Change in Body Weight from Age 20 Years Is a Powerful Determinant of the Metabolic Syndrome

Abstract

Background:

Higher body weight is a well-known determinant of the metabolic syndrome (MetS) and its components. It is however less well studied how the change in weight from age 20 years to middle age or old age affects MetS development.

Methods:

In the community-based EpiHealth (n = 19,000, age range 45 to 75 years, 56% females) and PIVUS (n = 1000, all aged 70 years, 50% females) studies, the participants were asked about their body weight at age 20 years. Data were collected to determine MetS prevalence (NCEP ATP III criteria).

Results:

In EpiHealth, the probability of having MetS increased fairly linearly with increasing weight from age 20 in the obese [odds ratios (OR) 1.04 per kg change in weight, 95% confidence interval (CI) 1.03–1.05, P < 0.0001], as well as in the overweight (OR 1.15, 95% CI 1.14–1.17, P < 0.0001) and normal-weight (OR 1.18, 95% CI 1.14–1.21, P < 0.0001), subjects after adjustment for age, sex, body mass index (BMI) at age 20, alcohol intake, smoking, education, and exercise habits. Also in the PIVUS study, the change in weight over 50 years was related to prevalent MetS (OR 1.08 per kg change in weight, 95% CI 1.06–1.10, P < 0.0001). In both studies, self-reported BMI at age 20 was related to prevalent MetS.

Conclusion:

Self-reported weight gain from age 20 was strongly and independently associated with prevalent MetS both in middle age or old age. Interestingly, this relationship was not restricted only to obese subjects. Our data provide additional support for the importance of maintaining a stable weight throughout life.

Introduction

Obesity, especially of the abdominal type, is one of the cornerstones of the metabolic syndrome (MetS). A relationship among a low birth weight, increased weight gain during childhood, and the development of MetS later in life has been described in several studies.^1

–4 It is however less well known if weight gain later in life is also associated with the development of MetS^5,6 and its individual components. Moreover, previous studies have not investigated the importance of weight gain in individuals with body mass index (BMI) in the normal range for MetS development.

Metabolically healthy obesity (MHO) is a term used to denote obese subjects without signs of MetS.⁷ Some factors, such as visceral obesity, inflammation, and low adiponectin levels, have been identified as distinguishing obese subjects with and without MetS.^8

–11 It is not known whether obese subjects with or without MetS differ regarding weight gain since early adulthood (age 20).

To study if the weight gain seen during the adult life span is associated with MetS development, we used data from two different population-based samples, one large-scale sample (the EpiHealth cohort study¹²), in which we could evaluate the impact of weight change from the age of 20 years for MetS prevalence in different BMI strata, and also the PIVUS study,¹³ in which the impact of weight gain from age 20 to age 70 was related to prevalent MetS at age 70 years.

Materials and Methods

The EpiHealth study

The overall objective of the EpiHealth cohort study is to provide a resource to study interactions between genotypes and lifestyle factors in a large cohort derived from the Swedish population in the age range 45 to 75 years regarding development of common degenerative disorders. The sample was included based on population-based randomization using an address database starting in 2011. EpiHealth is a two-center study in Uppsala and Malmö, and the participants were recruited by ordinary mail within these towns and in their surroundings (Uppsala Län and Region Skåne). At the time of the present analyses, the number of participants was 19,094. The participation rate was ∼20% of the invited populations. Sixty-nine percent of the participants were recruited in Uppsala, since the collection of data started earlier in Uppsala than in Malmö. The study was approved by the ethics committee and the participants gave informed consent.

The investigations were performed after at least 6 hr of fasting, using similar equipment at both sites. Data on medications for hypertension and diabetes were asked for. Blood pressure was recorded twice in the sitting position by an automatic device (Omron, Kyoto, Japan). Height was recorded, and waist circumference was measured at the umbilical level. A venous blood sample was taken to measure fasting plasma glucose, high-density lipoprotein cholesterol (HDL-C), and plasma triglycerides at the Uppsala hospital laboratory using an Architect Ci8200 analyzer (Abbott Laboratories, Abbott Park, IL) (Carlsson).

