Hostname: page-component-78c5997874-lj6df Total loading time: 0 Render date: 2024-10-31T23:16:46.233Z Has data issue: false hasContentIssue false

Lipid, protein and carbohydrate intake in relation to body mass index: an Italian study

Published online by Cambridge University Press:  01 March 2007

Giorgia Randi
Affiliation:
Istituto di Ricerche Farmacologiche ‘Mario Negri’, Milan, Italy
Claudio Pelucchi
Affiliation:
Istituto di Ricerche Farmacologiche ‘Mario Negri’, Milan, Italy
Silvano Gallus*
Affiliation:
Istituto di Ricerche Farmacologiche ‘Mario Negri’, Milan, Italy
Maria Parpinel
Affiliation:
Unit of Hygiene and Epidemiology, DPMSC – School of Medicine, University of Udine, Udine, Italy
Luigino Dal Maso
Affiliation:
Epidemiology Unit, Centro di Riferimento Oncologico, Aviano, Pordenone, Italy
Renato Talamini
Affiliation:
Epidemiology Unit, Centro di Riferimento Oncologico, Aviano, Pordenone, Italy
Livia SA Augustin
Affiliation:
Department of Nutritional Sciences, Faculty of Medicine, University of Toronto and Clinical Nutrition and Risk Factor Modification Centre, St Michael's Hospital, Toronto, Ontario, Canada
Attilio Giacosa
Affiliation:
Policlinico di Monza, Monza, Italy
Maurizio Montella
Affiliation:
Istituto Tumori ‘Fondazione Pascale’, Cappella dei Cangiani, Naples, Italy
Silvia Franceschi
Affiliation:
International Agency for Research on Cancer, Lyon, France
Carlo La Vecchia
Affiliation:
Istituto di Ricerche Farmacologiche ‘Mario Negri’, Milan, Italy Istituto di Statistica Medica e Biometria, University of Milan, Milan, Italy
*
*Corresponding author: Email gallus@marionegri.it
Rights & Permissions [Opens in a new window]

Abstract

Objective

To analyse the association between macronutrient intake and body mass index (BMI).

Design

A series of hospital-based case–control studies.

Settings

Selected teaching and general hospitals in several Italian regions.

Subjects

A total of 6619 subjects from the comparison groups of the case–control studies were included in the analysis.

Methods

We obtained data from a validated 78-item food-frequency questionnaire submitted between 1991 and 2002. For various macronutrients, the partial regression coefficient (variation of BMI (kg m− 2) per 100 kcal increment of energy intake) was derived from multiple linear regression models, after allowance for age, study centre, education, smoking habits, number of eating episodes and mutual adjustment for macronutrients.

Results

BMI was directly associated with protein intake among women only (β = 0.68) and with unsaturated fats in both genders (for monounsaturated fats β = 0.27 for men and 0.26 for women; for polyunsaturated fats β = 0.27 for men and 0.54 for women), and inversely related to carbohydrates (β = − 0.05 for men and − 0.21 for women) and number of eating episodes in both genders (β =  − 0.42 for men and − 0.61 for women) and to saturated fats among women only (β =  − 0.57).

Conclusions

These results confirm and provide convincing evidence that, after allowance for selected covariates including total energy intake, a protein-rich diet is not inversely related to BMI, and a carbohydrate-rich diet is not directly related to BMI.

Type
Research Article
Copyright
Copyright © The Authors 2007

Overweight and obesity are precursors of several diseases, including cardiovascular disorders, diabetes mellitus and also cancer at different sitesReference Kopelman1. The mechanism may be direct or through the pathways of the metabolic syndrome, a series of manifestations that often characterise obesity and include glucose intolerance, insulin resistance, dyslipidaemia and hypertensionReference Eckel, Grundy and Zimmet2.

