Body mass and the epidemic of chronic inflammation in early mid-adulthood

Abstract

Objectives.

Chronic inflammation is a potentially important mechanism through which social inequalities may contribute to health inequalities over the life course. Excess body fat contributes to chronic inflammation, and younger adults in the US have come of age during a pronounced secular increase in body mass index (BMI). We aim to document levels of chronic inflammation in a nationally representative sample of 33-to-44 year-old adults in the US, and to describe associations with BMI, race/ethnicity, and education.

Methods.

High sensitivity C-reactive protein (CRP) was measured in Wave V (2016–18) of the National Longitudinal Study of Adolescent to Adult Health, with complete data available for 4,349 participants. Sex-stratified weighted regression models were implemented to investigate CRP in association with education, race/ethnicity, and BMI.

Results.

Geometric mean CRP was 1.9 mg/L, and 35.4% of the sample had CRP > 3 mg/L. Females had significantly higher CRP than males. Body mass index was a strong positive predictor of CRP, and education level was negatively associated with CRP. Associations between education and CRP were substantially attenuated after adjusting for BMI. High risk CRP increased linearly with BMI even among the obese: 87.0 percent of females and 74.1 percent of males with class 3 obesity (BMI ≥40) were predicted to have high risk CRP > 3 mg/L.

Conclusions.

The obesity epidemic is producing an epidemic of chronic inflammation in early mid-adulthood in the US. Strong associations between BMI and chronic inflammation portend high risk for future disease—and inequitable distribution of disease—as the cohort ages.

Chronic inflammation is a well-established risk factor for cardiovascular disease, diabetes mellitus, chronic kidney disease, many cancers, autoimmune disease, depression, and neurodegenerative disorders (1). More than 50% of deaths globally can be attributed to these causes, and cardiometabolic disease alone kills more than 900,000 Americans each year (2). C-reactive protein (CRP) is a widely measured biomarker of systemic inflammation, and a meta-analysis of 54 prospective studies indicates that a one standard deviation increase in logCRP predicts increased risk for coronary heart disease (risk ratio (RR)=1.63), ischemic stroke (RR=1.44), and vascular mortality (RR=1.71) (3).

Prior research on inflammation and disease has focused primarily on older adults who are at increased risk for chronic degenerative diseases. At this stage in the life course, inflammation is understood to be a potential mechanism of aging, contributing directly to functional declines and pathogenesis across a wide range of tissues, which can in turn increase the burden of chronic inflammation in a positive feedback loop termed “inflammaging” (4). Less is known about the causes and consequences of chronic inflammation in younger adulthood. Filling this knowledge gap is particularly important because excess body fat is a primary contributor to chronic inflammation, and younger adults in the US have come of age during a pronounced secular increase in body mass index (BMI) (5). If rising rates of overweight and obesity are driving an epidemic of chronic inflammation in young/middle adulthood, then the degenerative process of inflammaging may be set in motion relatively early in the life course and increase the burden of morbidity and mortality in older adulthood. Socioeconomic and racial/ethnic inequalities in many health outcomes are widely recognized (6), and chronic inflammation is a potentially important pathway through which social inequalities may contribute to health inequalities over the life course (1). In the US, lower levels of education and income have been associated with elevated CRP (7–10), and higher CRP has been reported among individuals who self-identify as Black or Hispanic (11, 12). Parallel inequalities in overweight/obesity have also been documented and linked with chronic inflammation (13), although the extent to which body mass accounts for social inequalities in chronic inflammation in early mid-middle adulthood is not well established.

The objective of this report is to describe the level of chronic inflammation in early mid-adulthood in the US, and to document patterns of association with body mass. In addition, we document social inequalities in chronic inflammation, and evaluate the extent to which body mass accounts for inequalities associated with education level and race/ethnicity. We use data from a large, nationally representative sample of 33 to 44 year-old adults, representing a period in the life course when health is generally good and rates of chronic degenerative diseases are low. However, biological measures like CRP forecast future trajectories of morbidity and mortality, and generate insight into the factors contributing to chronic inflammation prior to the onset of clinical disease.

