in the NEP
from 1986 to 1989. As showed by the modi®cation of the
mean population cholesterol and triglycerides modi®cation,
the compliance to the dietetic suggestions was very good
during the NEP (Descovich, 1990).
High Risk Program. From 1989 to 1994, a specialized
staff performed risk strati®cation and, when necessary, in
agreement with the general practitioners, prescribed lipidlowering therapy (Gem®brozil, 600 mg twice daily and=or
cholestyramine, 3 g twice or three times daily, respectively)
to high-risk subjects. A total of 700 citizens (all involved in
Main characteristics of the Brisighella Study in the period 1980 ± 1992
Period
Subjects invited to take part
Subjects screened (historical cohort)a
New subjects enrolledb
Total subjects screened ( ˆ subjects screend ‡ new subjects enrolled)
Dietary recall (percentage vs total subjects screened)
a
events every 3 months and every 4 y a complete medical
check-up, nutritional habits record and fasting blood
sample. After 1986, several programs started to check
ef®ciency, costs and reliability of CHD primary and secondary prevention (school children and whole-population
nutritional education programs, general practitioner training
concerning therapeutic guide-lines, etc) (Gaddi et al, 1996;
Descovich et al, 1990a,b). The two main programs involved
in the current data analysis are the NEP and the High Risk
Program.
1980 ± 1984
1984 ± 1986
1986 ± 1988
2843.
2748.
2617.
2939.
2175.(100%)
1800.(82.7%)
1756.(80.7%)
1564.(71.9%)
0.
492.
1046.
1314.
2175.
2292.
3663.
2878.
1880.
1431.
2218.
1868.
(M ˆ 41%, F ˆ 59%) (M ˆ 44%, F ˆ 56%) (M ˆ 39%, F ˆ 61%) (M ˆ 48%, F ˆ 52%)
Percentage computed on the basis of 1980 ˆ 100% screened subjects.
Subjects who reached the minimum age for enrolment in the study (14 y) plus those who decided to participate subsequently.
b
European Journal of Clinical Nutrition
1988 ± 1992
Reducing lipid level improves calcium intake
A Gaddi et al
the NEP too) attended all control visits and complied with
the therapy until 1994. The main results of this program are
reported in Gaddi et al (1996).
Subjects
When the study started in 1972, all the free-living inhabitants of Brisighella aged over 14 y (4500 people) were
contacted directly or by mail and invited to take part in
the study; 2939 people (65.3%) agreed to participate. This
paper reports the nutritional data collected in the years 1980,
1984, 1988 and 1992. The nutritional parameters were
collected by means of the `7-day questionnaire' submitted
to 927 M and 953 F in 1980, 693 M and 738 F in 1984,
1109 M and 1109 F in 1988 and 967 M and 901 F in 1992.
We statistically processed only the data relative to the
651 M and 699 F who completed the questionnaire each
99
Figure 1 Seven-day questionnaire structure. Each questionnaire is formed by seven sheets (on the left: food eaten at every meal of the day (the subject is
invited to write the complete menu in descriptive form); on the right: daily amount of each food eaten).
European Journal of Clinical Nutrition
Reducing lipid level improves calcium intake
A Gaddi et al
100
Data processing and statistics
For each patient, the bromatological breakdown of the
foodstuffs was evaluated using the VIRGO program, a
part of the MIZAR program for medical statistics (Mannino
& Guidi, 1978). The food content of calcium, magnesium
and other nutrients was derived from foodstuff tables drawn
up by the Italian National Institute of Nutrition (Carnovale
et al, 1994) and was reported in terms of mean daily intake
and of the mean nutrient density. Current data processing
was performed on an ad hoc database, created using SPSS
6.1.2 software, also used for descriptive and con®rmatory
statistics. Simple factorial analysis of variance (ANOVA)
for paired samples, by `unique' non-hierarchical procedure,
evaluating all effects simultaneously, and multivariate
ANOVA were performed adopting the following categorical
variables: (a) `age of classes', (two for M and three for F, on
the basis of age and sex groups suggested by the NHI
Consensus Conference on Calcium RDAs, (Conference
NioHC, 1994); (b) sex; and (c) survey year (1980, 1984,
1988, 1992). Stepwise multiple regression was also performed (dependent variable: mean calcium intake, in
mg=day, or mean calcium density, in mg=1000 kcal day),
with probability of P ˆ 0.05 for both inclusion or removal of
each variable in=from the model. Logistic regression was
also employed, adopting as dependent variable the transformed calcium intake values, in the basis of calcium values
below which calcium balance becomes negative, ie
0 ˆ calcium intake 550 mg=day, 1 ˆ > 550 mg=day.
time. The adequacy of calcium and vitamin D intakes was
evaluated by grouping them according to the NIH RDAs
(Conference NioHC, 1994): men were divided into two age
groups (25 ± 65 and > 65 y) and women into three age
groups (25 ± 50, 51 ± 65 and > 65 y). Subjects aged below
25 were excluded because their number was too small to be
considered representative. The 1972 data are excluded
because they were collected with the ®rst experimental
version of the 7-day questionnaire. Table 1 illustrates the
number of people screened in the subsequent control visits
and summarizes the study structure during the period 1980 ±
1992 described in this paper.
