Unknown

The effect of diets high in protein or carbohydrate oninflammatory markers in overweight subjects A. Due,1 S. Toubro,1 S. Stender,2 A. R. Skov1 and A. Astrup11Department of Human Nutrition, Centre for Advanced Food Studies, The Royal Veterinary and Agricultural University,Rolighedsvej, Frederiksberg C, Denmark2Department of Clinical Biochemistry, Gentofte Hospital, Copenhagen University Hospital, Denmark Aim: Increased levels of inflammatory markers have been associated with weight gain and cardiovascular disease.
The aim of the present study was to investigate the effect of diets high in either carbohydrate or protein on theinflammatory markers C-reactive protein (CRP), haptoglobin and transferrin in plasma after weight loss.
Methods: Fifty overweight subjects [age: 18–56 years, body mass index (BMI): 27–34] were randomly assigned to anad libitum, fat-reduced diet (30% of energy, E%), either high in protein and low in carbohydrate (25 E and 45 E%respectively) or high in carbohydrate and low in protein (58 E and 12 E% respectively), during 6 months of strictlycontrolled dietary intervention with dietary counselling.
Results: An average reduction of 25% in CRP and an average increase of 20% in haptoglobin and transferrin wereseen in both groups, however, these changes were not significant. In cross-sectional analyses after the interventionCRP was associated with fat mass (r ¼ 0.323, p ¼ 0.03), and the changes in CRP were associated with various indices ofbody fatness (Dbody weight r ¼ 0.346, p ¼ 0.02). Changes in body fatness were positively associated with Dtransferrin(r ¼ 0.344, p ¼ 0.02) and nearly significantly associated with Dhaptoglobin (r ¼ 0.271, p ¼ 0.07) after 6 months. Multipleregression analysis showed no associations between dietary protein and carbohydrate content and serum CRP,haptoglobin or transferrin concentrations, and this remained unaltered after adjustment for weight change.
Conclusion: Dietary carbohydrate/protein ratio has no effect on inflammatory markers, but the study confirmed thatbody fatness is positively associated with levels of serum CRP.
Keywords: ad libitum diet, C-reactive protein, high-protein diet, inflammatory markers, low-carbohydrate diet, obesity, overweight Received 03 November 2003; returned for revision 11 June 2004; revised version accepted 28 June 2004 Arteriosclerosis is one of the major causes of morbidity tory markers such as sensitive C-reactive protein (CRP) and mortality in humans in industrialized countries.
have been shown to be independently predictive of future High levels of plasma total cholesterol are associated cardiovascular events [1,2]. A longitudinal study with 8 with increased vascular risk, and low-density lipoprotein years follow-up, including 27 939 apparently healthy (LDL)-cholesterol is currently the focus of guidelines women, found that CRP was a stronger predictor of CVD for the determination of the risk factor of cardiovascular than LDL [3]. Other acute phase proteins, ssuch as trans- diseases (CVDs). However, mild elevations of inflamma- ferrin and haptoglobin have also been shown to be Correspondence:Arne Astrup, Department of Human Nutrition, Centre for Advanced Food Studies, The Royal Veterinary and Agricultural University,Rolighedsvej 30, 1958 Frederiksberg C, Denmark.
E-mail:ast@kvl.dk Diabetes, Obesity and Metabolism, 7, 2005, 223–229 The effect of diets high in protein or carbohydrate involved in inflammation [4]. Like CRP, haptoglobin is a Table 1 Physical characteristics of subjects at baseline in positive acute phase protein as the plasma concentration the high-carbohydrate, low-protein (HCLP) and high- increases in inflammation, whereas transferrin is a negative acute phase protein, as plasma concentration decreases ininflammation [4]. CRP has been shown to decrease signifi- cantly after weight loss [5–7], whereas the influence on transferrin and haptoglobin is still unknown.
High dietary intakes of saturated fats and cholesterol contribute to the elevation of plasma cholesterol [8].
