Obesity Surgery

, 18:854

Systemic Inflammation and Cardiovascular Risk Factors: Are Morbidly Obese Subjects Different?


    • Department of GastroenterologyHospital das Clinicas
  • Patricia C. Marques
    • Nutrology Residency ProgramHospital das Clinicas
  • Luiz A. Bortolotto
    • Heart InstituteHospital das Clinicas
  • Jacob J. Faintuch
    • Department of Internal MedicineHospital das Clinicas
  • Ivan Cecconello
    • Department of GastroenterologyHospital das Clinicas
Research Article

DOI: 10.1007/s11695-008-9504-0

Cite this article as:
Faintuch, J., Marques, P.C., Bortolotto, L.A. et al. OBES SURG (2008) 18: 854. doi:10.1007/s11695-008-9504-0



Systemic inflammation is a hallmark of obesity as well as of other chronic diseases, usually indicating increased cardiovascular risk; however, studies with arterial documentation in morbid obesity are extremely scarce. Aiming to analyze correlation between inflammatory markers, pulse-wave velocity (PWV), and intima-media thickness (IMT), a prospective study was designed.


Morbidly obese patients [n = 29, age 46.3 ± 5.2 years, 82.8% females (24/29), BMI 44.9 ± 5.2 kg/m2] with C-reactive protein/CRP > 5 mg/l but free from trauma, infection, inflammation, or cancer were enrolled in this study. All were clinically stable candidates for elective bariatric operation. Variables included comorbidities, metabolic profile, inflammatory and coagulatory markers, and arterial morpho-functional indices.


Patients suffered from arterial hypertension (72.4%), metabolic syndrome (58.6%), and other comorbidities, but PWV and IMT were less aberrant than expected. Univariate correlation confirmed worse prognosis for those with metabolic syndrome and other accepted clinical risk factors. Multivariate confirmation was achieved for triglycerides (PWV) and D-dimer (IMT), but not for CRP, serum amyloid A, or neutrophil count, which were reversed in certain circumstances.


(1) Metabolic syndrome, hyperglycemia, hypertriglyceridemia and D-dimer were positively correlated with arterial measurements, whereas inflammatory and coagulatory markers often exhibited paradoxical association; (2) stratification confirmed that at certain levels of systemic inflammation or body mass index, acute phase proteins and other markers became unreliable or shifted signals; (3) when controlled for blood pressure, PWV was only moderately elevated, and IMT remained normal; (4) taken together, these findings are consistent with a unique interaction between adiposity, inflammation, and cardiovascular risk in seriously obese subjects.


Morbid obesityAtherosclerosisPulse wave velocityIntima media thicknessArterial stiffnessMetabolic syndromeArterial hypertensionDiabetesInflammationC-reactive protein


Cardiovascular disease is the leading cause of death in most countries, being responsible in the USA for as much as 40% of total mortality and up to US $430 billion of annual costs [1], and obesity is deemed a major predisposing factor for premature atherosclerosis, cardiovascular mortality, and also sudden death [2]. In the Nurses’ Health Study involving more than 100,000 women, obesity was directly related to all-cause mortality [3].

It has been debated for decades to what extent increased body mass index (BMI) is associated with a proinflammatory and prothrombotic phenotype [4, 5]. Adipokine over-expression and chronic microinflammation is not doubted anymore in severe obesity, and correlation with atherosclerosis, arterial hypertension, endothelial dysfunction, increased blood viscosity, insulin resistance, and various stigmata of the metabolic syndrome cluster has been demonstrated [4, 6, 7].

In spite of abundant publications about association between standard cardiovascular risk factors and BMI, including enormous populations such as the Framingham and the Framingham Offspring cohort study [7, 8], little is known about arterial structure and function measured by noninvasive procedures in morbidly obese subjects, especially concerning pulse-wave velocity (PWV) and intima-media thickness (IMT).

Clinical, biochemical, and epidemiologic information can be collected fairly easily in any group or setting; however, the same is not true for assessment of large-artery function and structure, which requires patient displacement and imaging methods. These are cumbersome diagnostic maneuvers in the case of subjects with huge adiposity of the neck or lower abdomen and limited mobility, particularly using equipment that has not been dimensioned for BMI > 40 kg/m2.

