SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusion
  9. Acknowledgments
  10. References

Objective

Sepsis, the syndrome of microbial infection complicated by systemic inflammation, is associated with significant morbidity and mortality. To determine if obesity increases risk of sepsis events.

Design and Methods

Data from the 30,239 subject population-based longitudinal cohort study REasons for Geographic and Racial Differences in Stroke (REGARDS) were used. Using measurements at the start of the study, we defined obesity using body mass index (BMI; <18.5 kg/m2 = underweight, 18.5-24.9 = normal, 25.0-29.9 = overweight, 30.0-39.9 = obese, ≥40 = morbidly obese) and waist circumference (WC; [male ≤102 cm or female ≤88 cm] = normal, [male >102 cm or female >88 cm] = obese). Over an 8-year observation period, we evaluated the association between obesity and subsequent sepsis events, adjusting for sociodemographic factors, health behaviors, chronic medical conditions, statin use, and high-sensitivity C-reactive protein.

Results

There were 975 incident sepsis events. Compared to those with a BMI of 18.5-24.9, sepsis risk was higher only for BMI ≥ 40 (hazard ratio [HR] 1.57, [1.16-2.14]). Risk of sepsis was associated with increased WC (HR 1.34 [1.14-1.56]). In a model with both BMI and WC, sepsis risk was associated with increased WC (HR 1.47 [1.20-1.79]) but not BMI.

Conclusions

Obesity is independently associated with future sepsis events. WC is a better predictor of future sepsis risk than BMI.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusion
  9. Acknowledgments
  10. References

Obesity remains one of the nation's most important public health problems, afflicting over one-third of US citizens [1]. Obesity has been attributed to increased mortality, and obese individuals are at greater risk for serious medical conditions such as cardiovascular disease and diabetes [2].

Sepsis, the syndrome of microbial infection complicated by systemic inflammation, is also a major public health problem associated with significant morbidity and mortality [3]. The substantial national burden of sepsis care in the US encompasses 750,000 hospital admissions, 570,000 Emergency Department visits, 200,000 deaths, and $16.7 billion in medical expenditures annually [4-6]. There are interesting clinical and pathophysiological connections between obesity and sepsis. In animal models, obesity is associated with exacerbated inflammatory responses [7, 8]. Sepsis often results in critical illness, and in these individuals obesity is associated with impairment of pulmonary function, antibiotic distribution, and insulin function [9, 10].

Prior studies have examined the connection between obesity and outcomes after hospitalization for sepsis or other critical illness [11-13]. However, there has been relative little attention directed toward the connection between obesity and the risk of future sepsis events. The latter point is important because efforts to reduce the public health impact of sepsis have focused primarily on optimizing hospital outcomes after the onset of disease rather than identification of the antecedent risk factors for developing the condition [14, 15]. As has been demonstrated for conditions such as cardiovascular disease and diabetes, the identification of obesity as an independent risk factor for sepsis would provide a target for sepsis risk prediction as well as a modifiable target for potential sepsis risk reduction. The objective of this study was to determine the association between baseline obesity and future risk of sepsis in community-dwelling individuals.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusion
  9. Acknowledgments
  10. References

Study design

The study was approved by the Institutional Review Board of the University of Alabama at Birmingham. This study utilized the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study, a national, population-based, and longitudinal cohort.

Selection of participants

Designed to evaluate reasons for geographic and racial variations in stroke mortality, REGARDS is one of the largest ongoing national cohorts of community-dwelling individuals in the USA, encompassing 30,239 individuals ≥45 years old [16]. REGARDS includes individuals from all regions of the continental USA. Participant representation oversampled the Southeastern USA, with 21% of the cohort originating from the coastal plains of North Carolina, South Carolina and Georgia (the “buckle” of the stroke belt), and 35% originating from the remainder of North Carolina, South Carolina and Georgia plus Tennessee, Mississippi, Alabama, Louisiana and Arkansas (the “stroke belt”). The cohort is 42% African American and 45% male, and 69% of individuals are >60 years old, and does not include Hispanics where stroke mortality disparities are small-to-non-existent.