Furthermore, a diet high in protein (25% of total energy) alters cardiovascular risk profiles in humans favourably compared to a diet with a normal protein level (12% of *Number of smokers in group (number of cigarettes smoked per day).
energy), lowering total cholesterol, LDL and triacylgly- There were no differences between groups by t-test.
cerols (TGs) and raising high-density lipoprotein (HDL)-cholesterol levels [8]. Little is known about the effect ofdietary intake on inflammatory markers. The glycaemic load of a diet has been shown to be positively associated A 50-square-metre shop with approximately 900 differ- with CRP [9], while high-fibre cereal foods [10] and ent food items was built at the Department of Human phytoestrogens [11] have not been seen to have an effect Nutrition. All food was collected from the shop during on CRP. However, intakes of long-chain polyunsaturated the first 6 months. A computerized system was devel- fatty acids derived from fish have been found to be oped for registration of the selection made at each visit to inversely associated with CRP [12], and moderate alco- the shop. All food items were bar-coded, and a dietician hol consumption has shown to reduce the concentration scanned the barcodes of all the chosen items to monitor macronutrient composition of the diet. When necessary, We have previously reported that an ad libitum the dietician assisted the subjects in altering the selec- high-protein, low-carbohydrate diet (HPLC) induces a tion made to meet the prescribed macronutrient com- greater weight loss than an ad libitum high-carbohydrate, position. The calculated energy content of the chosen low-protein diet (HCLP) during 6 months of dietary inter- groceries was not made known to the subjects. The diet- vention [14]. Based on the data from this study concen- ary intake at baseline was assessed by a 7-day dietary tration of inflammatory markers (CRP, haptoglobin and record. Compliance to the diets was measured by 24-h urinary nitrogen excretion. The experimental shop facil-ity and compliance are described in detail in earlierpublications [14,15].
Fifty overweight and obese (BMI, 27–34 kg/m2) men (n ¼ 12) and women (n ¼ 38), aged 18–56 years, were randomlyassigned to a fat-reduced diet (30% of energy, E%) either Body weight (BW) was measured, with subjects wearing high in protein and in carbohydrate (25 E and 45 E% only light clothing, weekly in both intervention groups respectively) or high in carbohydrate and low in protein on a decimal scale (Seca model 707, Copenhagen, (58 E and 12 E% respectively). The characteristics of the Denmark). Body composition was determined by a dual subjects are given in table 1. Macronutrient composition of energy X-ray absorptiometry (DXA) scanning (Hologic the diet in the intervention groups was strictly controlled 1000/W, software version 5.61; Hologic, Waltham, MA, but energy intake was ad libitum. The study was conducted USA) at baseline and after 6 months of dietary intervention.
as a 6-month, strictly controlled, dietary intervention with Subjects wore only underwear and cotton T-shirt during the provision of food from a purpose-built shop and with regu- scan. Fat mass (FM), fat-free mass and intra-abdominal adipose tissue (IAAT) were estimated from DXA scans.
Approval was obtained from the Municipal Ethical IAAT by the equation: IAAT (cm2) ¼ À208.2 þ 4.62 (sagittal Committee of Copenhagen and Frederiksberg. The study diameter, cm) þ 0.75 (age, years) þ 1.73 (waist, cm) þ 0.78 was performed in accordance with the Helsinki II (trunk fat%), given by Treuth et al. [16]. Waist and hip Declaration, and each subject signed an informed consent circumferences were measured using a tape measure, with document before the study commenced.
the subjects disrobed and in a standing posture.
Diabetes, Obesity and Metabolism, 7, 2005, 223–229 The effect of diets high in protein or carbohydrate The carbohydrate-E% increased in the HCLP group andremained unchanged in the HPLC group, and the differ- Blood samples were drawn into test tubes, for glucose ence between the groups was significant (p < 0.0001).
analysis into a tube containing EDTA/fluoride, and for After 6 months, the decrease in BW was 3.5 kg (0.8, 6.2) insulin, blood lipids and the inflammatory markers ana- and the loss of FM was 3.4 kg [1,2,5,7] higher in the HPLC lyses into tubes with no additives. The tubes (except for group than in the HCLP group (table 2).
insulin analyses) were kept on ice. Plasma glucose wasanalyzed by standard enzymatic methods (BoehringerMannheim GmbH Diagnostica, Copenhagen, Denmark).