Seven-day questionnaire
The 7-day questionnaire consists of two parts, both of which
were completed by the participants: the ®rst part concerns
the food eaten at every daily meal, and the second the daily
amount of each food component (Figure 1). The daily
amounts are expressed during either units of weight
(grams or hectograms) or colloquial terms (spoons, cups,
etc); when using colloquial terms, the corresponding mass
or volume was calculated by the dietician collecting the
questionnaire. Before starting, each participant was told
how to ®ll in the questionnaire and was asked to ®ll it in
day-by-day. However, a dietician was available to provide
explanations on the questionnaire and, if necessary, to make
visits to participants' houses to check that the questionnaire
was being ®lled in correctly and to provide further explanations. Finally, after the data collection, a dietician checked
each questionnaire during the control visit in order to ensure
that there were no discrepancies between parts 1 and 2. All
the food eaten and the relative quantities were encoded by
the dietician, printed out for the correction of any data input
errors and then transferred to the main database of the
Brisighella Study. This kind of dietary record meets the
characteristics required by the nutritional habit survey in
populations (Dwyer, 1994), particularly for the simplicity of
the descriptive part and the accuracy of the quantitative part,
avoiding typical disadvantages related to the use of a 24 h
recall (day-to-day variations of food consumption and other
bias sources; Bingham et al, 1988; Block, 1982).
Table 2
Food energy
The mean daily intake of food energy was fairly constant at
the four time points of the study, ranging with age from
1984 571 to 2386 622 kcal for F, and from 2426 716
to 2900 699 kcal for M (Table 2). Also, calories from
alcoholic beverages (mainly wine) do not remarkably
change throughout the observation period. MANOVA
detected a signi®cant effect of age and sex interaction on
alcohol calories (F ˆ 9.7, P ˆ 0.002) and on food energy
(F ˆ 66.9, P ˆ 0.001); age vs control year interaction proved
Mean daily energy intake (kcal=day, median and 95% CI for mean) per sex and age category at each Brisighella Study survey
1980
Men 25 ± 65 y (n ˆ 725)
Men >65 y (n ˆ 202)
Women 25 ± 50 y (n ˆ 394)
Women 51 ± 65 y (n ˆ 348)
Women >65 y (n ˆ 211)
a
Results
2831.000
(2849.034;
2385.000
(2361.816;
2297.000
(2324.515;
2232.000
(2228.196;
2011.000
(1983.637;
2951.008)
2560.026)
2447.704)
2346.878)
2126.363)
Signi®cant change in reply to the preceding survey (P < 0.05).
European Journal of Clinical Nutrition
1984
2787.000
(2828.833;
2463.500
(2462.437;
2272.000
(2249.841;
2242.000
(2200.196;
2026.000
(2083.002;
2947.374)
2665.551)a
2419.147)
2341.064)
2213.677)
1988
2834.000
(2842.803;
2404.000
(2396.073;
2295.000
(2315.089;
2241.000
(2226.356;
1976.000
(1944.481;
2942.541)
2557.813)
2444.270)
2346.800)
2080.506)
1992
2685.500
(2671.131;
2398.000
(2367.358;
2204.000
(2170.920;
2106.000
(2127.418;
1992.000
(1998.201;
2776.073)a
2519.916)
2280.200)a
239.429)a
2127.880)
Reducing lipid level improves calcium intake
A Gaddi et al
to have a less signi®cant effect (d.f. ˆ 12, P ˆ 0.07). Univariate analysis con®rms the age and gender effect on
calories intake (low caloric intake, both for alcohol and
food energy, in F and in the elderly), and that the survey
year has no effect on total food calories intake (F ˆ 1.94,
P ˆ 0.121). Multiple regression (dependent ˆ food calories)
indicates that the variance explained by the survey year was
lowest ( < 1%), while sex and age are responsible for 17%
of the residual variance. In all the surveys men's mean
weekly energy intake was always signi®cantly higher than
that of women (P < 0.001). Main food calories sources
(carbohydrates, fats and proteins) remained fairly constant
from 1980 to 1984, the ®rst accounting for approximately
52% of the energy intake, protein for approximately 15%,
and lipids for the remaining 33% in both M and F. A slight
reduction in total lipid intake (ÿ2%, approximately) was
observed in 1988 and not in 1992. However, remarkable
changes (P < 0.01) were observed from 1994 to 1988 (after
the NEP) in PUFA intake (values as week, mean, in g=day,
overall 1980 ˆ 10.94 4.22, 1984 ˆ 10.73 4.06, 1988 ˆ
12.74 5.05, 1992 ˆ 10.47 4.19) and in saturated fat
intake (38.14 11,97, 38.65 12.67, 35.40 12.53,
30.05 11.07, respectively).