In this preliminary study, selected vascular and systemic markers were investigated in bariatric candidates. The primary aim was to correlate systemic inflammation with arterial stiffness and thickening, the secondary objective being correlation according to different sets of comorbidities and clinical profile. We hypothesized that conventional links between systemic inflammation and morpho-functional surrogates of atherosclerosis, in mild and moderate obesity, might not apply to subjects with massive adiposity and conspicuous inflammation, albeit asymptomatic.

Indeed, the impact on elderly people of obesity, even of a serious degree, concerns more disability than mortality [9]. This finding supports the speculation that at extremes of body weight, analogously to extremes of age, one could be dealing with selected organisms, more resistant to adverse vascular mechanisms. The possibility that comorbidities are more destructive than obesity itself might also be entertained, in spite of the fact that they tend to be quite intimately intertwined.

These considerations are timely as both old age and excessive body mass have ceased to be fringe problems. Whether cardiovascular outlook is less dramatically affected by obesity than previously thought or not, one is dealing with an immense public health problem that merits the highest priority.

Unfortunately, those are not trivial theories to test or follow, and major cohort studies have barely touched the controversy. More restricted surveys have equally bypassed the question or simply endorsed conventional assumptions, this being to the best of our knowledge, the first prospective attempt to critically examine systemic inflammation and arterial stiffness in morbid obesity, with or without comorbidities.

Materials and Methods


Patients with morbid obesity applying for pre-bariatric clinical evaluation during a 12-month period (October 1, 2006 to September 30, 2007, n = 44) were prospectively submitted to clinical and metabolic screening.

Criteria for inclusion encompassed those for elective bariatric surgery, namely active males or females 18–65 years old with BMI > 40 kg/m2, or >35 kg/m2 with comorbidities, plus C-reactive protein (CRP) >5 mg/l.

Criteria for exclusion were alcohol or drug addiction, fever or infectious foci, cancer with or without chemo/radiotherapy, chronic inflammatory illness (arthritis, enteritis or colitis), transplantation or immunologic disease, trauma, surgery or hospitalization in the last 30 days, use of steroidal or non-steroidal anti-inflammatory or immune-modulating agents or antibiotics, history of myocardial infarction, stroke, or surgery for peripheral arteriopathy, and refusal to participate in the protocol.


No randomization was part of the experimental design, as no intervention was envisaged. Nevertheless, a control normotensive euglycemic group with normal BMI or overweight only (n = 16) had their arterial variables evaluated, in order to permit baseline interpretation of vascular findings in morbidly obese subjects equally free from hypertension (n = 8).


In order to search for possible turning points of inflammatory and obesity subtypes versus atherosclerotic signs, stratification for BMI < 45 or >45 kg/m2 as well as for CRP < 10 or >10 mg/l was undertaken.

Definitions and Medications

Metabolic syndrome was established according to NCEP criteria [10]; diagnosis of principal comorbidities was based on current treatment. Clinical records and medical prescriptions were carefully checked, and new patients were routinely referred to the appropriate outpatient services for review of diagnosis as well as of dietary and pharmacologic compliance regarding diabetes, hypertension, and dyslipemia.

Lifestyle and Physical Activities

All obese subjects undergoing treatment in Hospital das Clinicas are routinely encouraged to lead an active life, employ a healthy diet matched to energy expenditure, and avoid sedentarism. Targeted alimentation or exercise protocols were not in use during this investigation.


Only weight, height, and BMI were registered, as waist circumference and waist/hip ratio are considered unreliable in subjects with BMI > 40 kg/m2 [11].

General Features and Laboratory Assays

Variables included demographic information, arterial blood pressure, serum albumin, glucose, hemoglobin and general hematologic counts, lipid fractions, leptin, insulin, CRP, serum-amyloid A, prothrombin time, fibrinogen, D-dimer, and C3/C4 complement fractions.

Information concerning hospitalizations and surgical procedures, as well as previous and current diseases was also documented.

Atherosclerosis Markers

Common carotid artery intima-media thickness (IMT), diameter (DIAM) and distensibility (DIST), as well as carotid-femoral pulse-wave velocity (PWV) were documented. Carotid-femoral PWV was measured by applanation tonometry to quantify arterial stiffness, by means of accurate positioning of the transducers on the corresponding arteries. The Complior automatic device was employed (Colson, Garges les Gonesses, France).

