Monash University Department of Surgery, Alfred Hospital, Melbourne, Victoria, Australia

Correspondence to: J B Dixon, Monash University Department of Surgery, Alfred Hospital, Melbourne 3181, Australia. E-mail: john.dixon@med.monash.edu.au

OBJECTIVE: The aim of this study was to examine changes in lipid profile and markers of insulin resistance with increasing body mass index (BMI) in the range 34-77 kg/m². In addition we compare the lipid profiles of severely obese patients with those of the Australian community.

SUBJECTS AND METHODS: A total of 572 patients (85% F, 15% M) were assessed prior to gastric restrictive surgery. Conventional lipid profiles and markers of insulin resistance were measured. Lipids were compared with the Australian National Heart Foundation 1989 study (control group).

RESULT: There was no difference in mean total cholesterol levels between the obese group (5.52 mmol/l) and the control group (5.47 mmol/l). The mean total cholesterol levels in the obese group fell with increasing BMI (r=-0.13, P<0.01). Obese subjects had elevated fasting triglyceride levels 1.96 mmol/l (control group, 1.12 mmol/l, P<0.001), but levels did not change with increasing BMI (r=0.0, NS). HDL-C levels were lower, 1.21 mmol/l (control group 1.44 mmol/l, P<0.001), and decreased with increasing BMI (r=-0.20, P<0.01). LDL-C levels were lower in obese men (3.65 mmol/l vs control group 4.17 mmol/l, P<0.01) but not women and levels fell with increasing BMI (r=-0.15, P<0.05). For the obese group, markers of insulin resistance (fasting plasma glucose, HbA1c, fasting plasma insulin and C-peptide) all rose significantly with increasing BMI.

CONCLUSION: Raised total cholesterol is not a co-morbidity of severe obesity. There is a disparity between the conventional lipid measures and insulin resistance measures of the metabolic syndrome with increasing BMI. Conventional lipid measures may be poor indicators of dyslipidaemic risk in the severely obese.

International Journal of obesity (2001) 25, 793-797

morbid obesity; dyslipidemia; insulin resistance; body mass index; risk

Introduction

The prevalence of obesity in Australia and other western societies is approximately 15% and rising.¹ It is a major modifiable and independent risk factor for coronary heart disease² and is also associated with other major risk factors: hypertension, dyslipidaemia and impaired glucose tolerance.³ A central feature of this clustering of risk factors is insulin resistance and the cluster is referred to as the metabolic syndrome or insulin resistance syndrome. This syndrome is closely related to the presence of increased visceral obesity and the simple anthropometric measurement that relates closely with the syndrome is waist measurement.^4,6 Waist circumference measurements correlate well with body mass index (BMI) and cardiovascular risk. Subjects with android, central or upper body obesity, as expressed by a high waist-hip ratio, are at even greater risk.

The dyslipidaemia associated with obesity and the metabolic syndrome is now better defined and features include raised fasting triglyceride and low high-density lipoprotein cholesterol (HDL-C) as well as an increase in small dense low-density lipoprotein (LDL) particles. These small dense LDL particles are considered to be highly atherogenic.^3,7,12 This pattern B phenotype of LDL is associated with decreased insulin sensitivity. The above features of the dyslipidaemia, while clearly associated with one another, have all been shown to be independently associated with risk of coronary heart disease. The Framingham charts for estimating coronary heart disease risk have been modified in part to reflect some of these clustered features.¹³ The size of LDL particles is inversely related to BMI, fasting plasma insulin level and fasting triglyceride level.¹¹ The routine clinical test that has the best sensitivity and specificity for LDL particle size and density is fasting triglyceride. Tchernof et al¹² found a fasting triglyceride value greater than 1.9 mmol/l had an optimal sensitivity of 84% and a specificity of 83% for the dense LDL phenotype. They found an HDL value of less than 0.9 mmol/l had a sensitivity and specificity of 77 and 79%, respectively, for the dense LDL phenotype. Combining the two values did not improve sensitivity or specificity. Gardner et al¹⁴ found LDL size to be the best physiological differentiator of coronary artery disease status. However the total cholesterol:HDL ratio was a stronger non-physiological independent predictor.