101
Calcium intake
The absolute value of calcium intake (ACI, mg=day, mean
of 1 week) is reported in Table 3 (averages), while the
calcium nutrient density (CND, in mg=1000 kcal=day, mean
week values) is reported in Figure 2. In all the surveys,
mean weekly calcium intake was always signi®cantly higher
in mean than in women of all age classes (P < 0.01). The
mean calcium intake in 1980, 1984 and 1992 was signi®cantly lower than in 1988 and, on average, low in both sexes
Table 3 Calcium (mg=day, median and 95% CI for mean) and vitamin D daily food intakes (IU=day, median and 95% CI for mean) per sex and age
category at each Brisighella Study survey
Calcium daily dietary intake
Men 25 ± 65 y (n ˆ 518)
Men > 65 y (n ˆ 133)
Women 25 ± 50 y (n ˆ 259)
Women 51 ± 65 y (n ˆ 270)
Women > 65 y (n ˆ 170)
1980
1984
688.000
(721.2039;
775.2181)
643.500
(638.5822;
750.6356)
(647.9029;
712.8027)
661.500
(679.8505;
756.5575)
632.000
(629.7510;
710.7135)
713.000
(739.9045;
799.9449)
716.500
(705.6283;
827.3347)
(644.4670;
731.0634)
639.000
(636.5052;
702.7225)
642.000
(643.0697;
734.5025)
1988
951.000
(979.5596;
1035.8280)
809.000
(847.1015;
946.5083)
(893.7170;
965.4212)a
854.000
(840.1414;
905.7001)a
727.000
(734.1771;
822.8439)a
Vitamin D daily dietary intake
1992
706.500
(737.3657;
789.1659)a
691.500
(667.7409;
743.4245)
(656.3257;
714.6801)a
625.500
(640.3138;
698.1607)a
639.000
(640.7653;
720.2449)a
1980
1984
1988
1992
81.000
(96.2126;
115.7239)
77.500
(81.6085;
124.1440)
(81.5396;
105.0441)
69.000
(77.2471;
94.0977)
53.000
(60.1083;
72.1855)
86.000
(103.8749;
149.3379)
78.000
(76.2510;
127.2675)
(76.7261;
88.8609)
68.000
(72.000;
95.3927)
62.000
(62.5585;
94.2544)
86.000
(95.0040;
117.1460)
72.500
(73.5746;
100.9214)
(82.6433;
92.1858)
71.000
(72.0407;
90.9430)
57.000
(59.7978;
80.9954)
94.500
(107.3438;
131.0090)
65.500
(75.1551;
94.9528)
(90.9276;
104.4095)a
69.500
(63.3858;
119.3336)
57.000
(61.0667;
89.3400)
a
Signi®cant change in reply to the preceding survey (P < 0.05).
Figure 2
Median, quartile and range of calcium nutrition density in males, females and elderly of both sexes (mg=1000 kcal per day).
European Journal of Clinical Nutrition
Reducing lipid level improves calcium intake
A Gaddi et al
102
Figure 3 Overall distribution curve of calcium daily intake (mg=day) before the Nutrition Educational Program (1984, on the left) and after (1988, on the
right). The threshold of calcium intake under which calcium balance became negative (550 mg=day) is indicated from the arrow.
and at all the ages. Figure 3 shows the calcium intake
distribution in 1984 and 1988, before and after the
NEP start (P < 0.002). Before NEP, the 20 ± 40% of the
population had a daily calcium intake lower than calcium
losses (ie < 550 mg=day; Gaddi et al, 1996), while after
NEP the percentage was 16%. The right shift of calcium
distribution after NEP was equal for men and women and in
all age classes. The calcium distribution curves in 1980 and
1992 (not shown) overlap with the 1984 distributioin and
non-signi®cant differences were found both in overall and in
sub group analyses. The analysis of variance (performed on
the three main classi®cation factors: age, sex and control
year from 1980 to 1992) detected a signi®cant effect of sex
(ACI: higher values in males, F ˆ 4.77, P ˆ 0.029; CND:
higher values in females, F ˆ 13.28, P < 0.001), age (lower
calcium intake in the elderly; ACI: F ˆ 11.88;
CND ˆ 13.35, P < 0.001 for both) and the control year
(ACI: F ˆ 33.42, P < 0.001, CND: F ˆ 52.53, P < 0.001).
