The Association Between Diet Quality and Weight Change in Adults Over Time: A Systematic Review of Prospective Cohort Studies

  • Haya Aljadani
  • Amanda Patterson
  • David Sibbritt
  • Clare Collins
Chapter
Part of the Nutrition and Health book series (NH)

Abstract

Obesity is defined as an excess of body fat. Different methods can be used to determine the degree of adiposity, with body mass index (BMI) being a commonly used indirect method [1]. BMI is defined as weight in kilograms divided by the square of height, in metres. According to the WHO BMI classifications, obesity is defined as BMI ≥30 kg/m2, while overweight is defined as 25 ≤ BMI < 30 kg/m2 [2].

Keywords

Diet index Diet Quality Diet score Weight Change Weight Gain Obesity Longitudinal Cohort 

Abbreviations

BMI

Body mass index

DQI

Diet Quality Index

EI

Energy intake

FFQ

Food Frequency Questionnaire

FNRS

Framingham Nutritional Risk Score

MDP

Mediterranean dietary patterns

MDS

Mediterranean Diet Score

rMDS

Revised Mediterranean Dietary Score

TEI

Total energy intake

WC

Waist circumference

Key Points

  • In general women with unhealthy dietary intakes, as assessed by diet quality scoring tools, gain more weight over time compared to men with unhealthy dietary intake patterns.

  • Women with the poorest diet quality gain approximately an additional 300 g/year compared to those assessed as having the highest diet quality.

  • Assessing diet quality using differing diet scores or diet indexes can be used to evaluate the variation in annual weight gain. The strength of the relationship between diet quality and weight gain over time does vary depending on the tool used to assess the relationship.

  • However there are not enough studies using similar methods to allow a thorough examination of this, and further research is required using different diet quality assessment tools and approaches to determine how reliably they are able to predict for future weight gain in adults.

  • Although the current evidence base is not extensive and there is some inconsistency across the findings there is an inverse association between the healthiness of dietary intake and prospective weight gain.

  • More research of high methodological quality is needed to examine other lifestyle behaviours of participants with the poorest diet quality in order to try and identify other factors associated with consumption of unhealthy foods as well as opportunities to tailor interventions to prevent weight gain in adults.

  • Future research studies should consider confounders and interactions between variables when examining the relationship between diet quality and weight gain.

Introduction

Obesity is defined as an excess of body fat. Different methods can be used to determine the degree of adiposity, with body mass index (BMI) being a commonly used indirect method [1]. BMI is defined as weight in kilograms divided by the square of height, in metres. According to the WHO BMI classifications, obesity is defined as BMI ≥30 kg/m2, while overweight is defined as 25 ≤ BMI < 30 kg/m2 [2].

Obesity is associated with a decreased quality of life, increased morbidity and reduced life expectancy [3, 4]. Obesity increases the risk of many chronic diseases, including cardiovascular disease (CVD), diabetes, hypertension, metabolic syndrome and dyslipidaemia [5, 6]. It can also increase the risk of some cancers such as cancer of the breast, oesophagus, pancreas, colon and rectum, endometrium, kidney and potentially the gallbladder [7, 8, 9]. Studies show that obese people have lower levels of high-density lipoprotein (HDL) with higher levels of total cholesterol, triglycerides and apolipoprotein than nonobese people [10]. A relationship between body weight, body fat and bone mineral density (BMD) has also been suggested with some studies finding increased body fat and waist circumference with decreased BMD [11]. One such study examined 398 subjects aged 44.1 ± 14.2 years with a BMI 35.8 ± 5.8 kg/m2 and found a significant inverse correlation between BMI and BMD [12].

Obese women are more likely than nonobese women to experience morbidity and to die prematurely due to the adverse effects of overweight and obesity [7]. The same is true of men, especially men with excess abdominal adiposity, which is associated with premature mortality due to CVD [13], along with an increased prevalence of mental health conditions, such as depression [8].

The costs of the adverse effects of obesity on individuals, the health system and to society are enormous. A 2009 US report has predicted that from 2020 to 2025 about $208 billion will be attributed to the costs of lost worker productivity, morbidity and premature death [14]. Furthermore, 1.5 million life-years will be lost, with the total cost of medical care estimated at $46 billion for this same period [14]. In Australia, a recent report estimated that the annual (2004–2005) total direct cost in health care and non-health care per person increased from $1,472 per year (95 % CI, $1,204, 1,740) for those of normal weight to $2,788 per year (95 % CI, $2,542, 3,035) for those who were obese [15].

The prevalence of obesity has grown rapidly all over the world [16]. A 2006 estimate suggests that obesity affects at least 400 million adults worldwide with an additional 1.6 billion adults (age ≥15 years old) defined as overweight [2]. In Australia, more than half (54 %) of the adults were reported as overweight or obese in the 2004–2005 Australian National Health Survey [17] and this percentage increased to 61.4 % in the 2007–2008 Australian National Health Survey [18].

Today adults are more likely to gain weight at an earlier age than adults in the past [19]. Further, adults currently of an ideal body weight have a 50 % chance of becoming overweight and a 25 % chance of becoming obese over a period of 30 years [20]. Adults gain weight throughout life, especially at specific life stages. For example, men are more likely to gain weight after marriage, whereas women tend to gain weight during pregnancy and during the menopausal transition [21]. Also after giving up smoking and/or changing one’s place of residence, weight gain can occur and has been reported for both genders [21]. The specific causes of weight gain are complex and are due to more than one single factor. Epidemiological and experimental studies demonstrate that nutrition behavioural factors such as diet composition, portion size, types of food and eating patterns, as well as snacking lead to weight gain [19].

Diet is a major modifiable factor. Dietary manipulation assists people to successfully lose weight whereas unhealthy dietary patterns or poor dietary quality can contribute to weight gain [22, 23, 24]. An excessive calorie intake, above total energy requirements, contributes to obesity and overweight [2]. Some studies have reported that excessive dietary fat and protein intakes are associated with higher BMI. One such study followed a cohort of 31,940 healthy women aged 30–55 years for 8 years and found that prior weight loss and younger age were stronger predictors of subsequent weight gain than dietary intake and that calorie intake was significantly related to past weight gain but did not relate to future weight gain [25]. Other studies suggest that higher intakes of dietary fibre, carbohydrate, vegetables, fruits, vitamin C, carotene and caffeine can be inversely related to BMI, but the studies’ findings are mixed [25, 26].

Although major research efforts have been made, the contributions of dietary factors and eating behaviour to the development of excessive body weight have been particularly difficult to identify [27], partly due to the challenges in measuring dietary intake. Further, the contribution of dietary quality has not been studied extensively. Links between dietary intake and weight change are complex because of the multifactorial nature of obesity and overweight, including eating in excess of requirements, together with a poor lifestyle, including lack of physical activity [25, 28]. Further, the environment can affect eating behaviour. For example, people who live near fast-food restaurants have been shown to consume more fast food [29]. On the other hand, people who live near a supermarket tend to eat more healthful foods, especially fruit and vegetables [29]. There are also additional factors that contribute to obesity and overweight such as genetics, ethnicity and age [30, 31].