PIVUS study

Eligible were all subjects aged 70 years residing in the community of Uppsala, Sweden. The subjects were chosen from the register of community living, and they were invited in a randomized order from the start of the study in April 2001 to the last included subject in June 2004. The subjects received an invitation by letter within 1 month of their 70th birthday to standardize for age. Of the 2025 invited subjects, 1016 subjects were investigated giving a participation rate of 50.1%.¹³ The study was approved by the Ethics Committee of Uppsala University and the participants gave written informed consent.

The participants were asked to answer a questionnaire about their medical history, smoking, and regular medication. All participants were investigated in the morning after an overnight fast. No medication or smoking was allowed after midnight. Blood pressure was measured by a calibrated mercury sphygmomanometer to the nearest mmHg after at least 30 min of rest, and the average of three recordings was used. Glucose and lipids were assessed by standard techniques.

Definition of weight gain and MetS

In both studies, the participants were asked the question: What was your body weight at age 20 years? At the time of the actual examinations, weight was measured by an accurate hospital scale. Those data were used to calculate the weight change until the examination.

MetS was defined according to the NCEP ATP III criteria.¹⁴ Three of the following five criteria should be fulfilled: Blood pressure ≥130/85 mmHg or antihypertensive treatment, fasting plasma glucose ≥6.1 mM, serum triglycerides ≥1.7 mM, waist circumference >102 cm in men and >88 cm in women, and HDL-C <1.0 mM in men and <1.3 in women.

Statistics

A cubic restricted spline function with three knots (10th, 50th, and 90th percentile) for the variable weight change since age 20 was calculated for each of the two studies.

In the analysis versus age in EpiHealth, a linear regression model adjusting for sex was used. Linear regression analyses were also used to evaluate possible differences in weight change since age 20 and other obesity-related variables between subjects with and without MetS within each of the three BMI groups, adjusting for age (also with spline term), sex, BMI at age 20, alcohol intake, smoking, education level, and exercise habits. The odds ratios (OR) are given for a 1 kg change in body weight.

In the analyses versus prevalent MetS or the different MetS criteria (binary variables), logistic regression was used adjusting for age (also with spline term), sex, BMI at age 20, alcohol intake, smoking, education, and exercise habits. Logistic regression analyses were also used in PIVUS when relating change in weight since age 20 to prevalent MetS.

An interaction term between sex and change in weight since age 20 was primarily included in all analyses, but was removed if not significant.

Results

Basic characteristics of EpiHealth and PIVUS participants are displayed in Tables 1 and 2.

Table 1.

Basic Characteristics in the EpiHealth Sample

Variable	N	Mean (SD)
Age (years)	19,036	60.1 (8.5)
Sex (% female)	19,036	56
Height (cm)	19,036	171 (9)
Weight (kg)	19,036	77.2 (14.5)
Waist circumference (cm)	19,036	91 (12)
Hip circumference (cm)	19,035	102 (12)
BMI (kg/m²)	19,036	26.2 (4.0)
WHR	19,035	0.90 (0.08)
Fat mass (kg)	19,036	23.9 (8.5)
Diastolic blood pressure (mmHg)	19,035	82.0 (9.5)
Systolic blood pressure (mmHg)	19,035	134 (17)
Fasting glucose (mM)	18,807	5.9 (1.0)
Total cholesterol (mM)	18,804	5.8 (1.0)
HDL-C (mM)	18,805	1.5 (0.4)
LDL-C (mM)	18,799	3.7 (0.96)
Serum triglycerides (mM)	18,805	1.2 (0.7)
Antihypertensive treatment (%)	19,036	22
Antidiabetic treatment (%)	19,036	3.2
Number of MetS components	18,788	1.6 (1.2)
Prevalent MetS (%)	18,788	23
Glucose component (%)	18,807	24
HDL-cholesterol component (%)	18,805	17
Triglyceride component (%)	18,805	16
Waist component (%)	19,036	36
Blood pressure component (%)	19,035	71
Alcohol consumption (drinks per week)	17,643	2.5 (2.9)
Smoking (years of smoking)	14,329	9.1 (9.2)
Exercise habits (on a 5-level scale with 1 as totally sedentary and 5 athlete)	18,756	Level 1: 4%
		Level 2: 22%
		Level 3: 41%
		Level 4: 27%
		Level 5: 6%
Education level	16,609	<10 years: 18%
		10–12 years: 29%
		>12 years: 53%

BMI, body mass index; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; MetS, metabolic syndrome; WHR, waist/hip ratio.