Exact F-statistics for age ± sex interaction by multivariate
tests of signi®cance, were 19.0 (P < 0.001), with a slight
effect on ACI (univariate: F ˆ 4.99, P ˆ 0.025), and on
CND (univariate: F ˆ 3.68, P ˆ 0.055). Year ± sex interaction do not show any effect on both ACI and CND (F 1,2,
P 0.3), while year ± age interaction (F ˆ 2.52, P ˆ 0.003)
seems to in¯uence ACI (univariate F ˆ 3.7, P ˆ 0.001) and,
perhaps, CND (univariate F ˆ 1.91, P ˆ 0.075). Stepwise
multiple regression (P to enter 0.001, P to remove 0.01;
dependent ˆ ACI, ®le split by sex and survey year) suggested that the main predictors for calcium intake, explaining around 30% of ACI variance at every survey, are
phosphorous intake (positive association) and alcohol and
protein intakes (negative association). In 1980 and 1984 a
main effect is also detected for saturated fat intake
(r2 ˆ 0.38 and 0.44, respectively), while after 1984 carboEuropean Journal of Clinical Nutrition
hydrate intake also explains a slight proportion (3.5%) of
calcium intake variance (negative association). At every
survey, food cholesterol, magnesium, potassium (with negative sign) and vegetable ®bre (positive sign) appear in the
®nal model, but with a very modest contribution (0.01 or
less) to predict. Excluding from the computation the independent variables with low tolerance (and known to be
calcium-related) to avoid colinearity problems (particulary
evident between phosphorous, protein, total lipid, saturated
lipid and total energy intakes), carbohydrates and alcohol
calories showed a negative and relevant predictive value of
calcium intake (estimate for both between 5% and 14%).
Analysis on calcium values coded on the basis of the
550 mg=day cut-off con®rmed the in¯uence of the control
period, age and sex. The number of subjects with low
calcium intake was 687 (36%) in 1980, 485 (34%) in
1984, 355 (16%) in 1988 and rose to 678 (34%) in 1992
(Pearson test, overall, 263.8, d.f. ˆ 3, P < 0.0001). The
expected values were 552, 420, 644 and 588, respectively.
Vitamin D, phosphorus and magnesium intakes
Vitamin D intake is shown in Table 3: the mean values
appear low in all age classes, and are always lower in
women than in men and in the elderly with respect to
young subjects, with a minimum of 66 IU=day in elderly F
at the 1980 survey. Table 4 shows phosphorus intakes. As
expected, like energy iintake, the phosphorus intake in men
was greater than in women (P < 0.001 for all survey years).
Magnesium daily intake (Table 4) was also signi®cantly
higher in M than in F, ranging from 175 to 215 mg=day in
the former, and from 147 to 181 mg=day in the latter (t ˆ 8.9
to 13.3, P < 0.0001). Both nutrient intakes were lower than
RDA, particularly for magnesium, if we assume a recommended daily magnesium intake of 290 mg=day in F and
Reducing lipid level improves calcium intake
A Gaddi et al
103
Table 4 Magnesium and phosphorus daily dietary intakes (mg=day, median and 95% CI for mean) per sex and age category at each Brisighella Study
survey
Magnesium daily dietary intake
Men 25 ± 65 y (n ˆ 518)
Men > 65 y (n ˆ 133)
Women 25 ± 50 y (n ˆ 259)
Women 51 ± 65 y (n ˆ 270)
Women > 65 y (n ˆ 170)
1980
1984
206.000
(210.2325;
220.3882)
186.000
(183.7763;
206.7881)
156.000
(159.6395;
170.9595)
164.500
(168.0886;
183.1125)
149.000
(153.4498;
171.5265)
188.000
(194.6492;
207.7350)
180.500
(179.3336;
203.8763)
147.000
(150.2572;
166.0262)
150.000
(157.8426;
173.4577)
151.000
(156.6814;
181.1260)
1988
191.000
(191.6683;
200.1043)
166.500
(170.3233;
187.3190)
152.000
(152.6155;
163.4922)
153.000
(156.7010;
167.3037)
140.000
(142.1393;
156.3249)a
Phosphorus daily dietary intake
1992
199.000
(204.4170;
216.2857)a
194.500
(193.3827;
212.4302)a
164.000
(166.5738;
179.3347)a
172.000
(174.1861;
188.5537)a
164.000
(168.0195;
184.7731)a
1980
1984
1988
1992
1758.000
(1766.785;
1835.355)
1524.000
(1527.088;
1676.041)
1284.500
(1291.889;
1368.776)
1264.500
(1282.344;
1364.569)
1164.000
(1145.090;
1243.261)
1715.000
(1713.989;
1790.836)
1582.000
(1541.011;
1691.063)
1198.000
(1218.251;
1320.405)
1180.000
(1190.449;
1281.736)a
1118.000
(1126.974;
1234.684)
1745.000
(1761.505;
1828.978)
1498.000
(1510.247;
1631.948)
1319.000
(1337.740;
1419.581)a
1306.000
(1312.146;
1390.973)a
1107.000
(1132.199;
1225.978)
1545.000
(1546.768;
1613.023)a
1383.500
(1393.165;
1495.310)
1178.000
(1157.411;
1223.406)a
1168.000
(1170.479;
1239.256)a
1106.000
(1102.908;
1185.133)
a
Signi®cant change in reply to the preceding survey (P < 0.05).
350 mg=day in M, 93 ± 98% of our population had a low
intake. MANOVA (independent: age, sex and control year)
detected a signi®cant effect of age ± sex interaction, as
expected (for interaction, exact F ˆ 16.6, P < 0.001; univariate: for vitamin D F ˆ 4.15 and P ˆ 0.042, for magnesium F ˆ 31.3 and P < 0.001, and for phosphorus F ˆ 47.3
and P < 0.001), but not for the survey year plus sex, nor
survey year plus age (for interaction, respectively, F ˆ 1.6,
P ˆ 0.097 and F ˆ 1.23, P ˆ 0.195). Univariate ANOVA
con®rmed a signi®cant effect of sex on magnesium and
phosphorus (but not on vitamin D) and age on all three
dependent variables (P < 0.001 for all).