Little is known about the association between overall or total diet patterns, especially in terms of diet quality, and weight change over time. Diet quality is a measure of the quality of the whole diet and is a concept that aims to assess the quality of an individual’s overall eating patterns using various scores or indexes to assess how closely the individual’s usual diet is aligned with national dietary guidelines [32]. Therefore, the aim of this review is to identify the best available evidence on the association between diet quality and weight change over time, both short and long term, in cohort or case–control studies, and to summarise what is currently known in this area.

To obtain relevant published studies, a search was performed in two stages: the first stage was a search of four databases: MEDLINE, CINAHL, EMBASE and Scopus. This search was conducted to find cohort or case–control studies, published between 1970 and February 2011 in the English language. The second stage involved manual searches to find additional studies by means of the reference lists of all identified reports and articles.

Keywords included those relating to diet in the adult population (e.g. food patterns, eating behaviours, diet quality, diet index, diet variety, diet patterns) and weight status (e.g. weight gain, BMI, weight change, waist circumference).

The studies selected here include studies with a follow-up period in a longitudinal or case–control study and which assessed dietary intake and weight status in adults aged ≥18 years at baseline.

Dietary intake was the exposure variable with a range of methods used to measure dietary intake, including but not limited to 24-h recalls, multiple day dietary records, food frequency questionnaires (FFQ) or diet histories.

Weight status was the primary outcome and was measured by any of the following methods: weight (kg), BMI, waist circumference (WC) or % body fat.

Participants were all healthy adult males and females aged ≥18 years at baseline.

All longitudinal studies were included where the participants were followed up to determine the association between the overall dietary intake and weight change. A number of methods, including scoring techniques or indexes, were used to evaluate diet quality.

The selection of the included studies was carried out in two stages. The first stage was to retrieve all studies which met the previous criteria based on reading the title and abstract. The second stage was to thoroughly examine all of the retrieved papers. Seven studies were identified and assessed by two independent reviewers in the two stages. In the case of a disagreement about study selection in the first stage, that paper was retrieved. If any disagreement occurred in the second stage, a third independent reviewer was consulted to decide whether the paper should be included or not (Fig. 1.1).
Fig. 1.1

Flowchart of studies identified for inclusion in a systematic review of the relationship between diet quality and weight change in adults over time

Study quality was assessed using standardised critical appraisal instruments from the Joanna Briggs Institute Meta Analysis of Statistics Assessment and Review Instrument (JBI-MAStARI).

Figure 1.1 describes the flow of studies meeting the inclusion criteria for this systematic review. Table 1.1 summarises each study’s details (i.e. sample size, country and length of follow-up) and participant baseline characteristics. Table 1.2 reports the data extraction for dietary intake within the included studies and how the diet quality score is derived, including how each component or subscale is scored. Table 1.3 summarises the main outcome data related to absolute change in body weight or BMI during follow-up, the statistical analysis and any confounders for which results were adjusted. Table 1.4 reports the significance of the included studies and Table 1.5 the quality of the included studies.
Table 1.1

Characteristics of the studies that examined diet quality

 

References

Study design and cohort’s name

Population at baseline, N, age and gender

BMI at baseline

Country

F/U

Retention

Purpose

1

Quatromoni et al. [24]

Framingham Offspring and Spouse (FOSS) Study

M, F¨ = 2,245

BMI for M in lowest and highest scores of DQI were 27.4 ± 3.7 and 26.9 ± 3 (kg/m2), respectively; P for trend 0.4

USA

8 years

58 %

To assess the relationship between 8-year weight change among Framingham study participants and adherence to the fourth edition of the Dietary Guidelines measured using DQI

M = 990

And for F were 25.8 ± 5.8 and 25.7 ± 4.3, respectively; p for trend 0.0042

F = 1,255, aged 49–56 years

2

Kimokoti et al. [1]

Framingham Offspring and Spouse (FOSS) Study

M, F¨ = 1,515

Mean BMI = 24.7 ± 4.3; 26.8 ± 3.4 kg/m2 in F and M, respectively

USA

16 years

67 %

(1) Examined patterns of long-term weight change among Framingham men and women over 16 years. (2) Evaluate how diet quality compares with demographic, anthropometric, biological, clinical and other lifestyle factors in predicting weight change in our participants; and (3) impact of these factors, including smoking status, on the association between diet quality and weight change

M = 690

F = 825, aged ≥30 years

3

Wolongevicz et al. [33]

The Framingham Offspring and Spouse (FOSS) Study

F = 590 normal weight (BMI < 25 kg/m2), aged 25–71

Mean BMI across the FNRS tertiles were 22.3, 21.9 and 22 from the lowest to highest tertiles

USA

16 years

100 %

The relationship between diet quality and the development of overweight or obesity in women

4

Beunza et al. [34]

The Seguimiento Universidad de Navarra (SUN) cohort

F, M = 10,376

Mean BMIs were 23.4 ± 3.4, 23.4 ± 3.4, 23.4 ± 3.4 and 23.3 ± 3.4 across baseline MDS from the lowest to highest quartile of MDS

Spain

Mean F/U 5.7 ± 2.2 years

>90 % (for the first 24 months)

The correlation between MDSs and weight gain

Mean age 38 years

5

Sanchez-Villegas et al. [35]

The Seguimiento Universidad de Navarra (SUN) cohort

F, M = 6,319

Mean BMIs were 23.4 ± 3.4, 23.4 ± 3.4, 23.6 ± 3.4 and 23.4 ± 3.4 across baseline MDS from the lowest to highest quartile of MDS

Spain

28 months

90 %

The association between each component at baseline (score_1) and weight or BMI change as outcome. Also overall adherence to the MDP (quartiles of score_1 and outcome)

6

Mendez et al. [36]

The European Prospective Investigation into Cancer and Nutrition (EPIC)—Spain

F = 17,238

BMI of participants was >18 and <30 kg/m2

Spain

Mean of F/U 3.3 years

95 %

To examine whether adherence with MD patterns is associated with obesity incidence for 3 years of follow-up

M = 10,589

Aged 29–65 years

7

Romaguera et al. [37]

The European Prospective Investigation into Cancer and Nutrition—Physical Activity, Nutrition, Alcohol consumption, Cessation of cmoking, Eating cut of come And obesity (EPIC-PANACEA) project

F, M = 373,803

No data given—all categories of BMI included

Denmark, France, Germany, Greece, Italy, Netherlands, Norway, Spain, Sweden and UK

Median F/U 5 years

Unclear

The association between adherence to MDP and weight change and incidence of overweight or obesity

F = 270,348

M = 103,455

Aged 25–70 years

M male, F female, F/U follow-up

Table 1.2

Description of dietary intake methods used in included studies

 

References

Dietary intake method

Diet quality tool

How the score is derived using the diet quality tool

How the diet quality scores are calculated

1

Quatromoni et al. [24]

3 day estimated food records

Diet quality index (DQI)

5-point DQI to assess adherence to key US dietary recommendations

One DQI point was contributed for each of five nutrients if

 (1) Total fat (<30%kcal)

 (2) Saturated fat (<10 % kcal)

 (3) Carbohydrate (>50 % kcal)

 (4) Cholesterol (<300 mg/day)

 (5)Sodium (<2,400 mg/day)

DQI scores ranged from 0 to 5

Participants classified at baseline within gender across the five points of DQI scores

2

Kimokoti et al. [1]

3 day estimated food records

Framingham Nutritional Risk Score (FNRS)

19 nutrients

 (1) Total energy (kJ)