Table 2.

Basic Characteristics in the PIVUS Sample

Variable	N	Mean (SD)
Age (years)	1016	70.1 (0.1)
Sex (% female)	1016	50
Height (cm)	1016	168 (9)
Weight (kg)	1016	77.3 (14.4)
Waist circumference (cm)	1004	91 (11)
Hip circumference (cm)	1004	100 (8)
BMI (kg/m²)	1016	27.0 (4.3)
WHR	1004	0.90 (0.07)
Fat mass (kg)	—
Diastolic blood pressure (mmHg)	1012	78.6 (10.1)
Systolic blood pressure (mmHg)	1012	149 (22)
Fasting glucose (mM)	1013	5.3 (1.6)
Total cholesterol (mM)	1013	5.4 (1.0)
HDL-C (mM)	1013	1.5 (0.4)
LDL-C (mM)	1011	3.3 (0.8)
Serum triglycerides (mM)	1013	1.2 (0.6)
Antihypertensive treatment (%)	1007	31
Antidiabetic treatment (%)	1010	9.2
Number of MetS components	985	1.6 (1.1)
Prevalent MetS (%)	985	21
Glucose component (%)	1013	11
HDL-cholesterol component (%)	1013	23
Triglyceride component (%)	1013	17
Waist component (%)	1004	33
Blood pressure component (%)	1003	84
Alcohol consumption (gram alcohol per day)	861	2.5 (3.0)
Smoking (current smokers)	1015	11
Exercise habits (on a 4-level scale with 1 as totally sedentary and 4 athlete)	989	Level 1: 12%
		Level 2: 59%
		Level 3: 22%
		Level 4: 7%
Education level	1005	<10 years: 57%
		10–12 years: 18%
		>12 years: 25%

Change in body weight from age 20 versus age at examination in EpiHealth

The change in body weight from age 20 was significantly related to age at examination (P < 0.0001). The spline function showed this relationship to be nonlinear (P < 0.0001). The interaction term between sex and age regarding change in body weight from age 20 was highly significant (P = 0.0003). Supplementary Figure S1 (Supplementary Data are available online at www.liebertpub.com/met) shows that the youngest men in the study (age 45 years) showed more than 2 kg more weight gain than the youngest women. However, the women catch-up and at around 60 years of age, no difference between men and women is seen. After this age no further increase in change in body weight from age 20 was seen. If anything, a slight decline was seen above the age of 65 years.

Table 3 shows that subjects in the obese group with MetS were older and less likely to be female compared to the obese group without MetS. When adjusting for these characteristics (including the spline term for age), as well as for sex, alcohol intake, smoking, education, and exercise habits, BMI at age 20 was slightly lower (P < 0.05), while the weight gain since age 20 was substantially higher in the obese group with MetS (P < 0.0001) compared to the obese group without MetS.

Table 3.

Basic Characteristics, BMI at Age 20, BMI at Examination, and the Change in Weight from Age 20 Years Are Given as Mean (SD) in Different Combinations of BMI and the Metabolic Syndrome in the EpiHealth Study

	Normal-weight without MetS (n = 6381)	Normal-weight with MetS (n = 313)	Overweight without MetS (n = 4811)	Overweight with MetS (n = 1734)	Obese without MetS (n = 944)	Obese with MetS (n = 1483)
Age (years)	58.9 (8.5)	63.4 (7.2)	60.0 (8.5)	62.6 (7.6)	58.8 (8.4)	61.0 (8.0)
Sex (% females)	68	62	46	49	59	48
BMI at examination (kg/m²)	22.7 (1.6)	23.7 (1.1)^***	26.9 (1.3)	27.6 (1.3)^***	32.6 (2.7)	33.3 (3.1)^***
BMI at age 20 (kg/m²)	20.4 (1.9)	20.0 (2.0)^***	21.8 (2.3)	21.2 (2.4)^***	23.1 (2.9)	22.8 (3.1)^*
Change in weight (kg)	5.9 (6.4)	9.8 (6.3)^***	13.9 (7.1)	18.1 (7.4)^***	25.8 (10.7)	29.4 (11.1)^***

P < 0.05, ^*** P < 0.001 versus the corresponding group without MetS within each BMI group following adjustment for age (with spline term), sex, alcohol intake, smoking, education, and exercise habits.