IMT, carotid diameter and distensibility were registered with a high-definition echo-tracking apparatus (Wall Track System, Medical Systems, Arnhem, The Netherlands), coupled with a conventional, two-dimensional vascular echograph (Sigma 44 Kontrom Instruments, Watford, UK). These measurements were made on the right common carotid artery 1 cm below the bifurcation at the site of the distal wall. Diagnostic procedures were carried out between 2:00 and 4:00 p.m. with the patient recumbent and awake, by an experienced observer blinded to the clinical condition.

Biochemical Methods

All laboratory tests were conducted by automated methods with routine quality-control. Plasma lipids were estimated according to the guidelines of NCAP-III [10].

Ethical Considerations

Informed consent was given for the study, and this protocol was approved by the Internal Review Board of Hospital das Clinicas.

Statistical Analysis

Findings are presented as percentage or mean ± SEM. For comparison of qualitative findings (comorbidities) Chi-square test was selected whereas numerical results, previously checked for normality by Kolmogorov–Smirnov test, were compared by analysis of variance (ANOVA) followed by post-hoc Bonferroni test. Associations were confirmed by the Pearson correlation test and by multivariate regression analysis (SPSS for Windows, SPSS, Chicago, IL, USA, version 14.0). A p value of less than 0.05 was considered significant.


Preliminary Findings

Initially recruited patients (n = 44) underwent 15 exclusions because of CRP</ = 5 mg/l (n = 9) or inflammatory/infectious conditions (n = 6). As a consequence, 29 subjects were effectively enrolled in this protocol.

Age of the population was 46.3 ± 5.2 years, 82.8% were females (24/29), BMI was 44.9 ± 5.2 kg/m2 and excess body weight was 51.5 ± 19.5 kg. Comorbidities and other clinical findings are shown in Table 1. No subject had a history of myocardial infarction, stroke, lower limb amputation, or any modality of coronary or peripheral arterial revascularization or stenting. Nevertheless, angina and claudication were present in small proportions of the group (Table 1).
Table 1

Clinical features




Diabetes mellitus






Arterial hypertension



Metabolic syndrome



Previous smokersa



Occasional angina



Intermittent claudication



aTwo subjects admitted to current smoking of <10 cigarettes/day; All others had discontinued tobacco for at least 3 years

Biochemical and Hematologic Findings

Principal results concerning lipid fractions, inflammation and coagulation markers, as well as other tests can be found in Table 2. Values were mostly acceptable or close to the normal range, however C-reactive protein was obviously elevated, with 24.1% of the readings (7/29) above 20 mg/l.
Table 2

Laboratory assays





Met Synd

Prev. smokers

Age (years)

46.3 ± 5.2

51.3 ± 8.8*

49.6 ± 11.0

46.3 ± 11.3

44.5 ± 10.4



100% (7/7)

76.2% (5/21)



BMI (kg/m2)

