PAPER Relation of C-reactive protein to body fat distribution and features of the metabolic syndrome in Europeans and South Asians

Transcription

1 (2001) 25, ß 2001 Nature Publishing Group All rights reserved /01 $ PAPER Relation of C-reactive protein to body fat distribution and features of the metabolic syndrome in Europeans and South Asians NG Forouhi 1 *, N Sattar 2 and PM McKeigue 1 1 Epidemiology Unit, London School of Hygiene and Tropical Medicine, London, UK; and 2 University Department of Pathological Biochemistry, Glasgow Royal Infirmary NHS Trust, Glasgow, UK OBJECTIVE: To investigate the association between circulating C-reactive protein (CRP) concentrations and indices of body fat distribution and the insulin resistance syndrome in South Asians and Europeans. DESIGN: Cross-sectional study. SUBJECTS: A total of 113 healthy South Asian and European men and women in West London (age y, body mass index (BMI) kg=m 2 ). MEASUREMENTS: Fatness and fat distribution parameters (by anthropometry, dual-energy X-ray absorptiometry and abdominal CT scan); oral glucose tolerance test with insulin response; modified fat tolerance test; and CRP concentration by sensitive ELISA. RESULTS: Median CRP level in South Asian women was nearly double that in European women (1.35 vs 0.70 mg=1, P ¼ 0.05). Measures of obesity and CRP concentration were significantly associated in both ethnic groups. The correlation to CRP was especially strong among South Asians (P<0.01) for measures of central obesity (waist girth and visceral fat area), whereas BMI and percentage fat were more significantly associated with CRP in Europeans (P<0.05). In South Asians the associations of CRP with visceral fat area and waist girth persisted after adjustment for either BMI or percent fat (all, P<0.05). In age-, sex- and smoking-adjusted regression analyses CRP concentrations were significantly associated with fasting and 2 h insulin and lipid levels in both ethnic groups (P<0.05). When further statistical adjustment was made for visceral fat area these associations were abolished (P >0.15). CONCLUSION: We suggest that adiposity and in particular visceral adipose tissue is a key promoter of low-grade chronic inflammation. This observation may in part account for the association of CRP with markers of the metabolic syndrome. Future studies should confirm whether CRP concentrations are elevated in South Asians and whether losing weight by exercise or diet, or reduction in visceral fat mass, is associated with reduction in plasma CRP concentrations. (2001) 25, Keywords: central obesity; C-reactive protein; inflammation; metabolic syndrome; South Asian Introduction C-reactive protein (CRP) is an acute phase reactant which is a marker of inflammation in the body. Mild elevations of CRP concentrations even when within the clinically normal range are independently predictive of future cardiovascular *Correspondence: NG Forouhi, Specialist Registrar in Public Health, West Hertfordshire Health Authority, Tonman House, St Albans, Herts, AL1 3ER, UK. nita.forouhi@ha.wherts-ha.nthames.nhs.uk Received 13 October 2000; revised 16 February 2001; accepted 12 March 2001 events. 1,2 The causes(s) for increased CRP in apparently healthy individuals is not clear. Recently, elevated CRP levels have been associated with the features of the insulin resistance syndrome, namely abdominal obesity as assessed by waist girth, fasting glucose, hyperinsulinaemia and insulin sensitivity, high triglyceride and low HDL-cholesterol. 3 9 However, studies of the association of CRP with more detailed measures of body fatness, namely visceral and subcutaneous fat depots, are lacking. These are needed since visceral adipose tissue measures appear to be more closely linked with features of the metabolic syndrome and could be an important source for pro-inflammatory cytokine production in healthy individuals.