Insulin concentrations in serum were measured usingradio immunoassay (AutoDelfi Automatic Immunoassay Serum Concentrations of Inflammatory Markers, System, Wallac Oy, Turku, Finland). Homeostasis model assessment-insulin resistance (HOMA-IR) was calcu- At baseline, the mean CRP concentration was 2.7 mg/l (2.1–3.3) and 32.6% of subjects had a CRP concentration glucose (mmol/l)/22.5 [17]. Methods of analysis of TG, >3 mg/l. After 6 months of dietary intervention both HDL and cholesterol are described elsewhere [14]. LDL the HCLP and the HPLC group experienced a non- concentration was calculated: LDL ¼ Total cholesterol significant decrease in CRP of 0.8 mg/l (28%) (p ¼ 0.09) À HDL À (TG/2.2) [18]. CRP concentration was measured and 0.5 mg/l (21%) (p ¼ 0.06) respectively. The differ- using the CRP (Latex) ultra sensitive assay. Haptoglobin ence between the groups was not significant (table 3).
and transferrin were analyzed using the immuno- After 6 months, both haptoglobin and transferrin had turbidimetric method (Roche Diagnostic Systems, Basel, increased in both groups by 17% (p < 0.05) and 21% Switzerland). Serum for analyses of the inflammatory (p < 0.001) respectively. No differences between the markers was stored in a deep freezer at À80 C until the groups were found. These results remained unaltered analyses were conducted in the spring of 2002.
after adjustment for weight changes.
At baseline no differences were seen in fasting con- centrations of glucose, insulin, HOMA-IR or any of theblood lipids. After 6 months no significant changes in Forty-six subjects were included in the statistical ana- these blood concentrations were seen either within or lyses, as two subjects in each group dropped out during between the groups, except for a decrease in HDL the intervention. All results are given as means with 95% confidence intervals. Normal distribution was not CRP was associated with FM (r ¼ 0.323, p < 0.05) after always achieved, but the homogeneity of variance was 6 months, and the changes (Á) in CRP during the inter- achieved. Differences within groups were tested by vention were associated with ÁBW (r ¼ 0.346, p < 0.05).
paired t-test and between groups by analysis of variance (ANOVA) with baseline values and body weight changes as (r ¼ 0.382, p < 0.01) and IAAT (r ¼ 0.292, p < 0.05) at base- covariates when relevant. Correlations analyses were line, and BW (r ¼ 0.346, p < 0.05), waist circumference done by Pearson’s correlation coefficient for baseline, (r ¼ 0.323, p < 0.05), IAAT (r ¼ 0.398, p ¼ 0.01) and CRP 6 months values and changes from baseline to 6 months (r ¼ 0.493, p ¼ 0.001) after 6 months. Transferrin was separately. Stepwise multiple regression analysis was correlated to cholesterol (r ¼ 0.361, p < 0.05) and HDL performed to examine independent predictors of the (r ¼ 0.420, p < 0.01) and negatively correlated to body inflammatory markers. The level of significance was set weight (r ¼ À0.285, p ¼ 0.05) at baseline. After 6 months, at p < 0.05. All statistical tests were performed by SPSS changes in transferrin were positively correlated with version 11.0 (SPSS Chicago, IL, USA).
Multiple regression analyses showed that differences in ÁBW and ÁHDL explained 16% of the variation inÁCRP (r2 ¼ 0.164, p < 0.01) and that haptoglobin was Food Intake and Body Weight and Composition correlated with CRP at 6 months (r2 ¼ 0.226, p < 0.001).
During the 6 months of dietary intervention the protein- Correlation and multiple regression analyses showed E% increased in the HPLC group, while it decreased in no influence of protein-E%, carbohydrate-E%, fat-E%, the HCLP group, resulting in a significant difference in intake of energy or any of the blood parameters on CRP, protein-E% between the two groups (p < 0.0001) (table 2).
haptoglobin or transferrin after 6-month intervention.