 (2) Protein (% energy)

 (3) Total fat (% energy)

 (4) Monounsaturated fat (% energy)

 (5) Saturated fat (% energy)

 (6) Alcohol (% energy)

 (7) Cholesterol (mg/4,184 kJ)

 (8) Sodium (mg/4,184 kJ)

 (9) Carbohydrate (%energy)

 (10) Polyunsaturated fat (% energy)

 (11) Fibre (g/4,184 kJ)

 (12) Calcium (mg/4,184 kJ)

 (13) Selenium (μg/4,184 kJ)

 (14) Vitamin C (g/4,184 kJ)

 (15) Vitamin B-6 (g/4,184 kJ)

 (16) Vitamin B-12 (μg/4,184 kJ)

 (17) Vitamin E (g/4,184 kJ)

 (18) Folate (μg/4,184 kJ)

 (19) Beta-carotene (μg/4,184 kJ)

FNRS had an overall score computed from the mean score of 19 nutrients for each subjects within each gender

3

Wolongevicz et al. [33]

3 day dietary record

Framingham Nutritional Risk Score (FNRS)

19 nutrients

 (1) Total energy (kJ)

 (2) Protein (% energy)

 (3) Total fat (% energy)

 (4) Monounsaturated fat (% energy)

 (5) Saturated fat (% energy)

 (6) Alcohol (% energy)

 (7) Cholesterol (mg/4,184 kJ)

 (8) Sodium (mg/4,184 kJ)

 (9) Carbohydrate (%energy)

 (10) Polyunsaturated fat (% energy)

 (11) Fibre (g/4,184 kJ)

 (12) Calcium (mg/4,184 kJ)

 (13) Selenium (μg/4,184 kJ)

 (14) Vitamin C (g/4,184 kJ)

 (15) Vitamin B-6 (g/4,184 kJ)

 (16) Vitamin B-12 (μg/4,184 kJ)

 (17) Vitamin E (g/4,184 kJ)

 (18) Folate (μg/4,184 kJ)

 (19) Beta-carotene (μg/4,184 kJ)

Participants were categorised across tertiles of FNRS

4

Beunza et al. [34]

FFQ with 136 food item

MDS-Trichopoulou

There are nine food items—the amount of intake of these food are for M and F, respectively

 (1) Vegetables 550 and 500 (g/day)

 (2) Fruit and nuts 360 and 360 (g/day)

 (3) Legumes 9 and 7 (g/day)

 (4) Cereals, bread and potatoes 180 and 140 (g/day)

 (5) Ratio of monounsaturated fatty acids to saturated fatty acids 1.7 and 1.7 (g/day)

 (6) Moderate alcohol (10–15 g alcohol/day) for M and 50–25 g/day for F)

 (7) Fish 24 and 19 (g/day)

 (8) Meat and poultry 120 and 90 (g/day)

 (9) Dairy products 200 and 190 (g/m)

A maximum of 9 points total from the last components

MDS: low score range, 0–3; moderate score, 4–6; high score, 7–9. Participants were categorised into tertile

4

Beunza et al. [34]

FFQ with 136 food item

The Mediterranean Adequacy Index (MAI) proposed by Alberti-Fidanza et al. (2004)

Sum of percentage total energy from the following typical Mediterranean foods

 (1) Bread

 (2) Cereal

 (3) Legumes

 (4) Potatoes

 (5) Vegetables

 (6) Fruit

 (7) Fish

 (8) Red wine

 (9) Vegetable oils

MAI-Alberti-Fidanza Scores: ≤0.95 (lowest); >0.95 to ≤1.31 (moderate); >1.31 to ≤1.84 (high); >1.84 (highest)

    

Divided by sum of the percentage total energy from the following nontypical Mediterranean foods

 (1) Milk

 (2) Cheese

 (3) Meat

 (4) Eggs

 (5) Animal fat and margarines

 (6) Sweet beverages

 (7) Pastries

 (8) Cookies

 (9) Sugar

The researchers examined the relationship between these indexes and yearly weight change according to categorical (tertiles or quintiles) of indexes scores

   

Mediterranean Diet Quality Index (MDQI) proposed by Scali et al. (2001)

Each of the following items categorised into tertiles and scored from 1 to 3:

MDQI-Scali Scores: ≥11 (lowest); 8–10 (moderate); 5–7 (high); ≤4 (highest)

    

Positive (higher consumption scored 1, medium consumption scored 2, low consumption scored 3)

 (1) Vegetables and fruit (g/day)

 (2) Cereals (bread, pasta, rice and breakfast cereals) (g)

 (3) Olive oil (mL)

 (4) Fish (g)

Negative (higher consumption scored 3, medium consumption scored 2, low consumption scored 1)

 (1) Meat (g)

 (2) Saturated fatty acids (% total energy)

 (3) Cholesterol (mg); the 7 scores were summed (range 0–14): lower score = higher adherence to Mediterranean diet

 
   

Mediterranean dietary pattern (MDP) proposed by Sánchez-Villegas et al. (2002)

Energy-adjusted intakes of the following items:

Positive

 (1) Vegetables (g/day)

 (2) Fruit (g/day)

 (3) Legumes (g/day)

 (4) Cereals incl. bread and potato (g/day)

 (5) Ratio of monounsaturated fatty acids to saturated fatty acids

 (6) Moderate alcohol (g/day)—30 for men, 20 for women (not energy adjusted)

Negative

 (1) Meat and meat products (g/day)

 (2) Milk and dairy products (g/day)

 (3) Trans-fatty acids (g/day)

All items standardised as z-scores then summed (subtracting negative items) and transformed into % of adherence

Score range: 0–100 %

MDP-Sánchez-Villegas Scores: ≤44.23 (lowest); >44.23 to ≤49.47 (moderate); >49.47 to ≤54.97 (high); >54.97 (highest)

   

Diet score (DS) proposed by Panagiotakos et al. (2006)

Food items rated 0–5 (reverse for negative items) according to position in Mediterranean diet pyramid:

Positive

 (1) Vegetables (times/day or month)

 (2) Potatoes (times/day or month)

 (3) Fruit (times/day or month)

 (4) Legumes (times/day or month)

 (5) Non-refined cereals (times/day or month)

 (6) Olive oil (times/day or month)

 (7) Alcohol (times/day or month) (0 for >700 mL/day, 5 for <300 mL/day)

 (8) Fish (times/day or month)

 (9) Poultry (times/day or month)

Negative

 (1) Meat and meat products (times/day or month)

 (2) Full-fat dairy (times/day or month)

Score range: 0–55

MDS-Panagiotakos Scores: ≤29 (lowest); 30–33 (moderate); ≥34 (highest)

   

Mediterranean-Style Dietary Pattern Score (MSDPS) recently proposed by Rumawas et al. (2009)

0–10 rating for each item according to its closeness to the goals of the Mediterranean pyramid:

MSDPS-Rumawas Scores: ≤20 (lowest); >20 to 35 (moderate), >35 (highest)

    

 (1) Vegetables (6 servings/day)

 (2) Potatoes and other starchy roots (3 servings/week)

 (3) Fruit (3 servings/day)

 (4) Legumes, olives and nuts (4 servings/week)

 (5) Wholegrains (8 servings/day)

 (6) Sweets (3 servings/week)

 (7) Dairy products (2 servings/day)

 (8) Eggs (3 servings/week)