REGARDS enrolled participants during 2003-2007, obtaining baseline data for each participant using both phone interview and in-person evaluations. Baseline data included medical history, functional status, health behaviors, physical characteristics (height, weight), physiologic measures (blood pressure, pulse, electrocardiogram), and an inventory of medications. Each participant provided blood and urine specimens. Self-administered questionnaires evaluated diet, family history of diseases, psychosocial factors, and prior residences. On a semi-annual basis, the study contacted each participant to determine the date, location, and attributed reason for all emergency department visits and hospitalizations during the follow-up interval. If the participant died, the study team reviewed death certificates and related medical records and interviewed proxies to ascertain the circumstances of the participant's death.

Identification of sepsis events

Using infection taxonomies developed by Angus et al., we reviewed all reported hospitalizations and Emergency Department visits attributed by participants to a serious infection [4]. Two trained abstractors independently reviewed all relevant medical records to confirm the presence of a serious infection on initial hospital presentation, and the relevance of the serious infection as a major reason for hospitalization. The abstractors identified clinical and laboratory information from the first 28-hours of hospitalization a time period encompassing Emergency Department and up to one full day of inpatient treatment. The abstractors adjudicated discordances, with additional physician-level review as needed.

Consistent with international consensus definitions, sepsis consisted of presentation to the hospital with an infection plus two or more systemic inflammatory response syndrome (SIRS) criteria, including (1) heart rate >90 beats/minute, (2) fever (temperature >38.3°C or <36°C), (3) tachypnea (>20 breaths/min) or PCO2 <32 mmHg, and (4) leukocytosis (white blood cells [WBC] >12,000 or <4000 cells/mm3 or >10% band forms). Presentation to the hospital consisted of the time of Emergency Department triage or admission to inpatient unit (for participants admitted directly to the hospital). To allow for acute changes in the participant's condition during early hospitalization, we used vital signs and laboratory test results for the initial 28-hours of hospitalization. We did not include vital signs or laboratory findings at later time points, not did we include sepsis developing at later phases of hospitalization. We did not include organ dysfunction in the definition of sepsis. Initial review of 1349 hospital records indicated excellent inter-rater agreement for presence of a serious infection (kappa = 0.92) and the presence of sepsis (kappa = 0.90) upon hospital presentation.

Definition of obesity

Following standardized protocols, weight, height, and waist circumference (WC) were measured during initial subject examination at the beginning of the REGARDS study. Body mass index (BMI) was calculated as weight/height2 (kg/m2), and was categorized as underweight (<18.5 kg/m2), normal (18.5-25.0 kg/m2), overweight (25.0-29.9 kg/m2), obese (30-39.9 kg/m2), and morbidly obese (≥40 kg/m2) [17]. WC was determined with the subject standing and was measured midway between the lowest rib and the iliac crest, with normal WC defined as ≤102 cm for males and ≤88 cm for females and large WC as >102 cm for males and >88 cm for females [18].

Covariates

We considered covariates that may confound the relationship between obesity and sepsis, including sociodemographic characteristics, health behaviors, and chronic medical conditions. Sociodemographic characteristics included age, sex, race, geographic region, self-reported annual household income, and self-reported education (years of school). Geographic region was defined as participant residence in the stroke “buckle,” stroke “belt,” and elsewhere, included here to account for the sampling strategy used to recruit the REGARDS cohort [16]. Health behaviors included tobacco and alcohol use, and exercise. Smoking status use was defined as current, past and never. We defined alcohol use according to the National Institute on Alcohol Abuse and Alcoholism classification; i.e., moderate (1 drink per day for women or 2 drinks per day for men) and heavy alcohol use (>1 drink per day for women and >2 drinks per day for men) [19]. Participants reported the number of times per week of exercise (none, 1-3, 4 or more).