2 1328 South Asians (people of Indian, Pakistani, Bangladeshi or Sri Lankan descent) overseas have an increased risk of developing coronary heart disease (CHD). 10 Conventional risk factors fail to explain the excess CHD risk in South Asians compared with Europeans, but an enhanced predisposition for central obesity and insulin resistance accompanied by raised triglyceride and lowered HDL-cholesterol concentrations are likely to be contributing factors. 11,12 Whether inflammation also contributes to this increased CHD risk in South Asians is not known, but it is important to determine this since it may have therapeutic implications. The objectives of the study therefore were to test (i) if there is an association between circulating CRP concentration and a detailed panel of body fat measures in South Asians and Europeans, (ii) if, as a result, CRP concentration correlates with other measures of the insulin resistance syndrome (insulin and lipid levels), and (iii) whether CRP concentrations are elevated in South Asians compared to body mass index (BMI) matched Europeans. Methods We performed a cross sectional study of 113 healthy men and women. Written invitations with a covering letter by the general practitioner were sent to randomly eligible subjects from the registration lists of four general practitioners in West London. A short one-page questionnaire was sent to ascertain by direct enquiry the age, ethnic origin, past medical history and approximate BMI (by asking for height and weight). Healthy subjects who were aged y, had a BMI between 17 and 34 kg=m 2, and were of South Asian or European origin were recruited to participate. Ethnic origin was confirmed at the first visit on the basis of appearance, name and direct enquiry. The participants were frequencymatched for BMI to achieve roughly equal numbers of Europeans and South Asians in each of four categories of BMI (<23, , and >29.0 kg=m 2 ). Exclusion criteria were history of diabetes or CHD, drugs known to affect insulin resistance or lipids (including hormone replacement therapy or oral contraceptive use in women), fasting hypertriglyceridaemia (>3.5 mmol=l), and any chronic illness or weight-losing condition. Informed consent was obtained from all participants, and ethical approval for the study obtained from the local ethics committees. Measurements Anthropometric measurements included height (cm), weight (kg), and waist and hip circumferences (cm). From this BMI was calculated as weight=(height) 2 in kg=m 2. A 75 g standard oral glucose tolerance test 13 with insulin response was carried out. Percentage body fat was measured by dual energy X- ray absorptiometry. Visceral fat area (VFA) was measured by a single-slice (10 mm thickness) CT scan at the level of L4 L5 using a Toshiba X-speed scanner, with exposure time 2.7 s and exposure factors 160 ma and 120 kv. Participants were scanned in a supine position with arms stretched above their heads. The VFA was measured in the attenuation range of 7190 to 730 HU by drawing a line within the muscle wall surrounding the abdominal cavity. 14 Fat tolerance test was performed using a modification of Patsch s method, 15 with 65 g fat per square metre of body surface area, and fasting and 8 h postprandial samples collected for lipid estimations. During this visit fasting blood was also collected for CRP concentration measurement. Smoking was recorded as current, ex- and never smoker, but due to very small numbers of current smokers, it was coded in analyses as ever or never smoker. Social class was coded as manual or non-manual based on occupation. Laboratory analyses C-reactive protein was measured using a sensitive doubleantibody sandwich ELISA with rabbit antihuman C-reactive protein and peroxidase conjugated rabbit anti-human C- reactive protein. The assay was linear up to 5 mg=1 and logarithmic thereafter. The interassay and intra-assay coefficients of variation were less than 10% across the range of measured results. 16 Plasma glucose was measured by the oxidase method. Plasma insulin concentration was determined by a microplate-based chemiluminometric assay (Molecular Light Technology Research Limited, Cardiff, UK). Inter-assay coefficient of variation (CV) was less than 7%. Plasma triglyceride, cholesterol and HDL cholesterol were measured in EDTA plasma samples within 48 h. The triglyceride assay used Triglycerides N (WAKO, Alpha Laboratories) with intraassay CV of <1% and inter-assay CV <4%. The cholesterol assay used the enzymatic reagent CHOL (Boehringer Mannheim) with inter-assay CV of <2%. HDL was separated by precipitation of other lipoprotein fractions with manganese chloride and heparin solution. The inter and intra assay CVs were <5% and <1%, respectively. Statistical analyses As the distribution of CRP level was skewed, median values of CRP level were calculated, and groups compared by Wilcoxon rank sum test. Regression analysis was used to test associations between variables. Skewed variables were log (natural) transformed; these included CRP, triglyceride, HDL-cholesterol, insulin and VFA. In a series of multivariate analyses ln CRP was examined as the outcome variable in relation to age, smoking, social class and body fat distribution, insulin and lipids. All analyses were performed using STATA version 5.0 (STATACORP, Texas, USA). Results Table 1 summarises the clinical and metabolic features of the study participants. South Asian women had significantly higher levels of both percentage fat and visceral fat area