 (9) Olive oil (as the only fat)

 (10) Wine (3 glasses/day for M, 1.5 glasses/day for F)

 (11) Fish (6 servings/week)

 (12) Poultry (4 servings/week)

 (13) Meat (1 serving/week)

Further adjustment made according to proportion of total energy intake (TEI) provided by consumption of Mediterranean foods

Score range: 0–100

 

5

Sánchez-Villegas et al. [35]

FFQ with 136 food item

MDS

A maximum of 30 points total from the components

 (1) Cereals (g/day)

 (2) Vegetables (g/day)

 (3) Fruits (g/day)

 (4) Legumes (g/day)

 (5) Fish (g/day)

 (6) Nuts (g/day)

 (7) Olive oil (g/day)

 (8) Moderate red wine consumption (g/day)

 (9) Meat and meat products (g/day)

 (10) Whole-fat dairy products (g/day)

MDS categorised into quartiles: lower (<18); moderate (18–19); high (20–21); highest (≥22)

6

Mendez et al. [36]

Diet history with >600 items

MDS

A maximum of 8 points derived from the following components

 (1) Fish (g/MJ)

 (2) Vegetables (g/MJ)

 (3) Fruits and nuts (g/MJ)

 (4) Legumes (g/MJ)

 (5) Cereals (g/MJ)

 (6) Ratio of monounsaturated fat to saturated fat (g/MJ)

 (7) Moderate ethanol intakes (defined as 5–25 g/day for women, 10–50 g/day for men)

 (8) Meat (g/MJ)

MDS were categorised in three parts: low (0–3), medium (4–5) and high (6–8)

7

Romaguera et al. [37]

FFQ

A revised MDS

There are nine components of this score as follows

 (1) Vegetables

 (2) Legumes

 (3) Fruit and nuts

 (4) Cereals

 (5) Fish and seafood

 (6) Olive oil

 (7) Moderate alcohol consumption

 (8) Meat and meat products

 (9) Dairy products

Each component was measured in g/1,000 kcal to express ED

The rMED was classified into categories to reflect low (0–6 points), medium (7–10 points) or high (11–18 points) adherence to the MDP

Table 1.3

Description of the outcome measures for the included studies

 

References

The outcome

Statistical analysis

Result

Key finding

Confounders

1

Quatromoni et al. [24]

8-year wt gain in lb

Multivariate generalised estimating equations within gender using the mean of DQI score for all participants as the predictor

Participants in the highest DQI quintile gain less wt over 8 years (P for trend <0.01). The mean ± SD of wt gain in F in the highest DQI quintile was 3.3 ± 17.4 lb compared with 8.0 ± 13.0 lb gained in those within the lowest DQI quintile. The mean ± SD of wt gain in M in the highest DQI quintile was 2.7 ± 10.1 lb compared with 5.1 ± 13.3 lb those in the lowest DQI quintile

There is significant inverse association between the mean of DQI score for all participant and weight change in both adults; p = 0.026 and p = 0.008, respectively, for M and F

Age, BMI, physical activity, alcohol intake, smoking cessation, intentional change in diet, postmenopausal, energy intake

Note: Smoking cessation was an important predictor of weight gain, accounting for about a 5- to 9-pound difference in weight gain

2

Kimokoti et al. [1]

Wt change in kg

Multivariable linear regression analysis within gender

Baseline used FRNS as predictors for weight change during a 16-year follow-up

FNRS was not associated with weight gain in both M and F; p = 0.16 and p = 0.61, respectively

F who were former smokers and who were in the lowest tertile of FRNS gained an additional 5.2 kg compared with former smokers who were in the highest tertile of FNRS; P for trend = 0.03

FNRS was not a predictor for wt gain in either M or F

In F: age, wt, physical activity index, FNRS, former smoker and FNRS across smoking category

In M: age, wt, wt fluctuation and former smoking and nonsmokers

3

Wolongevicz et al. [33]

Incidence of overweight or obesity; BMI ≥25 kg/m2

Logistic regression model to estimate odds ratio of being overweight or obese

FNRS was associated with incidence of overweight or obesity (p for trend = 0.009). Women with lower diet quality were significantly more likely to become overweight or obese; OR, 1.76; CI, 95 %(1.16–2.69) times compared with those with highest diet quality

Higher FNRS was associated with a reduced chance of being overweight or obese

Age, physical activity and smoking status

4

Beunza et al. [34]

(1) Annual wt gain in kg

(2) Incidence of wt gain (≥3 kg/year or ≥5 kg/year) during the first 2 or 4 years of follow-up

(1) Multiple linear regressions were used to estimate the annual means change in body weight across categories of adherence the MD

(2) Logistic regression analysis for the first index only and to estimate OR for incidence of weight gain (≥3 or ≥5 kg) during the first 2 or 4 years of follow-up

 OR: CI 95 % (2 years) for ≥3 kg, 0.8 (070, 0.92)

(1) Participants in the highest tertile of MDS had the lowest average yearly wt gain relative to lowest tertile of MDS −059 kg/year; 95 % CI (−0.111, −0.008 kg/year)

(2) OR for incident of wt gain for highest tertile relative to lowest tertile

 OR: CI 95 % (4 years) for ≥3 kg, 0.80 (0.71, 0.91)

 OR: CI 95 % (2 years) for ≥5 kg, 0.76 (0.62, 0.92)

 OR: CI 95 % (4 years) for ≥5 kg, 0.76 (0.64, 0.90)

(1) There was a significant inverse association between all MDS and wt change

(2) Those in the highest tertile of MDS were less likely to have absolute wt gain relative to those in the lowest tertile

Age and sex, BMI, physical activity, sedentary behaviour, smoking, snacking and TEI

4

Beunza et al. [34]

(1) Annual wt gain in kg

(1) Multiple linear regressions were used to estimate the annual means change in body weight across categories of adherence the MD

MAI-Alberti-Fidanza Scores >1.84 (highest) had the highest mean wt gain −0.077 (−0.131, −0.022) compared with the lowest ≤0.95

There was a significant inverse association between all MDS and wt change except MDS-­Panagiotakos and MSDPS-Rumawas with p for trend 0.30, 0.41, respectively while all other MDS had p for trend <0.05

Age and sex, BMI, physical activity, sedentary behaviour, smoking, snacking and TEI

    

MDQI-Scali Scores ≤4 (highest) had the highest mean wt gain −0.102 (−0.194, −0.009) compared with ≥11 (lowest)

MDP-Sánchez-Villegas Scores >54.97 (highest) gained −0.061 (−0.116, −0.006) compared with ≤44.23 (lowest)

MDS-Panagiotakos ≥34 (highest) gained −0.029 (−0.079, 0.021) compared with ≤29 (lowest)

MSDPS-Rumawas Scores >35 (highest) gained −0.028 (−0.094, 0.038) compared with ≤20 (lowest)

  

5

Sánchez-Villegas et al. [35]

Wt in kg

(1) Linear regression models were used to assess the association between MD scores at baseline and wt change and BMI change during F/U time

(2) Logistic regression to examine the association between baseline of MDS and incidence of overweight or obesity during the F/U time

(1) Participants in the highest quartile of MDS gained less weight (+0.65 kg) (+0.59 to +0.80) compared with those in the lowest quartile (p for trend 0.291)