Evaluated chronic medical conditions included hypertension, diabetes, dyslipidemia, coronary artery disease, chronic kidney disease, and chronic lung disease. Hypertension consisted of systolic blood pressure ≥140 mm Hg, diastolic blood pressure ≥90 mm Hg, or the self-reported use of antihypertensive agents. Diabetes included a fasting glucose ≥126 mg/l (or a glucose ≥200 mg/l for those not fasting) or the use of insulin or oral hypoglycemic agents. Dyslipidemia included individuals with self-reported high cholesterol or the use of lipid lowering medications. A history of coronary artery disease consisted of individuals with a self-reported history of myocardial infarction, coronary intervention or baseline electrocardiographic evidence of myocardial infarction.

Chronic kidney disease consisted of an estimated glomerular filtration rate <60 ml/min/1.73 m2, calculated using the CKD-EPI equation [20]. Because REGARDS did not collect information on pulmonary conditions such as asthma and chronic obstructive pulmonary disease, we defined participant use of pulmonary medications as a surrogate for chronic lung disease. Obtained from each participant's medication inventory, pulmonary medications included beta agonists, leukotriene inhibitors, inhaled corticosteroids, combination inhalers, and other pulmonary medications such as ipatropium, cromolyn, aminophylline, and theophylline. We determined statin use through the participant's medication inventory.

Data analysis

We compared demographic, health behavioral, and clinical characteristics between BMI and WC categories using a chi-square test. We used a Cox proportional hazards model to calculate hazard ratios (HRs) and 95% confidence intervals (CI) for the association between elevated obesity and first episode of sepsis during follow-up. We defined person-time at risk as the time (days) from first in-person examination to the first episode of sepsis or the last follow-up interview, whichever came first.

We fit separate models for BMI and WC. We adjusted the models for demographic characteristics (age, sex, race, income, education, geographic region), health behaviors (smoking and alcohol use, exercise), chronic medical conditions (hypertension, diabetes, dyslipidemia, coronary artery disease, chronic kidney disease, chronic lung disease), and statin use. To test whether the proportionality assumption of the Cox model was met for both models, interactions with time for all variables in the model were included.

To further evaluate the robustness of the findings, we fit a model with both BMI and WC, as well as models of WC stratified by BMI. We examined variance inflation factor values to identify potential collinearity between WC and BMI. We also explored diabetes as a potential effect modifier by evaluating (BMI X diabetes) and (WC X diabetes) interactions.

Due to the time lag in observations and medical record retrieval, we could not review records for a portion of participants with reported hospitalizations for a serious infection. In a sensitivity analysis, we repeated assessment of the association between obesity measures and first sepsis event, excluding all data on individuals with unretrieved medical records. The sensitivity analysis excluded all hospital events for the participant – not just the individual unretrieved hospital event.

We considered P-values <0.05 to be statistically significant. We conducted all analyses using SAS v.9.3.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusion
  9. Acknowledgments
  10. References

Among the REGARDS participants, from February 5, 2003 through July 30, 2012, there were 2157 hospitalizations for a serious infection, including 1297 sepsis events. A total of 975 unique individuals experienced a sepsis event. Mean follow-up time was 4.6 years. Among the 975 incident sepsis events, the most common infection types were pneumonia, kidney and urinary tract infections, and abdominal infections (Table 1).

Table 1. Infection types associated with hospitalizations for sepsis. Includes first sepsis events for 975 individuals
Infection typeNumber of sepsis hospitalizations (n = 975) N (%)
Pneumonia427 (43.8)
Kidney and urinary tract infections155 (15.9)
Abdominal133 (13.6)
Bronchitis, influenza, and other lung infections84 (8.6)
Skin and soft tissue71 (7.3)
Sepsis63 (6.5)
Fever of unknown origin14 (1.4)
Unknown/other14 (1.4)
Surgical wound6 (0.6)
Catheter (IV/central/dialysis)5 (0.5)
Meningitis3 (0.3)

Mean BMI was 29.3 ± 6.2 kg/m2. Mean WC was 100.2 ± 13.7 cm for males and 92.9 ± 16.4cm for females. BMI and WC were higher among younger, female, and African American subjects. (Tables 2 and 3) Subjects with high BMI and WC were more likely to have chronic medical conditions.