44.9 ± 5.2

42.1 ± 4.1

44.8 ± 5.9

44.9 ± 6.3

44.4 ± 4.2

Tchol mg/dl

193.6 ± 45.0

200.6 ± 60.1

191.3 ± 44.5

204.0 ± 53.1

175.3 ± 47.6

HDL mg/dl

51.4 ± 12.9

47.0 ± 9.2

50.9 ± 14.6

45.8 ± 9.8

47.2 ± 11.2

LDL mg/dl

116.4 ± 39.5

120.1 ± 52.4

114.4 ± 40.0

129.0 ± 43.1

101.3 ± 37.4

VLDL mg/dl

24.4 ± 13.4

29.4 ± 17.7

24.1 ± 13.7

26.2 ± 15.5

26.8 ± 13.0

Trig mg/dl

143.9 ± 79.1

201.0 ± 97.9

150.9 ± 83.5

170.4 ± 88.2

133.8 ± 64.6


7,390 ± 1,997

7,699 ± 2,330

7,520 ± 2,122

7,744 ± 1,896

7,412 ± 2,571


4,145 ± 1,338

4,214 ± 1,236

4,119 ± 1,371

4,300 ± 1,331

3,960 ± 1,561

Hemogl g/dl

13.7 ± 1.5

13.5 ± 1.9

14.0 ± 1.6

14.4 ± 1.4

13.8 ± 1.8

Albumin g/dl

4.3 ± 0.3

4.4 ± 0.3

4.4 ± 0.3

4.4 ± 0.3

4.3 ± 0.3

CRP mg/l

12.8 ± 7.9

8.1 ± 3.6*

12.2 ± 7.5

10.1 ± 5.8

11.3 ± 8.0

SAA mg/l

12.9 ± 12.3

7.4 ± 5.5

10.9 ± 10.8

9.8 ± 8.6

7.3 ± 6.8

C3 mg/dl

150.7 ± 24.5

149.0 ± 22.2

147.9 ±  23.5

152.6 ± 22.6

146.7 ± 16.9

C4 mg/dl

29.8 ± 11.0

25.0 ± 7.7

28.9 ± 12.3

29.8 ± 11.1

29.1 ± 11.1

Gluc mg/dl

105.7 ± 31.2

131.7 ± 54.5*

109.6 ± 35.6

113.6 ± 38.2

111.6 ± 46.3

Insulin uU/ml

21.5 ± 15.8

28.8 ± 25.4

22.6 ± 17.6

26.0 ± 18.0

27.9 ± 22.8

Leptin ng/ml

42.9 ± 34.6

49.3 ± 57.0

45.8 ± 36.7

39.7 ± 40.8

50.3 ± 45.5

Proth % activ

105.3 ± 3.1

105.4 ± 0.8

105.1 ± 3.7

104.6 ± 3.9

104.3 ± 4.7

Fibrin mg/dl

446.8 ± 88.6

420.4 ± 68.4

457.8 ± 87.9

452.6 ± 86.9

421.9 ± 105.9

DD ng/ml

535.1 ± 379.4

373.2 ± 349.6*

578.9 ± 410.3

517.7 ± 399.4

502.6 ± 395.6

TChol Total cholesterol, Normal < 200; HDL HDL-cholesterol, Normal > 65 (females), >55 (males); LDL LDL-cholesterol, Normal < 130; VLDL VLDL-cholesterol, Normal < 40 ; Trig triglycerides, Normal < 150; WBC white blood cell count, Normal 4,000–10,000; Neutr neutrophil count, Normal 1,300–7,000; Hemogl Hemoglobin, Normal > 12 (females), >13 (males); CRP C-reactive protein, normal < 3; SAA serum amyloid A, Normal < 6.4; C3 complement fraction C3, Normal 90–180; C4 complement fraction C4, Normal 10–40 ; Proth % activ: Prothrombin time % activity, normal > 70%; Fibrin Fibrinogen, Normal 150–400; DD D-dimer, measured as Fibrinogen Equivalent Units, Normal < 500;

*P < 0.05

Diabetics were older and exhibited higher glucose concentration as expected, however C-reactive protein and D-dimer values were lower than in the remaining subjects (Table 2).

Atherosclerotic indices can be examined in Table 3. Differences again occurred in diabetics, in the form of elevated pulse-wave velocity and diminished distensibility.
Table 3

Arterial measurements






Prev. smokers

PWV (m/s)

10.0 ± 1.9

11.5 ± 1.8*

10.2 ± 2.0

10.9 ± 2.1

9.7 ± 1.8

IMT (mm)

0.66 ± 0.16

0.70 ± 0.10

0.68 ± 0.16

0.68 ± 0.17

0.63 ± 0.11

Diameter (mm)

7.20 ± 0.86

7.01 ± 0.79

7.34 ± 0.87

7.55 ± 1.00

7.00 ± 0.68

Dist (1/mmHg 102)

5.3 ± 1.7

4.5 ± 1.1*

5.3 ± 1.9

5.1 ± 1.8

5.4 ± 1.7

Dist Distensibility

*P < 0.05

Normotensive nondiabetic control subjects were included as a benchmark for vascular indices. PWV and diameter were increased, but not IMT. Distensibility was numerically different, but significance was not confirmed (Table 4).
Table 4

Comparison between normotensive cases (n = 8) and normotensive controls (n = 16)





Age (years)

40.6 ± 10.8

38.9 ± 7.0


BMI (kg/m2)

47.5 ± 7.0

29.1 ± 2.0

P = 0.000

PWV (m/s)

9.8 ± 1.4

8.4 ± 0.9

P = 0.009

IMT (mm)

0.61 ± 0.15

0.58 ± 0.09


Diameter (mm)

6.90 ± 0.61

6.37 ± 0.58

P = 0.06

Dist (1/mmHg 102)