(2) Participant in the highest quartile of MDS had smaller increase in BMI +0.23 (+0.12 to +0.33) compared with those in the lowest quartile of MDS (p for trend 0.279)

There was not a significant association between MDS and weight gain

Age, gender, BMI, smoking, physical activity, alcohol, EI, change in dietary habits (fruit, vegetables, meat, meat product, fish, olive oil and alcohol) and change in physical activity during F/U time

6

Mendez et al. [36]

(1) Incidence of obesity (BMI ≥30 kg/m2)

(2) Incidence of overweight (BMI ≥25 to <30 kg/m2)

Logistic regression models were used to estimate odds of becoming overweight or obese

(1) Overweight participants in the highest MDS were less likely to become obese

OR (95 % CI) of becoming obese for F and M, respectively

 +0.69, (0.54–0.89)

 +0.68, (0.53–0.89)

(2) There was no significant association between MDS and incidence of overweight

OR (95 % CI) of becoming overweight for F and M, respectively

 +0.99 (0.78, 1.25)

 +1.11 (0.81–1.52)

(1) There was inverse significant association between MDS and becoming obese

(2) High MD adherence was not associated with overweight incidence in women and men

Underreporting of dietary data, age, physical activity, education, centre, height, smoking status, season, follow-up time, changes in employment status during follow-up, use of special diets, parity and menopause in women and history of chronic diseases (cancer, diabetes or heart diseases) at baseline or follow-up

7

Romaguera et al. [37]

(1) Wt gain in 5-year F/U

(2) Overweight or obesity incidence in 5-years of F/U (BMI ≥25 kg/m2)

(1) Multiple linear regression between rMED and 5-year wt gain

(2) Logistic regression to examine the association between a 2-point increase in rMED and becoming overweight

(1) Those in the highest tertile of rMED had less wt gain −0.16 kg (CI 95 %: −0.24, −0.07) and −0.04 kg (−0.07, −0.02) for the combined results

(2) Overall results showed that a 2-point increase in the rMED was associated with becoming overweight after compounding (?) the result from all the cohorts

OR (CI 95 %): 0.97 (0.95, 0.99)

There is a significant inverse association between MDP and becoming overweight or obese

Sex, age, baseline BMI, follow-up time, educational, physical activity, smoking, TEI and misreporting of EI

WT weight, EI energy intake

Table 1.4

Conclusions and significance of included studies

 

References

Was there an inverse relationship between diet quality and body weight change?

Were the results significant?

1

Quatromoni et al. [24]

Y

Y

2

Kimokoti et al. [1]

N

N

However, there was inverse correlation between women who quit smoking in the lowest FNRS and higher weight gain in women only

Y in former smokers in the lowest FNRS tertile gain more weight than other in highest tertile

3

Wolongevicz et al. [33]

Y

Y

4

Beunza et al. [34]

Y in 6 diet quality indexes

Y in 4 indexes

N in two indexes

5

Sanchez-Villegas et al. [35]

Y

N

6

Mendez et al. [36]

Y of becoming obese

Y of becoming obese among the overweight only

N of becoming overweight

N of becoming overweight

7

Romaguera et al. [37]

Y

Y

Table 1.5

Quality assessment for included studies

 

References

(1) Is the sample representative of patients in the population as a whole?

Was everyone assessed at the same follow-up time?

Are the confounding factors identified and strategies to deal with them stated?

Are outcome assessed using objective criteria?

Was follow-up carried out over a sufficient time period?

Were the people who withdrew described and included in the analysis?

Were the outcome measured in reliable way?

Was appropriate statistical analysis used?

The total

1

Quatromoni et al. [24]

U

Y

Y

Y

Y

U

Y

Y

2U 6Y

2

Kimokoti et al. [1]

U

Y

Y

Y

Y

U

Y

Y

2U, 6Y

3

Wolongevicz et al. [33]

U

Y

Y

Y

Y

U

Y

Y

2U, 6Y

4

Beunza et al. [34]

N

N

Y

N

Y

U

Y

Y

1U, 3N, 4Y

5

Sanchez-Villegas et al. [35]

N

Y

Y

N

Y

U

Y

Y

1U, 2N, 5Y

6

Mendez et al. [36]

Y

Y

Y

N

Y

U

Y

Y

1U, 1N, 6Y

7

Romaguera et al. [37]

N

N

Y

N

Y

U

N

Y

1U, 4N, 2Y

 

The total

3U, 1Y, 3N

1N, 6Y

7Y

3Y 4N

7Y

7U

1N, 6Y

7Y

 

U unclear, N no, Y yes

Of 2,304 studies originally identified, seven met all inclusion criteria. These examined the association between overall diet quality and weight change, BMI or obesity incidence and used different methodologies to evaluate dietary patterns. One study [24] used the diet quality index (DQI), two studies [1] used the Framingham Nutritional Risk Score (FNRS) and four studies used different scoring methods to assess adherence to the Mediterranean dietary patterns (MDP).

The total number of participants across the two studies which used FNRS was 2,105 adults (≈67 % female) [1, 33]. Both of these studies were derived from the Framingham Offspring and Spouse Study (FOSS) in the USA, with 16 years of follow-up. The only study that used DQI was also derived from FOSS with 60 % of the participants being female [24]. Two studies were derived from the Seguimiento Universidad de Navarra (SUN) cohort, and different scoring methods were used to evaluate a number of Mediterranean Diet Scoring (MDS) approaches in these studies [34, 35]. Three studies were carried out in Spain [34, 35, 36]. Romaguera et al.’s [37] analysing data from the European Prospective Investigation into Cancer—Physical Activity, Nutrition, Alcohol consumption, Cessation of smoking, Eating out of home And obesity (EPIC- PANACEA) Project derived from six cohorts that were established in ten European countries to study the association between adherence to MDP and weight change, and incidence of overweight or obesity. Mean participant age across all included studies was ≥25 years and retention rates varied from 58 to 100 % (Table 1.1).

Dietary intake methods used in the seven included studies were 3-day estimated dietary records [21, 32, 33], FFQ [34, 35] and diet history [37] (Table 1.2).

Methods Used to Measure Diet Quality

To evaluate diet quality, the studies examined used a number of different indexes and scores which are described in detail in Table 1.2 and summarised below.

Quatromoni et al. [24] used the DQI to judge the quality of overall dietary intake, with a higher DQI representing greater adherence to the US national dietary guidelines. The DQI score ranges from zero to five, and the five points constitute adherence to percentage energy intake from total fat, saturated fat and carbohydrate (3 points), and the total intakes of cholesterol and sodium (2 points) [24]. The intake levels were set according to the US dietary guidelines [24] and are given in Table 1.2. Each DQI component is awarded a score of either zero or one. If consumption of the nutrient is optimal and within the recommended limits, then a score of one is awarded, otherwise it is scored as zero [38]. Therefore, higher DQI scores are associated with lower total fat, saturated fat, cholesterol and sodium intakes and higher intakes of carbohydrate, representing greater adherence to the Dietary Guidelines for Americans. Thus, it reflects a dietary pattern that contains food rich in carbohydrate and fibre such as fruits and vegetables and lower intakes of foods that are high in sodium and fat [21].

Two [32, 33] studies used the FNRS, which is based on intakes of 19 nutrient components (Table 1.2). These nutrients are classified into three groups according to their relationship with CVD risk as follows: (1) optimal intake profile of selected macronutrients including energy, protein, monounsaturated fat and polyunsaturated fat; (2) increased risk-related nutrients including total fat, saturated fat, cholesterol, alcohol and sodium; (3) protective nutrients including carbohydrate, dietary fibre, calcium, selenium, vitamin C, vitamin B-6, vitamin B-12, vitamin E, folate and beta-carotene [39]. The nutrients included in FNRS are scored in such a way that a person with a more desirable nutrient profile is awarded a lower score. For example, a lower fat or higher vitamin and mineral intake will attract a lower score. Similarly a less desirable nutrient profile is given a higher score, e.g. higher total fat or lower micronutrient intake. Higher monounsaturated fat intake was given a higher score, because the source of these fats was mostly animal products (e.g. beef fat) for participants in the Framingham study [1, 33]. The rank given to individual nutrients was aggregated to give an overall composite risk rank [1].

The Mediterranean Diet Score (MDS) was compiled by Trichopoulou et al. [40]. All four studies included in this review that used the MDS [34, 35, 36, 37] used different scoring methods to calculate adherence to the Mediterranean Diet Pattern (MDP). Each one referenced the MDS, even though they varied in the actual method. In general, the Mediterranean Diet Pattern includes regular consumption of vegetables, fruit, legumes, cereals, nuts, fish and foods rich in olive oil, as well as a low in intake of foods high in saturated fat, dairy products, meat and poultry [40].

The MDS as used in the paper by Beunza et al. [34] considered the intake of nine food items (Table 1.2), which were classified into (1) positive components (vegetables, fruit, nuts, legumes, cereals, moderate alcohol and fish) and (2) negative components (meat and poultry, dairy products) [34]. The score assigned to each component was based on both the nature of the food component and the quantity of that food consumed in relation to the median value for all subjects under consideration [41]. A score of zero was assigned to a positive component if the individual’s consumption was less than the gender-specific median consumption [42]. A score of one was awarded if the food item was positive and consumption was greater than the gender-specific median. The opposite procedure was followed in case of negative components. A score of zero was awarded if the consumption was greater than the gender-specific median and one if consumption of the negative component was less than the relevant median [34, 42].

The MDS as used by Sanchez-Villegas et al. [35] was defined a priori and considered the intake of ten food item components. These foods were classified into two groups: (1) beneficial food items which included cereals, vegetables, fruits, legumes, fish, nuts, olive oil and moderate red wine and (2) detrimental food items which included meat, meat products and whole-fat dairy products [35]. The total score range was 10–30 points. To compute the score for each participant, a rank system was applied to each component. First, each item was classified into tertiles and scored from 1 to 3 from the lowest to the highest tertile for beneficial food items and then 1–3 from the highest to lowest tertile for detrimental food components. The ten components scores were summed for each participant. Thus, a maximum of 30 points reflects the highest MDS and the greatest adherence to MDP, while a score of 10 points reflects the lowest adherence [35].

The MDS used by Mendez et al. [36] collected the dietary intake data from a diet history, which generated a list of approximately 600 food items. This MDS score is similar to a previous one [34] and classified food items into (1) beneficial foods including fish, vegetables, fruits, legumes, cereals and the ratio of monounsaturated saturated fat and (2) detrimental food items, including moderate alcohol intake and meat. Milk and dairy products were not considered in this index. A score of zero was assigned to a component if it was beneficial and if the individual’s consumption was less than the gender-specific median consumption of that particular food for all individuals in the study. The opposite procedure was followed in the case of detrimental foods, with a score of zero awarded if consumption was more than the gender-specific median and a score of one given if consumption was less than the relevant median [36].

Researchers for the Romaguera et al. [37] paper used a revised Mediterranean Diet Pattern [43] evaluated using a Relative Mediterranean Diet Score (rMed). This rMed differs from the original MDS and comprises nine nutritional components which characterise MDP. Components like vegetables, legumes, fruit, nuts, cereals, seafood and fish, moderate alcohol consumption and olive oil are rated as beneficial foods, while meat and meat products and dairy products are rated as detrimental. All food items in this score are expressed in units as grams/1,000 kcal. Each component in this index is scored as tertiles, except for alcohol and olive oil. Scores of 0, 1 and 2 were given from the lowest to the highest tertile for the beneficial food items. The detrimental components of meat/meat products and dairy were given 0, 1 or 2 from the highest to the lowest tertile. Thus, higher intakes of beneficial foods and lower intakes of detrimental foods contributed more to the score, reflecting greater adherence to Mediterranean patterns. For olive oil, a zero score was given to nonconsumers, 1 to those with consumption below the median level and 2 to those with consumption greater than or equal to the median. Regarding alcohol, a score of 2 was given to those with a moderate alcohol consumption, ranging from 10 to <50 g/day for men and 5 to <25 g/day for women. Consumption outside of this range for alcohol scored zero. The rMed score ranged from 0 (lowest adherence to MDP) to 18 (highest adherence to MDP). Further evaluation classified the score 0–6 as low adherence, 7–10 as medium and 11–18 as high adherence to MDP [43].

The Implications of Diet Quality on Weight Change

This systematic review aimed to synthesise the best evidence available on the relationship between diet quality and weight change in adults within cohort studies. There were only seven studies included that assessed dietary quality using a dietary quality score or index from an initial search generating 2,304 citations from a comprehensive search of the four most relevant databases, specific for the research question. This review indicates that the relationship between diet quality and weight gain is important. Using the DQI tool, the mean diet quality score at baseline was shown to be a strong predictor of prospective weight gain [24]. Another study [34] found that the risk of having a specific amount of weight gain (≥3 or ≥5 kg) during follow-up periods of 2 and 4 years was higher among those in the lowest tertile of MDS compared to those in the highest tertile. Moreover, there were significant associations between annual weight gain in adults using all four indexes of MDP [34]. However, in the same study, the authors also evaluated another two indexes to generate MDPs and found that there were no relationships between these two indexes and annual weight gain [34]. In addition, two other diet quality indexes [1, 35] found no relationship between diet quality and weight gain over time. More specifically, evaluation of FNRS at baseline with prospective weight gain within gender demonstrated a strong interaction with smoking status in both genders [1]. In this study women who quit smoking and also had the poorest diet quality, as evaluated by FNRS, had higher weight gain than those in the highest FRNS tertile. A significant association was found between having lower diet quality, as determined by FNRS or rMDS, and a higher risk of becoming overweight or obese [33, 37] Moreover, a significant risk of becoming obese was demonstrated among those who were overweight and in the lowest tertile of diet quality score, but there was no greater risk of becoming overweight among those with the poorest diet quality [36].

Quality of the Studies

From Table 1.5 it was identified that most studies have confounders and although this is a potential risk of bias, the majority adjusted for at least the major confounders in the statistical analysis. It was unclear from the population descriptions whether some of the researchers used a representative sample or not, but at least three studies used non-representative population samples. Three studies reported analyses were conducted in the USA [21, 32, 33] from the same cohort and three studies were carried out in Spain [34, 35, 36], so although they may each be representative within countries, internationally they are not.

Limitations of the Studies

Within the included studies, the reporting of confounders and statistical adjustment for them is the most common weakness. The most commonly reported confounders were similar and included age, BMI, physical activity, education and income (Table 1.3).