Table 2. Demographic and health behavioral characteristics between obese and non-obese subjects as measured by body mass index
  Body mass index (BMI – kg/m2) 
Characteristic<18.5 (n = 319)18.5-24.9 (n = 7091)25.0-29.9 (n = 11,057)30.0-39.9 (n = 9640)≥ 40.0 (n = 1859)P-valuea
  1. a

    Based on chi-square test.

Demographics
Age (%)
45-5410.311.310.412.317.7<0.001
55-6432.832.135.242.149.2 
65-7427.830.633.832.326.9 
75+29.126.020.613.26.3 
Gender (%)
Male30.244.153.041.022.3<0.001
Female69.855.947.059.077.7 
Race (%)
White63.269.662.550.336.1<0.001
African American36.830.437.549.763.9 
Education (%)
Less than high school18.210.511.514.117.8<0.001
High school graduate26.124.325.227.328.2 
Some college23.325.026.727.930.0 
College or higher32.440.236.730.724.0 
Income (%)
<$20k27.416.815.419.728.5<0.001
$20k-$34k27.023.224.324.526.1 
$35k-$74k16.729.030.730.124.8 
≥$75k9.117.517.414.19.0 
Refused19.813.512.211.611.5 
Geographic region (%)
Stroke Buckle17.321.120.221.322.30.02
Stroke Belt39.335.334.034.733.8 
Non-Belt/Buckle43.443.645.744.043.9 
Health behaviors
Tobacco use (%)
Current42.319.413.612.210.8<0.001
Past22.434.942.841.938.6 
Never35.345.843.745.950.6 
Alcohol use (%)
Heavy8.05.74.32.91.2<0.001
Moderate33.237.136.429.321.6 
None58.857.259.367.977.2 
Exercise
None42.931.230.138.349.4<0.0001
1-3 times/week27.134.237.137.033.6 
≥4 times/week30.034.632.824.717.0 
Chronic medical conditions
Chronic lung disease (%)12.38.28.110.013.5<0.001
Chronic kidney disease (%)28.822.422.227.232.4<0.001
Peripheral artery disease (%)4.42.42.22.01.80.01
Deep vein thrombosis (%)6.74.44.66.08.0<0.001
Diabetes (%)6.49.917.931.343.8<0.001
Stroke (%)9.26.36.46.36.60.35
Myocardial infarction (%)8.97.28.69.37.4<0.001
Coronary artery disease (%)17.916.418.419.015.5<0.001
Atrial fibrillation (%)9.98.98.38.810.30.06
Hypertension (%)37.445.256.269.580.4<0.001
Dyslipidemia (%)16.125.835.137.635.6<0.001
Elevated hsCRP (%)23.024.832.348.368.6<0.001
Other
Statin use (%)15.424.333.035.132.9<0.001
Table 3. Demographic and health behavioral characteristics between obese and non-obese subjects as defined by waist circumference (WC)
 Normal WC (n = 15,448)Large WC (n = 14,735)P-valuea
  1. Normal WC = Male ≤102 cm or female ≤88 cm. Large WC = male >102 cm or female >88 cm.

  2. a

    Based on chi-square test.