The prevalence of lipid abnormalities in the Australian community is well documented and a large proportion of the Australian community have lipid levels higher than recommended levels.¹⁵ Australian community levels of total cholesterol, fasting triglyceride and HDL-C were last reported in the National Heart Foundation risk prevalence study 1989.¹⁵ In this paper we use the results of this historic cohort study as community controls (CC).

It is clear that there are many features of morbid obesity that contribute to cardiovascular risk, with the metabolic syndrome being a major contributor. We have had the opportunity to examine routinely measured lipid profiles and markers of insulin sensitivity in a group of severely obese patients presenting for gastric restrictive surgery. The aim of the present study was to examine the trends in conventional lipid profile and markers of insulin sensitivity with increasing BMI in the range 34-77 kg/m², and additionally to compare the lipid profiles of these obese subjects with those of age- and sex-matched community levels (CC).

Patients and methods

Patients with a body mass index (BMI) greater than 34 kg/m², suffering significant medical, physical or psychosocial disabilities and who have attempted weight reduction by other means for at least 5 y were considered for entry into the a program of weight control through laparoscopic placement of an adjustable gastric band. A BMI of greater that 35 kg/m² is the usual criteria for surgery, but occasionally subjects with a BMI of 34 kg/m² are included. Preoperative laboratory tests included routine lipid studies, which initially included only a fasting total cholesterol and fasting triglyceride, and more recently low density lipoprotein cholesterol (LDL-C) and HDL-C. Assessment of insulin resistance was made by measuring fasting plasma glucose, B glycosylated haemaglobin Alc (HbA1c), fasting plasma insulin and C-peptide. All patients have had at least one of either fasting plasma insulin or C-peptide measured. These assays were preformed in an approved laboratory with internal and external quality control.

This obese group of patients was compared with age and sex matched historic controls taken from the National Heart Foundation (NHF) Risk Prevalence Study no. 3,1989,¹⁵ using the tables prepared by Bennett and Magnus.¹⁶ In this study 6097 subjects aged 20-64 y were randomly selected from six Australian capital cities. The mean weight and BMI for men and women in this study were 78.8 kg, 25.7 kg/m² and 64.8 kg, 24.7 kg/m² respectively. In all, 11.5% of men and 13.2% of women in this study were obese (BMI>30 kg/m²). A community control (CC) was generated for each subject and was given age- and sex-matched mean lipid levels. The NHF study provided mean lipid levels and distribution of lipids in various demographic groups living in Australian capital cities. Plasma insulin, C peptide levels, fasting plasma glucose and HbA1c levels were compared with the laboratory reference range. All patients give signed informed consent to the Lap-BandÒ surgery and the perioperative and post-operative requirements of this surgical weight loss procedure. These tests are performed as part of a routine pre-operative assessment. The hospital ethics committee has approved the questionnaires and protocols of the study as part of a larger study.

Statistical analysis

Mean and standard deviation demographic features of the cohort of obese patients are given. Patients were grouped in quartiles based on their BMI with mean and standard deviation characteristics of quartiles displayed. Linear regression analysis using the SPSS statistical package¹⁷ was used to assess correlation between BMI and laboratory results, with multivariate analysis allowing for any change in age or sex distribution with increasing BMI. Unpaired Student's t-tests were used to assess difference between subjects and historic community controls. In all assessments a P-value less than 0.05 was considered significant, however one in 20 of correlations at 0.05 level could occur randomly.

Results

The obese group consisted of 572 consecutive patients, 487 women and 85 men (85% F, 15% M). This group had a mean age of 40.5 y (s.d. 9.6, range 16-74) and BMI of 45.7 kg/m² (s.d. 7.6, range 34-77). Pre-operative total cholesterol and fasting triglyceride levels were performed on all patients. The most recent 272 consecutive patients also had HDL-C levels measured with LDL-C calculated from these if the fasting triglyceride was less than 4.6 mmol/l.¹⁸ Fasting plasma insulin levels were recorded on 317 patients and C-peptide on 417 patients.