Diabetes, Obesity and Metabolism, 7, 2005, 223–229 The effect of diets high in protein or carbohydrate Table 2 Changes from baseline in dietary composition and anthropometrical measures in the high-carbohydrate, low-protein(HCLP) and high-protein, low-carbohydrate (HPLC) groups after 6-month dietary intervention Values are means (95% CI); BMI, body mass index; FM, fat mass; IAAT, Intra-abdominal adipose tissue; NS, not significant; W/H, waist-to-hipratio.
*Test for difference of changes within groups by paired t-test.
†p-values were derived by ANOVA with baseline values as covariates.
‡Energy intake is adjusted for alcohol.
§p < 0.05.
††p < 0.0001.
The present findings indicate that the substitution of diet- low statistical power, and to the fact that a large number ary carbohydrate with protein for 6 months has no influ- of the subjects were still overweight (87%) or obese (24%) ence on the concentration of the inflammatory markers after 6 months of dietary intervention. In a similar study, CRP, haptoglobin or transferrin. At baseline, average CRP a 20% reduction in CRP was found after a mean weight was higher in both intervention groups compared to the loss of 8 kg induced by 16-week intervention of either general population. CRP usually exists at very low con- a high or medium protein diet [24]. Other studies have centrations in plasma, with 90% of individuals having a reported a similar (i.e. 26%) reduction in CRP after 7.9 kg CRP < 3 mg/l. This could be due to the obese state of the weight loss following 12 weeks of energy restriction [7] subjects, as obesity is associated with a higher concentra- and an impressive 70% reduction after 30.7 kg weight tion of CRP [7,19–23]. However, despite a significant loss following gastric banding [25]. In the present trial weight loss, the decrease in CRP was only borderline CRP was positively correlated with FM, which is in agree- significant in the two groups. This was probably due to ment with the general finding that the amount of body fat Diabetes, Obesity and Metabolism, 7, 2005, 223–229 The effect of diets high in protein or carbohydrate Table 3 Serum concentrations of inflammatory markers and blood lipids in the high-carbohydrate, low-protein (HCLP) andhigh-protein, low-carbohydrate (HPLC) groups after 6 months of dietary intervention Values are means (95% CI). CRP, C-reactive protein; HDL, high-density lipoprotein; NS, not significant; LDL, low-density lipoprotein ¼ Totalcholesterol À HDL À (TG/2.2); TG, triacylglycerol.
*HOMA-IR, Homeostasis model assessment-insulin resistance ¼ fasting insulin (mU/ml) Â fasting glucose (mmol/l)/22.5.
†Test for differences of changes within groups by paired t-test.
‡p-values were derived by ANOVA with baseline values as covariates.
§p < 0.05.
††p < 0.0001.
could play a role in the occurrence of high concentrations glycation end products (AGEs) (derived from long cook- of CRP [21,22,26]. However, no significant association ing time and cooking at high temperatures) have been was seen between CRP and IAAT or waist circumference shown to increase CRP by 35%, whereas CRP decreased (a crude index of abdominal obesity) as seen in other by 20% on a 6-week diet low in AGEs [29]. It is possible that a meat-rich diet, like the HPLC diet, could cause a The lack of a stronger association between the CRP and relatively higher CRP concentration compared to a HCLP diet. However, we have no registration of cooking pro- explained by other factors, such as seasonal variation, as cedures in the present study and therefore cannot assess the concentration of CRP and other inflammatory markers the possibility of an effect of food processing.
have been shown to differ significantly throughout the Cigarette smoking is a major cause of variation in CRP year [28]. However, no seasonal influence was detectable [30], but it is unlikely that smoking was a significant in the present trial. In addition, food processing methods factor in the present study as the number of smokers might have an influence on the concentration of inflam- was low and the subjects were asked not to change matory markers as a diet high in heat-generated advanced their smoking habits during the intervention.