One of the most important confounders that is rarely addressed is changes in dietary intake during the follow-up period, especially when follow-up times are extensive, as there is no guarantee of stability of dietary intake and behaviours over time. This means that weight gain could also have been influenced by other factors apart from baseline diet. There are other important confounders, such as misreporting of energy intake, and the only study to adjust for underreporting was Mendez et al. [36].

Further, use of a 3-day food record has been identified as commonly associated with underreporting of dietary intake [24].

The finding of this systematic review demonstrates that within cohort studies examining weight change over time, dietary intake has not commonly been categorised using a diet quality score or dietary index in adults. Due to the small number of studies in this area, further research is needed. Moreover, this is an important field warranting more research because diet is an important determinant of weight change. In addition, there is limited research identifying which are the most useful tools to assess diet quality as a predictor of weight change in adults within longitudinal studies.

The results from the studies in Table 1.1 are heterogeneous and use different methodologies to evaluate both dietary intake and the outcome of body weight. In general, four studies out of the total of seven found an inverse association between diet quality and weight change, both as absolute weight gain or as a change in weight status category. This means that consuming a higher quality diet, or one which is consistent with healthier eating habits and dietary intakes, that aligns more closely with national dietary guidelines does lead to smaller amounts of weight gain. However, two studies found this association in multivariate models only among specific groups of participants who were overweight at baseline or women who were former smokers and also had the poorest dietary intake as scored by FNRS or MDS [1, 36] (Table 1.4).

The first three studies in Table 1.1 [1, 24, 33] were evaluations using differing methodologies but performed on the same FOSS cohort. Each used different criteria for including participants; different methods to assess diet quality (DQI or FNRS), varied in how the weight change was defined; different methods to report the weight change outcomes; and differing approaches to the statistical analysis. However, across the studies, women with the poorest diet quality gained an additional 1,040 g compared to men with the poorest dietary intake over time [24].

On average, adults in the lowest diet quality category gain additional weight (59–1,090 g) compared to those who had eating patterns consistent with higher diet quality over follow-up periods of 1–8 years duration.

Conclusions

In conclusion, the current evidence is insufficient to set benchmarks for optimal diet quality in order to prevent weight gain. However, aspiring to diet quality scores that reflect adherence to national dietary guidelines appears prudent. Although we cannot determine whether lower diet quality, as assessed by scores or indexes, could definitely lead to higher weight gain due to the limitations of the evidence in this area currently, it is clear from the available evidence that there is a strong association between higher weight gain and poor diet quality. Further, smoking status, especially in women, and baseline BMI in adults were important confounders that should be considered in future studies, along with any interactions between them and dietary intake.

Recommendations for Researchers

Further studies are required that use a number of different methods within the one cohort to measure diet quality as predictors of weight gain in adults. This approach would ideally be repeated across a number of nationally representative population cohorts to examine how robust any relationships are internationally or whether any of the approaches is more applicable.

Although there have been a considerable number of studies that have examined the relationship between macronutrient or energy intakes and weight change over time, it is clear from this systematic review that there is a lack of studies examining overall diet quality or total dietary patterns. Such an approach would better characterise the relationship between weight change and diet quality in adults over time and facilitate development of food-based guidelines targeting the prevention of weight gain.