METHODS

Study population

The study uses data from the National Longitudinal Study of Adolescent to Adult Health (Add Health), a nationally representative cohort of adolescents recruited in 1994–95 (age 12–19 years) and followed into adulthood across five survey waves (14). The core sample included 20,745 adolescents at Wave 1. The current study uses data collected at Wave 5 in 2016–18, when participants were 33–44 years old. Venous blood samples were collected in the home for a nationally representative subset of 4,940 participants. Immediately after collection, tubes were stored at 4C for up to 2 hours, centrifuged and aliquoted, and overnight shipped to the laboratory for the quantification of CRP using a high sensitivity particle-enhanced immunonephelometric assay (15). All data were collected under conditions of informed consent, with protocols approved by the Institutional Review Board at the University of North Carolina, Chapel Hill.

Statistical analysis

The analytic sample included 4,349 participants with complete data. CRP results were log-transformed (base 10) to normalize the distribution. A series of multivariate regression models were implemented to identify the association between CRP and BMI, and to evaluate the extent to which overweight/obesity accounts for differences associated with race/ethnicity, and education. Regression models were run separately by sex. Ordinary least squares regression was used to predict log₁₀ CRP, and logistic regression was used to predict a dichotomous indicator of high-risk CRP (CRP> 3 mg/L) (16). Wave V biomarker sample weights were used to account for sampling probabilities and generate nationally representative estimates, and variance estimators adjusted for the clustered design (17).

Since acute elevations in inflammation can obscure CRP as a measure of chronic inflammation, we controlled for whether respondents reported infectious disease symptoms (e.g., cold or flu-like symptoms; fever; night sweats; nausea, vomiting, or diarrhea) in the two weeks preceding blood collection (15). Participants reporting these symptoms did not differ with respect to age, BMI (continuously measured, or categorized), foreign born, level of attained education, or race/ethnicity. Females (based on reported sex assigned at birth) were more likely to report infectious symptoms than males (OR = 1.39, 95% CI: 1.14, 1.71).

For race/ethnicity, participants were asked to respond to the question “What is your race or ethnic origin?” If multiple categories were checked then respondents were asked to identify the group with which they “most strongly identify.” Education was defined as a categorical variable, based on highest level of attainment at the time of the survey. Foreign-born was indicated using a dichotomous variable indicating whether the respondent was a first-generation immigrant. Height and weight were measured by field examiners (or taken from self-report for the small number not measured during the in-home examination), and BMI was calculated as kg/m². Additional covariates that may influence levels of inflammation were also considered, including the use of anti-inflammatory medication and prescription contraceptives (among females), current smoking (defined as having smoked cigarettes in the 30 days prior to the in-home interview), and pregnancy at the time of blood collection (among females).

RESULTS

Mean age in the analytic sample was 37.9 years, with relatively equal proportion of females (50.3%) and males (49.7%). 71.3% of the sample self-identified as white, 8.8% as Hispanic, 17.2% as Black, and 2.6% as Asian/Pacific Islander. Due to small cell sizes, people identifying as any other category (or combination of categories) were omitted from the sample. 4.0% of respondents were first-generation immigrants. Geometric mean CRP for the sample was 1.9 mg/L, with females having significantly higher CRP than males (2.3 vs. 1.5 mg/L; t = 7.73; p<0.001) (Table 1). Mean BMI was 31.3 kg/m² among females and 30.8 among males, with 25.2% of females and 33.2% of males classified as overweight (BMI 25.0 to 29.9), and 49.2% of females and 46.8% of males as obese (30.0 and higher).

Table 1.

Geometric Mean of CRP (mg/L), by Sex, Education, BMI Category, and Race/Ethnicity in a Nationally Representative Sample of Early Mid-Adults in the US, 2016–18.