There was no difference in the total cholesterol level between the obese group and the control group. Mean total cholesterol level for the obese group was 5.52 mmol/l. Age- and sex-matched Australian community levels were 5.47 mmol/l (Table 1). Men had a slightly lower and women a higher total cholesterol than matched controls. There was a significant fall in mean total cholesterol with increasing BMI (r=-0.13, P<0.0l) (Table 2). The significance was not altered when corrected for any change in sex and age distribution with increasing BMI. The levels of lipids in each quartile are shown in Table 2.

The mean fasting triglyceride levels for the obese group were 75% higher than the controls (Table 1). The mean level for the group was 1.96 mmol/l (CC, 1.12 mmol/l, P<0.001). The difference was most evident in the obese female patients who had values 84% above controls (1.92 mmol/l vs CC, 1.04 mmol/l, P<0.001). For obese males, there was a 43% difference (2.23 mmol/l vs CC, 1.56 mmol/l, P<0.001) due primarily to higher values in control males. Despite this major difference in triglyceride level with obesity, analysis of the quartiles and variance (Table 2) showed no change in fasting triglyceride levels from BMI 34 to 77 kg/m².

Overall HDL-C levels were 16% lower than controls with men 11% and women 17% lower. Regression analysis showed a significant trend to lower HDL-C levels with increasing BMI (r=-0.2, P<0.01) (Table 2).

LDL-C levels in the obese group were lower than controls but only significantly lower in men (l2%; Table 1). The mean LDL-C for obese men was 3.65 mmol/l and controls 4.17 mmol/l (P<0.01). As for TC and HDL-C there was a significant fall in LDL-C with increasing BMI (r=-0.15, P<0.05; Tab1e 2).

Obese patients were found to have a significant fall in total cholesterol, HDL-C and LDL-C with increasing BMI. There was no change in the fasting triglyceride level. Indicators of insulin sensitivity were similarly examined for changes with increasing BMI (Table 3). All parameters, fasting plasma glucose, HbA1c, fasting plasma insulin and C-peptide, showed a significant positive relationship with BMI (Table 3).

Discussion

A raised total cholesterol level is often included as a comorbidity in morbid obesity. Recently the Bariatric Analysis and Reporting Outcome System (BAROS) has been proposed as a uniform way of assessing progress after bariatric surgery.¹⁹ It lists high cholesterol as a major health comorbidity. The American Obesity Association defines a comorbidity as any condition associated with obesity that (a) usually worsens as the degree of obesity increases and (b) often improves as the condition is treated.^19,20

Our study shows that raised total cholesterol is not a comorbidity, as defined by the American Obesity Association in morbidly obese men and women. Accepting the weakness of historical population control values in this study, it is clear that a high triglyceride and low HDL-C are characteristics of obesity dyslipidaemia, raised levels of total cholesterol and LDL-C are not. In fact there is a significant fall in total cholesterol with increasing BMI above a BMI of 34 kg/m². We have also found no change in total cholesterol with weight loss associated with the adjustable gastric band placement a gastric restrictive procedure (Dixon & O'Brien, unpublished data) and others have consistent findings.^21,22 Modest reductions in total cholesterol occur with some dietary methods of weight loss, however it appears that the nature of the diet and associated exercise may be critical to any fall in total cholesterol.^23,24 Significant falls in cholesterol do occur after malabsorbtive weight loss procedures.^25,27

The Framingham heart study²⁸ ranked body weight third most important predictor of coronary heart disease in males, after age and dyslipidaemia. A similar effect of body weight is seen in women. The Framingham charts used to assess coronary heart risk do not include hypertriglyceridemia, obesity, LDL particle size, or markers for insulin resistance other than frank diabetes. It was felt the metabolic syndrome would be expressed in the 'major risk factors'.¹³ Morbidly obese patients as a group are at risk, a risk that may not be reflected in standard risk profiles. Our study shows that the total cholesterol:HDL-C ratio for example may not reflect risk in morbidly obese men who as a group have a similar ratio to the community controls. This group has a high BMI, a raised fasting triglyceride, and raised fasting insulin levels; all independent CHD risk factors and all associated with the small dense atherogenic LDL particles.¹¹