5.6 ± 1.4

6.7 ± 2.4


Dist Distensibility

Hypertension had a preference for older age. Diabetes, or at least hyperglycemia, also tended to be associated with older age, as well as with hypertriglyceridemia, but C-reactive protein displayed negative correlation (significance not achieved). Metabolic syndrome tended to correlate with its diagnostic components, namely lipid fractions and insulin, but also with male gender. Two inflammatory indices were negatively linked to this diagnosis. Male gender was analogously predictive of negative inflammatory or coagulatory markers (Table 5).
Table 5

General correlations of conventional clinical and metabolic variables


Positive correlation

Negative correlation


Hypertension r = 0.454, P = 0.02

BMI r = −0.550, P = 0.003

Glucose r = 0.335, P = 0.07

Male gender

Hemoglobin r = 0.539, P = 0.004

CRP r = −0.425, P = 0.03

Metabolic syndrome r = 0.383, P = 0.004

Neutrophil r = −0.356, P = 0.05


HDL r = −0.585, P = 0.000


Proth r = −0.523, P = 0.005

Met Synd

Insulin r = 0.344, P = 0.06

CRP r = −0.421, P = 0.03

LDL r = 0.402, P = 0.03

Neutrophil r = −0.421, P = 0.03

Triglycerides r = 0.407, P = 0.03

HDL r = −0.547, P = 0.003


Triglycerides r = 0.415, P = 0.03

CRP r = −0.348, P = 0.06

Correlation of inflammatory and coagulatory cells and proteins between themselves was often present, as expected, but two lipid fractions, namely HDL and VLDL, also reached statistical significance in some tests (Table 6). There were no negative associations except for those already signaled in Table 5.
Table 6

Correlations of inflammatory markers and coagulation measurements


Positive correlation

C-reactive protein

HDL r = 0.430, P = 0.04

SAA r = 0.416, P = 0.03

Proth r = 0.380, P = 0.05

Fibrinogen r = 0.550, P = 0.003

Serum amyloid A

D-dimer r = 0.371, P = 0.05

WBC count

VLDL r = 0.500, P = 0.008

Prothrombin activity r = 0.336, P = 0.07

Fibrinogen r = 0.330, P = 0.07

Neutrophil count

Prothrombin activity r = 0.371, P = 0.05

Prothrombin activity

Fibrinogen r = 0.491, P = 0.008

Each of the four measurements presented its own associations. However, age and metabolic syndrome were involved with two indices. In this sense, male gender and hypertension were associated with higher diameter, hyperglycemia with PWV, metabolic syndrome with PWV and diameter, and older age with PWV and IMT.

The profile of inflammatory variables was incongruously favorable. Only D-dimer, which is an index of coagulation/fibrinolysis, was associated with a dangerous measurement, IMT, a finding confirmed by the multivariate test. In contrast, serum amyloid A and neutrophil count were protective, exhibiting positive correlations with a desirable arterial index (distensibility). Most notable was C-reactive protein, which achieved strongly negative significance with regard to PWV, again revealing an advantageous prognosis (Table 7).
Table 7

Principal correlations of arterial indices







r = 0.551, P = 003

r = 0.631, P = 0.000



r = 0.441, P = 0.03




r = −0.364, P = 0.05




r = 0.377, P = 0.05



r = 0.420, P = 0.04


C-reactive prot.

r = −0.426, P = 0.04


Serum amyloidA


r = 0.421, P = 0.04

Neutrophil count


r = 0.414, P = 0.04

Male gender


r = 0.682, P = 0.000


Metabolic synd

r = 0.471, P = 0.01


r = 0.477, P = 0.01




r = 0.524, P = 0.005


The list of independent factors related to arterial indices, which were confirmed by multivariate analysis, is transcribed here: (1) PWV—age (p = 0.03, mean square 5.42), triglycerides (p = 0.01, mean square 8.08), and systolic blood pressure (p = 0.006, mean square 15.7); (2) IMT—age (p = 0.001, mean square 15.58), D dimer (p = 0.009, mean square 9.54); (3) Diameter—male gender (p = 0.01, mean square 9.01); (4) Distensibility—diastolic blood pressure (p = 0.002, mean square 10.1).