	Female	Male	Total
W5 Education
< HS, HS, or GED	3.30	1.96	2.41
Some College	2.63	1.50	1.99
College	1.80	1.26	1.53
W5 BMI Category
Normal/Underweight (BMI<25)	0.85	0.85	0.85
Overweight (BMI 25 to 29.9)	1.78	1.07	1.34
Obese I (BMI 30 to 34.9)	2.82	1.70	2.13
Obese II (BMI 35 to 39.9)	4.17	2.63	3.39
Obese III (BMI≥40)	8.33	5.32	7.00
Race/Ethnicity
White	2.14	1.47	1.78
Black	3.28	1.69	2.35
Hispanic	2.35	1.51	1.86
Asian/Pacific Islander	1.46	0.79	1.09
Total	2.29	1.49	1.85

Note: “Some College” includes some vocational school, completed vocational training, and associate or junior college degree, as well as some four-year or community college.

In multivariate regression models, BMI was a strong positive predictor of CRP, among both females (Table 2) and males (Table 3). In models that do not adjust for BMI, lower levels of education were associated with higher CRP for women and men (Tables 2 and 3, model 2). In addition, CRP was significantly higher for Black women in comparison with white women. For males, there were no significant differences in CRP across race/ethnic groups. The inclusion of BMI eliminated the Black-white difference in CRP for women, and attenuated the association with education for both women and men (Tables 2 and 3, model 3). In general, these patterns of association did not change substantively after adjusting for pregnancy, smoking, prescription contraceptive use, and anti-inflammatory medications (Tables 2 and 3, model 4). One exception, however, was the positive association between CRP and the lowest level of education among males, which was reduced and not statistically significant when adjusting for these covariates.

Table 2.

Ordinary Least Squares Regression Results Predicting Log₁₀ CRP in Females in a Nationally Representative Sample of Early Mid-Adults in the US, 2016–18.

	Model 1	Model 2	Model 3	Model 4
BMI (Wave 5)	0.04*** [0.04,0.04]		0.04*** [0.03,0.04]	0.04*** [0.04,0.04]
Infectious Symptoms	0.11*** [0.06,0.17]	0.13*** [0.06,0.19]	0.11*** [0.05,0.16]	0.09** [0.04,0.15]
Age		−0.02 [−0.03,0.00]	−0.02* [−0.03,−0.00]	−0.01 [−0.02,0.00]
Race/Ethnicity (Reference=White)
Black		0.17*** [0.11,0.24]	0.00 [−0.07,0.06]	0.00 [−0.06,0.07]
Hispanic		0.05 [−0.06,0.16]	0.00 [−0.10,0.09]	0.00 [−0.09,0.09]
Asian/Pacific Islander		−0.07 [−0.25,0.12]	−0.03 [−0.18,0.12]	−0.02 [−0.17,0.13]
Education (Reference=College)
< HS, HS, or GED		0.25*** [0.16,0.34]	0.08* [0.00,0.16]	0.10* [0.02,0.18]
Some College		0.15*** [0.08,0.22]	0.03 [−0.02,0.09]	0.05 [−0.00,0.10]
Foreign-born		−0.13 [−0.34,0.08]	−0.06 [−0.23,0.12]	−0.05 [−0.21,0.10]
Pregnant				0.41*** [0.29,0.53]
Smoked				0.05 [−0.02,0.12]
Prescription Contraceptive Use				0.31*** [0.23,0.39]
Anti-Inflammatory Use				0.00 [−0.04,0.05]
Constant	−0.87*** [−0.97,−0.78]	0.76* [0.15,1.37]	−0.27 [−0.75,0.21]	−0.48* [−0.92,−0.04]
R-squared	0.35	0.06	0.36	0.40
N	2599	2599	2599	2599

^*p<.05

^**p<.01

^***p<.001

Note: 95% CI’s shown in brackets. “Some College” includes some vocational school, completed vocational training, and associate or junior college degree.

Table 3.

Ordinary Least Squares Regression Results Predicting Log₁₀ CRP in Males in a Nationally Representative Sample of Early Mid-Adults in the US, 2016–18.