Discussion
This study shows that the Brisighella population daily
intakes of calcium, vitamin D and magnesium are below
the recommended allowances and that one-third of the
population has a food calcium intake that is not even
suf®cient to counterbalance inevitable calcium losses.
Moreover, in 1980, 22.5% of older men and 21.5% of
older women (9.5% and 8.3%, respectively, after NEP in
1986 ± 1989) had a multiple de®cit in food intake of
calcium, vitamin D and magnesium. These daily intakes
were calculated from data obtained by the 7-day questionnaires: this dietary record seems to be accurate as long as
the subjects are willing to cooperate, as also shown in some
controlled clinical trials (Gaddi et al, 1990). It meets the
characteristics required by the nutritional habit survey in
populations (Dwyer, 1994), particularly for the simplicity of
the descriptive part and the accuracy of the quantitative
part, avoiding typical disadvantages related ot the use of a
24 h recall (day-to-day variations of food consumption and
other bias sources; Dwyer, 1994; Bingham et al, 1988;
Block, 1982). However, the missing data of calcium intake
from drinking water might cause an underestimation of total
calcium daily intake; water calcium was not evaluated in
the Brisighella study because of: (a) the dif®culty of
devising a correct questionnaire for water consumption (to
correct this possible bias we therefore assessed the data on
the calcium contained in water collected from three sampling points in the Brisghella water supply system at
different times: the concentrations of calcium were very
low (always 100 mg=l at different times and different
sampling points). Even supposing that the water calcium
can be assimilated and used like the food calcium, we
calculated that, on average, there could be an underestimation of around 90 mg=day for M and 75 mg=day for F. The
adequacy of daily intakes in the Brisighella population was
estimated by comparison with the NIH RDAs, because
these are the most recent and because no comparable
European document exists. Italian tables, although not
supported by large epidemiological studies, provide similar
or slightly lower values (Harper, 1994). However, it is not
always clear whether RDAs are proposed as safe intakes for
individuals or as appropriate average intakes for groups
(SocietaÁ Italiana Nutrizone Umana, 1996), but, at the same
time, the RDA serves as a standard for establishing health
policy and preventing health problems deriving from nutritional inadequacy. In our opinioin, at present it is not
possible to de®nitely establish the individual probability
that a given nutrient is inadequate. Therefore, according to
Harper, we consider the RDAs as `those amounts suf®cient
to meet the physiological needs of practically all healthy
persons in a speci®ed group' (Harper, 1994). As for calcium
intake, we preferred to adopt the value of 550 mg=day,
which represents the threshold below which the calcium
balance becomes negative and is therefore certainly correlated with an increase in the risk of disease. Adopting this
cut-off point, nutritional advice given during the NEP
reduced the number of subjects needing calcium supplement by one-half (from a mean of 35% before 1986 to 16%
in 1988).
European Journal of Clinical Nutrition
Reducing lipid level improves calcium intake
A Gaddi et al
104
It has been suggested that both protein and sodium affect
calcium balance, possibly by increasing calcium excretion
(Shapses et al, 1995) and sodium (Devine et al, 1995);
however their intake is closely related to that of calcium. In
the Brisighella population a correlation between protein
intake and calcium intake (r2 ˆ 0.231 ± 0.506, P < 0.001)
was found; however, when calcium values are transformed
into a dichotomous variable ( or < 550 mg=day), the
intakes of saturated fats and phosphorous, but not of
proteins, became signi®cant predictors (at the various
years of control in both genders). This suggests that saturated fat intake may be a marker of the consumption of dairy
products rather than proteins themselves, since their animal
quota (from meat), which is not linked to a greater intake of
calcium, may constitute a confusing variable, particularly in
subjects with the lowest calcium intakes.
Con¯icting results have been published concerning the
effect of calcium intake on bone density, but a recent
metanalysis (Welten et al, 1995) suggested that calcium
intake can at least partially explain the bone density variability. Moreover, an analysis of studies evaluating the effect
of exercise on bone density (Specker, 1996) showed that, at
least for spine bone mineral density, a calcium intake of
more than 1000 mg=day is necessary before exercise can
exert a positive effect on bones. Finally, both crosssectional (Matkovic et al, 1979) and perspective (Holbrook
et al, 1988) studies showed a higher risk of hip fractures in
subjects with lower calcium intakes, while guidelines for
osteoporosis treatment and prevention suggests dietary calcium supplementations (Castelo-Branco, 1998). Various
studies (Scaccini et al, 1992; Howarth et al, 1995) have
tried to understand what factors in¯uence calcium intake.