Diabetes, Obesity and Metabolism, 7, 2005, 223–229 The effect of diets high in protein or carbohydrate A significant increase in haptoglobin within both 5 Davi G, Guagnano MT, Ciabattoni G et al. Platelet acti- groups, after the dietary intervention and a significant cor- vation in obese women: role of inflammation and relation between haptoglobin and the anthropometrical oxidant stress. JAMA 2002; 288 (16): 2008–2014.
measures IAAT and waist circumference, were found at 6 Esposito K, Pontillo A, Di Palo C et al. Effect of weight 6 months. This does not correspond to the fact that hapto- loss and lifestyle changes on vascular inflammatory mar-kers in obese women: a randomized trial. JAMA 2003; globin is a positive acute phase protein and was expected to decrease after weight loss. Moreover, we found an 7 Heilbronn LK, Noakes M, Clifton PM. Energy restriction increase in transferrin after weight loss and a negative and weight loss on very-low-fat diets reduce C-reactive correlation between transferrin, FM and IAAT. This find- protein concentrations in obese, healthy women. Arterio- ing is in accordance with the concept that transferrin is a scler Thromb Vasc Biol 2001; 21 (6): 968–970.
negative acute-phase protein, decreases in inflammation 8 Wolfe BM. Potential role of raising dietary protein intake and is expected to increase after weight loss in obesity [4].
for reducing risk of atherosclerosis. Can J Cardiol 1995; Furthermore, we found no significant improvements in the concentrations of glucose and insulin or insulin 9 Liu S, Manson JE, Buring JE, Stampfer MJ, Willett WC, resistance or any of the blood lipids despite a significant Ridker PM. Relation between a diet with a high glycemic weight loss. This could also be due to the fact that a large load and plasma concentrations of high-sensitivityC-reactive protein in middle-aged women. Am J Clin number of the subjects were still overweight or obese and that our measurements were made after a new 10 Jenkins DJ, Kendall CW, Augustin LS et al. Effect of wheat bran on glycemic control and risk factors for car- In conclusion, this study confirms that the inflamma- diovascular disease in type 2 diabetes. Diabetes Care tory markers CRP, haptoglobin and transferrin do respond to changes in body fatness but are independent of dietary 11 Jenkins DJ, Kendall CW, Connelly PW et al. Effects of changes of carbohydrate and protein.
high- and low-isoflavone (phytoestrogen) soy foods oninflammatory biomarkers and proinflammatory cyto-kines in middle-aged men and women. Metabolism This research was supported by The Danish Research 12 Madsen T, Skou HA, Hansen VE et al. C-reactive protein, and Development Programme for Food Technology dietary n-3 fatty acids, and the extent of coronary artery (FØTEK), The Federation of Danish Pig Producers and disease. Am J Cardiol 2001; 88 (10): 1139–1142.
13 Sierksma A, van der Gaag MS, Kluft C, Hendriks HF. Mod- Slaughterhouse, Danish Dairy Research Foundation erate alcohol consumption reduces plasma C-reactive and The Danish Livestock and Meat Board. Master Food protein and fibrinogen levels; a randomized, diet-controlled a/s, Kellogg’s a/s, Daloon a/s, Hatting sBrød a/s, Haribo intervention study. Eur J Clin Nutr 2002; 56 (11): 1130–1136.
a/s, Pingvin a/s, Galle & Jessen a/s and Malaco a/s con- 14 Skov AR, Toubro S, Ronn B, Holm L, Astrup A. Random- tributed to the food selection. We thank the project staff ized trial on protein vs carbohydrate in ad libitum fat of the Department of Human Nutrition.
reduced diet for the treatment of obesity. Int J Obes RelatMetab Disord 1999; 23 (5): 528–536.
15 Skov AR, Toubro S, Raben A, Astrup A. A method to achieve control of dietary macronutrient composition in 1 Ridker PM, Cushman M, Stampfer MJ, Tracy RP, ad libitum diets consumed by free-living subjects. Eur J Hennekens CH. Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men.