It is still unclear how to control energy intake in order to achieve and maintain a reduction of body mass index (BMI). In particular, the most appropriate combination of macronutrients in the diet has not yet been identifiedReference Astrup3. Low-carbohydrate diets (e.g. Atkins and Protein Power) propose a reduction of carbohydrate intake to 5–10% of total energy, and an increase of protein intake up to 30% and of fats up to 65%Reference Freedman, King and Kennedy4. Such diets have been reported to achieve a decrease of ad libitum caloric intake, possibly due to the anorexic effect of proteinsReference Weigle, Breen, Matthys, Callahan, Meeuws and Burden5, and a clinically significant weight loss which, however, is associated with the duration of the diet and the restriction of energy intake, not with carbohydrate restriction itselfReference Astrup, Meinert Larsen and Harper6. More important, subsequent long-term effects on nutritional status and body composition after such diets have not been documentedReference Astrup, Meinert Larsen and Harper6.

A study by Trichopoulou et al. Reference Trichopoulou, Gnardellis, Benetou, Lagiou, Bamia and Trichopoulos7 conducted in Greece between 1994 and 1999 found a positive association between protein intake and BMI, in both men and women. That study also suggested that neither fats (saturated or monounsaturated) nor carbohydrates play a major role in BMI determination. Likewise, the findings from the US Prostate Cancer Prevention Trial showed that energy from fat but not from carbohydrates was associated with obesity in a cohort of 15 266 middle-aged and elderly menReference Satia-Abouta, Patterson, Schiller and Kristal8. Further, a randomised trial on 34 individuals followed for 12 weeks showed that a high-carbohydrate diet consumed ad libitum, with no attempt at energy restriction or change in energy intake, results in losses of body weight and body fat for both gendersReference Hays, Starling, Liu, Sullivan, Trappe and Fluckey9.

In order to analyse the relationship between BMI and macronutrient intake in another Mediterranean population, we repeated the analysis of Trichopoulou et al.7 in a data set collected in Italy between 1991 and 2002.

Methods

Overall, 6619 subjects (3090 men and 3529 women; median age 58, range 17–82 years) were included in the present analysis. These were derived from the comparison groups of a series of case–control studies on cancer at several sites and recruited by trained interviewers from several teaching and general hospitals in four Italian areas: greater Milan and the provinces of Pordenone, Gorizia, Padua and Forlì (North-eastern Italy); the province of Genova (North-western Italy); the province of Latina (Central Italy); and the urban area of Naples (Southern Italy)Reference Gallus, Talamini, Giacosa, Montella, Ramazzotti and Franceschi10.

These subjects were admitted to hospitals for a wide spectrum of acute non-neoplastic conditions, unrelated to known or potential risk factors for cancers: 23% had traumatic conditions (mostly fractures and sprains), 31% non-traumatic orthopaedic disorders (mostly lower back pain and disc disorders), 18% acute surgical conditions (mostly abdominal, such as acute appendicitis or strangulated hernia) and 28% other miscellaneous illnesses such as eye, ear, nose, throat and dental disorders. Fewer than 4% of subjects approached refused to be interviewed, and response rates did not vary across hospitals and geographical areas.

In the hospital setting, trained interviewers administered to the patients a structured questionnaire which included information on age, education and other socio-economic factors, occupational and leisure-time physical activity, smoking habits, alcohol intake and coffee consumption. The subjects' diet during the 2 years before hospital admission was investigated through a validated and reproducible 78-item food-frequency section of the questionnaireReference Decarli, Franceschi, Ferraroni, Gnagnarella, Parpinel and La Vecchia11Reference Franceschi, Barbone, Negri, Decarli, Ferraroni and Filiberti13. We used an Italian food composition database to estimate energy and nutrient intakeReference Salvini, Parpinel, Gnagnarella, Maisonneuve and Turrini14. A detailed section of the questionnaire was prepared to collect information on self-reported height and weight measures 1 year prior to the interview. BMI was computed as weight divided by the square of height (Kg m-2).

Data analysis

Multiple linear regression was applied to assess the effects of macronutrients and ethanol on BMI. We considered two models separately for men and women. The first model included adjustments for age, study centre, education, smoking habits, total energy intake and number of eating episodes per day, and the second one further mutual adjustments for the other macronutrients, including alcohol. Both analyses were performed for all the subjects and for the subset of subjects who had never been on a weight reduction diet.