They have shown that: (a) the majority of subjects with low
calcium intakes are unaware that their calcium consumption
is insuf®cient; (b) 20 ± 30% of the interviewed subjects
believe that dairy products are bad for their health, since
they contain a lot of cholesterol and calories; (c) almost 15%
of the interviewed subjects have been warned against dairy
products by their doctors. In effect, according to other
authors (Adult treatment panel II, 1994), our study con®rms
a close association between calcium intake and saturated fat
intake in 1980, 4 y before the NEP, but it also shows the
possibility of increasing calcium intake without a further
increase of saturated fats. It is important to underline that
the right shift of calcium intake appears to be the result of
actual increase in calcium intake, independently from calcium supplementation (in Brisighella pharmacy and drugstores there was no signi®cant increase observed in the
calcium supplement sales in the 1984 ± 1992 period) since
the daily intake of energy did not follow the same trend. It is
interesting to note that all the models used (multiple
regression with continuous dependent variable, both total
calcium and nutrient density, and logistic regression with
calcium cut-off at 550 mg=day) show that, after the NEP, the
saturated fats also lose their correlation with calcium intake,
probably because of the increase in the intake of low-fat
dairy products. However, apart from speculation, our data
from the 1988 survey clearly show that nutritional education
European Journal of Clinical Nutrition
which reduced serum lipids (Menotti et al, 1994) not only
did not reduce calcium intake, but actually increased it, even
though the NEP was speci®cally aimed at decreasing the
excessive consumption of saturated fats and cholesterol.
This shows, beyond all reasonable doubt, that the adoption
of correct nutritional guidelines, substantially similar to
those of the AHA (Adults treatment panel II, 1994), does
not give rise to negative effects but, instead, tends to
encourage most people to follow correct and complete
diets. The NIH RDA for vitamin D is 200 IU=day, but
recent studies suggested that, especially in the elderly,
higher amounts ( > 400 ± 500 IU=day) are needed
(Dawson-Hughes et al, 1995). In youngsters and adults,
skin cell synthesis is considered suf®cient to produce 400 IU
of vitamin D every day as long as the exposure to sunlight is
adequate (Lawson, 1979). However, skin cell synthesis of
vitamin D in the elderly is more likely to be insuf®cient,
since they tend to spend less time outdoors and the ability of
their skin to synthesize vitamin D3 is reduced (Holick et al,
1989). As a consequence, hypovitaminosis D is frequent not
only among the elderly living at higher latitudes (McKenna
et al, 1985), but also among those living in those European
countries which are said to enjoy the sun (Van der Wielen et
al, 1995). Elderly people and those who use heavy sunscreen should have 400 ± 800 UI vitamin D=day (Murray,
1996). In conclusion, at least in older people, vitamin D
concentration must be supplemented by vitamin D intake.
Our data clearly show that vitamin D intake decreases with
age and that only a mean of 66 ± 103 IU=day of vitamin D
comes from the diet and, furthermore, that the people with
the lowest calcium intake also have the lowest vitamin D
intake. Since a low calcium intake is associated with an
increased synthesis of 1,25-dihydroxy-vitamin D, and therefore with a rapid decrease in 25-hydroxy-vitamin D (Clements et al, 1987), vitamin D intake in our population appears
to be inadequate, although we believe that more studies are
needed to de®ne some RDAs for vitamin D which can be as
indisputably reliable as those recently proposed for calcium.
Moreover, on the basis of the Brisighella experience, the
compensation of low vitamin intake through diet seems to
be dif®cult, especially with non-speci®c nutrition education
programs directed to the whole population. However, the
results' consistence in vitamin D dietary assessment is
partially limited by the great variation in the occurrence
of this vitamin in different foodstuffs. Moreover, it is
necessary to remember that the vitamin D RDA is greatly
in¯uenced from the individual mean exposure to the sun,
which is hardly quanti®able.
Finally, recently it was proposed that a diet rich in
calcium could contribute to reduce the risk of cerebrovascular disease (Yang, 1998; Suter, 1999; Abbott et al, 1996)
and coronary heart disease (Klor et al, 1997). Considering
that the main aim of the Brisighella study is to lower the
cardiovascular disease-related mortality (Dormi et al, 1999),
the observed results agree with this ®nality.
In conclusion, this epidemiological study shows that
calcium, vitamin D and magnesium intakes may be low in
a very high percentage of apparently healthy subjects, even
Reducing lipid level improves calcium intake
A Gaddi et al
in a country with high socio-economic conditions and where
the basic diet seems to be varied and complete. A full onethird of the population may have a calcium intake below the
minimum amount needed to balance inevitable calcium
losses. This problem may be particularly relevant to the
elderly of both sexes and in menopausal women; in the
young, it seems more dif®cult to predict what the long-term
dietary habits will be. However, a warning to subjects of all
ages does seem advisable.
Our conclusion is that a well-conducted collective educational program, aimed at lowering food saturated fats and
cholesterol intakes, can improve the calcium intake; however,
in order to maintain their ef®cacy on nutritional habit changes,
these kinds of programs must become an ongoing item.
References
Abbott RD, Curb, JD, Rodriguez BL, Sharp DS, Burch®eld CM & Yano K
(1996): Effect of dietary calcium and milk consumption on risk thromboembolic stroke in older middle-aged men. The Honolulu Heart
Program. Stroke 27, 813 ± 818.