References

  1. 1.
    Kimokoti RW, Newby PK, Gona P, Zhu L, Jasuja GK, Pencina MJ, et al. Diet quality, physical activity, smoking status, and weight fluctuation are associated with weight change in women and men. J Nutr. 2010;140:1287–93.PubMedCrossRefGoogle Scholar
  2. 2.
    World Health Organization. Obesity and overweight. http://www.who.int/mediacentre/factsheets/fs311/en/. Accessed Mar 2011.
  3. 3.
    Kushner RF, Foster GD. Obesity and quality of life. Nutrition. 2000;16:947–52.PubMedCrossRefGoogle Scholar
  4. 4.
    Peeters A, Barenderegt JJ, Willekens F, Johan PM, Mamun AA, Bonneux L. Obesity in adulthood and its consequences for life expectancy: a life-table analysis. Ann Intern Med. 2003;138:24–32.PubMedCrossRefGoogle Scholar
  5. 5.
    Cameron AJ, Dunstan DW, Owen N, Zimmet PZ, Barr EL, Tonkin AM, et al. Health and motality consequences of abdominal obesity: evidence from AusDiab study. Med J Aust. 2009;191:202–8.PubMedGoogle Scholar
  6. 6.
    Lijing LY, Daviglus ML, Liu K, Stamler J, Wang R, Pirzada A, et al. Midlife body mass index and hospitalization and mortality in older age. JAMA. 2006;295:290–8.Google Scholar
  7. 7.
    Hu FB. Overweight and obesity in women: health risks and consequences. J Womens Health. 2003;12:163–72.CrossRefGoogle Scholar
  8. 8.
    Sanchez-Villegas A, Adriano MP, Beunza JJ, Guillen-Grima F, Toledo E, et al. Childhood and young adult overweight/obesity and incidence of depression in the SUN Project. Obesity. 2010;18:1443–8.PubMedCrossRefGoogle Scholar
  9. 9.
    World Cancer Research Fund International/American Institute for Cancer Research. Food, nutrition, physical activity, and the prevention of cancer: a global perspective. http://www.dietandcancerreport.org/downloads/chapters/chapter_08.pdf. Accessed Apr 2011.
  10. 10.
    Hu D, Hannah J, Gray RS, Jablonski KA, Henderson JA, Robbins DC, et al. Effects of obesity and body fat distribution on lipids and lipoproteins in nondiabetic American Indians: the Strong Heart Study. Obesity. 2000;8:411–21.CrossRefGoogle Scholar
  11. 11.
    Lee DC, Kim KC, Shin DH, Lee SY, Im JA. Relation between obesity and bone mineral density and vertebral fractures in Korean postmenopausal women. Yonsei Med J. 2010;51:857–63.PubMedCrossRefGoogle Scholar
  12. 12.
    Greco EA, Fornari R, Rossi F, Santiemma V, Prossomariti G, Annoscia C, et al. Is obesity protective for osteoporosis? Evaluation of bone mineral density in individuals with high body mass index. Int J Clin Pract. 2010;64:817–20.PubMedCrossRefGoogle Scholar
  13. 13.
    Larsson B, Svardsudd K, Welin L, Wilhelmsen L, Bjorntorp P, Tibblin G. Abdominal adipose tissue distribution, obesity, and risk of cardiovascular disease and death: 13 year follow up of participants in the study of men born in 1913. Br Med J. 1984;288:1401–4.CrossRefGoogle Scholar
  14. 14.
    Lightwood J, Bibbins-Domingo K, Coxson P, Wang C, Williams L, Goldman L. Forecasting the future economic burden of current adolescent overweight: an estimate of the Coronary Heart Disease Policy Model. Am J Public Health. 2009;99:2230–7.PubMedCrossRefGoogle Scholar
  15. 15.
    Colagiuri S, Lee CMY, Colagiuri R, Magliana D, Shaw JE, Zimmet PZ, et al. The cost of overweight and obesity in Australia. Med J Aust. 2010;192:260–4.PubMedGoogle Scholar
  16. 16.
    Kolotikin RL, Meter K, Williams GR. Quality of life and obesity. Obes Rev. 2001;2:219–29.CrossRefGoogle Scholar
  17. 17.
    Australian Burean of Statistics. Overweight and obesity in adults, Australia, 2004–05. (ABS4719.0). http://www.abs.gov.au. Accessed May 2011.
  18. 18.
    Australian Burean of Statistics. National Health Survey: summary of results 2007–2008. (ABS 4364.0). http://www.abs.gov.au. Accessed May 2011.
  19. 19.
    McCrory MA, Suen VMM, Roberts SB. Biobehavioral influences on energy intake and adult weight gain. J Nutr. 2002;132:3830S–4.PubMedGoogle Scholar
  20. 20.
    Tang JW, Kushner RF, Thompson J, Baker DW. Physician counseling of young adults with rapid weight gain: a retrospective cohort study. BMC Fam Pract. 2010;11:31.PubMedCrossRefGoogle Scholar
  21. 21.
    Yanovski JA, Yanovski SZ, Sovik KN, Nguyen TT, O’Neil PM, Sebring NG. A prospective study of holiday weight gain. N Engl J Med. 2000;342:861–7.PubMedCrossRefGoogle Scholar
  22. 22.
    Drapeau V, Despras J-P, Bouchard C, Allard L, Fournier G, Leblanc C, et al. Modifications in food-group consumption are related to long-term body-weight changes. Am J Clin Nutr. 2004;80:29–37.PubMedGoogle Scholar
  23. 23.
    Foreyt JP, Goodrick GK. Evidence for success of behavior modification in weight loss and control. Ann Intern Med. 1993;119:698–701.PubMedCrossRefGoogle Scholar
  24. 24.
    Quatromoni PA, Pencina M, Cobain MR, Jacques PF, D’Agostino RB. Dietary quality predicts adult weight gain: findings from the Framingham Offspring Study. Obesity. 2006;14:1383–91.PubMedCrossRefGoogle Scholar
  25. 25.
    Colditz GA, Willett WC, Stampfer MJ, London SJ, Segal MR, Speizer FE. Patterns of weight change and their relation to diet in a cohort of healthy women. Am J Clin Nutr. 1990;51:1100–5.PubMedGoogle Scholar
  26. 26.
    Bes-Rastrollo M, Martinez-Gonzalez MA, Sanchez-Villegas A, de la Fuente Arrillaga C, Martinez JA. Association of fiber intake and fruit/vegetable consumption with weight gain in a Mediterranean population. Nutrition. 2006;22:504–11.PubMedCrossRefGoogle Scholar
  27. 27.
    Schulz M, Kroke A, Liese AD, Hoffmann K, Bergmann MM, Boeing H. Food groups as predictors for short-term weight changes in men and women of the EPIC-Potsdam cohort. J Nutr. 2002;132:1335–40.PubMedGoogle Scholar
  28. 28.
    Fogelholm M, Kukkonen-Harjula K. Does physical activity prevent weight gain—a systematic review. Obes Rev. 2000;1:95–111.PubMedCrossRefGoogle Scholar
  29. 29.
    Kapinos KA, Yakusheva O. Environmental influences on young adult weight gain: evidence from a natural experiment. J Adolesc Health. 2011;48:52–8.PubMedCrossRefGoogle Scholar
  30. 30.
    Loos RJF, Bouchard C. Obesity—is it a genetic disorder? J Intern Med. 2003;254:401–25.PubMedCrossRefGoogle Scholar
  31. 31.
    Roberts SB, Williamson DF. Causes of adult weight gain. J Nutr. 2002;132:3824S–5.PubMedGoogle Scholar
  32. 32.
    Wirt A, Collins CE. Diet quality–what is it and does it matter? Public Health Nutr. 2009;12:2473–92.PubMedCrossRefGoogle Scholar
  33. 33.
    Wolongevicz DM, Zhu L, Pencina MJ, Kimokoti RW, Newby PK, D’Agostino RB, et al. Diet quality and obesity in women: the Framingham Nutrition Studies. Br J Nutr. 2010;103:1223–9.PubMedGoogle Scholar
  34. 34.
    Beunza JJ, Toledo E, Hu FB, Bes-Rastrollo M, Serrano-Martínez M, Sánchez-Villegas A, et al. Adherence to the Mediterranean diet, long-term weight change, and incident overweight or obesity: the Seguimiento Universidad de Navarra (SUN) cohort. Am J Clin Nutr. 2010;92:1484–93.PubMedCrossRefGoogle Scholar
  35. 35.
    Sanchez-Villegas A, Bes-Rastrollo M, Martinez-Gonzalez M, Serra-Majem L. Adherence to a Mediterranean dietary pattern and weight gain in a follow-up study: the SUN cohort. Int J Obes. 2006;30:350–8.Google Scholar
  36. 36.
    Mendez MA, Popkin BM, Jakszyn P, Berenguer A, Tormo MJ, Sanchez MJ, et al. Adherence to a Mediterranean diet is associated with reduced 3-year incidence of obesity. J Nutr. 2006;136:2934–8.PubMedGoogle Scholar
  37. 37.
    Romaguera D, Norat T, Vergnaud A, Mouw T, May AM, Agudo A, et al. Mediterranean dietary patterns and prospective weight change in participants of the EPIC-PANACEA project. Am J Clin Nutr. 2010;92:912–21.PubMedCrossRefGoogle Scholar
  38. 38.
    Drewnowski A, Henderson SA, Driscoll A, Rolls BJ. The Dietary Variety Score: assessing diet quality in healthy young and older adults. J Am Diet Assoc. 1997;97:266–71.PubMedCrossRefGoogle Scholar
  39. 39.
    Kimokoti RW, Newby PK, Gona P, Zhu L, Campbell WR, D’Agostino RB, et al. Stability of the Framingham Nutritional Risk Score and its component nutrients over 8 years: the Framingham Nutrition Studies. Eur J Clin Nutr. 2012;66(3):336–44.PubMedCrossRefGoogle Scholar
  40. 40.
    Trichopoulou A, Costacou T, Bamia C, Trichopoulos D. Adherence to a Mediterranean diet and survival in a Greek population. N Engl J Med. 2003;348:2599–608.PubMedCrossRefGoogle Scholar
  41. 41.
    Willett WC, Sacks F, Trichopoulou A, Drescher G, Ferro-Luzzi A, Helsing E, et al. Mediterranean diet pyramid: a cultural model for healthy eating. Am J Clin Nutr. 1995;61:1402S–6.PubMedGoogle Scholar
  42. 42.
    Simopoulos AP, Visioli F. More on Mediterranean diets. New York: Basel/Karger; 2007.Google Scholar
  43. 43.
    Romaguera D, Norat T, Mouw T, May AM, Bamia C, Slimani N, et al. Adherence to the Mediterranean diet is associated with lower abdominal adiposity in European men and women. J Nutr. 2009;139:1728–37.PubMedCrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2013

Authors and Affiliations

  • Haya Aljadani
    • 1
    • 2
  • Amanda Patterson
    • 1
  • David Sibbritt
    • 3
  • Clare Collins
    • 1
  1. 1.School of Health SciencesThe University of NewcastleNewcastleAustralia
  2. 2.University of King Abdul-AzizJeddahSaudi Arabia
  3. 3.Centre for Clinical Epidemiology and Biostatistics, School of Medicine and Public HealthThe University of NewcastleNewcastleAustralia

Personalised recommendations