It may be argued that the insulin resistance-dyslipidemia syndrome is the most prevalent cause of coronary artery disease.⁷ It certainly is the condition most closely associated with the increasing obesity in the western world, including Australia. It would appear that in any standard risk assessment some marker for the metabolic syndrome other than hypertension and frank diabetes should be included. Tchernof¹² showed that the best clinical marker for small LDL particles was the fasting triglyceride level (FTG), a readily available routine laboratory test. FTG is also closely related to obesity and insulin resistance. Laboratory tests associated with increasing BMI and insulin resistance include fasting glucose, HbA1c, plasma insulin or C-peptide. Even more readily available are measurements of weight, BMI, waist circumference or waist-to-hip ratio.

We have shown that markers of insulin resistance all increase significantly with increasing BMI in the range 34-77 kg/m². Standard lipid measures on the other hand show no unfavourable change with increasing BMI in this range. There appears to be a disparity between the insulin resistance markers and the conventional lipid markers for the metabolic syndrome as BMI increases. The fall in total cholesterol and subfractions with increasing BMI is interesting and raises the possibility of a metabolic change associated with increasing BMI. For example, hepatic dysfunction associated with increased obesity and hyperinsulinaemia may alter cholesterol levels. Perhaps it simply represents small dense cholesterol-depleted LDL particles seen with the metabolic syndrome. An examination of apolipoprotein B levels and LDL particle size may clarify this.

Increasing obesity and the associated metabolic syndrome is very much a lifestyle and a modern environmental problem. This study confirms the high prevalence of metabolic problems of those with morbid obesity and demonstrates total cholesterol levels and conventional lipid measures are of limited value in assessing risk in this group.

1 Australian Institute of Health and Welfare. Australia's health 1998: the sixth biennial health report of the Australian Institute of Health and Welfare. Australian Institute of Health and Welfare: Canberra 1998; AIHW Cat No Aus 10, p 151,

2 Eckel RH. American Heart Association call to action: obesity as a major risk factor for coronary heart disease. Circulation 1998; 97: 2099-2100, MEDLINE

3 Davidson M. Clinical implications of insulin resistance syndromes. Am J Med 1995; 99: 420-426, MEDLINE

4 Pouliot M, Despres J, Lemieux S et al. Waist circumference and abdominal saggital diameter: best simple anthropometric indexes of abdominal visceral adipose tissue accumulation and related cardiovascular risk in men and women. Am J Cardiol 1994; 73: 460-468, MEDLINE

5 Sattar N, Tan C, Han T, Forster L, Lean M, Shepherd J, Packard C. Associations of indices of adiposity with atherogenic lipoprotein. Int J Obes Relat Metab Disord 1998; 22: 432-439, MEDLINE

6 Han T, Van Leer E, Seidell J, Lean M. Waist circumference action levels in the identification of cardiovascular risk factors: prevalence study in a random sample. Br Med J 1995; 311: 1401-1405,

7 Despres J. The insulin resistance-dyslipidemia syndrome: the most prevalent cause of coronary artery disease. Can Med Assoc J 1993; 148: 1339-1340,

8 Gaudet D, Vohl M, Perron P et al. Relationship of abdominal obesity and hyperinsulinemia to angiographically assessed coronary artery disease in men with known mutations in the LDL receptor gene. Circulation 1998; 97: 871-877, MEDLINE

9 Lamarche B, Dagenais G, Despres J. Triglycerides and small, dense low density lipoprotein. JAMA 1998; 280: 1990-1991, MEDLINE

10 Despres J, Lamarche B, Mauriege P, Cantin B, Dagenais G et al. Hyperinsulinemia as an independent risk factor for ischemic heart disease. New Engl J Med 1996; 334: 952-957, MEDLINE