Demographics
Age (%)
45-5412.111.3<0.0001
55-6436.039.2 
65-7431.033.1 
75+20.816.4 
Gender (%)
Male55.034.3<0.001
Female45.065.7 
Race (%)
White65.451.4<0.001
African American34.648.6 
Education (%)
Less than high school10.215.0<0.001
High school graduate24.427.4 
Some college26.127.6 
College or higher39.329.9 
Income (%)
<$20k14.522.0<0.001
$20k-$34k22.925.6 
$35k-$74k31.327.7 
≥$75k19.012.3 
Refused12.412.3 
Geographic region (%)
Stroke Buckle21.020.80.40
Stroke Belt34.335.0 
Non-Belt/Buckle44.744.2 
Health behaviors
Tobacco use (%)
Current15.913.3<0.001
Past39.341.0 
Never44.945.6 
Alcohol use (%)
Heavy4.93.1<0.001
Moderate38.328.1 
None56.868.9 
Exercise
None28.340.8<0.0001
1-3 times/week36.635.3 
≥4 times/week35.123.9 
Chronic medical conditions
Chronic lung disease (%)7.710.7<0.001
Chronic kidney disease (%)21.128.5<0.001
Peripheral artery disease (%)2.32.20.39
Deep vein thrombosis (%)4.26.4<0.001
Diabetes (%)12.731.8<0.001
Stroke (%)5.87.1<0.001
Myocardial infarction (%)7.79.3<0.001
Coronary artery disease (%)16.919.0<0.001
Atrial fibrillation (%)8.29.4<0.001
Hypertension (%)49.469.6<0.001
Dyslipidemia (%)29.837.4<0.001
Elevated hsCRP (%)26.749.7<0.001
Other
Statin use (%)28.234.8<0.001

After adjustment for confounders, only the highest BMI category (morbid obesity) was independently associated with increased sepsis risk (adjusted HR 1.57, 95% CI 1.16-2.14) (Table 4). Large WC was independently associated with increased sepsis risk (adjusted HR 1.34, 95% CI: 1.15-1.56). For both models the interactions with time were not statistically significant, suggesting that the proportionality assumption was satisfied.

Table 4. Hazard ratios (HRs) and 95% confidence intervals (CI) for the association between obesity and first sepsis episode
Measure of obesityNRisk of sepsis (per 1000)Unadjusted HR for sepsis (95% CI)-separate models for BMI and WCaAdjusted HR for sepsis (95% CI)-separate models for BMI and WCbAdjusted HR for sepsis (95% CI)-separate models for BMI and WCbAdjusted HR for sepsis (95% CI)-model including both BMI and WC
  1. BMI, Body mass index; WC, Waist circumference.

  2. a

    Adjusted for age, sex, race, geographic region.

  3. b

    Adjusted for age, race, gender, geographic region, income, education, smoking status, alcohol use, exercise, statin use, hsCRP level, hypertension, diabetes, dyslipidemia, coronary artery disease, chronic kidney disease, and chronic lung disease.

Body mass index (BMI)
<18.5 kg/m231840.91.66 (0.95-2.91)2.00 (1.14-3.53)1.50 (0.81-2.77)1.56 (0.84-2.88)
18.5-24.9 kg/m2709128.3RefRefRefRef
25.0-29.9kg/m211,05729.21.00 (0.84-1.19)1.11 (0.92-1.33)1.01 (0.83-1.23)0.87 (0.71-1.08)
30.0-39.9 kg/m2964034.11.20 (1.01-1.43)1.52 (1.26-1.84)1.10 (0.89-1.35)0.81 (0.62-1.06)
≥40.0 kg/m2185947.91.78 (1.39-2.29)2.96 (2.26-3.89)1.57 (1.16-2.14)1.14 (0.81-1.62)
Waist circumference (WC)
Male ≤ 102 cm or Female ≤ 88 cm15,44825.4RefRefRefRef
Male > 102 cm or Female > 88 cm14,73539.41.60 (1.41-1.82)1.88 (1.64-2.16)1.34 (1.15-1.56)1.47 (1.20-1.79)

In a model with both BMI and WC as independent variables, large WC – but not BMI – remained independently associated with sepsis (HR 1.47, 95% CI 1.20-1.79). The tolerance was 0.42 and the variance inflation factor was 2.4, suggesting no problems of collinearity between WC and BMI.

We repeated the analysis examining the association between WC and risk of sepsis, stratified by BMI. We observed that WC was associated with increased sepsis risk for BMI 25-29.9 and 30-39.9 but not BMI<25 or BMI≥40. (Table 5) Interactions between BMI and diabetes as well as WC and diabetes were not significant, suggesting no effect modification by diabetes.