Adults treatment panel II (1994): Expert panel on the detection, evaluation
and treatment of high blood cholesterol in adults. Summary of second
report of expert panel on the detection, evaluation and treatment of high
blood colesterol in adults. Circulation 89, 1329 ± 1345.
Attili AF, Caruli N, Roda E et al (1995): Epidemiology of gallstone disease
in Italy: prevalence data of the Multicenter Italian Study of Cholelithiasis
(MICOL). Am. J. Epidemiol. 141, 158 ± 165.
Bingham SA, Nelson M, Paul AA et al (1988): Methods for data collection
at the individual level. In Manual on Methodology for Food Consumption Studies, eds. ME Cameron, WA Van Staveren, pp 98 ± 115. Oxford:
Oxford University Press.
Block G (1982): A review of validations of dietary assessment methods.
Am. J. Epidemiol. 115, 492 ± 505.
Broadus AE (1993): Physiological functions of calcium, magnesium and
phosphorus and mineral ion balance. In Primer on the Metabolic Bone
Diseases and Disorders of Mineral Metabolism, ed. MJ Favus, pp 41 ±
45. New York: Raven.
Carnovale E, Marletta L & Miuccio F (1994): Tabelle di composizione degli
alimenti, aggiornamenti al 31.12.94. Roma: Istituto Nazionale della
Nutrizione Ed..
Castelo-Branco C (1998): Management of osteoporosis. An overview.
Drugs Aging 12(Suppl), 1S25 ± S32.
Christakos S (1989): Vitamin D-dependent calcium binding proteins:
chemistry, distribution, functional considerations, and molecular biology. Endocr. Rev. 10, 3 ± 8.
Clements MR, Johnson L & Fraser DR (1987): A new mechanism for
induced vitamin D de®ciency in calcium deprivation. Nature 325,
62 ± 65.
Conference NioHC (1994): Optimal calcium intake. NIH Consensus
Development Panel on Optimal Calcium Intake. J.A.M.A. 272, 1942 ±
1948.
Dawson-Hughes B, Harris SS, Krall EA et al (1995): Rates of bone loss in
postmenopausal women randomly assigned to two dosages of vitamin D.
Am. J. Clin. Nutr. 61, 1140 ± 1145.
Descovich GC (1990): The Brisighella Heart Study: an interim report.
Eur. Heart J. 11, S32 ± 73.
Descovich GC, Mannino G & Lenzi S (1974): Relationships between
ischaemic heart disease, diet and lipid spectrum in a group of homogeneous
populatioins in Emilia Romagna. Giorn. It. Cardiol. 4, 373 ± 381.
Descovich GC, D'Addato S, Dormi A et al (1990a): The Brisighella heart
study report from 1984 to 1989. In Atherosclerosis and Cardiovascular
Diseases, eds. GC Descovitch, A Gaddi, GL Magri, S Lenzi, pp 622 ±
631. Dordrecht: Kluwer Academic.
Descovitch GC, Gaddi A, Minard A et al (1990b): The Brisghella Heart
Project. In Biotechnology of Dyslipoproteinemias. Application in Diagnosis and Control, eds. C Lefant, A Albertini, R Paoletti, AL Catapano,
pp 311 ± 317. New York: Raven Press.
Devine A, Criddle AR, Dick IM et al (1995): A longitudinal study of the
effect of sodium and calcium intakes on regional bone density in
postmenopausal women. Am. J. Clin. Nutr. 62, 740 ± 745.
Dormi A, Cicero AFG, Favali, D, Mascitti M, Toma M & Gaddi A (1999):
Cardiovascular mortality and risk factors trend in a large rural population: the Brisighella Heart Study. Atherosclerosis 144(Suppl 1), S140
(abstract).
Dwyer TD (1994): Dietary assessment. In Modern Nutritioin, eds. ME
Shils, JA Olson, M Shike, pp 842 ± 60. Philadelphia: Lea & Febiger
ERICA Project (1988): The CHD risk map of Europe. The 1st report of the
WHO-ERICA Project. Eur. Heart J. 9, 1 ± 35.
Gaddi A, Ciarrocchi A, Matteucci A et al (1990): Dietary treatment of
familial hypercholesterolemia: differential effects of dietary soy protein
according to the apolipoprotein E phenotypes. Am. J. Clin. Nutr. 53,
1191 ± 1196.
Gaddi A, Galetti C, Dormi A et al (1996): Hypolipidemic treatments in the
elderly: the `Brisighella study' experience. In Drugs Affecting Lipid
Metabolism, eds. AM Gotto, R Paoletti, LC Smith, AL Catapano, AS
Jackson, pp 447 ± 459. Dordrecht: Kluwer Academic.
Ghijsen W & Van Os CH (1982): 1,25-dihydroxyvitamin D3 regulates
ATP-dependent calcium transport in basolateral plasma membranes of
rat enterocytes. Biochem. Biophys. Acta 689, 170 ± 175.