Table 3 also shows that the difference in weight gain between subjects with and without MetS was not only restricted to obese subjects but also that this difference was prominent in the normal-weight and overweight groups (adjusting for age with spline term, sex, alcohol intake, smoking, education, and exercise habits). As for the obese groups, BMI at age 20 was slightly lower in the normal-weight and overweight groups with MetS compared to the corresponding group without MetS.

Change in body weight versus prevalent MetS in EpiHealth

Figure 1 shows that the probability of having MetS among the obese increased fairly linearly with increasing weight from age 20 (following adjustment for age, sex, and BMI at age 20), as indicated by the nonsignificant spline term (P = 0.34). This relationship was highly significant [OR 1.04 for a 1 kg change in weight, 95% confidence interval (CI) 1.03–1.05, P < 0.0001] when adjusted for age, sex, BMI at age 20, alcohol intake, smoking, education, and exercise habits. Also BMI at age 20 was significantly related to prevalent MetS in that model (OR 1.07 for a 1 kg change in weight, 95% CI 1.03–1.11, P < 0.0001). No interaction was seen between sex and change in weight from age 20 regarding prevalent MetS (P = 0.84).

FIG. 1.

Relationship between the change in weight since age 20 years (including a spline term in the calculations) and probability of having the MetS at the EpiHealth examination (P < 0.0001). Dashed lines indicate 95% CI. CI, confidence interval; MetS, metabolic syndrome.

Even closer relationships between change in weight and prevalent MetS were found in the overweight (OR 1.15 for a 1 kg change in weight, 95% CI 1.14–1.17, P < 0.0001) and normal-weight (OR 1.18, 95% CI 1.14–1.21, P < 0.0001) subjects when adjusted for age, sex, BMI at age 20, alcohol intake, smoking, education, and exercise habits. Also in these two groups, BMI at age 20 was related to prevalent MetS (OR 1.33, 95% CI 1.27–1.39, P < 0.0001 in the overweight group and OR 1.39, 95% CI 1.25–1.54, P < 0.0001 in the normal-weight group).

Change in body weight versus different MetS criteria in EpiHealth

The change in weight from age 20 years was significantly related to all of the five MetS criteria when adjusted for age, sex, BMI at age 20, alcohol intake, smoking, education, and exercise habits. Table 4 shows that the relationship was strongest for the waist criteria. The strengths of the associations were not very different between the age- and sex-adjusted models and the models also adjusted for alcohol intake, smoking, education, and exercise habits. In these models, also BMI at age 20 was significantly related to, respectively, MetS criteria (P < 0.0001).

Table 4.

Relationships Between the Change in Weight from Age 20 Years and the Five Metabolic Syndrome Criteria (Binary Variables) Using Two Levels of Adjustments (Age, Sex, and BMI at Age 20), or Multiadjusted in the EpiHealth Study (Age, Sex, BMI at Age 20, Alcohol Intake, Smoking, Education, and Exercise Habits)

	Age, sex, and weight at BMI 20-adjusted		Multiadjusted
Variable	OR (95% CI)	P	OR (95% CI)	P
Blood pressure	1.06 (1.05–1.06)	<0.0001	1.06 (1.05–1.07)	<0.0001
Glucose	1.05 (1.04–1.05)	<0.0001	1.05 (1.04–1.05)	<0.0001
HDL-C	1.06 (1.05–1.07)	<0.0001	1.05 (1.05–1.06)	<0.0001
Waist	1.34 (1.32–1.35)	<0.0001	1.33 (1.31–1.34)	<0.0001
Triglycerides	1.06 (1.06–1.07)	<0.0001	1.05 (1.05–1.06)	<0.0001

The OR are for a 1 kg change in body weight.

CI, confidence interval; OR, odds ratios.

Change in body weight versus prevalent MetS in PIVUS

In PIVUS, the mean weight change from age 20 to 70 years was 14.4 kg (SD 11.6), being similar in men and women (P = 0.92 for difference). A very similar relationship was found for the change in weight since age 20 and MetS prevalence in the PIVUS, as seen in the EpiHealth study when adjusting for sex, BMI at age 20, alcohol intake, smoking, education, and exercise habits (OR 1.08 for a 1 kg change in weight, 95% CI 1.06–1.10, P < 0.0001, Fig. 2), with no evidence of a sex interaction (P = 0.30). Also BMI at age 20 was significantly related to prevalent MetS at age 70 (OR 1.29, 95% CI 1.18–1.40, P < 0.0001).