Table 5. Hazard ratios and 95% confidence intervals or the association between elevated waist circumference and first sepsis episode, stratified by body mass index category
  Body mass index (kg/m2) 
Waist circumference<25.0a25.0-29.930.0-39.9≥40.0
  1. All models adjusted for age, race, gender, geographic region, income, education, smoking status, alcohol use, exercise, statin use, hsCRP level, hypertension, diabetes, dyslipidemia, coronary artery disease, chronic kidney disease, and chronic lung disease. BMI, Body mass index.

  2. a

    BMI <18.5 and 18.5-24.9 combined into a single category due to the low number of BMI<18.5 with abnormal waist circumference.

Male ≤ 102 cm or female ≤ 88 cmReferentReferentReferentReferent
Male > 102 cm or female > 88 cm1.65 (0.99-2.73)1.42 (1.10-1.83)1.80 (1.12-2.90)0.58 (0.08-4.33)

There were 1157 participants with reported serious infection hospitalizations that had not yet been reviewed or adjudicated, a figure expected to yield an additional 300 first sepsis events. Analyses excluding these individuals revealed results similar to the primary analysis. (Appendix) Compared with the remainder of the cohort, the excluded individuals exhibited similar BMI and WC. Compared with the remainder of the cohort, excluded individuals were older (P = 0.002), more likely to be female (P = 0.02), exhibited lower income (P = 0.006), reported more alcohol use (P = 0.004), reported less exercise (P = 0.02), and were more likely to have hypertension (P < 0.001), dyslipidemia (P = 0.04), coronary artery disease (P = 0.001), atrial fibrillation (P < 0.001), deep vein thrombosis (P = 0.002), chronic kidney disease (P < 0.001), chronic lung disease (P < 0.001), and elevated C-reactive protein (P < 0.001).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusion
  9. Acknowledgments
  10. References

Prior studies have evaluated the association between obesity and the outcomes of individuals suffering from sepsis or other critical illness [11-13]. Our study extends upon these findings, suggesting a connection between obesity at a stable phase of health and the development of future sepsis events.

There are plausible connections between the hypothesized pathophysiologic features of obesity and susceptibility to future sepsis events. One hypothesized mechanism of obesity suggests that increased adiposity induces a chronic inflammatory state characterized by increased cytokine production by adipocytes or macrophages infiltrating adipose tissue [21]. Exaggerated inflammatory response to microbial infection is a prominent feature of sepsis [22]. Yende et al. described associations between baseline inflammatory markers (IL-6, TNF7-alpha) and future risk of pneumonia, suggesting that individuals with a chronic hyperinflammatory state may be at increased risk for future infection or sepsis events [23]. Adipose tissue secretes proinflammatory adipokines such as such as interleukin-6, tumor necrosis factor-alpha and calcitonin, which are commonly associated with sepsis pathophysiology [24]. Adipocytes also express Toll-like receptors, which are responsive to endotoxin. We have previously reported that individuals with elevated baseline TNF-alpha are increased risk of future sepsis events [25].

Another hypothesized mechanism of obesity is the presence of systemic lipotocixity resulting from adiposity, leading to the production of toxic metabolites and over-activation of oxidative pathways [26]. Oxidative stress and high-lipid concentrations may lead to apoptosis and endothelial dysfunction [26, 27]. The endothelium plays a prominent role in immune response and sepsis pathophysiology, facilitating leukocyte trafficking, activation of coagulation and increased vascular leakage [28]. An animal study suggests that obesity exacerbates sepsis-induced microvascular dysfunction [7]. While endothelial cell activation plays a prominent role in acute sepsis, we have identified that chronically elevated markers of endothelial cell activation may predict future episodes of sepsis [25].