3 Table 1 Clinical and metabolic features of the study participants 1329 Men Women European South Asian P European South Asian P Number BMI (kg=m 2 ) 26.1 (0.7) 25.7 (0.6) NS 24.9 (0.7) 25.5 (0.6) NS Percentage fat by DEXA 22.6 (1.4) 25.3 (1.2) NS 34.9 (1.4) 39.7 (1.0) Waist girth (cm) 88.4 (1.9) 88.2 (1.8) NS 75.9 (1.4) 78.6 (1.6) NS Visceral fat area a (cm 2 ) (93.5, 144.0) (105.2, 145.2) NS 67.0 (58.2, 77.0) 85.6 (72.9, 102.9) Fasting insulin a (mu=ml) 4.9 (3.9, 6.1) 8.9 (6.9, 11.5) (4.0, 5.4) 6.0 (4.7, 7.6) h insulin a (mu=ml) 11.2 (7.9, 16.1) 40.9 (25.7, 65.1) < (13.6, 22.2) 35.2 (25.2, 49.0) Fasting triglyceride a (mmol=l) 1.58 (1.31, 1.92) 1.72 (1.45, 2.04) NS 1.24 (1.06, 1.44) 1.27 (1.02, 1.58) NS 8 h triglyceride a (mmol=l) 1.97 (1.56, 2.48) 2.51 (1.96, 3.21) NS 1.45 (1.17, 1.79) 1.39 (1.06, 1.83) NS HDL-cholesterol a (mmol=l) 1.39 (1.25, 1.55) 1.26 (1.16, 1.38) NS 1.56 (1.43, 1.71) 1.51 (1.40, 1.62) NS CRP (mg=l) 0.92 ( ) 1.07 ( ) NS 0.70 ( ) 1.35 ( ) Values are mean and s.e.m., except for log (natural) transformed variables ( a ) where geometric means and 95% Cl are given. For CRP level, median values and the interquartile range are given; NS, not significant, P>0.10. compared to European women. All South Asians had significantly higher levels of insulin. The mean age ranged from 45.5 (s.e.m. 0.8) to 47.6 (s.e.m. 0.9) years across the four groups. Prevalence of being an ever smoker was 64 and 59% in European men and women, respectively, and 21 and 0% in South Asian men and women, respectively. Median CRP level in South Asian women (1.35 mg=l ( )) was nearly double that in European women (0.70 mg=1 ( ); P ¼ 0.05). Measures of obesity and CRP concentration were significantly associated, as shown in Table 2. The association was especially strong among South Asians for measures of central obesity (waist girth and visceral fat area), whereas BMI and percentage fat were more significantly associated with CRP in Europeans. Moreover, in South Asians but not Europeans the associations of CRP with visceral fat area and waist girth persisted after correction for either BMI or percentage fat (all, P<0.05). Social class was not significantly associated with CRP in either ethnic group in this study. For the entire study group, with CRP level as the dependent variable and putting into the model age, sex, ethnic group, smoking status, percentage fat and visceral fat area, the only independent correlate of CRP level was visceral fat area (beta ¼ 0.87, P ¼ 0.005). The latter (VFA) was also the only independent correlate of CRP level if BMI was in the model instead of percentage fat. In age-, sex- and smoking-adjusted regression analyses CRP concentration was associated with insulin and lipid levels in both ethnic groups. For the entire study group Table 3 shows that there was an independent association between CRP level and fasting and 2 h insulin and fasting and 8 h triglyceride and fasting HDL concentration in analyses adjusted for age, sex, ethnic group and smoking. Further adjustment for total percentage body fat reduced the significance of the associations between CRP level and insulin and lipids in most cases to 0.05 < P < 0.10 but did not completely abolish them except for HDL-cholesterol. When statistical adjustment was made for visceral fat area the significance of the associations was abolished in all cases (P>0.15). The results for the association between CRP concentration and HOMA (homeostasis model assessment, as a marker for insulin resistance) were virtually Table 2 Determinants of C-reactive protein concentration among Europeans and South Asians. Age-, sex- and smoking-adjusted regression coefficients are given for log (natural) C-reactive protein as the dependent variable Europeans South Asians Models adjusted for ageþ sex þ smoking b s.e. P b s.e. P Age (y) (adjusted for sex þ smoking only) Body mass index (kg=m 2 ) Percentage fat Waist girth (cm) Visceral fat area a (cm 2 ) Visceral fat area a þ percentage fat Visceral fat area a þ BMI Waist girth þ percentage fat Waist girth þ BMI a Log (natural) transformed.