11 Sherrard B, Simpson H, Cameron J, Wahi S, Jennings G, Dart A. LDL particle size in subjects with previously unsuspected coronary heart disease: relationship with other cardiovascular markers. Atherosclerosis 1996; 126: 277-287, Article MEDLINE

12 Tchernof A, Lamarche B, Prud'homme D et al. The dense LDL phenotype: association with plasma lipoprotein levels, visceral obesity, and hyperinsulinemia in men. Diabetes Care 1996; 19: 629-637, MEDLINE

13 Grundy SC. Primary prevention of coronary heart disease: guidance from framingham: a statement for healthcare professionals from the AHA Task Force on Risk Reduction. Circulation 1998; 97: 1876-1887, MEDLINE

14 Gardner C, Fortmann S, Krauss R, Ronald M. Association of small low-density lipoprotein particles with the incidence of coronary artery disease in men and women. JAMA 1996; 276: 875-881, MEDLINE

15 Committee RFPSM. Risk Factor Prevalence Study: Survey no. 3 1989. National Heart Foundation and Australian Institute of Health: Canberra. 1990;

16 Bennett S. Trends in cardiovascular risk in Australia. Med J Aust 1994; 161: 519-527, MEDLINE

17 SPSS for Windows, 10.0.5 edn. SPSS. Inc.: Chicago, IL 1999;

18 Tietz textbook of clinical chemistry, 2nd edn. JB Lippincott: Philadelphia, PA 1986;

19 Oria H, Moorhead M. Bariatric analysis and reporting outcome system (BAROS). Obes Surg 1998; 8: 487-499, MEDLINE

20 Shape up America! Organization and the American Obesity Association. Guidelines for the treatment of adult obesity. Bethesda, MD 1996;

21 Borson-Chazot F, Harthe C, Teboul F, Labrousse F, Gaume C, Guadagnino L, Claustrat B, Berthezene F, Moulin P. Occurrence of hyperhomocysteinemia 1 year after gastroplasty for severe obesity. J Clin Endocrinol Metab 1999; 84: 541-545, MEDLINE

22 Manzato E, Zambon S, Zambon A et al. Lipoprotein sub-fraction levels and composition in obese subjects before and after gastroplasty. Int J Obes Relat Metab Disord 1992; 16: 573-578, MEDLINE

23 Nicklas B, Katzel L, Bunyard L, Dennis K, Goldberg A. Effect of an American Heart Association diet and weight loss on lipoprotein lipids in obesity. Am J Clin Nutr 1997; 66: 853-859, MEDLINE

24 Mori TA, Bao DQ, Burke V, Puddey IB, Watts GF, Beilin LJ. Dietary fish as a major component of a weight-loss diet: effect on serum lipids, glucose, and insulin metabolism in overweight hypertensive subjects. Am J Clin Nutr 1999; 70: 817-825, MEDLINE

25 Brolin R, Kenler H, Wilson A, Kuo P, Cody R. Serum Lipids after gastric bypass surgery for morbid obesity. Int J Obes 1990; 14: 939-950, MEDLINE

26 Cowan G, Buffington C. Significant changes in blood pressure, glucose, and lipids with gastric bypass surgery. World J Surg 1998; 22: 887-992,

27 Scott H, Dean R, LeQuire V et al. Alterations in plasma lipid concentrations in normal and hyperlipidemic patients with morbid obesity before and after jejunoileal bypass. Am J Surg 1978; 135: 341-347, MEDLINE

28 Hubert HB, Feinleib M, McNamara PM, Castelli WP. Obesity as an independent risk factor for cardiovascular disease: a 26-year follow-up of participants in the Framingham Heart study. Circulation 1983; 67: 968-977, MEDLINE

1 Mean lipid levels in severely obese patients compared with age and sex matched. Australian community control (CC) values from the risk factor prevalence study¹⁵ (mean levels (s.d.))

2 Lipid changes in quartiles based on BMI (mean values (s.d.))

3 Glucose metabolism changes in quartiles based on BMI (mean levels (s.d.))