	Model 1	Model 2	Model 3	Model 4
BMI (Wave 5)	0.03*** [0.03,0.04]		0.03*** [0.03,0.04]	0.03*** [0.03,0.04]
Infectious Symptoms	0.13*** [0.06,0.20]	0.13*** [0.06,0.20]	0.13*** [0.06,0.20]	0.11** [0.04,0.18]
Age		0.01 [−0.00,0.03]	0.01 [−0.00,0.02]	0.01 [−0.00,0.03]
Race/Ethnicity (Reference=White)
Black		0.04 [−0.05,0.12]	0.02 [−0.05,0.10]	0.02 [−0.05,0.10]
Hispanic		0.00 [−0.11,0.11]	−0.03 [−0.12,0.06]	−0.01 [−0.10,0.08]
Asian/Pacific Islander		−0.23 [−0.46,0.01]	−0.12 [−0.30,0.06]	−0.13 [−0.31,0.05]
Education (Reference=College)
< HS, HS, or GED		0.17*** [0.10,0.25]	0.12*** [0.05,0.19]	0.07 [−0.00,0.13]
Some College		0.06 [−0.01,0.13]	0.03 [−0.03,0.08]	0.01 [−0.05,0.07]
Foreign-born		−0.03 [−0.27,0.22]	−0.04 [−0.18,0.09]	−0.02 [−0.16,0.12]
Smoked				0.13*** [0.07,0.19]
Anti-Inflammatory Use				0.04 [−0.02,0.11]
Constant	−0.88*** [−1.00,−0.76]	−0.34 [−0.92,0.24]	−1.32*** [−1.85,−0.78]	−1.38*** [−1.91,−0.85]
R-squared	0.27	0.05	0.28	0.30
N	1750	1750	1750	1750

^*p<.05

^**p<.01

^***p<.001

Note: 95% CI’s shown in brackets. “Some College” includes some vocational school, completed vocational training, and associate or junior college degree.

A cut-off value of CRP>3 mg/L identifies individuals at high risk for the development of cardiovascular disease (16). Overall, 35.4% of the sample had CRP>3, with a higher percentage of females (43.8) than males (26.9) in the high risk group. In a fully adjusted weighted logistic regression model, a one unit increase in BMI was associated with a 19% increase in the odds of CRP>3 mg/L among females (OR=1.19, CI: 1.16, 1.21) and 15% increase among males (OR=1.15, CI: 1.12, 1.18; Tables S1 and S2). Figure 1 breaks this association down by BMI category, and presents the predicted probability of high-risk CRP, holding covariates at their sex-specific means. For individuals who were not overweight or obese (BMI<25), 11% of females and 11% of males were predicted to have CRP>3 mg/L, in comparison with 67 percent of obese females and 41 percent of obese males. When the BMI distribution was divided into categories of obesity, a strong dose-response pattern became clear, with no evidence of ceiling effects at the highest levels of BMI: 88 percent of females and 74 percent of males with class 3 obesity (BMI of 40 or higher) were predicted to have CRP>3 mg/L.

Figure 1.

Predicted probability of CRP > 3mg/L by BMI category and sex in a nationally representative sample of early mid-adults in the US, 2016–18.

Models above included a variable for the presence of infectious symptoms in the two weeks prior to blood collection to adjust for acute elevations in CRP. As a robustness check, we ran the multivariate regression models excluding all participants reporting infectious symptoms or who had CRP>10 mg/L. Although log₁₀ CRP was lower in this subsample, conclusions regarding main findings were substantively unchanged (Table S3, Table S4, Figure S1). Of note, coefficients for two sociodemographic characteristics—some college education and reporting to be Asian/Pacific Islander—became statistically significant for males in these models (Table S4, Model 2). For females, we ran robustness checks excluding participants who were pregnant at the time of CRP collection; main conclusions from multivariate regression analyses were substantively unchanged in these models.

DISCUSSION

Social inequalities in health in the US are large, and for many outcomes getting larger (6). Chronic inflammation is an established clinical marker of risk for a wide range of chronic degenerative diseases of aging, and its measurement in a population of early middle age adults identifies risk for future morbidity and mortality at a relatively healthy stage in the life course. We document a large education gradient in CRP in this cohort, strong associations between BMI and chronic inflammation that account for much of this gradient, and a high prevalence of high-risk CRP that portends increased risk for future disease as the cohort ages.