4 1330 Table 3 Relationship between measures of the metabolic syndrome and CRP concentration for the entire study population (with log (natural) CRP as the dependent variable and each of insulin and lipid measurements as independent variables in turn) Independent variables b s.e. P Fasting insulin (mu=ml) a b c h insulin (mu=ml) a b c Fasting triglyceride (mmol=l) a b c h triglyceride (mmol=l) a b c HDL-cholesterol (mmol=l) a b c a adjusted for age, sex, ethnic group and smoking; b adjusted for age, sex, ethnic group, smoking and percent fat; c adjusted for age, sex, ethnic group, smoking and visceral fat area. identical to the results for fasting insulin (data not shown). None of the subjects had diabetes, based on history or the glucose tolerance test. There were eight subjects with impaired glucose tolerance in this study group and they were not excluded from the analyses presented. However, when the analyses were restricted to only those with normoglycaemia (n ¼ 105), the results were unchanged and in the same direction as reported. Discussion The present study is important for several reasons. First, this is as far as we are aware one of the first studies relating CRP concentration to a detailed panel of markers of adiposity including visceral fat and percentage fat measures. We noted strong correlations between all markers of adiposity and CRP in both ethnic groups. However, there was a difference in pattern between the two ethnic groups. In South Asians but not Europeans, visceral fat area and waist girth were strong (P < 0.01) correlates with CRP whereas BMI and percentage fat only showed trends of association (P < 0.10). Interestingly, both visceral fat area and waist girth were associated with CRP concentration even after adjustment for either BMI or percentage fat in South Asians. These latter two observations suggest that, at least in this ethnic group, visceral adipose tissue may be an important contributor to lowgrade chronic inflammation. These data accord with preliminary observations of Despres et al. 17 They noted in a cross-sectional study of 168 men that highest CRP concentrations were observed in those with both obesity and a high accumulation of visceral adipose tissue. In other words, visceral fat was associated with CRP independently of total adiposity. However, they report this independent association in a group of men of presumed European origin, while we found the independence of association only in South Asians in our study. This difference could be due to the smaller numbers in our study (57 Europeans) compared with theirs (n ¼ 168), the fact that their study was restricted to men but ours also included women, and that their study sample was generally more obese (mean BMI 30.5, s.d. 4 kg=m 2 ) than our European sample (mean BMI 25.5, s.d. 3.5 kg=m 2 ). There is clearly a need for other groups to perform such studies as well to see if these associations can be replicated so that comparisons can be made. Second, we were able to confirm the association of CRP with markers of the metabolic syndrome in our population. Indeed, in line with the recent work by several groups, we noted that CRP correlated (positively) with insulin measures, triglyceride (fasting and 8 h levels) and (inversely) with HDLcholesterol. 3 9 Interestingly, addition of percentage fat mass to these analyses weakened, but did not abolish completely, many of the correlations presented. In contrast, addition of visceral fat area into these analyses abolished all significant associations (P>0.15). The above observations add weight to the argument that adipose tissue-derived cytokine expression (TNF-a, IL-6), perhaps particularly from visceral adipose tissue, is an important contributor to low-grade chronic inflammation, just as it plays a prominent role in the metabolic syndrome. In other words, accumulation of visceral adipose tissue is a key factor underpinning up-regulation of both features of the metabolic syndrome and low-grade chronic inflammation. In this respect it is noteworthy that visceral fat is drained by the portal vein to the liver and the liver is the principal site for both CRP production, under cytokine stimulus, and for the synthesis of triglyceride-rich lipoproteins, the latter from the substrate non-esterified fatty acids (NEFAs). Moreover, cytokines can stimulate adipocyte lipolysis directly and promote de novo hepatic fatty acid synthesis. 