Prior studies have pointed to associations between obesity and diabetes as well as diabetes and infection risk [29]. Specifically, defects in neutrophil function, including abnormalities in adhesions, chemotaxis, and intracellular killing, have been observed in diabetes and may be heighten the risk of infections and sepsis. In this context, one would expect obesity to act as a surrogate marker for diabetes. However, our study found that obesity was independently associated with sepsis, even after accounting for the confounding influence of diabetes. Furthermore, on examination of obesity and diabetes interactions, we did not find any evidence of effect modification. Therefore, the findings of our study cannot be completely attributed to the coexistence of diabetes.

An interesting observation was that WC was a stronger predictor of future sepsis events than BMI. This finding is not surprising since WC and sagittal diameter are better predictors of dyslipidemia, metabolic syndrome, cardiovascular disease, sudden cardiac death, and all-cause mortality than BMI [30-32]. In a study of 403 intensive care unit patients (including one-third with sepsis or septic shock), sagittal abdominal diameter was a stronger predictor of death than BMI [32]. WC may better reflect central abdominal and visceral obesity, which have stronger connections with cardiovascular and metabolic abnormalities than general obesity [10]. For example, in a study of men with similar BMI, viscerally obese individuals exhibited lower levels of adiponectin, which plays a key role in glucose regulation and fat catabolism [33]. Compared with subcutaneous fat, visceral fat expresses more proinflammatory cytokines; as discussed previously, a chronic hyperinflammatory state may be associated with sepsis risk. CRP levels also appear to be elevated in individuals with abdominal obesity; we have previously found that elevated high sensitivity CRP is associated with sepsis risk [34, 35].

Obesity is a modifiable condition. Therefore, the most important question raised by our study is whether weight reduction could lower the future risk of sepsis events. Exercise and weight reduction have been demonstrated to reduce the risk of medical conditions such as diabetes and cardiovascular disease [36, 37]. Weight loss may also alter metabolic profiles. Shai et al. showed that weight loss increased adiponectin and reduced high sensitivity C-reactive protein (hsCRP), which we have shown to be associated with sepsis risk [35, 38]. Pharmacotherapy may help to prevent or alter the course of obesity-related cardiovascular disease in obese individuals and could potentially play a similar role in sepsis [10]. While often difficult to achieve or maintain, the associations identified in this study offer additional factors for motivating weight reduction.

Limitations

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusion
  9. Acknowledgments
  10. References

Due to time lags in event reports and record retrieval, there were 1157 individuals with unretrieved medical records for reported serious infection hospitalizations. However, we repeated the analysis excluding these individuals and found similar results. Because REGARDS is not a surveillance study, we likely did not detect all sepsis events. However, there is no reason to believe that misclassification of sepsis events occurred between BMI or WC groups. Hence, our reported associations likely reflect underestimates of the true association.

We did not examine severity variants of sepsis such as severe sepsis and septic shock because these conditions often develop later in the hospital course; however, it is possible that associations between obesity and various forms of sepsis may differ than those reported here. We also did not examine death after sepsis; prior studies suggest that obesity may be protective against death after the development of sepsis [11, 13]. Our objective, however, was to evaluate the associations with the risk of developing sepsis – not sepsis recovery.

By design, the REGARDS cohort includes only African Americans and whites, and thus these results may not generalize to other ethnic groups. History of cancer was not ascertained by REGARDS, which may represent an important risk factor for subsequent sepsis. Also, our study was able to detect the presence of chronic medical conditions but not their level of severity. Our analysis utilized baseline obesity measurements; we could not assess the effect of changes in weight, BMI or WC over time.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusion
  9. Acknowledgments
  10. References

In this study obesity was associated with increased risk of future sepsis. WC was a better predictor of sepsis risk than BMI. Weight reduction or the control of its sequelae may provide options for sepsis risk prediction, mitigation or prevention.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusion
  9. Acknowledgments
  10. References

The authors thank the other investigators, the staff, and the participants of the REGARDS and REGARDS-Sepsis studies for their valuable contributions. A full list of participating REGARDS investigators and institutions can be found at http://www.regardsstudy.org.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusion
  9. Acknowledgments
  10. References