In order to shed some light on the possible influence of exaggerated body weight or unbridled systemic inflammation on the performance of vascular markers, correlations were analyzed again after stratification according to BMI and CRP (Table 8).
Table 8

Correlations according to BMI and CRP status






CRP < 10 mg/L

TChol r = 0.551, P = 0.03

DD r = 0.787, P = 0.000

DD r = 0.575, P = 0.02

DD r = −0.642, P = 0.009

LDL r = 0.538, P = 0.03

MetSynd r = 0.434, P = 0.07

Trig r = 0.512, P = 0.04


MetSynd r = 0.522, P = 0.04


CRP > 10 mg/L

HDL r = −0.690, P = 0.01

DD r = 0.546, P = 0.05

Neutr r = 0.559, P = 0.05

Neutr r = 0.702, P = 0.008

CRP r = −0.456, P = 0.09

MetSynd r = 0.484, P = 0.08

WBC r = 0.770, P = 0.002


CRP r = −0.551, P = 0.05

BMI < 45 kg/m2

Neutr r = 0.624, P = 0.01

DD r = 0.593, P = 0.02

Insulin r = −0.491, P = 0.05

HDL r = 0.542, P = 0.0.03

PCR r = −0.670, P = 0.008

CRP r = −0.599, P = 0.03


SAA r = −0.497, P = 0.05


HDL r = −0.591 P = 0.02

BMI > 45 kg/m2

Trig r = 0.598, P = 0.03

Glucose r = 0.682, P = 0.01

Neutr r = 0.732, P = 0.005

DD r = 0.618, P = 0.03

Glucose r = 0.888, P = 0.000

Fibrinogen r = 0.632, P = 0.03

Fibrinogen r = 0.583, P = 0.04

MetSynd r = 0.592, P = 0.003

WBC r = 0.648, P = 0.02

WBC r = 0.727, P = 0.005


Insulin r = 0.570, P = 0.05

SAA r = 0.498, P = 0.08


SAA r = 0.497, P = 0.08


MetSynd r = 0.697, P = 0.009


BMI < 45: n = 16/29, 55.2% ; BMI > 45: n = 13/29, 44.8% ;CRP < 10: n = 16/29, 55.2% ; CRP > 10: n = 13/29, 44.8%TChol Total cholesterol; Trig triglycerides; MetSynd metabolic syndrome; DD D-dimer; Neutr: neutrophil count; WBC white blood cell count

C-reactive protein confirmed its negative correlation with PWV only when BMI < 45, and both CRP and SAA predicted positive prognosis in the same group regarding arterial diameter. SAA only partially hinged on BMI, as patients with BMI > 45, exhibited positive correlation for both an undesirable (diameter) and a valuable measurement (distensibility).

Inconsistencies in the predictive properties of D-dimer were also unveiled. Conventionally linked to a pro-inflammatory and pro-thrombotic state, as here evidenced with regard to IMT and diameter in various circumstances, it became associated with improved distensibility in patients with BMI > 45, whereas the opposite was demonstrated when CRP < 10.

CRP associations were also partially susceptible to CRP plasma levels. No significant results were observed with CRP < 10. With higher concentrations, a mixed profile emerged, when negative correlations with both a deleterious (PWV) and a beneficial result (distensibility) occurred.


There are compelling reasons to believe that morbid obesity does not simply represent an additional step in the sequence of adiposity. Drapeau et al. called attention to the fact that visceral obesity, as measured by waist circumference, correlated poorly or not at all with biochemical tests of glucose and lipid physiology in severely obese women. Reinforcing the suspicion that such anthropometric documentation was quite useless in patients with BMI > 40 kg/m2, they conclude that it conducted to overdiagnosis of metabolic syndrome, as defined by the NCEP ATP III guidelines, by 72% [11].

Livingston et al. underline that inconsistencies between obesity and risk factors for CVD are not mere artifacts but part of a truly uneven profile. According to information retrieved from the Third National Health and Nutrition Examination Survey (NHANES III) database, important indices such as insulin, C-peptide, apolipoprotein B, non-high-density lipoprotein cholesterol, low-density lipoprotein, and serum triglycerides peak at the BMI range of 30–40 kg/m2. Thereafter, they diminish [12], potentially generating a flat or even inverted-U matrix.

In the current protocol, BMI exhibited virtually no correlation with biochemical, inflammatory or arterial variables, in disagreement with established experience in lesser forms of obesity [1, 2, 5, 6].

Other prognostic indicators were fairly useful in the current experience including CRP and neutrophil count. However, sometimes displaying reverse correlation, not a direct one, with metabolic syndrome or diabetes (Table 4).