18,19 Elevations in NEFA flux, in turn, may be linked to impaired hepatic and peripheral insulin sensitivity. 20 These combined observations would also explain the correlation of insulin sensitivity to CRP, as recently noted by Festa and colleagues, 9 who suggested that low grade chronic inflammation is likely a part of the insulin resistance syndrome. Interestingly, the thiazolidinedione group of insulinsensitising drugs has recently been shown to decrease several low-grade inflammatory markers in people with type 2 diabetes. 21 Third, we report for the first time increased CRP concentrations in healthy South Asian women relative to European women frequency-matched for BMI. The increased CHD risk in South Asians overseas compared with other groups is not fully accounted for by traditional risk factors. A higher inflammatory response in part mediated by increased visceral fat mass could be an additional risk factor contributing to the increased risk of CHD in South Asian women, and larger studies in both South Asian men and women are

5 needed to expand on our results. If confirmed, then CRP concentration may be useful in assessing the extent of excess risk of CHD in South Asians. Prospective studies are needed to address this potential. Furthermore, a possible next step might be to evaluate the efficacy of aspirin therapy in reducing the high CHD risk in this group, since Ridker et al 1 have demonstrated that the efficacy of aspirin in reducing the incidence of myocardial infarction appears to be directly related to the level of CRP. Finally, the observation of an association between 8 h triglyceride and CRP level in this study is novel and suggests a link between inflammation and postprandial fat clearance. This may be a direct effect of cytokines (IL-6 or TNF-a) on the activity of lipoprotein lipase 18,22 (the endothelial enzyme responsible for catabolism of triglyceride-rich lipoproteins), or alternatively may reflect the association of both factors with insulin resistance and visceral adiposity. In conclusion, we suggest that adiposity, and perhaps in particular visceral adipose tissue, is a key promoter of lowgrade chronic inflammation. Further studies are required to confirm this observation, which may in part account for the association of CRP with markers of the metabolic syndrome. Future studies should assess whether losing weight by exercise or diet, or therapeutic modalities such as PPARg agonists associated with reductions in visceral fat mass, are associated with reductions in plasma CRP concentrations. Acknowledgements This study was supported by a Wellcome Training Fellowship in Clinical Epidemiology for NGF. We thank Dr Jaspal Kooner for facilitating the study at Ealing Hospital, the general practitioners who helped with recruitment, the Radiology Department and Dr Bill Svensson at Ealing Hospital for CT data, and Mr U Bhonsle for DEXA scan data. References 1 Ridker PM, Cushman M, Stampfer MJ, Tracy RP, Hennekens CH. Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. New Engl J Med 1997; 336: Ridker PM, Buring JE, Shih J, Matias M, Hennekens CH. Prospective study of C-reactive protein and the risk of future cardiovascular events among apparently healthy women. Circulation 1998; 98: Haverkate F, Thompson SG, Pyke SD, Gallimore JR, Pepys MB. Production of C-reactive protein and risk of coronary events in stable and unstable angina. European Concerted Action on Thrombosis and Disabilities Angina Pectoris Study Group. Lancet 1997; 349: Koenig W, Sund M, Frohlich M, Fischer HG, Lowel H, Doring A, Hutchinson WL, Pepys MB. C-Reactive protein, a sensitive marker of inflammation, predicts future risk of coronary heart disease in initially healthy middle-aged men: results from the MONICA (Monitoring Trends and Determinants in Cardiovascular Disease) Augsburg Cohort Study, 1984 to Circulation 1999; 99: Mendall MA, Patel P, Ballam L, Strachan D, Northfield TC. C- Reactive protein and its relation to cardiovascular risk factors: a population based cross sectional study. Br Med J 1996; 312: Tracy RP, Psaty BM, Macy E, Bovill EG, Cushman M, Cornell ES, Kuller LH. Lifetime smoking exposure affects the association of C- reactive protein with cardiovascular disease risk factors and subclinical disease in healthy elderly subjects. Arterioscler Thromb Vasc Biol 1997; 17: Pickup JC, Mattock MB, Chusney GD, Burt D. NIDDM as a disease of the innate immune system: association of acute-phase reactants and interleukin-6 with metabolic syndrome X. Diabetologia 1997; 40: Yudkin JS, Stehouwer CD, Emeis JJ, Coppack SW. C-reactive protein in healthy subjects: associations with obesity, insulin resistance, and endothelial dysfunction: a potential role for cytokines originating from adipose tissue? Arterioscler Thromb Vasc Biol 1999; 19: Festa A, D Agostino RJ, Howard G, Mykkänen L, Tracy RP, Haffner SM. Chronic Subclinical Inflammation as Part of the Insulin Resistance Syndrome: the Insulin Resistance Atherosclerosis Study (IRAS). Circulation 2000; 102: McKeigue PM, Miller GJ, Marmot MG. Coronary heart disease in South Asians overseas: a review. J Clin Epidemiol 1989; 42: McKeigue PM, Shah B, Marmot MG. Relation of central obesity and insulin resistance with high diabetes prevalence and cardiovascular risk in South Asians. Lancet 1991; 337: McKeigue PM, Ferrie JE, Pierpoint T, Marmot MG. Association of early-onset coronary heart disease in South Asian men with glucose intolerance and hyperinsulinemia. Circulation 1993; 87: WHO Study Group on Diabetes Mellitus. Diabetes mellitus: report of a WHO study group. World Health Organization Technical Report Series 727. WHO: Geneva; Seidell JC, Björntorp P, Sjöström L, Sannerstedt R, Krotkiewski M, Kvist H. Regional distribution of muscle and fat mass in men new insight into the risk of abdominal obesity using computed tomography. Int J Obes 1989; 13: Patsch JR, Miesenböck G, Hopferweiser T, Mühlberger V, Knapp E, Dunn JK, Gotto AM, Patsch W. Relation of triglyceride metabolism and coronary artery disease: studies in the postprandial state. Arterioscler Thromb 1992; 12: Highton J, Hessian P. A solid phase enzyme immunoassay for C- reactive protein: clinical value and the effect of rheumatoid factor. J Immunol Meth 1984; 68: Després JP, Pascot A, Bergeron J, Lemieux I, Dumont M, Almeras N, Nadeau A, Prud homme D. Abdominal obesity: the critical correlate of elevated plasma C-reactive protein levels associated with features of the insulin resistance syndrome in men (Abstract.) Atherosclerosis 2000; 151: Greenberg AS, Nordan RP, McIntosh J, Calvo JC, Scow RO, Jablons D. Interleukin 6 reduces lipoprotein lipase activity in adipose tissue of mice in vivo and in 3T3 L1 adipocytes: a possible role for interleukin 6 in cancer cachexia. Cancer Res 1992; 52: Chajek-Shaul T, Friedman G, Stein O, Shiloni E, Etienne J, Stein Y. Mechanism of the hypertriglyceridemia induced by tumor necrosis factor administration to rats. Biochim Biophys Acta 1989; 1001: Bjorntorp P. Fatty acids, hyperinsulinemia, and insulin resistance: which comes first? Curr Opin Lipidol 1994; 5: Ebeling P, Teppo AM, Koistinen HA, Viikari J, Ronnemaa T, Nissen M, Bergkulla S, Salmela P, Saltevo J, Koivisto VA. Troglitazone reduces hyperglycaemia and selectively acute-phase serum proteins in patients with Type II diabetes. Diabetologia 1999; 42: Saxena U, Witte LD, Goldberg IJ. Tumour necrosis factor induced release of endothelial cell lipoprotein lipase. Arteriosclerosis 1990; 10: