Visser M, Bouter LM, McQuillan GM, Wener MH, Harris TB. Elevated C-Reactive Protein Levels in Overweight and Obese Adults. JAMA. 1999;282(22):2131-2135. doi:10.1001/jama.282.22.2131
Author Affiliations: Institute for Research in Extramural Medicine, Faculty of Medicine (Dr Visser and Prof Bouter), Vrije Universiteit, Amsterdam, the Netherlands; Epidemiology, Demography, and Biometry Program, National Institute on Aging, National Institutes of Health, Bethesda, Md (Drs Visser and Harris); National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, Md (Dr McQuillan); and the Departments of Laboratory Medicine and Medicine, University of Washington, Seattle (Dr Wener).
Context Human adipose tissue expresses and releases the proinflammatory cytokine
interleukin 6, potentially inducing low-grade systemic inflammation in persons
with excess body fat.
Objective To test whether overweight and obesity are associated with low-grade
systemic inflammation as measured by serum C-reactive protein (CRP) level.
Design and Setting The Third National Health and Nutrition Examination Survey, representative
of the US population from 1988 to 1994.
Participants A total of 16,616 men and nonpregnant women aged 17 years or older.
Main Outcome Measures Elevated CRP level of 0.22 mg/dL or more and a more stringent clinically
raised CRP level of more than 1.00 mg/dL.
Results Elevated CRP levels and clinically raised CRP levels were present in
27.6% and 6.7% of the population, respectively. Both overweight (body mass
index [BMI], 25-29.9 kg/m2) and obese (BMI, ≥30 kg/m2)
persons were more likely to have elevated CRP levels than their normal-weight
counterparts (BMI, <25 kg/m2). After adjustment for potential
confounders, including smoking and health status, the odds ratio (OR) for
elevated CRP was 2.13 (95% confidence interval [CI], 1.56-2.91) for obese
men and 6.21 (95% CI, 4.94-7.81) for obese women. In addition, BMI was associated
with clinically raised CRP levels in women, with an OR of 4.76 (95% CI, 3.42-6.61)
for obese women. Waist-to-hip ratio was positively associated with both elevated
and clinically raised CRP levels, independent of BMI. Restricting the analyses
to young adults (aged 17-39 years) and excluding smokers, persons with inflammatory
disease, cardiovascular disease, or diabetes mellitus and estrogen users did
not change the main findings.
Conclusion Higher BMI is associated with higher CRP concentrations, even among
young adults aged 17 to 39 years. These findings suggest a state of low-grade
systemic inflammation in overweight and obese persons.
Adipose tissue previously was considered a passive storage depot for
fat but is now known to play an active role in metabolism.1,2
Among the recently discovered compounds expressed in human adipose tissue
is the proinflammatory cytokine interleukin 6 (IL-6).3,4
Moreover, IL-6 produced in the adipose tissue of healthy humans is released
into the circulation.4,5 Adipose
tissue is estimated to produce about 25% of the systemic IL-6 in vivo.4 Because of the inflammatory properties of IL-6, including
the stimulation of acute-phase protein production in the liver,6,7
the release of IL-6 from adipose tissue may induce low-grade systemic inflammation
in persons with excess body fat.
A sensitive marker for systemic inflammation is the acute-phase C-reactive
protein (CRP). In a meta-analysis of 7 prospective studies, elevated serum
CRP concentration was shown to predict future risk of coronary heart disease.8 C-reactive protein levels well below the conventional
clinical upper limit of normal of 1 mg/dL have been associated with a 2- to
3-fold increase in risk of myocardial infarction, ischemic stroke, peripheral
arterial disease, and coronary heart disease mortality in healthy men and
This study tested whether overweight and obesity are associated with
low-grade systemic inflammation as measured by serum CRP concentration.
The study included 16,616 adult participants of the Third National Health
and Nutrition Examination Survey (NHANES III), 1988-1994. NHANES III was conducted
by the National Center for Health Statistics of the Centers for Disease Control
and Prevention.14 The survey had a complex,
stratified, multistage probability-cluster design for selecting a sample of
approximately 40,000 persons representative of the noninstitutionalized civilian
US population. Children younger than 5 years, persons aged 60 years or older,
Mexican American persons, and non-Hispanic blacks were sampled at higher rates
than others. Eighty-one percent of all eligible adults consented to an initial
interview in their household. Of the 20,050 persons aged 17 years or older
who were interviewed, 18,162 were subsequently examined in a mobile examination
center or in their homes. Persons with missing data on height, body weight,
or serum CRP level (n = 1239) and pregnant women (n = 307, validated by urine
pregnancy test) were excluded, leaving 16,616 persons (7938 men and 8678 women)
available for the statistical analyses.
Body weight and height were measured using standardized procedures.15 Body mass index (BMI) was calculated as weight in
kilograms divided by the square of height in meters and used as an indicator
of body fat.16,17 The 1998 clinical
guidelines18 were used to define overweight
(BMI, 25-29.9 kg/m2) and obesity (BMI ≥30 kg/m2).
Waist circumference was measured at the level of the high point of the
iliac crest and the circumference at the level of maximum extension of the
buttocks.15 The waist-to-hip ratio, calculated
as waist circumference divided by hip circumference, was used as an indicator
of abdominal visceral fat.19
Serum specimens for the measurement of CRP were stored at −70°C
and analyzed within 2 months after phlebotomy. C-reactive protein was analyzed
using a modification of the Behring Latex-Enhanced CRP assay on the Behring
Nephelometer Analyzer System (Behring Diagnostics, Westwood, Mass) (M.H.W.,
Phyllis R. Daum, MT [ASCP], G.M.M., unpublished data, 1999). Both within-
and between-assay quality control procedures were used and the coefficient
of variation of the method was 3.2% to 16.1% through the period of data collection.
The assay could detect a minimal CRP concentration of 0.22 mg/dL, and values
below this level were classified as undetectable. The assay was designed primarily
to detect inflammation and was included as part of the NHANES III cohort to
help detect inflammation as a confounding variable for interpretation of nutrition
markers. Because most individuals had values less than the minimum detectable
concentration, CRP is treated as a categorical rather than a continuous variable.
Race was defined by self-report as non-Hispanic white, non-Hispanic
black, or Mexican American. People outside these categories were classified
as other. Smoking status was based on self-report and categorized as never,
former, or current smoking. All persons with a serum cotinine concentration
of more than 57 nmol/L (10 ng/mL)20 as measured
by high-performance liquid chromatography and atmospheric-pressure chemical
ionization tandem mass spectroscopy21 were
categorized as current smokers, irrespective of self-report. Inflammatory
disease prevalence was determined through self-report of physician-diagnosed
conditions (chronic bronchitis, asthma, emphysema, and rheumatoid arthritis)
and self-report of "having a cold" in the past few days. A serum tube dilution
latex fixation test for rheumatoid factor was assessed in persons aged 60
years or older.22 All persons with a positive
test result (≥1:40 titer) were categorized as having rheumatoid arthritis
or a related inflammatory disorder, irrespective of self-report. Cardiovascular
disease included self-reported physician-diagnosed myocardial infarction and
stroke and angina as assessed by the Rose Angina Questionnaire.23
Diabetes mellitus was defined as self-reported physician-diagnosed diabetes
mellitus with insulin use or, in the case of undiagnosed diabetes mellitus,
a fasting plasma glucose level of at least 6.99 mmol/L (126 mg/dL).24,25 Estrogen use was based on self-report,
categorized as contraceptive medications (oral or implant) or estrogen replacement
The study population was divided into 2 categories based on CRP concentration,
undetectable (<0.22 mg/dL) and elevated (≥0.22 mg/dL). The population
was also divided into 2 categories based on the conventional clinical cut
point for inflammation, a CRP concentration of more than 1.00 mg/dL. Two outcome
variables were defined: elevated CRP level (≥0.22 mg/dL), which was compared
with undetectable CRP, and clinically raised CRP level (>1.00 mg/dL), which
was compared with CRP level of no more than 1.00 mg/dL. Within each sex, the
relationship between BMI and CRP concentration category was examined by multiple
logistic regression analysis. We calculated odds ratios (ORs) and 95% confidence
intervals (CIs) for BMI as a categorical variable according to the clinical
guidelines, with normal weight (BMI <25 kg/m2) as the reference
category, and for BMI as a continuous variable, expressed per 5-kg/m2 (about 1 SD) increment. Moreover, ORs per SD increment of waist-to-hip
ratio (0.1 units) were calculated. Adjustments were made for potential confounders,
including age, race, smoking status, estrogen use, inflammatory disease, and
other diseases associated with low-grade inflammation, including cardiovascular
diabetes mellitus.28 To assess potential effect
modification by age, smoking status, disease status, or estrogen use, the
analyses were repeated, restricted to young (aged 17-39 years), healthy non–estrogen-using
nonsmokers. Odds ratios do not approximate risk ratios when the prevalence
of the outcome variable in the study population is greater than 10%.29 The calculated OR for elevated CRP concentration
therefore should not be interpreted as a risk ratio. Analyses were performed
using SAS (SAS Institute Inc, Cary, NC) and SUDAAN (Research Triangle Institute,
Research Triangle Park, NC) and incorporated sampling weights to account for
oversampling and nonresponse to the household interview and examination.30 Variance estimates were calculated with SUDAAN, incorporating
the complex sampling design of NHANES III.30
Elevated CRP levels (≥0.22 mg/dL) were present in 21.8% of men and
33.1% of women, and clinically raised CRP levels (>1.00 mg/dL) in 4.4% and
8.9%, respectively. Other characteristics of the study population are shown
in Table 1.
With increasing BMI, the prevalence of elevated CRP level increased
in both men and women (Figure 1).
However, with increasing BMI the prevalence of clinically raised CRP level
increased among women only; the prevalence was 4.0% (95% CI, 3.3%-4.8%) in
normal-weight women, 7.7% (95% CI, 6.4%-9.4%) in overweight women, and 20.2%
(95% CI, 18.1%-22.5%) in obese women.
Obese men were 2.13 times more likely and obese women 6.21 times more
likely to have elevated CRP levels compared with their normal-weight counterparts
(Table 2). Per 1-SD increase in
BMI, men were 1.38 and women were 2.04 times more likely to have elevated
CRP levels. Among women, BMI was also associated with clinically raised CRP
levels. Obese women were 4.76 times more likely to have clinically raised
CRP levels compared with normal-weight women. Per 1-SD increment in BMI, women
were 1.69 times more likely to have clinically raised CRP levels.
The waist-to-hip ratio was independently associated with both elevated
and clinically raised CRP levels in men and women. Per 1-SD increase in waist-to-hip
ratio, men were 1.41 and women were 1.21 times more likely to have elevated
CRP levels (Table 2). The OR for
clinically raised CRP levels per 1-SD increase in waist-to-hip ratio was 1.36
in men and 1.28 in women.
The association between BMI and CRP was also investigated after stratification
by age group (young = 17-39 years; middle-aged = 40-59 years; old = ≥60
years). Among women, the association between BMI and CRP categories was influenced
by age group. Older obese women were less likely to have elevated or clinically
raised CRP levels than young obese women. A similar effect modification by
age group in women was observed using BMI as a categorical variable. No effect
modification by age group was observed in men.
To avoid any potential effect modification by age, inflammatory disease,
cardiovascular disease, diabetes mellitus, current smoking, or estrogen use,
the analyses were repeated restricted to healthy, nonsmoking, non–estrogen-using
persons aged 17 to 39 years. The positive association between BMI category
and elevated CRP level remained statistically significant after adjustment
for age, race, smoking status (never and former smoking only), and waist-to-hip
ratio (Table 3). In this restricted
analysis, BMI also remained positively associated with clinically raised CRP
levels among women.
Previous studies in middle-aged and elderly persons have reported a
positive association between BMI and CRP concentration.12,26,27
However, in these age groups, the association may have been confounded by
disease. Rheumatoid arthritis, diabetes mellitus, and cardiovascular disease
are prevalent diseases in older persons and are associated with both obesity31- 33 and increased CRP
We carefully controlled for inflammatory disease and other factors known to
influence CRP concentrations. A higher prevalence of low-grade systemic inflammation
was observed in overweight and obese persons compared with normal-weight persons.
Most importantly, our study extends these findings to young adults aged 17
to 39 years, in whom the prevalence of any confounding subclinical disease
is generally very low. Of interest is our observation that the distribution
of body fat is associated with CRP concentration independent of BMI. A high
waist-to-hip ratio, indicative of a large amount of abdominal visceral fat,
was associated with low-grade systemic inflammation in men and women.
Our results, together with the evidence of previous studies, have important
implications for the health risks of overweight and obese individuals, including
those at young ages. Based on NHANES III data, we estimated that 53.9% of
US adults aged 17 years or older are overweight or obese. Overweight, obesity,
and a large waist-to-hip ratio pose a considerable health risk, including
cardiovascular health.33,35- 37
Low-grade systemic inflammation has been shown to increase the risk for cardiovascular
Some of the increased risk for cardiovascular disease in overweight and obese
persons may be explained by our observation that increased CRP concentrations
are more prevalent in these persons.
C-reactive protein concentrations well below the conventional clinical
upper limit of normal of 1 mg/dL have been associated with a 2- to 3-fold
increase in risk of myocardial infarction, ischemic stroke, and peripheral
arterial disease in healthy men and women.9- 13
In addition, elevated CRP levels are predictive of cardiac complications in
patients with unstable angina or myocardial infarction38,39
and CRP induces the production of tissue factor, a potent procoagulant, in
monocytes.40 Moreover, elevated CRP concentrations
are associated with increased coronary heart disease mortality and total mortality.9,41
Approximately 25% of circulating IL-6 is estimated to be released by
human subcutaneous adipose tissue in vivo,2
and IL-6 stimulates the production of acute-phase proteins in the liver.6,7 This might explain the observed associations
between BMI and CRP. In vitro, human abdominal visceral adipose tissue releases
more IL-6 compared with subcutaneous adipose tissue,5
possibly explaining our observation that a higher waist-to-hip ratio, after
adjustment for BMI and several confounders, was independently associated with
elevated CRP level.
Body mass index is an important clinical indicator of overweight and
obesity,18 but its use as an indicator of body
fatness has limitations. At a similar BMI, women have more body fat than men.42 This difference was reflected in our data, showing
a higher prevalence of elevated and clinically raised CRP levels in women
compared with men in overweight and obese persons (Figure 1). The higher prevalence of elevated and clinically raised
CRP levels among obese women compared with obese men could also be due to
by the fact that women were more likely to be extremely obese: a BMI of 35
to 40 kg/m2 was prevalent among 3.4% of men and 6.4% of women,
and a BMI of 40 kg/m2 or more was present among 1.7% of men and
3.6% of women. Both phenomena might also explain why BMI was associated with
clinically raised CRP levels in women but not men.
Persons with a normal body weight (BMI <25 kg/m2) were
used as the reference group. However, this group included a small percentage
(1.3% of men and 3.8% of women) of underweight persons (BMI <18.5 kg/m2) who might be more likely to be in poor health, with associated higher
CRP concentrations. However, when the analyses were repeated after exclusion
of underweight people in the reference group, similar results were obtained.
Because the lower detection limit of the CRP assay was 0.22 mg/dL, serum
CRP level was used as a categorical variable. It is unlikely that the use
of a more sensitive assay would have changed the conclusions of the study.
The association between obesity and CRP concentration was observed regardless
of the CRP cut point that was used (≥0.22 or >1.00 mg/dL). Second, although
the cut point of 1.0 mg/dL has been used in clinical studies, more recent
epidemiological studies have shown an increased risk for cardiovascular disease
at CRP levels of 0.2 mg/dL and higher.9- 13
We used a single CRP measurement that may not accurately reflect long-term
inflammation status. The biological variability of CRP is substantial, with
reported values ranging from 10.6% to 63.0%.43- 46
However, because random misclassification due to biological variability will
lead to underestimation of true associations, this limitation is unlikely
to explain our findings.
Measurements of the serum concentration of IL-6 were not available in
the present study. Although the results support the hypothesis that IL-6 produced
by the adipocytes increase CRP concentration, direct assessment of IL-6 concentration
is needed in future studies to further test this hypothesis.
In conclusion, the results of this large-scale cross-sectional study
show that higher BMI is associated with higher CRP concentrations that could
not be explained by inflammatory disease or other factors or diseases known
to increase CRP concentrations. Because these associations also were observed
among young adults aged 17 to 39 years, subclinical disease is unlikely to
explain our findings. These data suggest that a state of low-grade systemic
inflammation is present in overweight and obese persons.