A few inflammatory and one coagulatory marker were endowed with associations with vascular measurements, but counter-intuitively, they occasionally pointed towards diminished risk factors instead of noncompliant arteries [2, 4, 13]. CRP negatively correlated with PWV and SAA was positively linked to distensibility, when the opposite was expected (Table 7). Another example was neutrophil count, which curiously portended good arterial distensibility. Even D-dimer, which presented the proper signal for the whole group and was confirmed at multivariate analysis (Table 7), aimed at the erroneous target for certain population strata (Table 8).

Conventional bulwarks did not fail, confirming that for most purposes, these men and women with their array of comorbidities were not intrinsically different from subjects with less massive obesity studied elsewhere. The classic atherogenic cluster of metabolic syndrome and its component feature of hyperglycemia, and particularly of hypertriglyceridemia which stood out on multivariate analysis, never switched sides and continued to move in parallel with vascular structural and physiologic deterioration. The same could be affirmed for older age, male gender and high blood pressure, on the basis of both uni- and multivariate tests (Tables 6 and 7) [57].

This was an asymptomatic or reasonably well-compensated population, with the distinguishing feature that acute-phase markers were obviously elevated to as high as 20 mg/l or more, in the case of C-reactive protein. In the light of exclusion criteria that focused all common infectious and inflammatory diseases, there remains little doubt that visceral adipose tissue was the overarching source of an adipokine/cytokine/chemokine-mediated systemic phenomenon.

This unfavorable milieu notwithstanding, when controlled for glucose and hypertension PWV values were only moderately deranged and IMT not at all. Contrary to ordinary assumptions, far-advanced obesity including an overt inflammatory phenotype was compatible with healthier blood vessels than hitherto suspected.

As alluded to, the other remarkable finding was that at least in some contexts, pro-inflammatory cells and biomolecules could be negatively correlated to early signs of atherosclerosis. This radically conflicts with the roles of the components of the cytokine net, with emphasis on adiponectin but not excluding C-reactive protein, which are incriminated in aberrations of endothelial NO production and pathologic neointima formation, with possible troubles also in blood viscosity, platelet activation, coagulation and fibrinolysis [2, 4, 1317].

It is often criticized that cardiovascular prognostic interest of C-reactive protein has been overestimated in the past, nevertheless significant correlation with arterial stiffness and intima-media thickness continues to be reported in various disease groups, albeit not those with morbid obesity [1317]. It should not be overlooked that the same obesity-related peptides such as leptin, resistin, IL-6, TNF-alpha, and IL-17 that activate acute-phase proteins and cells, as well as coagulation and fibrinolysis factors in the liver, are known to fuel the expression of endothelial adhesion molecules, proteases, and other mediators, thus precipitating a pro-atherothrombotic vascular environment [4, 7, 1417].

CRP itself is present in atherosclerotic plaques, and experimentally this biomolecule stimulates oxidized LDL uptake by macrophages, as well as intracellular cholesterol ester accumulation [15]. CRP could also flag atheroma instability and intraplaque hemorrhage, as recently emphasized in the carotid artery of elderly hypertensive subjects [16].

Quite a few groups attribute pathogenesis of cardiovascular events in subjects with low-degree systemic inflammation to progressive immunoninflammatory attack of the endothelium, particularly within the context of metabolic syndrome and obesity [2, 4, 1317].

No explicit evidence of incongruous or erratic association between acute-phase molecules and vascular status at high levels of obesity and microinflammation, or of relatively unhindered vessels in spite of the theoretically ominous milieu, could be found in the literature. Stratification of the patients according to BMI and to CRP levels failed to elicit any signature patterns. Nonetheless, it reinforced the pattern of cardiovascular unpredictability of this class of obesity. As shown in Table 8, certain biochemical markers were able to migrate from positive outlook to lack of significance or even reverse association with arterial findings, depending on the analyzed category.

One single experience could be found with bariatric candidates which stratified subjects according to low and high cardiovascular risk, and measured arterial elasticity albeit by a different procedure. Examination of the pertinent results substantially corroborates the paradigm shift [18]. Correlation between C-reactive protein and elasticity was not the target of that investigation. Yet, it is informed that baseline value for the protein was 15.6 ± 10.9 g/l in the low-risk population, and 13.0 ± 7.2 g/l in the riskier patients, approximately the same range as here observed. Elasticity was found out to be, respectively, 16.3 ± 6.3 versus 13.6 ± 4.7 (large vessel) and 9.0 ± 5.1 versus 6.3 ± 2.7 ml/mmHg*100 (small artery).