Results

Almost 20% of men reported frequent or irregular dieting, while the percentage among women was 31%. Median values of BMI (kg m-2) were 26.0 for men and 24.8 for women, and were 25.7 and 24.0, respectively, after exclusion of subjects who had been dieting.

Table 1 shows the cut-off points of quartiles of total energy intake (kJ) and of percentage of energy intake from specific macronutrients, including alcohol, among men and women. On average, carbohydrates accounted for ~47% of energy intake (median value: 46.5% for men and 48.3% for women), lipids for 30% (29.4 and 32.3%), proteins for 15% (14.5 and 16.4%), and alcohol varied greatly between genders (8.9% in men and 0.8% in women). After exclusion of subjects who had been dieting, the dietary nutrient composition was similar, but the cut-offs for energy intake were higher.

Table 1 Cut-off points of quartiles of total energy intake (in kJ) and of percentage of energy intake (macronutrient/total energy intake %) from specific macronutrients by gender*

* Note that the values denote the cut-off points of quartiles of energy intake from the different nutrients. Therefore, the figures in the columns do not add up to 100% and the sums increase, as they should, from the first to the third cut-off of quartiles.

Table 2 shows the partial regression coefficients and the corresponding 95% confidence intervals giving a change of BMI for a 418.4 kJ (100 kcal) increment of energy intake for each macronutrient according to the two models applied to all subjects (upper part of the table) and then to subjects who had never been dieting (lower part of the table), separately among men and among women. Proteins were not associated with BMI among men: the estimates were − 0.06 among all men and − 0.01 for men with no restricted dietary habits, after mutual adjustment for other macronutrients. A positive association was shown between proteins and BMI among women in all models. The estimates varied from 0.68 among all women to 0.58 among women with no restricted dietary habits, after mutual adjustments for other macronutrients. Carbohydrates were moderately and inversely associated with BMI in all models for both genders. The coefficients in the second model were − 0.05 and − 0.06 among men and − 0.21 and − 0.16 among women. Saturated fats were inversely associated with BMI among women only, after mutual adjustment for other macronutrients. The estimates were − 0.57 for all women and − 0.54 for women with no restricted dietary habits. No association was found among men. Unsaturated fats were positively associated with BMI in all models for both genders. Estimates for monounsaturated fats were 0.27 both among all men and after exclusion of men who have been on a diet; corresponding values for women were 0.26 and 0.36. Similarly, for polyunsaturated fats, the corresponding estimates were 0.27 and 0.20 among men and 0.54 and 0.38 among women. Alcohol consumption was moderately associated with BMI among men (β = 0.09) and among women who had not been dieting (β = 0.13). The number of eating episodes per day was inversely related to BMI for both genders. The estimates varied between − 0.42 and − 0.40 for men, and − 0.61 and − 0.14 for women, respectively among all subjects and after exclusion of subjects who had been dieting.

Table 2 Partial regression coefficients (β) and 95% confidence intervals (95% CI) of variation of body mass index by an increase of 418.4 kJ (100 kcal) of energy intake from selected macronutrients (bold font is used for statistically significant estimates)

* Adjusted for age, study centre, education, smoking habit, total energy intake and number of eating episodes per day.

Adjusted for age, study centre, education, smoking habit, number of eating episodes per day and mutual adjustment for other macronutrients (proteins, carbohydrates, and saturated, monounsaturated and polyunsaturated lipids).

Mutual adjustment model included vegetable and animal proteins, sugars and starch, and saturated, monounsaturated and polyunsaturated lipids.

Discussion

Our study showed that a protein-rich diet was associated with BMI only among women, in broad agreement with some studies from GreeceReference Trichopoulou, Gnardellis, Benetou, Lagiou, Bamia and Trichopoulos7 and the USAReference Satia-Abouta, Patterson, Schiller and Kristal8, Reference Hays, Starling, Liu, Sullivan, Trappe and Fluckey9. When we separated animal and vegetal proteins, vegetable proteins were inversely associated with BMI among men only. Although vegetable proteins may offer metabolic advantages over animal proteinsReference Jenkins, Kendall, Marchie, Faulkner, Wong and de Souza15, there is as yet no evidence suggesting that vegetable proteins per se may affect body weight differently from animal protein.

Inverse associations were found between BMI and carbohydrates, but our results do not show a different effect according to the type of carbohydrates, as suggested by a recent study where a positive association was found between BMI and the glycaemic index (GI), but not with total carbohydrate intakeReference Ma, Olendzki, Chiriboga, Hebert, Li and Li16.

The inverse association between the number of eating episodes per day and BMI is in agreement with previous results by Ma et al. Reference Ma, Bertone, Stanek, Reed, Hebert and Cohen17, supporting the hypothesis that increasing the number of eating episodes per day is inversely related to obesity, independently from energy intake and nutrient composition. Increasing the frequency of meals mimics a low GI diet and has been shown to produce similar metabolic effects to those of low GI dietsReference Jenkins, Kendall, Augustin and Vuksan18.

BMI was directly associated with mono- and polyunsaturated fats in both genders, but inversely with saturated fat among women only, in agreement with the report by Trichopoulou et al. Reference Trichopoulou, Gnardellis, Benetou, Lagiou, Bamia and Trichopoulos7 Monounsaturated fats (such as olive oil) are the main type of fats consumed in Italy, and it is possible that an overconsumption of these fats could result in weight gain.

Alcohol intake was directly associated with BMI in this study, but inversely in the Greek studyReference Trichopoulou, Gnardellis, Benetou, Lagiou, Bamia and Trichopoulos7. This apparent discrepancy may be due to higher alcohol consumption in Italy, especially among men (median values of percentage of energy from alcohol was 8.9% for Italian and 3.2% for Greek men).

Physical activity is inversely related to overweightReference Eckel, Grundy and Zimmet2. In order to consider its potential effect, we repeated the analysis adjusting for occupational and leisure-time physical activity, but estimates were not materially different.

The strengths and weaknesses of the present study include those typical of a cross-sectional design, which discloses the degree of the association between BMI and macronutrients, but may leave open issues on causal interpretation. Furthermore, subjects admitted to hospital may not be representative of a healthy population with respect to BMIReference Zang and Wynder19. Our interviewers, however, specifically asked for average weight in the 2 years before admission to hospital and medical records could be checked to verify this. Moreover, great attention was paid in excluding subjects with chronic conditions leading to long-term modifications of diet. On the other hand, it is known that subjects tend to underestimate weight and overestimate heightReference Niedhammer, Bugel, Bonenfant, Goldberg and Leclerc20Reference Villanueva22, but different self-reporting of weight and height across the strata of the various nutrients considered is unlikely. Moreover, distributions by height and weight in our subjects were similar to those from population surveys in ItalyReference Gallus, Colombo, Scarpino, Zuccaro, Negri and Apolone23. BMI is a measure of body mass uncorrelated to heightReference Benn24 and is used to measure overweight (BMI ≥ 25 kg m− 2) and obesity (BMI ≥ 30 kg m− 2) according to standard criteria25. BMI does not, however, distinguish between body fat and lean body mass. The positive association between protein and BMI might therefore represent a direct association between protein intake and lean mass, since the same BMI could be shared by people with different body fat content and distribution. However, on a population level, other measurements of body fat, including waist circumference, percentage body fat mass, waist-to-height ratio and waist-to-hip ratio, had little advantage over BMI in the prediction of obesity-related metabolic riskReference Wannamethee, Shaper, Morris and Whincup26, Reference Bosy-Westphal, Geisler, Onur, Korth, Selberg and Schrezenmeir27.

Among other strengths of this study are the large number of subjects analysed and the collection of extensive dietary information using a validated food-frequency questionnaireReference Decarli, Franceschi, Ferraroni, Gnagnarella, Parpinel and La Vecchia11Reference Franceschi, Barbone, Negri, Decarli, Ferraroni and Filiberti13.

In conclusion, the results found in this Italian population support those reported by a Greek studyReference Trichopoulou, Gnardellis, Benetou, Lagiou, Bamia and Trichopoulos7, and provide evidence that a diet rich in protein is not inversely related to BMI in Mediterranean populations.

Acknowledgements

The authors thank Mrs I Garimoldi for editorial assistance. This work was conducted with the contribution of the Italian Association for Cancer Research (AIRC/FIRC), the Italian League Against Cancer and the Italian Ministry of Education (PRIN 2005).

References

1Kopelman, PG. Obesity as a medical problem. Nature 2000; 404: 635–43.CrossRefGoogle ScholarPubMed
2Eckel, RH, Grundy, SM, Zimmet, PZ. The metabolic syndrome. Lancet 2005; 365: 1415–28.CrossRefGoogle ScholarPubMed
3Astrup, A. The satiating power of protein – a key to obesity prevention? American Journal of Clinical Nutrition 2005; 82: 12.CrossRefGoogle ScholarPubMed
4Freedman, MR, King, J, Kennedy, E. Popular diets: a scientific review. Obesity Research 2001; 9(Suppl. 1): 1S40S.Google Scholar
5Weigle, DS, Breen, PA, Matthys, CC, Callahan, HS, Meeuws, KE, Burden, VR, et al. . A high-protein diet induces sustained reductions in appetite, ad libitum caloric intake, and body weight despite compensatory changes in diurnal plasma leptin and ghrelin concentrations. American Journal of Clinical Nutrition 2005; 82: 41–8.Google Scholar
6Astrup, A, Meinert Larsen, T, Harper, A. Atkins and other low-carbohydrate diets: hoax or an effective tool for weight loss? Lancet 2004; 364: 897–9.CrossRefGoogle ScholarPubMed
7Trichopoulou, A, Gnardellis, C, Benetou, V, Lagiou, P, Bamia, C, Trichopoulos, D. Lipid, protein and carbohydrate intake in relation to body mass index. European Journal of Clinical Nutrition 2002; 56: 3743.CrossRefGoogle ScholarPubMed
8Satia-Abouta, J, Patterson, RE, Schiller, RN, Kristal, AR. Energy from fat is associated with obesity in US men: results from the Prostate Cancer Prevention Trial. Preventive Medicine 2002; 34: 493501.Google Scholar
9Hays, NP, Starling, RD, Liu, X, Sullivan, DH, Trappe, TA, Fluckey, JD, et al. . Effects of an ad libitum low-fat, high-carbohydrate diet on body weight, body composition, and fat distribution in older men and women: a randomized controlled trial. Archives of Internal Medicine 2004; 164: 210–7.CrossRefGoogle ScholarPubMed
10Gallus, S, Talamini, R, Giacosa, A, Montella, M, Ramazzotti, V, Franceschi, S, et al. . Does an apple a day keep the oncologist away? Annals of Oncology 2005; 16: 1841–4.CrossRefGoogle Scholar
11Decarli, A, Franceschi, S, Ferraroni, M, Gnagnarella, P, Parpinel, MT, La Vecchia, C, et al. . Validation of a food-frequency questionnaire to assess dietary intakes in cancer studies in Italy. Results for specific nutrients. Annals of Epidemiology 1996; 6: 110–8.CrossRefGoogle ScholarPubMed
12Franceschi, S, Negri, E, Salvini, S, Decarli, A, Ferraroni, M, Filiberti, R, et al. . Reproducibility of an Italian food frequency questionnaire for cancer studies: results for specific food items. European Journal of Cancer 1993; 29: 2298–305.CrossRefGoogle Scholar
13Franceschi, S, Barbone, F, Negri, E, Decarli, A, Ferraroni, M, Filiberti, R, et al. . Reproducibility of an Italian food frequency questionnaire for cancer studies. Results for specific nutrients. Annals of Epidemiology 1995; 5: 6975.CrossRefGoogle ScholarPubMed
14Salvini, S, Parpinel, M, Gnagnarella, P, Maisonneuve, P, Turrini, A. Banca Dati di Composizione degli Alimenti per Studi Epidemiologici in Italia. Milano: European Institute of Oncology, 1998.Google Scholar
15Jenkins, DJ, Kendall, CW, Marchie, A, Faulkner, DA, Wong, JM, de Souza, R, et al. . Direct comparison of a dietary portfolio of cholesterol-lowering foods with a statin in hypercholesterolemic participants. American Journal of Clinical Nutrition 2005; 81: 380–7.CrossRefGoogle ScholarPubMed
16Ma, Y, Olendzki, B, Chiriboga, D, Hebert, JR, Li, Y, Li, W, et al. . Association between dietary carbohydrates and body weight. American Journal of Epidemiology 2005; 161: 359–67.Google Scholar
17Ma, Y, Bertone, ER, Stanek, EJ 3rd, Reed, GW, Hebert, JR, Cohen, NL, et al. . Association between eating patterns and obesity in a free-living US adult population. American Journal of Epidemiology 2003; 158: 8592.CrossRefGoogle Scholar
18Jenkins, DJ, Kendall, CW, Augustin, LS, Vuksan, V. High-complex carbohydrate or lente carbohydrate foods? American Journal of Medicine 2002; 113(Suppl. 9B): 30S–7S.CrossRefGoogle ScholarPubMed
19Zang, EA, Wynder, EL. The association between body mass index and the relative frequencies of diseases in a sample of hospitalized patients. Nutrition and Cancer 1994; 21: 247–61.Google Scholar
20Niedhammer, I, Bugel, I, Bonenfant, S, Goldberg, M, Leclerc, A. Validity of self-reported weight and height in the French GAZEL cohort. International Journal of Obesity and Related Metabolic Disorders 2000; 24: 1111–8.CrossRefGoogle ScholarPubMed
21Spencer, EA, Appleby, PN, Davey, GK, Key, TJ. Validity of self-reported height and weight in 4808 EPIC-Oxford participants. Public Health Nutrition 2002; 5: 561–5.CrossRefGoogle ScholarPubMed
22Villanueva, EV. The validity of self-reported weight in US adults: a population based cross-sectional study. BMC Public Health 2001; 1: 11.Google Scholar
23Gallus, S, Colombo, P, Scarpino, V, Zuccaro, P, Negri, E, Apolone, G, et al. . Overweight and obesity in Italian adults 2004, and an overview of trends since 1983. European Journal of Clinical Nutrition, 2006; 60: 1174–9.CrossRefGoogle Scholar
24Benn, RT. Some mathematical properties of weight-for-height indices used as measures of adiposity. British Journal of Preventive and Social Medicine 1971; 25: 4250.Google ScholarPubMed
25World Health Organization (WHO). Obesity: Preventing and Managing the Global Epidemic. Technical Report Series No. 894. Geneva: WHO, 2000.Google Scholar
26Wannamethee, SG, Shaper, AG, Morris, RW, Whincup, PH. Measures of adiposity in the identification of metabolic abnormalities in elderly men. American Journal of Clinical Nutrition 2005; 81: 1313–21.CrossRefGoogle ScholarPubMed
27Bosy-Westphal, A, Geisler, C, Onur, S, Korth, O, Selberg, O, Schrezenmeir, J, et al. . Value of body fat mass vs anthropometric obesity indices in the assessment of metabolic risk factors. International Journal of Obesity (London) 2006; 30: 475–83.CrossRefGoogle ScholarPubMed
Figure 0

Table 1 Cut-off points of quartiles of total energy intake (in kJ) and of percentage of energy intake (macronutrient/total energy intake %) from specific macronutrients by gender*

Figure 1

Table 2 Partial regression coefficients (β) and 95% confidence intervals (95% CI) of variation of body mass index by an increase of 418.4 kJ (100 kcal) of energy intake from selected macronutrients (bold font is used for statistically significant estimates)