FIG. 2.

Relationship between the change in weight since age 20 years (including a spline term in the calculations) and probability of having the MetS at the PIVUS examination at age 70 years (P < 0.0001). Dashed lines indicate 95% CI.

Since co-morbidities might have an impact on the relationship between the change in weight and prevalent MetS, we added information on prevalent myocardial infarction, stroke, heart failure, and diabetes to the model, but the estimate (HR) did not change from the model not including these co-morbidities.

We also included data on some commonly used medications at the time of the examination that might change weight [beta-blockers, calcium antagonists, ACE-inhibitors, diuretics, statins, insulin, oral antidiabetics, and hormone replacement therapy (in women only)], but the inclusion of these medications in the model did not have any impact on the estimate (HR) of the relationship between the change in weight and prevalent MetS.

Discussion

The present study showed that the change in body weight from age 20 years to both middle age and older age was a major determinant of prevalent MetS. So was also BMI at age 20. Furthermore, the pronounced impact of the change in body weight from age 20 on MetS was not restricted to those being obese at the examination at middle age or later in life, but was also seen in overweight and normal-weight individuals.

Comparison with the literature

Only a few studies have been devoted to studying the impact of the change in body weight over a prolonged time on MetS. In a cross-sectional study conducted in 664 Japanese men, the change in weight since age 20 years was related to the presence of MetS and all of its components.⁶ This was also seen in normal-weight men. In another cross-sectional study of 3342 men and women, a weight gain since age 20 was associated with increased risk of prevalent MetS even in nonobese subjects later in life.⁵ The present study expands our knowledge in that it included both a cross-sectional part and a longitudinal part and clearly showed that the change in body weight from age 20 to both middle age or older age was a major determinant of MetS using both approaches. Furthermore, the impact of weight change on MetS prevalence was not restricted to the obese, but was even more powerful in the overweight and normal weight.

Weight gain at different times in life

According to the Developmental Origins of Health and Disease (DOHaD) hypothesis,¹⁵ malnutrition or other environmental influences could lead to a low birth weight, but an increased growth during childhood that later could result in obesity and obesity-related diseases. And indeed, a low birth weight has been associated with presence of MetS at adulthood in several studies.^1

–4 Another study supporting this idea was conducted in 280 young adults (18–24 years). More weight gain during the first 3 months of life was related to an increased number of MetS components and presence of MetS in young adulthood.¹⁶

In the present study we evaluated the change in weight from age 20 to mid life and older age. If that is “only” a manifestation of the DOHaD hypothesis or not is not known since we do not have data on birth weight or growth at childhood, but as we found that BMI at age 20 is also related to future MetS development it is likely that a rapid catch-up in weight between birth and age 20 is of importance.

It could also be that different MetS components are sensitive to weight gain at different time points throughout life. In a study of primary care patients between 60 and 75 years of age, the weight gain that was linked to hypertension was mainly seen in the fourth and fifth decades with no relationship between weight at age 20 and hypertension. In contrast, diabetics were more obese already at age 20 and also gained excess weight between 20 and 40 years.¹⁷

Thus, while a DOHaD effect of the MetS is likely to exist, weight gain later in life seems to be of importance for the development of the MetS.

Metabolically healthy obesity

Although the MetS prevalence was higher in obese individuals than in normal-weight and overweight subjects in the EpiHealth sample, the impact of the change in weight from age 20 on MetS was at least as strong in the two nonobese groups as in the obese. Thus, although MHO is characterized by a less pronounced increase in body weight since age 20 than the obese with the MetS, this relationship is not restricted to the obese.

Strengths and limitations

Among the major strengths of the present study is the large cross-sectional EpiHealth study that has an excellent power for this kind of analysis in subgroups of BMI. Another strength is the use of the PIVUS study, which could reproduce the EpiHealth findings also at a higher age.