Definitely, as one moves from low to high cardiovascular risk both CRP and elasticity measurements numerically decrease, suggesting a positive correlation between them that should actually be negative, as elasticity bodes well for cardiovascular prognosis.

An obesity paradox, with diminished cardiovascular morbidity and mortality, has been documented for a number of years in a variety of contexts including chronic heart failure, patients submitted to coronary interventions, after myocardial infarction, and with hypertension and coronary disease [19, 20].

Circumstantial events could partially underlie this conflict, especially a lead-time bias, as seriously obese individuals could eventually develop symptoms at a younger age or receive medical attention earlier, when response is arguably more favorable [19, 20]. It is worth mentioning that in the vast epidemiological investigation of Fonarow et al., encompassing over 100,000 subjects, even after correction for age, sex, blood pressure, and other variables, obese patients with acute heart failure died less. This study is further relevant for including subjects with BMI up to 60 kg/m2 [21].

Reverse epidemiology between selected cardiovascular risk factors including obesity, and end-stage renal patients receiving dialysis, those with advanced malignancies, and individuals with advanced age, is also registered [7, 22].

Again, a time discrepancy between competing risk factors, resulting in an asymmetric clinical course, has been hypothesized to explain the better outcome in those with massive adiposity [19, 20, 22].

In none of the alluded to surveys has an inflammation paradox, or an atypical relationship between acute-phase biomolecules and arterial stiffness, been reported neither in obese nor in non-obese populations. Only moderate influence of race and ethnicity on C-reactive protein, adhesion molecules and other cardiovascular risk factors has been highlighted so far [1].

As an alternative, the nominally more favorable survival of obese patients suffering from those diseases has been ascribed to the poorer clinical course of the less generously nourished subjects, albeit not necessarily cachectic, on account of an exacerbated malnutrition–inflammation complex [22].

The relatively limited number of patients in the current protocol, which was a consequence of the strict selection criteria but anyway prevented additional stratifications, is a weak point of the investigation. It could theoretically conduct to loss of linearity in the relationship between C-reactive protein and other indicators from one side, and outcome results from the other, due not to clinical reasons but on account of the small-sample bias.

In this sense, great care was exercised in the selection of the population according to the inclusion and exclusion criteria, and also in the preliminary testing for Gaussian distribution of the findings by means of the Kolmogorov–Smirnov analysis.

Absolute homogeneity was not achieved because of the intrinsic clinical and metabolic variability of bariatric candidates, depending on their demographics and comorbidities. Yet, the fact that conventional risk factors maintained the expected profile, as partially established also by multivariate correlation, and only inflammatory markers diverged, points toward a credible result.

In the light of the current findings, it is improbable that the uncoupling of inflammatory and clinical cardiovascular risk factors in certain classes of obesity is due to merely circumstantial factors. Further elucidation will depend on careful follow up, before and after bariatric intervention, including documentation of adiponectin, cytokines, chemokines, and adhesion molecules, along with tests for endothelial and arterial performance, as well as registration of long-term course of cardiovascular events. That will be essential to provide more insight about pathophysiology of the prothrombotic and proinflammatory reaction, and its actual repercussions on cardiovascular health in all bariatric categories, both below and above the traditional BMI threshold of 40 kg/m2.


(1) Metabolic syndrome, hyperglycemia, hypertriglyceridemia and D-dimer were positively correlated with arterial measurements, whereas inflammatory and coagulatory markers often exhibited paradoxical association; (2) stratification confirmed that at certain levels of systemic inflammation or body mass index, acute phase proteins and other markers became unreliable or shifted signals; (3) when controlled for blood pressure, PWV was only moderately elevated in the presence of severe obesity and IMT remained normal; (4) taken together, these findings are consistent with a unique interaction between adiposity, inflammation, and cardiovascular status in morbidly obese subjects.


The valuable assistance of Marcio Mancini, MD, PhD, Luiz A. M. Cesar, MD, PhD, Lilian M. Horie, RD, Alfredo Halpern, MD, PhD and Artur B. Garrido Jr, MD, PhD, is appreciated.

Copyright information

© Springer Science + Business Media, LLC 2008