Gross M and Kumar R (1990): Physiology and biochemistry of vitamin
D-dependent calcium binding proteins. Am. J. Physiol. 259, 195 ± 199.
Haines CJ, Chung TK, Leung PC, Leung DH, Wong MY & Lam L (1994):
Dietary calcium intake in postmenopausal Chinese women. Eur. J. Clin.
Nutr. 48, 591 ± 594.
Harper AE (1994): Recommended dietary intakes: current and future
approaches. In Modern Nutrition, eds. ME Shils, JA Olson, M Shike,
pp 1475 ± 1490. Philadelphia: Lea & Febiger.
Holbrook TL, Barrett-Connor E & Wingard D (1988): Dietary calcium and
risk of hip fracture: 14-year prospective population study. Lancet 706,
1046 ± 1049.
Holick MF, Matsuoka LY & Wortsman J (1989): Age, vitamin D, and solar
ultraviolet radiation. Lancet 88, 1104 ± 1105.
Horwath CC, Parnell WR, Birkbeck JB, Wilson N, Russell D & Herbison P
(1991): Life in New Zealand: Hillary Commission Report: Nutrition, Vol
IV. Otago: University of Otago Press.
Horwath CC, Govan CH, Campbell AJ, Busby W & Scott V (1995): Factors
in¯uencing milk and milk product consumption in young and elderly
women with low calcium intakes. Nutr. Res. 15, 1735 ± 1745.
Kor HU, Hauenschild A, Holbach I, Schnell-Kretschmer H & Stroh S (1997):
Nutrition and cardiovascular disease. Eur. J. Med. Res. 2, 243 ± 257.
Lawson DEM (1979): Relative contributioins of diet and sunlight to
vitamin D state in the elderly. Br. Med. J. 2, 303 ± 305.
Mannino G & Guidi G (1978): MIZAR un programma interattivo guidato di
statistica sanitaria, pp 76 ± 104. Roma: Informatica Dedalo edizioni.
Matkovic V, Kostial K, Simonovic I, Buzina R, Brodarec A & Nordin BEC
(1979): Bone status and fracture rates in two regions of Yugoslavia.
Am. J. Clin. Nutr. 32, 540 ± 549.
McKenna MJ, Freaney R, Meade A & Muldowney FP (1985): Hypovitaminosis D and elevated serum alkaline phosphatase in elderly Irish
people. Am. J. Clin. Nutr. 41, 101 ± 109.
Menotti A, Farchi G, Seccareccia F and the RIFLE Research Group (1994):
The prediction of coronary heart disease mortality as a function of major
risk factors in over 30,000 men in the Italian RIFLE pooling project (risk
factors and life expectancy). A comparison with the MRFIT primary
screeners. J. Cardiovasc. 1, 263 ± 270.
Murray TM (1996): Prevention and management of osteoporosis: consensus
statements from the Scienti®c Advisory Boad of the Osteoporosis
Society of Canada. 4. Calcium nutrition and osteoporosis. C.M.A.J.
155, 935 ± 939.
Pennington J and Young BE (1991): Total Diet Study nutritional elements
1982 ± 1989. J. Am. Diet. Assoc. 91, 179 ± 183.
Scaccini C, Sette S, Mariotti S, Verdecchia A and Ferro-Luzzi A (1992):
Nutrient adequacy of dietary intakes of elderly. Age Nutr. 3, 41 ± 47.
105
European Journal of Clinical Nutrition
Reducing lipid level improves calcium intake
A Gaddi et al
106
Shapses SA, Robins SP, Schwartz EI & Chowdhury H (1995): Short-term
changes in calcium but not protein intake alter the rate of bone readsorption in healthy subjects as assessed by urinary pyridinium cross-links
excretion. J. Nutr. 125, 2814 ± 1821.
SocietaÁ Italiana Nutrizione Umana (1996): Livelli di assunzione raccomandati di energia e nutrienti Ð LARN, pp 112 ± 117. Roma: SINU ed.
Speckler BL (1996): Evidence for an interaction between calcium intake
and physical activity on changes in bone mineral density. J. Bone Miner.
Res. 10, 1539 ± 1544.
Suter PM (1999): The effects of potassium magnesium, calcium and ®ber
on risk of stroke. Nutr. Rev. 57, 84 ± 88.
European Journal of Clinical Nutrition
Tucker KL & Hannan MT (1999): Potassium, magnesium, and fruit and
vegetable intakes are associated with greater bone mineral density in
elderly men and women. Am. J. Clin. Nutr. 69, 727 ± 736.
Van der Wielen RP, Loewick MRH, van den Berg H et al. (1995): Serum
vitamin D concentratioins among elderly people in Europe. Lancet 346,
207 ± 210.
Welten DC, Kemper HC, Post GB & van Staveren WA (1995): A metaanalysis of the effect of calcium intake on bone mass in young and
middle aged females and males. J. Nutr. 125, 2802 ± 2813.
Yang CY (1998): Calcium and magnesium in drinking water and risk of
death from cerebrovascular disease. Stroke 29, 411 ± 414.