A drawback of this study is that weight at age 20 is self-reported and therefore less reliable when measured at the clinic. However, the fact that the mean change in weight from age 20 to 70 was very similar in the two samples included gives some confidence that those self-reported data have adequate validity. Any bias introduced by the lack of weight measurements at age 20 would likely drive the associations to the null.

In the EpiHealth study the participant rate is low (20%). This is deleterious if the aim of the study was to investigate the prevalence of a disorder such as MetS. When studying associations between physiological variables, like in the present study, a low participation rate is no major problem if the ranges of the variables are similar to what is seen in the background population, and this is fulfilled in the EpiHealth cohort.

Conclusion

Weight gain from age 20 was strongly and independently associated with prevalent MetS in middle age or old age. Interestingly, this relationship was present also in normal-weight individuals. Our data provide additional support for the importance of maintaining a stable weight throughout life.

Footnotes

Author Disclosure Statement

No conflicting financial interests exist.

References

Barker

, Hales

, Fall

, et al. Type 2 (non-insulin-dependent) diabetes mellitus, hypertension and hyperlipidaemia (syndrome X): Relation to reduced fetal growth. Diabetologia, 1993; 36:62–67.

Levy-Marchal

, Czernichow

. Small for gestational age and the metabolic syndrome: Which mechanism is suggested by epidemiological and clinical studies?. Horm Res, 2006; 65(Suppl 3):123–130.

Meas

, Deghmoun

, Alberti

, et al. Independent effects of weight gain and fetal programming on metabolic complications in adults born small for gestational age. Diabetologia, 2010; 53:907–913.

Valdez

, Athens

, Thompson

, et al. Birthweight and adult health outcomes in a biethnic population in the USA. Diabetologia, 1994; 37:624–631.

Suzuki

, Akamatsu

, Long-term weight gain is related to risk of metabolic syndrome even in the non-obese. Diabetes Metab Syndr, 2014; 8:177–183.

Zhang

, Tamakoshi

, Yatsuya

, et al. Long-term body weight fluctuation is associated with metabolic syndrome independent of current body mass index among Japanese men. Circ J, 2005; 69:13–18.

Karelis

. To be obese—Does it matter if you are metabolically healthy?. Nat Rev Endocrinol, 2011; 7:699–700.

Ahl

, Guenther

, Zhao

, et al. Adiponectin levels differentiate metabolically healthy vs unhealthy among obese and nonobese white individuals. J Clin Endocrinol Metab, 2015; 100:4172–4180.

Hwang

, Hayashi

, Fujimoto

, et al. Visceral abdominal fat accumulation predicts the conversion of metabolically healthy obese subjects to an unhealthy phenotype. Int J Obes (Lond), 2015; 39:1365–1370.

10.

Perez-Martinez

, Alcala-Diaz

, Delgado-Lista

, et al. Metabolic phenotypes of obesity influence triglyceride and inflammation homoeostasis. Eur J Clin Invest, 2014; 44:1053–1064.

11.

Shaharyar

, Roberson

, Jamal

, et al. Obesity and metabolic phenotypes (metabolically healthy and unhealthy variants) are significantly associated with prevalence of elevated C-reactive protein and hepatic steatosis in a large healthy Brazilian population. J Obes, 2015; 2015:178526.

12.

Lind

, Elmståhl

, Bergman

, et al. EpiHealth: A large population-based cohort study for investigation of gene-lifestyle interactions in the pathogenesis of common diseases. Eur J Epidemiol, 2013; 28:189–197.

13.

Lind

, Fors

, Hall

, et al. A comparison of three different methods to evaluate endothelium-dependent vasodilation in the elderly: The Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS) study. Arterioscler Thromb Vasc Biol, 2005; 25:2368–2375.

14.

Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on

Detection

, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). Final report. Circulation, 2002; 106:3143–3421.

15.

Barker

, The origins of the developmental origins theory. J Intern Med, 2007; 261:412–417.

16.

Kerkhof

, Leunissen

, Hokken-Koelega

. Early origins of the metabolic syndrome: Role of small size at birth, early postnatal weight gain, and adult IGF-I. J Clin Endocrinol Metab, 2012; 97:2637–2643.

17.

Jancso

, Halmy

, Rurik

. Differences in weight gain in hypertensive and diabetic elderly patients primary care study. J Nutr Health Aging, 2012; 16:592–596.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.05 MB