16 Treuth MS, Hunter GR, Kekes-Szabo T. Estimating intra- N Engl J Med 1997; 336 (14): 973–979.
abdominal adipose tissue in women by dual-energy X-ray 2 Ridker PM, Buring JE, Shih J, Matias M, Hennekens CH.
absorptiometry. Am J Clin Nutr 1995; 62 (3): 527–532.
Prospective study of C-reactive protein and the risk of 17 Haffner SM, Miettinen H, Stern MP. The homeostasis future cardiovascular events among apparently healthy model in the San Antonio Heart Study. Diabetes Care women. Circulation 1998; 98 (8): 731–733.
3 Ridker PM, Rifai N, Rose L, Buring JE, Cook NR. Compari- 18 Friedewald WT, Levy RI, Fredrickson DS. Estimation of son of C-reactive protein and low-density lipoprotein the concentration of low-density lipoprotein cholesterol cholesterol levels in the prediction of first cardiovascular in plasma, without use of the preparative ultracentrifuge.
events. N Engl J Med 2002; 347 (20): 1557–1565.
4 Gabay C, Kushner I. Acute-phase proteins and other sys- 19 Ford ES. Body mass index, diabetes, and C-reactive temic responses to inflammation. N Engl J Med 1999; 340 protein among US adults. Diabetes Care 1999; 22 (12): Diabetes, Obesity and Metabolism, 7, 2005, 223–229 The effect of diets high in protein or carbohydrate 20 Hak AE, Stehouwer CD, Bots ML et al. Associations of 25 Laimer M, Ebenbichler CF, Kaser S et al. Markers of C-reactive protein with measures of obesity, insulin chronic inflammation and obesity: a prospective study resistance, and subclinical atherosclerosis in healthy, on the reversibility of this association in middle-aged middle-aged women. Arterioscler Thromb Vasc Biol women undergoing weight loss by surgical intervention.
Int J Obes Relat Metab Disord 2002; 26 (5): 659–662.
21 Lemieux I, Pascot A, Prud’homme D et al. Elevated 26 Forouhi NG, Sattar N, McKeigue PM. Relation of C- C-reactive protein: another component of the athero- reactive protein to body fat distribution and features of thrombotic profile of abdominal obesity. Arterioscler the metabolic syndrome in Europeans and South Asians.
Thromb Vasc Biol 2001; 21 (6): 961–967.
Int J Obes Relat Metab Disord 2001; 25 (9): 1327–1331.
22 Tchernof A, Nolan A, Sites CK, Ades PA, Poehlman ET.
27 Chan DC, Watts GF, Barrett PH, Beilin LJ, Mori TA.
Weight loss reduces C-reactive protein levels in obese Effect of atorvastatin and fish oil on plasma high- postmenopausal women. Circulation 2002; 105 (5): sensitivity C-reactive protein concentrations in individuals with visceral obesity. Clin Chem 2002; 48 (6 Part 1): 23 Yudkin JS, Stehouwer CD, Emeis JJ, Coppack SW.
C-reactive protein in healthy subjects: associations with 28 Khaw KT, Woodhouse P. Interrelation of vitamin C, in- obesity, insulin resistance, and endothelial dysfunction: fection, haemostatic factors, and cardiovascular disease.
a potential role for cytokines originating from adipose tissue? Arterioscler Thromb Vasc Biol 1999; 19 (4): 29 Vlassara H, Cai W, Crandall J et al. Inflammatory medi- ators are induced by dietary glycotoxins, a major risk 24 Farnsworth E, Luscombe ND, Noakes M, Wittert G, factor for diabetic angiopathy. Proc Natl Acad Sci USA Argyiou E, Clifton PM. Effect of a high-protein, energy- restricted diet on body composition, glycemic control, 30 Danesh J, Whincup P, Walker M et al. Low grade and lipid concentrations in overweight and obese hyper- inflammation and coronary heart disease: prospective insulinemic men and women. Am J Clin Nutr 2003; 78 study and updated meta-analyses. BMJ 2000; 321 (7255): Diabetes, Obesity and Metabolism, 7, 2005, 223–229

Source: http://anettedue.dk/PDF/CRPDOM2005WEB.pdf