Our results point toward the accumulation of body mass as a key driver of high levels of chronic inflammation, as well as social inequities in CRP. Visceral adipose tissue produces several pro-inflammatory molecules, including the cytokines IL6 and TNFa, both of which upregulate the CRP production and increase the level of systemic inflammation (18). The accumulation of body mass and duration of obesity during early life are increasingly important risk factors for early onset of cardiometabolic diseases among recent cohorts of young adults (13, 19, 20). The obesity epidemic began in the early 1980s as a period-based phenomenon that affected children and adults of all ages. However, the health consequences have followed a cohort-specific pattern such that those who were younger when the epidemic occurred have been exposed to obesity for a longer period relative to previous cohorts of young adults (21, 22). Add Health cohort members were born in the late 70s and early 80s and have therefore spent their entire lives exposed to obesogenic environments.

A growing body of research shows that both the prevalence of obesity and the rate of increase in obesity with age are higher for more recent adult cohorts than older cohorts born before the 1980s (23, 24). Indeed, nearly half the Add Health cohort can be classified as obese, and young to middle age adults today are more likely than previous cohorts to have comorbidities associated with longer durations of obesity such as hypertension, diabetes, and physical limitations typically associated with older ages (25). Our study underscores the contribution of obesity to chronic inflammation earlier in the life course, with linear increases in high-risk CRP with BMI even among the obese: At a BMI of 40 or higher, 80 percent of 33–44 year-olds have CRP>3 mg/L. Our results suggest that the obesity epidemic is producing an epidemic of chronic inflammation in the US.

This prevalence of high-risk CRP in younger adulthood is alarming, and it forecasts a future wave of cardiovascular and cerebrovascular diseases as the cohort ages into older adulthood. Meta-analysis of long-term prospective studies, using records from participants without a prior history of vascular disease, indicates that a geometric mean CRP of 3 mg/L more than doubles the risk of coronary heart disease, and increases the risk of stroke more than 1.5 times, in comparison with CRP of 1.0 mg/L or lower (3). Risk of all-cause mortality increases 75% for individuals with high-risk CRP (26). These analyses also indicate that CRP increases steadily with age, nearly doubling, on average, as young adults move from their 30s and into their 60s. If these patterns hold for the future, thirty years from now more than 55% of Add Health participants will have CRP>3 mg/L. And as BMI increases with age, along with CRP, it will be important to consider the implications for morbidity and mortality across the upper ranges of BMI, and not only among individuals categorized as overweight and obese.

Consistent with prior epidemiological and demographic studies, we document significant differences in CRP across gender, race/ethnicity, and level of education. For example, analyses of data from the National Health and Nutrition Examination Survey (NHANES) report significantly higher CRP for individuals with lower education, and for Blacks in comparison with whites (27–29). In particular, our study underscores education as an important social determinant of chronic inflammation in early middle adulthood. This is particularly true for females, where the magnitude of association between CRP and education is substantially larger than for males. The addition of BMI, however, attenuates the education gradient in CRP for both females and males. It also eliminates the Black-white gap among females. Thus, the effects of education on chronic inflammation are not operating exclusively through body mass, and future research should consider the pathways through which education, race/ethnicity, and gender may intersect to influence trajectories of health into middle adulthood.

A limitation of our analysis is the use of a single CRP measure to characterize baseline levels of systemic inflammation. However, we adjust for acute inflammation associated with infection, and our conclusions are robust to the exclusion of participants with infectious symptoms around the time of blood collection. In addition, our cross-sectional analysis precludes identifying factors associated with changing levels of inflammation, but still provides the baseline for future studies and the opportunity to track levels of inflammation and disease onset in this prospective cohort going forward. An important contribution of our study is the documentation of high levels and early onset of chronic inflammation in a large, nationally representative cohort of young to middle age adults, whereas prior research has focused primarily on older adults, when diseases of aging are more prevalent and may be an underlying cause of chronic inflammation.

Levels of chronic inflammation are high among young and middle-age US adults

Body mass predicts chronic inflammation and largely explains differences by education

The obesity epidemic is producing an epidemic of chronic inflammation in the US

High levels of chronic inflammation portend increased risk for future disease